Phase I Massacre and Pedernales Watersheds

Transcription

Phase I Massacre and Pedernales Watersheds
 Revegetation and transboundary natural resources management project: Phase I Massacre and Pedernales Watersheds Final Report February 16, 2012 Prepared by: Alexandra Morel Kevin Tschirhart Paola Kim-­‐Blanco Joseph Muhlhausen Earth Institute, Columbia University EXECUTIVE SUMMARY 4 KEY FINDINGS SOCIO-­‐ECONOMIC BASELINE DATA LAND COVER MAPPING (1986-­‐2010) CONCLUSIONS SUGGESTED FOLLOW-­‐UP RESEARCH 4 4 4 5 5 INTRODUCTION 6 THE FRONTERA VERDE PROJECT MAIN AIMS OF STUDY MASSACRE WATERSHED PEDERNALES WATERSHED 6 8 9 10 METHODS 10 SOCIO-­‐ECONOMIC DATA LAND-­‐COVER MAPPING 10 11 RESULTS 11 SOCIO-­‐ECONOMIC DATA HOUSEHOLD SIZE EDUCATION COSTS MAIN SOURCE OF ENERGY WATER SOURCES ENERGY AND WATER INSECURITY FOOD INSECURITY HEALTH ISSUES WEALTH HOUSEHOLDS CONCERNS LAND-­‐COVER MAPPING (1986-­‐2010) MASSACRE WATERSHED PEDERNALES WATERSHED CHANGE IN MANGROVE EXTENT 11 11 12 14 17 20 25 32 33 41 45 45 51 56 LIVELIHOODS 58 OVERVIEW OF THE MAIN SOURCES OF INCOME REPORTED SEASONAL DEPENDENCE ON INCOME SOURCES SEASONAL FOOD INSECURITY, FARMING AND ENERGY VULNERABILITY FARMING AND LIVESTOCK AGRICULTURE AS A MAIN SOURCE OF INCOME LIVESTOCK AS A LIVELIHOOD STRATEGY CHARCOAL PRODUCTION STRONG DISPARITIES ACROSS AREAS AND COMMUNITIES A SLIGHTLY MORE INTENSE AND RECENT ACTIVITY IN THE SOUTHEAST A STRONG DECREASE IN WOOD RESOURCES OVER THE PAST 2 YEARS SPECIES OF TREES AND COLLECTION METHODS FISHING BASIC DEMOGRAPHICS A REDUCTION IN THE NUMBER OF FISH DEPENDENCE ON TRADITIONAL FISHING EQUIPMENT FISHERMEN APPEAR TO BE BETTER ORGANIZED IN THE SOUTHEAST INCIDENCE OF FISHING ACROSS THE BORDER DIFFERS BY WATERSHED FRONTERA VERDE BRIGADE MEMBERSHIP 58 62 63 65 65 67 67 67 69 73 74 77 77 80 81 82 82 83 Frontera Verde Final Report 2 SELECTION PROCESS FOR MEMBERSHIP PREVIOUS LAND USE OFTEN FORESTED ALREADY ASSESSMENT OF MEMBERSHIP CRITERIA 83 84 85 INTERACTION AND CONFLICTS WITH DOMINICANS 88 BORDER CROSSING MAIN REASONS FOR CROSSING THE BORDER ISSUES AND CONFLICTS IN THE BORDERLAND TENSIONS AND FRUSTRATIONS ON THE HAITIANS SIDE CONFLICTS AND FEELINGS ACROSS WATERSHEDS A STRONG SENTIMENT OF FEELING ABANDONED AND RESIGNATION RECOMMENDATIONS MADE BY HAITIANS FOR AVOIDING CONFLICTS 88 89 91 94 96 99 100 CONCLUSIONS 102 INTERVENTIONS SHOULD FOCUS ON REDUCING FOOD INSECURITY SEASONAL AND SPATIAL VARIABILITY IN VULNERABILITY SHOULD BE CONSIDERED IMPROVING LIVELIHOODS IN HAITI WILL REDUCE BORDER TENSIONS GREATER CERTAINTY IN BORDER GOVERNANCE WOULD BENEFIT HAITIANS SUGGESTED FOLLOW-­‐UP RESEARCH 102 102 103 103 103 APPENDIX 1-­‐1 105 CONCEPTUAL MAPPING OF SOURCES OF INCOME AND A HOUSEHOLD’S MAIN CONCERNS FOOD INSECURITY SCORES AND CATEGORIES WEALTH QUARTILES 105 106 107 APPENDIX 1-­‐2 108 REFERENCES 109 Frontera Verde Final Report 3 Executive Summary Key findings Socio-­‐economic baseline data We present analysis concerning gender of household head, household size, education costs, livelihood strategy, energy/water sources and insecurity, food insecurity, household concerns, asset wealth and health as well as how some of these factors interact to influence incidence of food insecurity or dependence on charcoal production. Our results suggest that households that have lower asset values and higher food insecurity scores are significantly more likely to produce charcoal. We also describe characteristics we were able to discern of our target livelihood strategies for this study, namely: farming, livestock, fishing, Frontera Verde tree planting and charcoal production. Our analysis of the Frontera Verde criteria for choosing brigade members reveals that this project may be inadvertently choosing the wealthiest members of the community, thereby limiting the environmental benefits of the project. Finally, we discuss the findings from our qualitative questions on border conflicts between Haitians and Dominicans, from the Haitian perspective only. The results of this study suggest that there are significant differences between the two watersheds sampled, with the Southeast exhibiting lower incidences of food insecurity and generally higher asset values. However, within the Pedernales watersheds the three eco-­‐zones (mountainous, mid-­‐altitude and coastal) exhibit very different patterns and future interventions in this region will need to take these into account. Massacre on the other hand has some differences; however, they are not easily described by proximity to roads and rivers or by dividing communities into those with Frontera Verde brigades or not. Assessment of seasonal interaction between energy, food insecurity and farm income revealed some interesting patterns. These results are indicating that households have different vulnerabilities throughout the year and may be relying on different income streams depending on what strategies are available to them. Land cover mapping (1986-­‐2010) Land cover classification using k-­‐means unsupervised classification of Landsat 5 images was performed for both watersheds. Three time periods are presented, 1986, 2001 and 2010, largely limited by when relatively, cloud-­‐free images of a similar time period were available. Results for Massacre suggest a complicated interplay of less-­‐dense and intensive agriculture as well as semi-­‐natural pastureland use, with both sides of the border exhibiting low forest cover. This may be an artifact of confusion of the classification algorithm between lightly forested and agricultural areas or a viable result. However, the change detection results do not reveal a clear pattern of deforestation or areas of particularly rapid land cover change. For Pedernales, the pattern appears to be clearer. The Dominican side of the border exhibits higher forest cover, as expected, although there is an “arc of deforestation” pushing into the Dominican side that may or may not be driven by Haitian charcoal production from the mid-­‐altitude region of this watershed. The increase in this deforestation appears to be occurring between the 2001 and 2010 time periods, which was consistent with surveys of the mid-­‐altitude region where respondents cited more recent charcoal production activities. In terms of mangrove area, we found that the areas are relatively small and have stayed relatively stable over the time period assessed. We did, however, notice a reducing in mangrove extent near the town of Jackzil, which is consistent with the results of our qualitative data collection in the Massacre watershed. Conclusions The following bullets summarize the main conclusions of this report: 1. Interventions should focus on reducing food insecurity 2. Seasonal and spatial variability in vulnerability should be taken into account for project design 3. Improving livelihoods in Haiti will reduce border tensions 4. Greater certainty in border governance would benefit Haitians Suggested Follow-­‐up Research 1. A more robust on the ground mapping effort for Massacre. This would focus on identification of steep slopes and relatively bare soils, which are contributing to significant soil erosion that has silted the Massacre River and negatively impacted water quality and fisheries down stream. 2. Development of a monitoring program to track the progress of planted seedlings or enforce their management by the parcel’s landowner. 3. Better understanding needed regarding the timing and scale of charcoal production for sale and gathering of firewood for energy use. 4. Finally, the most straightforward follow up research, would be to perform a similar study on the Dominican side of the border to compare our results. Frontera Verde Final Report 5 Introduction The Frontera Verde Project The degraded state of the environment and the increasing depletion of forest, marine and fresh water resources along the border zone of Haiti and the Dominican Republic are widely recognized as important factors in increasing disaster vulnerability and fuelling tensions between the two countries. At the same time, the ecological interconnections between transboundary watersheds could also represent an opportunity to enhance confidence building and cooperation for natural resource management and sustainable development between the two countries. The Frontera Verde project is part of an overall initiative aiming to reforest the border area between Haiti and the Dominican Republic. It is supported by the Government of Norway and involves the cooperation between several institutions including the Haitian and Dominican Ministries of the Environment, UNDP, UNEP and WFP. This effort is a continuation of an existing and much lauded reforestation effort across the Dominican Republic known as Quisqueya Verde, which was launched in 1997. The model for Frontera Verde is similar to that used in the Dominican Republic; however it has only been operating since May 2011. The managers of this effort decided to begin efforts in two watersheds located along the border between Haiti and the Dominican Republic. Massacre is the most northern watershed in Haiti and Pedernales the most southern (Figure 1). Figure 1 Geographic reference of Massacre and Pedernales watersheds in Haiti. Source: Center for International Earth Science Information Network, CIESIN, 2012. This study is intended to provide baseline data on the socio-­‐economic conditions of the Massacre and Pedernales watersheds as well as past land cover history of these two areas where reforestation efforts have been started through the Frontera Verde Project. We sampled 13 communities in the Massacre watershed and 7 in the Pedernales watershed (Figure 2). The hiring of brigades and planting of trees has already begun in Frontera Verde Final Report 6 the Massacre watershed and is only in the planning stages for Pedernales watershed. The reason for this discrepancy appears to be due to different management teams in each watershed with different levels of preparedness and organization on the ground. The criteria for choosing brigades for each watershed are also slightly different. The basic requirements for a Frontera Verde brigade member are: 1) be 20 to 55 years old 2) be living in the watershed 3) have a valid national identification form (CIN) 4) have a good moral conduct: “bonnes vie et moeurs” 5) show strong leadership 6) be in good physical and mental health 7) be disciplined and respectful of the rules 8) be a farmer Figure 2 Massacre (left) and Pedernales (right) watersheds. Source: Center for International Earth Science Information Network, CIESIN, 2012. Each brigade is to consist of 10 members, three of whom must be women and the head of the brigade must be a woman (capataz). The brigadiers are expected to work from 7 am until 3 pm, with an hour break for lunch. They 1) prepare the land, 2) collect and prepare manure, 3) plant the seedlings, 4) visit the plantations, 5) replace seedlings that are not growing, 6) help with nursery preparation and management, 7) generally care for the plantations and 8) help with organizing reforestation days. Due to the overwhelming interest in Massacre, the team there decided to provide the additional requirement that any brigade member must be able to volunteer a minimum of 1.25 carreau of land for the project. The ramifications of this requirement is one aspect of the project this baseline assessment would like to address, as it may be removing a critical number of needy households to this cash for work project. At the same time, this is seen as a useful way of guaranteeing areas of land to be reforested as Frontera Verde Final Report 7 well as providing an incentive for brigade members to be invested in the survival of planted seedlings. Conversely, in the Pedernales watershed, the target area for reforestation is within a 200-­‐meter band along the Pedernales River, where practicable. For this watershed, few if any Haitians can claim to own any land along this boundary and therefore Frontera Verde managers are concerned there may be little incentive for brigade members to ensure the survival of planted seedlings. Finally, the project was suffering from a lack of funds as PNUD was awaiting the replacement of the previous Haitian minister. The new minister has agreed to replace the previous memorandum of understanding, whereby the Haitian Ministry of the Environment will no longer have direct access to the bank account of the project. Main Aims of Study In order to provide useful background information for both the current reforestation intervention and future planned interventions in these two watersheds, this study intends to shed light on the main sources of income for households, the level of food insecurity, the seasonal nature of vulnerability and how each of these interacts with household size and relative wealth. We selected five sub-­‐groups by main income strategy to target for more specific data collection. These included: agriculturalists, fishermen, Frontera Verde brigade members, landowners for the reforestation effort and charcoal producers or transporters. After the field campaign, we realized that the brigade members and landowners were often the same people and therefore those two groups have been combined. Fishermen and charcoal producers were targeted as they were seen as two groups that may have competing uses for natural resources. In the case of charcoal being produced from mangroves or even charcoal production contributing to deforestation of hillsides and the resultant soil erosion, these activities negatively impact the fisheries closest to the Haitian coasts. Much of the data on these groups was collected using qualitative survey methods. Agriculturalists were targeted as this is the most common source of income for Haitians and a requirement to be a brigade member; however, it is very dependent on climatic and environmental conditions and therefore can contribute to a household’s vulnerability should harvests be poor or during the seasons when it is not the most productive. In order to be interviewed as any of these target groups a household had to either rank the income strategy as the most important for them or state it contributed more than 50% of their total earnings. This became problematic for collecting data on charcoal production, as few households listed this as either their primary source of income or a significant proportion of their overall income. Therefore an exception was made for this category and if a household cited this as a source of income, they were asked the qualitative survey questions. Out of a total of 224 household surveys performed, we collected 127 surveys on agriculture, 55 surveys on charcoal production (though 63 households identified themselves as charcoal producers), 9 surveys on fishing, 6 surveys of brigade members/landowners and 2 surveys of charcoal transporters. There were additional household surveys where members cited one of these strategies as a contribution of their income, but we did not ask them targeted survey questions on this income strategy. One oversight was not to ask specific questions about commerce as this was a significant source of income for several Frontera Verde Final Report 8 households and significantly more common in Pedernales watershed, where agricultural prevalence was much lower. Massacre Watershed In the previously submitted interim report we gave more detailed information on each of the communities chosen, so we will only briefly describe how and why we chose each community sampled and how we grouped them for subsequent analysis. For Massacre, we sampled 13 communities, 12 within the watershed boundary and a 13th along the coast, Jackzil, for additional fishermen sampling (Table 1). Distance of each community from roads and rivers was calculated, the median value found and then each community was assigned a rank relative to the median: 1 if less than the median for distance from roads and rivers; 2 if either less than the median distance from roads or rivers and 3 if greater than the median distance to both. A distance from the border rank was calculated the same way; however, only two ranks were calculated: 1 if less than the median value and 2 if greater than the median distance. The number of surveys performed was based on the relative size of each community. In the smallest communities only 8 households were sampled, while 16 were sampled in the largest. Small to medium communities had 10 households surveyed and in medium to large communities 14 surveys were performed. Each household was randomly selected, except where specific income groups were being targeted (e.g. to be sure to have fishermen or brigade members sampled). Table 1 Breakdown of communities sampled in Massacre Watershed. Community Chevalier Lamine Welch Acajou Savane au Lait Savane longue Merand Meiac Acul des pins La Fleur Manket Dumas Jackzil Reason chosen FV Brigade FV Brigade FV Brigade Randomly Selected High incidence of Charcoal Production High incidence of Charcoal Production Remote, Rice region Coastal community Far from border Far from border Will have FV Brigade Far from border Coastal community No of Surveys 14 10 14 8 10 10 14 14 16 8 16 10 8 Rank of Distance from Roads & Rivers 2 3 1 1 Rank of Distance from Border 1 2 1 1 3 2 2 2 3 2 2 3 1 2 2 1 1 2 2 1 2 2 Frontera Verde Final Report 9 Pedernales Watershed In the interim report we gave more detailed information on each of the communities chosen, so we will only briefly describe how and why we chose each community sampled and how we grouped them for subsequent analysis. We sampled 7 communities in Pedernales, divided into three eco-­‐zones. The criteria were different for this watershed as, unlike Massacre, there were distinctly different ecological conditions along the border and it is a much more sparsely populated watershed. The three distinct eco-­‐zones were mountainous (characterized by coffee production), mid-­‐altitude (with significant charcoal and agricultural production) and dry coastal (dominated by commerce with the Dominican Republic and some fishing). The breakdown of community, eco-­‐zone and number of surveys in each is presented in Table 2. Due to little data on relative size of each community, the same number of surveys was performed for each. Table 2 Breakdown of communities sampled in the Pedernales Watershed. Community Saline Zone de refugies Zone d’irrigation Bony Banane Garreau Plateau de Cedre Eco-­‐zone Dry coastal Dry coastal Dry coastal Mid-­‐altitude Mid-­‐altitude Mountainous Mountainous No of surveys 10 10 10 10 10 10 10 Methods This study involved a two-­‐pronged approach, which involved collecting socio-­‐economic data from households and performing land cover analysis from satellite imagery. We will offer some analysis as to how these two factors interact, although additional environmental sampling will be necessary to make firmer conclusions. Socio-­‐economic data The majority of the field campaign was spent collecting household survey data. The general survey consisted of over 300 questions covering information on: household size; geographical location of the household; whether a household was paying school fees; what were the main sources of household income and ranking their importance; the household’s level of food insecurity; free-­‐listing and ranking of a households main concerns/problems; the household’s main source of energy; distance to and main sources of the household’s water for drinking and cooking; listing of the household’s assets for grouping into wealth quartiles; listing of the household’s livestock if applicable; the household’s access to and ownership of land; any conflicts the household may have with Dominicans and finally specific questions on the household’s main income strategy if it was one of the targeted groups. From these data a number of Frontera Verde Final Report 10 indices were generated for subsequent analysis, which are further described in the Appendix. Land-­‐cover mapping Three time periods were analyzed for this study, namely 1986, 2001 and 2010 using Landsat 5 or TM (Thematic Mapper) data. For land cover classification, an unsupervised k-­‐means algorithm was used, which began with the identification of 30 classes. The main classes used were based on CNIGS’s classification system, which in total consists of: urban areas; agricultural areas; semi-­‐natural areas; natural (forested) areas; non-­‐
vegetated (bare soil) areas and water surfaces. In some cases, the classification system of CNIGS was too specific to be identified using Landsat’s 30 meter resolution imagery. Therefore, for the Massacre watershed land cover map we used the classes: water, forest, agriculture-­‐less dense, agriculture, semi-­‐natural and bare soil. For Pedernales the classes were kept, from the previous report, as: forest, agriculture and pasture (equivalent to semi-­‐natural). Using higher resolution aerial photographs from 2010, the classification results were tested for classification accuracy. The same unsupervised technique was utilized for the 1986 and 2001 mosaics; however, without concurrent aerial photographs it was more difficult to reliably assess the accuracy of each classification. Change detection was then performed to show areas of deforestation and forest gain between each time period. Results Socio-­‐economic data We will begin by presenting basic demographic data across both watersheds and, where applicable, across communities. Household size The size of the average household1 in both watersheds is slightly over 7, with no significant variations across sub-­‐watersheds (Figure 3). However, there are strong disparities by community, the household size varying from 5.4 in Acajous to 10 in La Fleur (Figure 4). 1 The household size is the number of total household members, including parents, children, and grandparents. Frontera Verde Final Report 11 &'()*+$,-$.,'/*.,01$(*()*+/$
2,'/*.,01$/34*$)5$678*+/.*1$9:;##<=$$
+"!#
!"#$
!"#$
!"%$
,-./0#
1/22/345#
657548/052#
*"!#
)"!#
("!#
'"!#
&"!#
%"!#
$"!#
!"!#
Figure 3 Average household size across the two watersheds. (F=0.03, Pr=0.8614) .("+%*(,-'+/0%'$1'2(##"3/41''
!"#$%&'()'*("+%*(,-'#%#$%&+'
(%"'$
(*"'$
('"'$
('"'$
)"'$
&"'$
#"'$
!"#$
!"%$
&"'$
&")$
#"%$
,"'$
)"*$
#")$
&"+$
&"#$
#"%$
&")$
&")$
#"!$
#"%$
#",$
#"+$
&"'$
%"'$
*"'$
01
$: .8<0
L<A
1
01 A<G. $
M5
$:
1$
A1 0$
N6
G<1
O. 7$
0.
01
;8
.D
$
1.
O
6$
5
:1 0 P
$
$E
1:
A1
Q. 7$
AA1
.6
$
K5
K5
-.
/.
0
23 1$
23
.4
5
68
9 : 6 7$
17
9;
<0
7
=6 $
>
.7
$
?.
$@ 8
16
?. A$
>
<0
B 1$
.0
C1
D$
B
1<
B .3$
1A
.0
EF :1$
-.
1/
/.
.
01 8<1A
$? 5 $
0G
61
H $
18
I. 3F$
3J
61
7< 8
$
'"'$
Figure 4 Distribution of household size across all communities sampled (F=1.50, Pr=0.0889) Education costs In both watersheds, almost 9 out of 10 households have to pay for schooling as well as for additional school cost (Figure 5). This was considered an important factor to ask, as often the need to pay for school fees is highlighted as a reason for producing charcoal or selling livestock. As paying school fees was an extremely common practice, it is unclear whether we will be able to discern the effect of this household expense with charcoal production. Frontera Verde Final Report 12 /(+,"*(-.,'0123%4')(#',$*((-3%4'
5%67789'
!!#'%&
+""#"$%
"(#(%&
!"#$"%&'()'*(+,"*(-.,'
!"#$%&
'*#(%&
*(#(%&
)*#(%&
(#(%&
+,-./&
0.11.234&
546437./41&
!"#"$%
!"#&$%
''#&$%
,-./0%
1/22/345%
657548/052%
*)#"$%
)"#"$%
()#"$%
"#"$%
Figure 5 Proportion of households paying for education (Pearson chi2(1) =0.0872, Pr=0.768) and having to pay for additional education costs, such as books and uniforms (Pearson chi2(1) =0.2140, Pr=0.644) by watershed. In villages selected for the Frontera Verde project in the Massacre watershed, the proportion of households paying for schooling and additional school costs is significantly higher than in non-­‐intervention villages (98% relative to 83%). In the Pedernales watershed, households in the mountainous area are significantly less likely to have to pay for schooling and additional school cost (75% compared to 97% and 95% for the coastal and mid-­‐altitude eco-­‐zones, respectively). This is due to an NGO intervention. As expected, the larger the household, the more likely it has to pay for schooling. While less than 70 percent of households with fewer than 4 persons pay for schooling, more than 90 percent of them have to pay for school when they contain at least 6 people (Figure 6). Moreover, the larger the size of the household, the more likely participants are to spontaneously mention education as an area concern (results not shown). Household size and education concern (n=229)
Mention education/education cost as one of the main household concern
Linear(Mention education/education cost as one of the main household concern)
80.0%
66.2%
70.0%
60.4%
60.0%
50.0%
42.5%
40.0%
30.0%
21.1%
20.0%
10.0%
0.0%
Less than 4
5-6 people
7-8 people
9 and more
Number of household members
Figure 6 Household size and whether the household mentioned paying for school fees as a main concern (Pearson chi2(3)=23.4581, Pr<0.001) Percentage of households
!"#$"%&'()'*(+,"*(-.,'
$((#(%&
/(+,"*(-.,'0123%4')(#'1..35(%1-',$*((-'
$(,&,'6%7889:'
Frontera Verde Final Report 13 Main source of energy Primary energy for cooking Both the Massacre and Perdernales watersheds rely heavily on biomass, the primary source of energy used for cooking being firewood and charcoal, while mixed sources of energy and LP gas are essentially non-­‐existent. However, there are differences between the two watersheds regarding energy use. Firewood, the most basic and inexpensive supply of energy available, accounts for 83 percent of the energy used for cooking in the Massacre area, compared to 71 percent in Pedernales (Figure 7). While respondents from the Massacre watershed almost exclusively gather firewood from the field, 16 percent of the people interviewed in Pedernales buy firewood from the market. This is a significant difference between the watersheds and it is unclear whether this is due to different income levels, less access to land for gathering firewood in Pedernales or due to another reason. Primary energy for cooking
meals - Massacre (n=159)
Mix
1%
Primary energy for cooking
meals - Pedernales (n=70)
LP gas
1%
Charcoal
16%
Firewood
Charcoal
27%
Firewood
Charcoal
Charcoal
LP gas
LP gas
Mix
Firewood
83%
Firewood
71%
Mix
Figure 7 Primary source of cooking fuel for each watershed (Pearson chi2(3)=6.7707, Pr =0.080) Frontera Verde Final Report 14 Figure 8 Use of firewood density maps. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Figure 8 shows where the concentrations of firewood users in both watersheds are. As it can be observed, the great majority of households sampled depend on firewood as the primary source of cooking fuel, without any spatial pattern associated to it. On the other hand, charcoal appears to be consumed less, and at specific localities from both watersheds. In the case of the Massacre, the area where the two fishing communities sampled appear to have the largest concentration of charcoal consumers at the household level (Figure 9). Interestingly, more than half of the localities sampled in Massacre report no charcoal consumption at all. Frontera Verde Final Report 15 Figure 9 Charcoal use density map. (Left) Massacre watershed (right) Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Similar to Massacre, charcoal consumers in Pedernales are also concentrated around fishing communities, and close to localities where households reported charcoal production as one of the family’s sources of income (see Figure 69—income from charcoal). In addition, semi-­‐urban areas like Anse-­‐à-­‐Pitre also show high concentration of charcoal consumers. In addition, respondents from the Pedernales watershed were more likely to cross the border in order to fetch their fuel, with 33 percent reporting crossing the border while only 10 percent did so in Massacre. Primary energy for lighting There is no electricity grid to speak of in either watershed, with the primary energy used for lighting being kerosene. In the Northeast, kerosene is preferred by over 90 percent of the respondents. In the Southeast, the use of alternative sources of energy like candles, solar panels or flash lamps, appears to be more widespread (Figure 10). Frontera Verde Final Report 16 !#/01#2'"%"#32')(#'-/3*4%3'1$#(,,'51&"#,*".,'6%7889:'
G4114=/.&
!"#$"%&'()'*(+,"*(-.,'
'""#"%&
(,#"%&
H.5./246.1&
!"#$%&
('#)%&
,"#"%&
+,#"%&
')#*%&
$#!%&
)#*%&
'#!%&
)#*%&
"#$%&
'#)%&
'#)%&
+#!%&
"#"%&
-./01.2.&
34256.1&
7064/&842.6&
9641:&64;8&
<6.=>/?=?>@& A.=:4/B.4C6.&
C4D./?.&
E/@&=.66&
C4D./?.1&
F02.&
Figure 10 Primary energy for lighting across both watersheds (Pearson chi2(7)=22.38, Pr=0.002 ) In the Perdernales watershed, sources of energy for lighting differ strongly across eco-­‐
zones. Kerosene is used by more than 80 percent of the households surveyed in the coastal and middle altitude area, which is certainly related to the proximity and the relatively easy access to the Pedernales market. On the other hand, only 45 percent of the households surveyed in the mountainous areas use kerosene for lighting, while 35 percent use candles and 10 percent use flash lamps. Limited access to kerosene might be partly related the poor roads that make supplying of kerosene more difficult and costly than in other areas. Water sources Main source of drinking water Sources of drinking water diverge across watersheds (Figure 11). The prevalence and accessibility of public taps or tube wells in most Massacre communities might explain the higher proportion of respondents drinking ground water in the Northeast. When tap stands are not available or too far away from households, families obtain their water from rivers, even though they generally understand the higher risk of contamination. The low quality of the water (as in Meiac, for instance) can also explain why some households prefer obtaining water from alternative sources. On the other hand, the proximity and apparent higher quality of the water in the Southeast leads to higher surface water consumption in the Pedernales watershed. The technical director of the Frontera Verde project in Pedernales noted that the source of the river is in the Dominican Republic, where deforestation is minimal, and perhaps that is why the quality of water was significantly better than in the Northeast. Bottled water does not appear to be a viable alternative in either watershed. Frontera Verde Final Report 17 1'2%).*-#$")*+)0#2%32%(45'&"#)'$#*..)5'&"#.,"0.)
6%7889:)
!"#$"%&'(")*+),*-.",*/0.)
A4BB49.6&
C626.14>6B&
",#,%&
),#,%&
!"#$%&
!!#"%&
!,#,%&
+,#,%&
'+#'%&
'(#"%&
',#,%&
(,#,%&
*,#,%&
)#'%& !#"%&
'#*%& +#'%&
:4;1&3456.&
</=>62?541@&
3456.&&
,#,%&
-./012&3456.&
70.8496&3456.&
Figure 11 Main source of drinking by watershed, (Pearson chi2(6)= 11.8156, Pr=0.066) In Massacre, respondents living in the communities that have not been targeted for the Frontera Verde program are significantly more likely to drink ground water, such as from a dug tubewell. While people living in targeted communities show a higher incidence of drinking surface water (Figure 12). In the Pedernales watershed, different eco-­‐zones exhibit significant variation in water drinking habits. The dry and relatively arid climate, found in the coastal areas, combined with a better distribution system (public taps, tubewells), has greater ground water consumption. On the other hand, the apparent higher quality and proximity of fresh surface water might explain why the respondents of the middle and mountainous areas use only surface or rainwater for drinking purposes (Figure 13). Main source of drinking-water - Brigades
vs. Non brigades - Massacre
No brigades
Percentage of households
80%
Brigades
73%
70%
60%
60%
50%
40%
33%
30%
19%
20%
7%
10%
5%
0%
Ground water
Surface water
Rain water
0%
2%
Bottled/tank water
Figure 12 Mains source of drinking water in Massacre divided by communities currently with brigades and those without (Pearson chi2(3) = 28.5820 Pr = 0.000). Frontera Verde Final Report 18 Main source of drinking-water across
eco-zones - Pedernales (n=70)
Coastal area
Middle altitude area
Mountainous area
Percentage of households
120%
100%
90%
100%
80%
80%
60%
40%
20%
0%
0% 0%
Ground water
7%
3%
Surface water
10%
10%
0%
0% 0%
Rain water
Bottled/tank water
Figure 13 Main source of drinking water for communities in Pedernales by eco-­‐zone (Pearson chi2(6)=63.8675, Pr<0.001). Main source of water for cooking/bathing When it comes to water use for cooking or bathing, the proportions are reversed, with more than 65 percent of people from Massacre using surface water, while a slim majority of respondents use ground water in Pedernales (Figure 14). 1'2%).*-#$")*+)3'&"#)+*#)$**42%(5)6'&,2%()'$#*..)
3'&"#.,"0.)7%899:;)
A4BB49.6&
!"#$"%&'(")*+),*-.",*/0.)
C626.14>6B&
$'#(%&
(*#*%&
$*#*%&
'(#)%&
'!#"%&
'*#*%&
,*#*%&
+*#*%&
!"#$%&
++#+%&
!*#*%&
))#+%&
)*#*%&
*#*%&
*#*%&
-./012&3456.&
70.8496&3456.&
:4;1&3456.&
!#'%&
</=>62?541@&3456.&&
Figure 14 Main source for water used for cooking or bathing across watersheds (Pearson chi2(3)=12.0161, Pr=0.007). Frontera Verde Final Report 19 Energy and water insecurity Energy vulnerability Approximately seven in ten people from both watersheds have experienced days with a lack of energy to meet their family's needs in the past 12 months. Sixty-­‐seven percent of households in Massacre reported having periods of insufficient energy while 71 percent of households in Pedernales reported the same. The level of vulnerability is constant across the two watersheds, but seems to be partly predicted by the proximity of main transport infrastructure or by eco-­‐zone (Figure 15). In both the Massacre and Pedernales watersheds, the more remote the area, the more likely people are to suffer from lack of energy. Only 55 percent of the households in the close area have had insufficient energy to meet their family’s needs, compared to 79 percent in the remote area. Similarly, while 60 percent of households have experienced days with insufficient energy in the Pedernales watershed, this proportion rises up to 85 percent in the mountainous eco-­‐zone. As a result, energy vulnerability appears to be strongly correlated with accessibility issues. 1%.-2$3"%&)"%"#(4)'$#*..)"$*56*%".)5)
!"0"#%'/".)
1%.-2$3"%&)"%"#(4)'$*#03%()&*)5#*6373&4)
83&,)#*'0.9#3:"#.);)<'..'$#")
$(#(%&
'$#(%&
)(#(%&
'(#(%&
!"#$%&
!(#(%&
;21&
"(#(%&
,(#(%&
+(#(%&
*(#(%&
('#"$%
-"#"$%
)$#'%&
!"#$"%&'(")*+),*-.",*/0.)
!"#$"%&'(")*+),*-.",*/0.)
-(#(%&
&'#"$%
("#"$%
&"#"$%
!"#"$%
!"#"$%
'"#"$%
,"#"$%
<51%
+"#"$%
*"#"$%
)"#"$%
"#"$%
./01203%0450%
(#(%&
./012&
3425672&
8290:2&
678835%0392:85% 6/:;207;/:1%0450%
0450%
Figure 15 Whether a household suffered insufficient energy by distance measures for Massacre (Pearson chi2(2) = 5.4411, Pr=0.066) and eco-­‐zone for Pedernales (Pearson chi2(4) = 4.5561, Pr = 0.336). Frontera Verde Final Report 20 Figure 16 Energy insecurity density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Figure 16 shows where the concentrations of energy insecure households are, per km2. Results obtained from these maps are consistent with the relatively high levels of energy insecurity observed across the two watersheds. No specific spatial patterns are observed. Water vulnerability (both drinking and bathing/cooking water) About four in ten people in both watersheds have suffered from a lack of water availability for either drinking or bathing/cooking over the past 12 months. However, while this level is consistent across the Massacre sub-­‐watersheds, there is dramatic divergence in Pedernales. In fact, the mountainous area appears to be significantly more exposed to water insecurity than the two other eco-­‐zones, where only 20 to 30 percent of households reported suffering from water shortages (Figure 17). Frontera Verde Final Report 21 1"2*#3%(),'45%()5%.-6$5"%&)7'&"#))
'$#*..)"$*89*%".)8)!"0"#%'/".):%;<=>)
'(#"$%
!"#$"%&'(")*+),*-.",*/0.)
,"#"$%
'"#"$%
+"#"$%
("#"$%
*"#"$%
!"#"$%
!"#"$%
&"#"$%
&"#"$%
)"#"$%
"#"$%
-./01/2%/34/%
567724%/281974%/34/%
5.9:1/6:.90%/34/%
Figure 17 Whether a household suffered insufficient water by eco-­‐zone in Pedernales (Pearson chi2(2) = 14.7917 Pr = 0.001). Figure 18 shows where households who reported water insecurity issues are concentrated. In the case of Massacre, areas near Acajou reveal the largest density of households reporting water insecurity. In the case of Pedernales, localities within the high altitude areas appear to be the most water insecure. Further research is needed to understand if these cases are artifacts of population density2, or if water insecurity is associated with the quantity or quality of water available in the region. 2 The population density data set from MINUSTAH (2010) provides qualitative estimates only. Frontera Verde Final Report 22 Figure 18 Water insecurity density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Seasonality of energy and water insecurity As with income sources, households reported highly seasonal insecurities for both water and energy. Figure 19 shows two distinct peaks in energy shortage, one in April -­‐ June, another in September -­‐ November. These peaks are also consistent across both watersheds. Frontera Verde Final Report 23 1*%&,.)23&,)3%.-4$3"%&)"%"#(5)'$#*..)2'&"#.,"0.)
?<:*5$@+A"%BC$
/*DD*90-$@+A"(BC$
E-F-0+*5-D$@+A%(C$
'%#$
!"#$"%&'(")*+),*-.",*/0.)
'(#$
'!#$
6%5)7*%&,.8)
/*DD*90-G$'!#$
E-F-0+*5-DG$%(#$$
!%#$
!"#$
!"#$
!(#$
"%#$
"(#$
&#$
%#$
%#$
(#$
)*+$
,-.$
/*0$
120$
/*3$
)4+$
)45$
146$
7-2$
89:$
;<=$
>-9$
Figure 19 Percent of total household stating which months they had insufficient energy. On the other hand, the lack of water corresponds to the dry season, from December to April (Figure 20). People interviewed in the Pedernales watershed are significantly more likely to report insufficient water access from January to March than in Massacre, suggesting a stronger impact of the climate variability and the dry season in the Southeast. 1*%&,.)23&,)3%.-4$3"%&)2'&"#))
'$#*..)2'&"#.,"0.)
@=;+6#A,BCDE#
!"#$"%&'(")*+),*-.",*/0.)
)'"#
0+FF+:1.#A,BG'E#
H.1I.1,+6.F#A,B$DE#
&%"#
5%6)7*%&,.8)
0+FF+:1.J#(!"#
H.I.1,+6.FJ#G"##
&'"#
%'"#
!!"#
!'"#
$%"#
$'"#
('"#
'"#
*+,#
-./#
0+1#
231#
0+4#
*5,#
*56#
257#
8.3#
9:;#
<=>#
?.:#
Figure 20 Percentage of households by watershed claiming they suffered from insufficient water. Frontera Verde Final Report 24 Food insecurity Food insecurity score !""#$%&'()*+,-.$/)"+($0123$4$56789$
Both watersheds suffer from a state of chronic food insecurity. We generated a household food insecurity score from a series of questions regarding their ability to access food over the previous month. A detailed description of our methodology is available in Appendix 1-­‐2. The food insecurity scores ranges from 6.9 for Pedernales to 8.0 for Massacre (Figure 21), with an average 94 to 98 percent of the respondents being moderately to severely food insecure. :;(+2<($=(;(=$">$>""#$,&'()*+,-.$>"+$(2)?$
@2-(+'?(#$$
,"##%
!"##%
'"##%
&"##%
+"##%
!"#$%
$"##%
&"!'%
*"##%
)"##%
("##%
#"##%
-.//.012%
324215.62/%
Figure 21 Difference in food security scores between watersheds. The result appears to be significant; however, the sampling in both watersheds was not statistically representative (t = -­‐2.7492, Pr =0.0065). Food insecurity status The calculation of the different levels of food insecurity is described in Appendix 1-­‐1. Figure 22 shows the breakdown of households into each category. Very few households were considered moderately food insecure, making differentiation with this parameter quite difficult. Frontera Verde Final Report 25 Massacre (n=159)
Pedernales (n=70)
2%
5.71
19%
32.86
61.43
79%
Food Secure
Moderately Food Insecure
Severely Food Insecure
Food Secure
Moderately Food Insecure
Severely Food Insecure
!"#$"%&'(")*+),*-.",*/0.)
Figure 22 Breakdown of levels of food insecurity by watershed (Pearson chi2(2) = 7.4902, Pr = 0.024). In the Pedernales area, there appears to be significant differences in terms of food security across eco-­‐zones, with a higher prevalence of people severely food insecure in the middle altitude and mountainous areas (Figure 23). 1**0)."$-#2&3)/"4"/.)'$#*..)"$*56*%".)5)
!"0"#%'/".)7%89:;)
;..7$<4=934$
5.743/142>$;..7$?:04=934$
<4@4342>$;..7$?:04=934$
,!"!#$
)&"!#$
)!"!#$
(&"!#$
(!"!#$
&!"!#$ &!"!#$
&!"!#$
+!"!#$
*!"!#$
%&"!#$
%!"!#$
%!"!#$
'&"!#$
'!"!#$
!"!#$
!"!#$
!"!#$
-./01/2$/34/$
567724$/281974$/34/$
5.9:1/6:.90$/34/$
Figure 23 Breakdown of food insecurity levels by eco-­‐zone in Pedernales (Pearson chi2(4) = 16.2791 Pr = 0.003). Figure 24 shows where the concentration of food insecure households in both watersheds is, by km2. Although some minor spatial differences can be observed among areas on each watershed, in general, all areas present relatively high concentrations of food insecure households, regardless of their geographic position. Both datasets are Frontera Verde Final Report 26 depicted over agriculture land use density layers, derived from LandSat 2010. The visual comparison between concentrations of land use and food insecurity is inconclusive for both cases. The two watersheds present cases with relative high concentrations of food insecurity in relation to high concentrations of agricultural land use. Further research is needed to understand household economic dynamics, and food consumption expenditures. Figure 24 Food insecurity density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Seasonal food insecurity Levels of food security vary throughout the year, with two peaks of food insecurity for each watershed. The first one occurs at the end of the dry season in March – April in the Massacre area and slightly later, in May – June, in the Pedernales watershed. The second food insecurity peak is common to the two watersheds and takes place in October and November, with a stronger impact in the Massacre watershed (Figure 25). It is possible that the later peak in food insecurity is due to households having to pay school fees and, therefore, does not exhibit a seasonal difference between watersheds. Frontera Verde Final Report 27 1*%&,.)23&,)/'$4)*+)+**0)
'$#*..)2'&"#.,"0.)
A><,7$B-C!!"D$
1,EE,;2/$B-C*)FD$
G/H/2-,7/E$B-CIID$
)(#$
!"#$"%&'(")*+),*-.",*/0.)
")#$
"(#$
"&#$
5%6)7*%&,.8)
1,EE,;2/J$*!#$
G/H/2-,7/EJ$!"#$$
%!#$
%)#$
%!#$
!'#$
%(#$
!"#$
!)#$
!(#$
*)#$
*(#$
)#$
(#$
+,-$
./0$
1,2$
342$
1,5$
+6-$
+67$
368$
9/4$
:;<$
=>?$
@/;$
Figure 25 Percentage of households reporting insufficient food during specific months of the year. In the Massacre area, communities that have been chosen for the Frontera Verde project and will eventually have brigades are more likely to report a lack of food during the food insecurity peaks. In Pedernales, the mountainous area in particular is severely affected by the lack of food in May and June (Figure 26). )1*%&,.)23&,)/'$4)*+)+**0)5)6#3('0".))
>?$13B9-C0D$E.F(&(G$
+&#$
A>,B<,7$,2/,$C-D!EF$
'!#$
7%8)9*%&,.:)
>?$13B9-C0DI$(*#$
H3B9-C0DI$+#$$
()#$
*&#$
"&#$
"&#$
1GHH7/$,7I<6H/$,2/,$C-D"*F$
*)#$
!"#$"%&'(")*+),*-.",*/0.)
!"#$"%&'(")*+),*-.",*/0.)
'&#$
1*%&,.)23&,)/'$4)*+)+**0)
'$#*..)"$*56*%".)5)!"0"#%'/".)
H3B9-C0D$E.F*+G$
!%#$
!"#$
!&#$
)&#$
(&#$
')#$
1>-<,G->6B$,2/,$C-D!)F$
()#$
7%8)9*%&,.:)
A>,B<,7$,2/,J$!&#$$
1GHH7/$,7I<6H/J$!&#$$
1>-<,G->6BJ$!'#$
%'#$
&)#$
%"#$
%)#$
!&#$
!"#$
!)#$
!)#$
")#$
&#$
,-.$
/01$
2-3$
453$
2-6$
,7.$
,78$
479$
:05$
;<=$
>?@$
A0<$
)#$
+,-$
./0$
1,2$
342$
1,5$
+6-$
+67$
368$
9/4$
Figure 26 (left) Percentage of households in Massacre communities that currently have a Frontera Verde brigade and those that do not, who listed specific months they did not have sufficient food. (right) Percentage of households in each eco-­‐zone in Pedernales that listed specific months when they did not have sufficient food. Frontera Verde Final Report 28 :;<$
=>?$
@/;$
Food insecurity and household size While household size does not significantly predict food insecurity levels in the Massacre area, family size seems to be a slightly stronger influence of food insecurity in the Pedernales area (Figure 27), with households under 6 people being less likely to be affected by food insecurity issues. !""#$*7%&'()*+,$/7#$8"(%&8"9#$%*:&$;$<&#&)7/9&%$-71=56$
;668%<2.-=>:</?%.=6:-%
,<2-1:%@;668%<2.-=>:</?%.=6:-A%
!""#$%&'()*+,$%'")&$-./0$1$23456$
+'"''%
*"''%
)"''%
#"!(%
#"))%
#"''%
&"''%
!"#$%
&"''%
!"''%
,-..%/012%3%
!4&%5-657-%
#4)%5-657-%
*%128%96:-%
Figure 27 Relationship between food insecurity and household size for Pedernales (F=2.12; Pr= 0.1066). Characteristics of food insecure households The regression coefficients reported in Table 3 measure the relationship between the food insecurity scores (on a scale from 0 to 15) and four categories of independent variables (livelihood strategies, demographics/location, land ownership and wealth quartile). The coefficients offer the level of influence and the nature of these possible determinants on levels of food insecurity, with model 4 (full model) providing the effects of all four categories of factors.3 Commerce seems to be the livelihood strategy least likely to lead to food insecurity (p<0.01). Net of other factors, engaging in commerce reduces food insecurity scores by 1 point. Farming (defined as generating at least 40 percent of the total household income) and charcoal production appear to have a strong positive 3 The Variance Inflation Factor (Mean VIF=1.39) was calculated and no apparent multicollinearity
issues were detected in the model. Taken together, our variables explain about 21% of the variance in
food insecurity level (R2 =0.213).
Frontera Verde Final Report 29 relationship with food insecurity scores in the first three models; however, there become statistically insignificant when we control for wealth quartiles. In terms of demographics and watershed, being from the northeast and having a higher number of children relate positively with higher food insecurity scores (p<0.01), while gender seems to be independent from food insecurity levels. Similarly, the size of the land households have access to do not relate with food insecurity. Membership to wealth quartiles also affects a household’s food insecurity scores. Belonging to the richer half of the population (the two highest wealth quartiles, in terms of value of assets possessed) is negatively related to food insecurity (p<0.05). Likewise, there is a positive relationship between spontaneously mentioning a lack of money as an area of concern and the likelihood of being food insecure (p<0.01). This is relevant to the following section as buying food is the most widespread strategy (74 percent of households) for coping with food deficiency. Therefore, the less well-­‐off households do not have adequate financial resources to buy food when they need too and hence suffer greater food insecurity. In sum, Haitians from the Massacre watershed, not engaged in commerce, with a larger number of children, belonging to the two lower wealth quartiles, are the most likely to score high on the food insecurity scale. Frontera Verde Final Report 30 Table 3 Multiple regression analysis of the influence of household characteristics on a household’s food insecurity score. Impact of sources of livelihood, demographics and location, land ownership, and wealth level on
Food insecurity scores (linear regression)
Model 1
Model 2
Model 3
Model 4
0.940**
0.852**
0.783*
0.603
(0.43)
(0.42)
(0.44)
(0.43)
0.257
0.169
0.151
0.114
(0.43)
(0.43)
(0.44)
(0.43)
0.155
0.134
0.073
-0.009
(0.58)
(0.57)
(0.58)
(0.58)
0.495
0.355
0.343
0.135
(0.74)
(0.74)
(0.74)
(0.73)
-1.138**
-1.247***
-1.260***
-1.012**
(0.44)
(0.43)
(0.44)
(0.43)
0.957**
0.838*
0.820*
0.504
(0.45)
(0.45)
(0.45)
(0.45)
-0.372
-0.625
-0.587
-0.639
(0.65)
(0.64)
(0.65)
(0.63)
0.053
0.050
-0.077
(0.41)
(0.42)
(0.41)
0.172**
0.177**
0.262***
(0.09)
(0.09)
(0.09)
1.259***
1.293***
1.251***
(0.45)
(0.46)
(0.45)
-0.176
-0.144
(0.58)
(0.59)
0.325
0.331
(0.63)
(0.62)
-0.162
-0.152
(0.60)
(0.59)
Livelihood strategies
Farming (>40%)
Livestock owner
Tree planting
Fishing
Commerce
Charcoal production
Salaried job
Demographics / Location
Female
Nb of children
Northeast region
Land ownership
Reference category: Less 0.5 kawo
0.5 to 1 kawo
1 to 2 kawo
More than 2 kawo
Wealth level
Reference category: Lowest wealth quartile
Second wealth quartile
-0.605
(0.56)
Third wealth quartile
-1.197**
(0.59)
Highest wealth quartile
-1.579**
(0.61)
Lack of money
1.064***
(0.41)
Constant
7.370***
(0.56)
6.102***
(0.70)
Observations
210
210
R-squared
0.099
0.15
Notes:
Dependent variable: Food insecurity scores (0 to 15 scale)
*** p<0.01, ** p<0.05, * p<0.1
6.152***
(0.76)
6.304***
(0.87)
210
0.154
210
0.213
Strategies to cope with food insecurity In order to cope with food insecurity, most households declare purchasing food as their main strategy, which might be indicating it is cheaper for these households to purchase outside food than produce it themselves. Selling assets, reducing the number of meals or the size of the meals are alternative ways to meet the food needs, especially in the Pedernales watershed. A marginal proportion of households did not cite any strategy for dealing with their lack of food access (Figure 28). Frontera Verde Final Report 31 1&#'&"(2".)&*)$*3")42&,)&,")/'$5)*+)+**0)'$#*..)4'&"#.,"0.)
K23320/4$L;M%"(N$
(+#$
-474/;2@43$L;M,,N$
!"#$
!%#$
!"#$"%&'(")*+),*-.",*/0.)
!+#$
,+#$
"+#$
)(#$
&+#$
)'#$
'(#$
)+#$
')#$
'+#$
(#$
'%#$
%"#$
%&#$
%+#$
+#$
&%#$
*#$
&#$
-./01234$ 83393:2;04$ <667$56/$
5667$
=6/>$
(#$
%)#$
"#$
%#$+#$
)#$
+#$
%#$'#$
,#$
&#$
?4@@9;A$ B47.04$ B47.04$ F231$56/$ <667$6;$ -/67.04$ G34$6=;$ H:14/$
J6$
2334:3$ ;.CD4/$65$ 39E4$C42@3$ =6/>$
0/479:$ 012/062@$ A2/74;$ 3:/2:4AI$ 3:/2:4AI$
C42@3$
Figure 28 Strategies households utilized to cope with lack of food by watershed. Food insecurity and education / food insecurity and livestock No significant relations have been found between food insecurity and paying for schooling or food insecurity and livestock ownership. Health issues We did not specifically ask households information on the main diseases they or their family members suffer from. We collected some data on disease prevalence from the head of household’s free listing of their main household concerns and only in some cases did enumerators ask for the specific diseases they were suffering from. Therefore, these data we present are only a subset of the diseases these households may be experiencing; however, we felt they were still important to include. Frontera Verde Final Report 32 More than one out of four respondents spontaneously mentioned fever when speaking about health issues. Flu (14%), headache and cholera (12%), and diarrhea (10%) were other health concern regularly mentioned (Figure 29). We did not ask the source of these diseases, nor what they specifically meant by fever. 1'2%),"'/&,)2..-".)#"3*#&"0)'4*%()3"*3/")4"%5*%2%()6'0)
,"'/&,)7%89:;<)
J>K>-#
!'"#
J.4#
(*"#
E>,6,9:>#
(!"#
I:3.>-,#
(!"#
<1,--:3>,#
()"#
E1;:#=.336#A->004->#
'"#
H83C,9:#A,15#
&"#
B:>4C,G0C0#
%"#
D>>8:#
$"#
F,.,-1,#
$"#
E>,-8#A-3=.>C#
$"#
D:/A:316#
!"#
B>0A1-,83-/#A-3=.>C#
!"#
?->,08@9:>08#A,15#
!"#
<1,=>8>0#
!"#
789:15;#
!"#
23456#
!"#
+,-,./010#
!"#
)"#
%"#
()"#
(%"#
!)"#
!"#$"%&'(")*+),*-.",*/0.)
!%"#
$)"#
Figure 29 Main diseases mentioned by households listing bad health as a household concern. Wealth We collected data on assets, livestock and land access and ownership. We decided to derive the wealth quartiles solely from assets owned and not livestock, as the inclusion of livestock skewed the results and were not correlated with other household characteristics. The description for how we derived these wealth quartiles can be found in Appendix 1-­‐1. Our results suggest that households in the Pedernales watershed seem to be better off than households in Massacre. Fifty-­‐nine percent of the households in Pedernales fall into the two higher wealth quartiles, compared to only 46 percent in Massacre. Conversely, 26 percent of the households from Massacre belong to the poorest wealth quartile, compared to only 16 percent for Pedernales (Figure 30). Frontera Verde Final Report 33 1"'/&,)2-'#3/".)'$#*..)4'&"#.,"0.)5%67789)
&)),%
+),%
!*#+%
!(#&%
!"#$"%&'(")*+),*-.",*/0.)
-),%
(),%
!!#)%
"),%
89:;30<%=3/7<;%>?/2@73%
*&#$%
A;925%=3/7<;%>?/2@73%
'),%
B31C65%=3/7<;%>?/2@73%
!(#(%
$),%
DC=30<%=3/7<;%>?/2@73%
*),%
!'#(%
!"#$%
&'#(%
),%
./00/123%
!),%
&),%
435326/730%
Figure 30 Breakdown of households into wealth categories for each watershed. If we look at the differences across sub-­‐watersheds, there are no significant wealth differences across communities that have or will have brigades and others in Massacre. However, the proximity to rivers, roads and markets seems to partially predict how wealthy communities are. In fact, closer communities are systematically better off than the more remote ones (Figure 31). For reference for the Frontera Verde project we provide a breakdown of wealth quartiles by community sampled for Massacre and Pedernales watersheds (Figure A.1 in Appendix 1-­‐2). 1"'/&,)2-'#3/".)'$#*..)"$*45*%".)4)!"0"#%'/".)
1"'/&,)2-'#3/".)'$$*#04%()&*)5#*6474&8)9):'..'$#")
!""(#
!++,#
!*#
!"#$"%&'(")*+),*-.",*/0.)
%+,#
&'#
>86?109#@15.9?#AB54C.1#
)+,#
$+,#
!+,#
+"(#
!$#
(+,#
'+,#
,"(#
$)#
*+,#
&+,#
!"#
!*#
&'#
$(#
$%#
D?847#@15.9?#AB54C.1#
E1;/:7#@15.9?#AB54C.1#
F/@109#@15.9?#AB54C.1#
!$#
!"#
$"#
!"#$"%&'(")*+),*-.",*/0.)
"+,#
-./01#
2314561#78095:;1#
<1=/91#
%"#
)"(#
&"(#
;6<=401#>4/21=#?9/3824#
$$#
'"#
'"#
@=637#>4/21=#?9/3824#
A4B.:7#>4/21=#?9/3824#
'"(#
C.>401#>4/21=#?9/3824#
$"(#
%$#
!&#
$"#
!"(#
!"#
+,#
$&#
*"(#
%"(#
$"#
$$#
!"#
"(#
-./01/2#/34/#
567724#/281974#/34/# 5.9:1/6:.90#/34/#
Figure 31 Breakdown of households in each wealth category by distance measure in Massacre (Pearson chi2(6) = 19.7009 Pr = 0.003) and eco-­‐zone in Pedernales (Pearson chi2(6) = 10.5599 Pr = 0.103). Frontera Verde Final Report 34 If we looked at the breakdown of incidence of each asset by watershed we found that the incidence of the most valuable assets is generally higher in the Pedernales area (motorcycle, improved cook stove, cell phones) (Figure 32). Also, the high incidence of cell phone use is a promising sign for possible energy interventions involving pre-­‐paid phone cards for community-­‐level solar-­‐powered electricity projects. 1*-.",*/0)'.."&.)2"%"&#'3*%)#'&".)'$#*..)4'&"#.,"0.)5%67789)
@.44.;10#
H0<019.704#
G0164090#7.C8#
&!"#
(&"#
('"#
F077#82690#
!+"#
B169#
E.<>6#
'!"#
!!"#
$'"#
BC81630<#;66A4/630#
$)"#
-.77#;76;A#
))"#
'$"#
$%"#
@6/61;?;70#
=>;?;70#
&"#
:.80#10;61<01#
$%"#
$%"#
567.1#8.9074#
&"#
$%"#
-./01#2.1304/#
!"#
%"#
*)"#
),"#
'$"#
=.D01?#/61;2#
+!"#
'("#
$("#
&"#
$%"#
'%"#
)%"#
!%"#
,%"#
*%"#
(%"#
!"#$"%&'(")*+),*-.",*/0.)
+%"#
&%"#
$%%"#
Figure 32 Percentage of households for each watershed reporting owning various assets. Frontera Verde Final Report 35 Figure 33 Asset index density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. An asset index was built based on the data collected from assets, weighting for the number of persons in the household. The index follows Filmer and Pritchett's asset index formula: Aj= f1 x (aj1 -­‐ a1) / (s1) + f2 x (aj2 -­‐ a2) / (s2) + ……+ fN x (ajN -­‐ aN) / (sN) where: -­‐Aj is the asset for household j -­‐f represents the principal component analysis score. There are 8 components, explaining 80 percent of the joint variance of the input assets. -­‐(aj-­‐a1)/s1 is the standardization of the asset, by subtracting the mean and standard deviation (a = mean, s= st.dev). Index values were represented in quartiles and represented in Figure 31. Figure 33 shows the concentration of the asset index values for both watersheds. As it can be observed, there is no spatial pattern associated with the concentration of either high or low asset index values. Wealth distribution and food insecurity Unsurprisingly, the wealthier the household the lower their food insecurity score; however, this provided some indication that these indices may be consistent for describing household characteristics (Figure 34). Frontera Verde Final Report 36 !""#$%&'()*+%,-$')"+('$1)+"''$9(1:,.$;*1+<:('$/&3==>8$
!""#$%&'()*+%,-$').""+($/012$3$45678$
B..=$?<10;78?2C$1;.80$
)"*$
!"#$
-?<038$DB..=$?<10;78?2C$1;.80E$
!"%$
!"*$
&"'$
(")$
&"*$
("*$
#"*$
,"*$
'"*$
+"*$
%"*$
*"*$
-./012$/03425$
6738940$
:0;.<=$/03425$
6738940$
>5?8=$/03425$
6738940$
@?A5012$/03425$
6738940$
Figure 34 Food insecurity relative to wealth categories for all households. Land accessibility and ownership Overall, 97 percent of households have access to a parcel of land. Eighty-­‐five percent of people either own all of this land or at least part of it in Massacre, as compared to 81 percent in Pedernales (Figure 35). Only a minority has access to a piece of land without being at least the partial owner of the land, and there are no significant differences across sub-­‐watersheds. For the purposes of describing land as a source of wealth, we focus on land ownership for this report. Frontera Verde Final Report 37 Land ownership Massacre (n=156)
No
15%
Land ownership
Pedernales (n=67)
No
19%
Yes - All of
it
55%
Yes - Part
of it
30%
Yes - Part
of it
21%
Yes - All of
it
60%
()%*+,-%
Figure 35 Breakdown of landownership categories by watershed (Pearson chi2(2) = 2.1441 Pr = 0.342). Households in the Southeast reported slightly larger average size of accessible and owned land than in the Northeast (although these differences are not significant). On average, land ownership is about 1.28 kawo (1.65 ha) in Massacre and 1.46 kawo (1.88 ha) in Pedernales (Figure 36). ./01+20%3450%-6%7+)8%+10+%-,)08%
9:%,+;013<08%=)>!'$?%
&"!!#
!"&'%
%"$!#
!"#$%
%"!!#
!"$!#
!"!!#
'())(*+,#
-,.,+/(0,)#
Figure 36 Average size of land a household owns by watershed (N.B.: 1 kawo (carreau) = 16 tales = 1.29 hectares). There are more households in the Massacre watershed that own a small amount of land less than 0.5 kawo (0.65 ha), while 2 out of 3 household in the Pedernales area have access to land of 1.29 ha and larger (Figure 37). There were slight differences in land owned by communities with FV brigades and without, though not significant, as well as across eco-­‐zones in Pedernales, which was slightly more significant (Figure 38). Frontera Verde Final Report 38 123")*+)&,")/'%0)*4%"0)*5"#)&,")6'.&)7"'#)8%9:;<=)
1/66/723&
!"#$"%&'(")*+),*-.",*/0.)
!+#"%&
839325/:36&
!"#$%&
!"#"%&
*'#(%&
*+#"%&
*"#"%&
'*#*%&
'+#"%&
*"#'%&
'(#!%&
'"#$%&
'"#"%&
)!#"%&
)'#"%&
)+#"%&
)"#"%&
+#"%&
"#"%&
"&,-&"#+&./0-&
"#+&,-&)&./0-&
)&,-&'&./0-&
1-23&,4/5&'&
./0-&
Figure 37 Percentage of households that owned different sized land parcels by watershed (Pearson chi2(3) = 5.8604 , Pr = 0.119) 123")*+)&,")/'%0)*4%"0)*5"#)&,")6'.&)7"'#)8)9'..'$#")
:%;<=>?)
!"#$"%&'(")*+),*-.",*/0.)
()")$%
!)")$%
;178.92:%
8/19.1:&1451&
!!"#$%
#!"*$%
#)")$%
#!"*$%
&'"($%
&)")$%
&!"'$%
'("&$%
')")$%
'("&$%
')"'$%
!"#$"%&'(")*+),*-.",*/0.)
5,%6178.92:%
123")*+)&,")/'%0)*4%"0)*5"#)&,")6'.&)7"'#)8)!"0"#%'/".)
9%:;<=)
,-#-%&
3;<<:5&1:=.><5&1451&
!"#$%&
!-#-%&
('#*%&
"+#$%&
(-#-%&
"-#-%&
)-#-%&
'-#-%&
3/>7.1;7/>9&1451&
""#"%&
""#"%&
)"#'%&
''#'%&
+#"%&
'!#(%&
',#*%&
',#*%&
*#*%&
-#-%&
)")$%
)%+,%)"(%-./,%
)"(%+,%'%-./,%
'%+,%&%-./,%
0,12%+3.4%&%-./,%
-&./&-#!&012/&
-#!&./&'&012/&
'&./&)&012/&
Figure 38 (left) Percentage of households owning different sized parcels of land by communities with FV brigades or without for Massacre (Pearson chi2(3) = 4.3017 Pr = 0.231). (right) Percentage of households owning different sized parcels of land by communities across eco-­‐zones in Pedernales (Pearson chi2(6) = 10.3025, Pr = 0.112) Interaction between wealth measures There are limitations to the wealth quartiles presented here, as the assets used as inputs were not weighted. Therefore, in order to test the robustness of this measure, Figures 39 and 40 present comparisons of land ownership and livestock value across wealth quartiles for both watersheds. It is interesting to note that the area of land owned differs between the two watersheds, with it steadily increasing across wealth quartiles in Massacre but the highest wealth quartile exhibits among the lowest land ownership in Frontera Verde Final Report 39 3/45&.617&)&012/&
!"#$%&#'()*#'+,'-%$.#/'+01#2'34%0+5'
Pedernales. Again, livestock ownership appears to mirror the increasing wealth quartile in Massacre, while Pedernales exhibits more of a “U” shape. Here the lowest wealth quartile apparently owns the largest value of livestock and then values increase again after this. While not definitive, these figures seem to suggest that wealth measures may differ slightly between the two watersheds, as commerce activity is more prevalent in Pedernales income from this may be overcoming the need for land ownership and/or livestock to provide adequate financial support for a household. 6%12'()*#'+01#2'%.$+(('0#%/78'9:%$;/#('31<==>5'
&"*!$
&"#'$
&"#!$
&"(!$
&"+!$
&"'($
&"'!$
&"!!$
!"%%$ !"%)$
&"!!$
!"*+$
A2002:7/$
!"*!$
!"##$
B/</7;23/0$
!"#!$
!"+!$
!"'!$
!"!!$
,-./01$./2314$ 9/:-;<$./2314$ =4>7<$./2314$ ?>@4/01$./2314$
562783/$
562783/$
562783/$
562783/$
Figure 39 Interaction between land area owned and wealth quartile for both watersheds. 2"-*)3"%0&($%-)&"4&%'()*#"+,&$+5"**&6)$%#3&
7-$58%)*&.9:;;<1&
(#$!!!"
!"#$%&%'()*#"+,&($%-)&./0*1&
(!$!!!"
'#$!!!"
'!$!!!"
&#$!!!"
>/--/74,"
&!$!!!"
?,9,48/0,-"
%#$!!!"
%!$!!!"
#$!!!"
!"
)*+,-."+,/0.1"
23/450,"
6,7*89"+,/0.1"
23/450,"
:1;49"+,/0.1"
23/450,"
<;=1,-."+,/0.1"
23/450,"
Figure 40 Interaction between livestock value and wealth quartile for both watersheds. Frontera Verde Final Report 40 Households concerns In terms of prevalence, food insecurity, lack of money, bad health and education are the most frequently identified worries. The preoccupations spontaneously mentioned first are issues of house ownership (the household does not own the house they live in and often have to pay rent), lack of credit and money, and the bad state of their house. In terms of severity, these issues are perceived to be the most severe, along with the issues of death and access to healthcare. Table 4 shows the incidence of each household concern, while Table 5 lists the household concerns in order of rank and severity assigned. Table 4 Ranking of incidence of different household concerns. Incidence of preocupation (%)
Massacre Pedernales
(n=159)
(n=70)
Total
(n=229)
Ranking of preocupation
Massacre Pedernales
Severity of preocupation (1 to 5)
Total
Massacre Pedernales
Total
Food insecurity
72%
53%
66%
8
6
5
4.3
4.3
4.3
Lack of money
56%
59%
57%
3
2
3
4.5
4.8
4.6
Bad health
62%
39%
55%
6
11
6
4.1
4.0
4.1
Education
53%
43%
50%
7
10
7
4.1
3.7
4.0
Bad housing
30%
31%
30%
4
5
4
4.1
4.3
4.2
Water insecurity
30%
14%
25%
13
12
13
3.9
4.4
3.9
No bathroom
14%
6%
12%
16
14
17
3.7
4.8
3.9
Lack of work
8%
19%
11%
10
8
10
4.2
4.5
4.3
10%
9%
10%
2
3
2
4.4
4.2
4.4
Energy
9%
6%
8%
15
17
18
3.7
3.8
3.7
Natural disasters
7%
9%
7%
14
13
15
4.0
4.2
4.1
Theft
6%
6%
6%
12
16
14
4.0
4.3
4.1
House ownership
3%
11%
5%
1
1
1
5.0
4.9
4.9
Death
4%
6%
5%
5
15
9
4.8
4.6
4.7
Access to care
4%
4%
4%
9
7
8
4.7
4.5
4.7
Debts
6%
0%
4%
11
18
11
4.2
-
4.2
Clothes
6%
1%
4%
17
4
16
3.8
3.0
3.7
Poor insfrastructures
1%
6%
2%
18
9
12
2.0
4.3
3.8
Livestock issue
3%
0%
2%
18
19
19
2.5
-
2.5
Lack of credit
Table 5 Household concerns by average severity ranking. Incidence of preocupation (%)
Massacre Pedernales
(n=159)
(n=70)
House ownership
Total
(n=229)
Ranking of preocupation
Massacre Pedernales
Severity of preocupation (1 to 5)
Total
Massacre Pedernales
Total
3%
11%
5%
1
1
1
5.0
4.9
4.9
Lack of credit
10%
9%
10%
2
3
2
4.4
4.2
4.4
Lack of money
56%
59%
57%
3
2
3
4.5
4.8
4.6
Bad housing
30%
31%
30%
4
5
4
4.1
4.3
4.2
Food insecurity
72%
53%
66%
8
6
5
4.3
4.3
4.3
Bad health
62%
39%
55%
6
11
6
4.1
4.0
4.1
Education
53%
43%
50%
7
10
7
4.1
3.7
4.0
Access to care
4%
4%
4%
9
7
8
4.7
4.5
4.7
Death
4%
6%
5%
5
15
9
4.8
4.6
4.7
Lack of work
8%
19%
11%
10
8
10
4.2
4.5
4.3
Debts
6%
0%
4%
11
18
11
4.2
-
4.2
Poor insfrastructures
1%
6%
2%
18
9
12
2.0
4.3
3.8
30%
14%
25%
13
12
13
3.9
4.4
3.9
Theft
6%
6%
6%
12
16
14
4.0
4.3
4.1
Natural disasters
7%
9%
7%
14
13
15
4.0
4.2
4.1
Clothes
6%
1%
4%
17
4
16
3.8
3.0
3.7
14%
6%
12%
16
14
17
3.7
4.8
3.9
Energy
9%
6%
8%
15
17
18
3.7
3.8
3.7
Livestock issue
3%
0%
2%
18
19
19
2.5
-
2.5
Water insecurity
No bathroom
Frontera Verde Final Report 41 Similar to the analysis of income strategies, we developed graphs of incidence of household worry named, its rank and assigned severity. There is a more detailed description of the methodology used in Appendix 1-­‐1. The results for the Massacre and Pedernales watersheds are presented in Figures 41 and 42. Figure 43 provides a breakdown of household concerns by eco-­‐zone of Pedernales. Lack of money is consistently cited as the most pressing household concern. $"&&%
&",&%
&")&%
!"#$%&'()*+,(-*.+/01+
&"+&%
:');+$<+
)%*-,&+
!"!!%
&"!&%
!"$(%
:');+$<+
A$%;+
>*?&3+
B'&4%'8+
-,3'3&*%3+
!"#$%
@'&*%+
,(3*)4%,&5+
C-4)'F$(+
#")'%
B$+
?'&9%$$"+
&"(&%
&"&&%
&"&&%
!"$&%
2$$-+
,(3*)4%,&5+
D9*E+
&"#&%
&"$&%
6'-+
7*'8&9+
!"'$%
!"$#%
&"*&%
&"'&%
:');+$<+
"$(*5+
6'-+
7$43,(=+
G8$&9*3+
C(*%=5+
&"$&%
&"(&%
&"#&%
&"!&%
&"'&%
&"*&%
&"+&%
&")&%
&",&%
$"&&%
!(),-*()*+,(-*.+/!1+
Figure 41 Map representing the incidence/importance/severity of the main household concerns for the Massacre Watershed (n=159). Frontera Verde Final Report 42 1.00
House
ownership
0.90
Lack of
money
4.88
0.80
4.17
Lack of
credit
Importance index (P)
0.70
0.60
0.50
4.36
0.30
4.17
0.20
4.60
4.32
Education
4.46
4.25
4.75
4.25
Food
insecurity
4.27
Lack of
Poor
infrastructures work
0.40
4.80
Bad
Housing
3.96
Water
insecurity
Natural
disasters
Lack of
Death credit
3.70
Bad
Health
0.10
Theft
0.00
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Incidence index (I)
Figure 42 Map representing the incidence/importance/severity of the main household concerns for the Pedernales Watershed (n=70). Frontera Verde Final Report 43 2$34*5$6-+)$()*%(4+7$%+)$'4&'6+'%*'+8+0*-*%('6*4+/(9:;1+
#"))%
A')B+$7+
)%*-,&+
)"+)%
!"#$%&'()*+,(-*.+/01+
>'-+
2$34,(?+
F))*44+
&$+)'%*+
!"))%
(",(%
)"')%
A')B+$7+
"$(*=+
>'-+
2*'6&5+
!"#$%
<$$-+
,(4*)3%,&=+
!"()%
!"'*%
C'&*%+
,(4*)3%,&=+
!"')%
&"'(%
!"')%
G-3)'H$(+
@*'&5+
)"&)%
0$$%+
,(7%'4&%3)&3%*4+
("'(%
!"#)%
A')B+$7+
I$%B+
)"()%
)"!)%
'"))%
!"&&% D5*E+
D'&*%+
,(4*)3%,&>+
(")(%
)",)%
)"&)%
)"!)%
)"')%
)"()%
)"*)%
)"+)%
)"$)%
#"))%
!(),-*()*+,(-*.+/!1+
)"#)%
)",)%
&"))%
?'-+
=')>+$7+ 2$34,(@+
)%*-,&+
!"#$%&'()*+,(-*.+/01+
)"!)%
!"*)%
)"((%
&"&&%
("*(%
F(*%B>+
("$(%
("&(%
("!(%
(")(%
("+(%
("'(%
(",(%
;$$-+
,(4*)3%,&<+
'"))%
E6$&5*4+
!"*(%
B$+
C'&5%$$"+
B'&3%'6+
-,4'4&*%4+
!"''%
=')>+$7+
H$%>+
!"''%
F-3)'G$(+
!"&&%
?'-+
2*'6&5+
)"')%
&"))%
)"$)%
!"&)%
)"+)%
("((%
("((%
E'&3%'6+
-,4'4&*%4+
!"#$%
'"*(%
)"&)%
H))*44+
&$+)'%*+
=')>+$7+
"$(*<+
)"()%
)"*)%
A'-+
2$34,(B+
!"&&%
!(),-*()*+,(-*.+/!1+
2$34*5$6-+)$()*%(4+7$%+"$3(&',($34+'%*'+/(89:1+
2$34*+
$H(*%45,#+
&"'&%
!"$#%
?')@+$7+
G$%@+
)"((%
!")(%
+"))%
F-3)'9$(+
A'-+
2*'6&5+ &")(%
C*'&5+
("!(%
("*(%
)"#)%
!"&(%
!"&&%
("$(%
)"#)%
)"))%
)"))%
=$$-+
,(4*)3%,&>+
!"((%
("+(%
("&(%
)",)%
?')@+$7+
"$(*>+
!"#$%
("#(%
!"(*%
)"*)%
2$34*+
$G(*%45,#+
!"+&%
!"#$%&'()*+,(-*.+/01+
)"$)%
2$34*5$6-+)$()*%(4+7$%+",-8'69&3-*+'%*'8+0*-*%('6*4+/(:;<1+
*"((%
2$34*+
$I(*%45,#+
A*'&5+
!"))%
)"))%
!"))%
)"))%
)"+)%
'"()%
0$$%+
,(7%'4&%3)&3%*4+
D'&*%+
,(4*%)3%,&<+
)"$)%
)"')%
)"!)%
)"&)%
)"*)%
)"()%
)",)%
)"#)%
+"))%
!(),-*()*+,(-*.+/!1+
Figure 43 Map representing the incidence/importance/severity of the main household concerns for the all three eco-­‐zones of the Pedernales Watershed. Frontera Verde Final Report 44 ("#(%
*"((%
Land-­‐cover Mapping (1986-­‐2010) Land cover mapping was very challenging for both watersheds, though particularly for the Northeast. The method used was to assign discreet land cover classes; however, in the case of Haiti, it is often difficult to reliably differentiate different forms of land management as the terrain can be characterized by intensive multi-­‐cropping, fallows, grassed areas for cattle grazing, etc. This was certainly a challenge for 30-­‐meter resolution imagery, therefore for the purposes of this study, the interest was in identifying changes in forest cover. For the Northeast even 1986 appears to have had little forest cover; hence, a slightly different classification strategy needed to be utilized for this watershed. Also, classification accuracy could only really be calculated for 2010 imagery with accuracies hovering around 50-­‐60 percent, which are not reliable enough to make strong conclusions in patterns of land cover change. Massacre Watershed Figures 44 to 46 present the land cover classification for the Massacre watershed for 1986, 2001 and 2010. All images were acquired during the months of December to February, in order to minimize confusion of land covers due to seasonal differences. As this watershed appears to have had a much longer history of land use change than the Southeast, the land cover classification system developed by CNIGS was used, because simple land cover classes like forest, pasture and agriculture did not adequately capture the changes occurring. As described in the methods section, six land cover classes were used, which consisted of: water, forest, agriculture, less dense agriculture, semi-­‐natural (pasture/savannah) and bare soil. The classification accuracies for each class are presented in Table 6. The differentiation between these land cover classes was often difficult as they are considered to be somewhat continuous. From Table 6 it is evident that forest and bare soil areas are the most effectively identified, while the three other classes may be too similar to each other to be classified reliably at 30-­‐meter resolution. Certainly, a more thorough land cover mapping exercise, which was not practicable within the time frame and budget of this project, would be beneficial for the Frontera Verde project. Currently, the identification of reforestation sites is entirely opportunistic, depending on which brigade members are willing to donate their parcels of land. It is clearly evident, however, that while the Haitian side of the border shows a complex patchwork of agricultural and agro-­‐forestry activities, the Dominican side does not have significantly more forest cover and also has large areas of intensive agriculture (e.g. rice). As there is documented migration of Dominicans out of this area (Murray, 2010), increasing numbers of Haitians are being hired to work on these fields. It does not appear that these activities are being transparently regulated with many Haitians complaining of being robbed or arrested when they return to Haiti after performing this migrant work (see Section on conflicts). Frontera Verde Final Report 45 Table 6 Classification accuracy for classes of Massacre 2010. Overall accuracy is 53.6% and the κ coefficient is 0.41. Class Producers Accuracy (%) Users Accuracy (%) Forest 56.3 72 Agriculture 37.6 51.1 Agriculture-­‐less dense 57.7 50.7 Semi-­‐natural 48.7 48.7 Bare soil 72.7 54.7 Figure 44 Land cover classification for Massacre Watershed for 2010. Classes are listed in the legend. The outline of the watershed is given in white and the communities sampled are labeled (overall accuracy = 53.6% and κ=0.41). Frontera Verde Final Report 46 Figure 45 Land cover classification for Massacre Watershed for 2001. Classes are listed in the legend. The outline of the watershed is given in white and the communities sampled are labeled. Unfortunately, this image suffers from a lot of missing data due to clouds and hazy conditions. Many the areas appearing as water are in fact cloudshade. Frontera Verde Final Report 47 Figure 46 Land cover classification for Massacre Watershed for 1986. Classes are listed in the legend. There large black areas are due to cloud cover and the outline of the watershed is given in white. Figure 47 provides a land cover change analysis for the three time periods. For ease of presentation bare soil and agriculture areas were combined as well as less dense agricultural and semi-­‐natural areas. It is difficult to discern any clear patterns, just that there are few areas of forest cover and those that were identified in 2010 have changed very little in extent. Table 7 provides a breakdown of the percent changes for each land cover class between all three time-­‐periods. From this table it appears that forest area has increased significantly; however, this is due to changes in very small areas and therefore does not indicate areas of forest recovery. There are also large changes reported in each of the other land cover classes. This is evidence of the difficulty in differentiating between these relatively continuous land cover classes rather than definitive shifts in semi-­‐natural areas to agriculture and vice versa. Frontera Verde Final Report 48 Figure 47 Change detection for Massacre watershed between 1986 and 2010. Areas that remained pasture, agriculture and forest are in the same colors while additional colors have been assigned for areas of forest loss and forest gained. Frontera Verde Final Report 49 Table 7 Land cover class change analysis for Massacre watershed for 1986 to 2001 and 2001 to 2010. Percent change from 1986 to 2001 Forest Agriculture Semi-­‐
Natural Bare Soil Agriculture-­‐
less dense Forest 21.6 5.6 7.6 3.5 4.1 Agriculture 25.2 30.7 23.0 21.1 23.3 Agriculture-­‐
less dense 21.9 32.3 30.7 24.1 31.0 Semi-­‐
natural 2.0 4.6 5.6 9.3 8.6 Bare soil 10.3 13.3 20.7 32.9 19.7 Class Change 72.7 83.0 -­‐77.0 86.3 -­‐24.2 Percent change from 2001 to 2010 Forest 32.1 11.1 2.2 4.1 8.5 Agriculture-­‐
less dense 27.2 32.1 31.9 29.8 39.6 Agriculture 16.0 26.9 12.0 14.7 18.2 Semi-­‐
natural 14.3 13.9 34.1 26.2 19.9 Bare soil 5.6 13.6 18.5 23.8 12.1 Class Change 66.0 -­‐17.9 180.5 -­‐28.8 13.1 Frontera Verde Final Report 50 Pedernales Watershed Mapping the land cover of Pedernales watershed was more straightforward than for Massacre. Three main categories were chosen: forest, pasture/savannah and agriculture. These are considered to be relatively continuous categories of land cover, with somewhat arbitrary cutoffs depending on the unsupervised grouping of pixels using K-­‐
means classification. As many Haitians have trees on their plots of land, while intercropping underneath, it is difficult to get reliable differentiation of forest from pasture. Pasture is considered to be similar to the semi-­‐natural class from Massacre, where there may be a few trees present or mainly grass covered. Land cover classifications of agriculture are considered to be areas of low to no vegetative cover, rather than areas of actual active cultivation. Figures 48 to 50 present land cover maps for 1986, 2001 and 2010. It is evident here that forests are significantly more intact on the Dominican side of the border, which is what would be expected. Also, from 1986 to 2010 land cover on the Haitian side appears to be becoming increasingly denuded. Table 8 presents the breakdown of classification accuracy for each land cover class. It appears to be relatively robust except for classifying areas of pasture/semi-­‐natural, indicating this is a very complex land cover class. Table 8 Classification accuracy for classes of Pedernales 2010. Overall accuracy is 64.3% and the κ coefficient is 0.49. Class Producers Accuracy (%) Users Accuracy (%) Agriculture 85.0 60.8 Forest 70.3 90.0 Pasture 32.9 69.7 Frontera Verde Final Report 51 Figure 48 Land cover classification for Pedernales watershed in 2010. Land cover classes are provided in the legend and the watershed outline in white. The communities sampled are labeled (overall accuracy = 64.3% and κ=0.49). Frontera Verde Final Report 52 Figure 49 Land cover classification for Pedernales watershed in 2001. Land cover classes are provided in the legend and the watershed outline in white. The communities sampled are labeled. Frontera Verde Final Report 53 Figure 50 Land cover classification for Pedernales watershed in 1986. Land cover classes are provided in the legend and the watershed outline in white. Communities sampled are labeled. Figure 51 presents the change detection analysis of the three land cover maps. From this image it is possible to see the clear areas of deforestation that have been occurring both in the mountainous areas north of the watershed and on the Dominican side near the mid-­‐altitude regions. It appears the areas of forest loss may be closest to the mid-­‐
altitude regions of Pedernales where the communities of Bony and Banane cited charcoal production as one of their main sources of income (see Section on charcoal production). At the same time, the areas of clear forest loss are several kilometers from the border; therefore, it seems unlikely it would be due mainly to Haitian activities. Table 9 presents the results of the change detection analysis and shows that the rate of forest loss appears to have accelerated after 2001, which would be consistent with the results of our qualitative analysis of charcoal producers in the mid-­‐altitude region. Also, pasture areas appear to be reducing with an increase in agricultural areas, perhaps indicative of an intensification of agricultural activities in this watershed. Frontera Verde Final Report 54 Figure 51 Change detection for Pedernales watershed between 1986 and 2010. Areas that remained pasture, agriculture and forest are in the same colors while additional colors have been assigned for areas of forest loss and forest gained. Frontera Verde Final Report 55 Table 9 Land cover class change analysis for Pedernales watershed for 1986 to 2001 and 2001 to 2010. Percent change from 1986 to 2001 Forest Forest 70.3 Pasture 35.7 Agriculture 14.6 Pasture 10.1 33.7 24.8 Agriculture 5.3 20.8 47.9 Class Change (%) 1.9 4.5 -­‐14.4 Percent change from 2001 to 2010 Forest 64.5 19.6 8.4 Pasture 7.8 16.5 6.6 Agriculture 16.3 54.8 74.8 Class Change (%) -­‐16.7 -­‐54.6 72.5 Change in Mangrove extent The mangrove was extracted from Landsat 5 images (1986, 2001 and 2010). The images were radiometrically corrected and orthorectified, using an IDL script based on the maximum likelihood algorithm. We excluded areas of high elevation, non-­‐vegetated zones using a Normalized Difference Vegetation Index (NDVI) and then selected mangrove areas using a band ratio (B5-­‐B7/B5+B7). Finally we ran a maximum likelihood analysis based on the mangrove sample isolated in step 3. Our results indicated little change in mangrove areas (Table 10). The change in mangrove area appears to have begun only in the second time period (2001 and 2010) indicating it is a more recent phenomenon. Table 10 Change detection of mangrove areas for selected time periods. Year 1986 2001 2010 Area (km2) 125.2 136.4 131.1 Change (%) 8.946 -­‐3.886 Frontera Verde Final Report 56 Ü
Legend
1
Mangrove 2010
0
Mangrove 2001
0
Mangrove 1986
0
15
30
60
90
120
Kilometers
Figure 52 Change in mangrove area from 1986 to 2010 for all of Haiti.
Overall there is only minimal variation of the mangrove cover in Haiti (Figure 52). It is difficult to identify the causes of these changes because of the small variations. Most of the mangroves (~90%) of Haiti are located in two main areas: • The Northwestern part of Haiti close to the Dominican border in the Department Nord Est and Nord in the communes of Caracol, Limonade, Quartier Morin and Cap Haitien • The Department of Artibonite in the Commune of Gonaives, l'Estere, Grande saline and Desdunes Our methods did not reveal mangrove extent near the Pedernales watershed, which is consistent with what we observed during the field campaign. Frontera Verde Final Report 57 Legend
Legend
1
1
Mangrove 2010
Mangrove 2010
Ü
0
Mangrove 2001
0
Mangrove 1986
Limite Commune
Nan PiguelGONAIVES
0
Mangrove 2001
0
Mangrove 1986
Limite Commune
L'ESTERE
DESDUNES
QUARTIER MORIN
GRANDE SALINE
DESSALINES / MARCHANDES
LIMONADE
Jackzil
TERRIER ROUGE
CARACOL
0
1.25
2.5
5
7.5
Ü
SAINT-MARC
PETITE RIVIERE DE L'ARTIBONITE
0
10
Kilometers
1.5
3
6
9
Figure 53 (left) Northwestern Department and (right) Department of Artibonite showing change in mangrove area from 1986 to 2010. Even if the overall trend in mangrove coverage is stable, locally we observe a reduction of the mangrove extent inland since 1986 around Jackzil. (Figure 53) Reasons for this reduction may be related to an extension of cultivated land around mangrove areas or logging of mangrove trees for fuelwood or construction. Livelihoods Overview of the main sources of income reported Agriculture is by far the major source of income in the Massacre watershed. Nine out of 10 respondents mention farming as an income-­‐generating activity. It is ranked as the most important activity and contributes, on average, 44.5 percent of a household’s total income. Commerce stands out as the second activity, with close to one in two respondents mentioning commerce as a source of income. In terms of importance, commerce occupies the second rank and also contributes close to 40 percent of the total income. If livestock has a high incidence, respondents tend to consider it as a less important activity. Most of the other activities have lower incidence and varying importance (Figure 54). Frontera Verde Final Report 58 12
Kilometers
1.00
0.90
0.80
Agriculture
Migrate to
DR
Commerce
44.5%
Importance index (P)
0.70
65.0%
39.3%
0.60
40.8%
0.50
Fishing
32.5%
Salt
production
0.40
Salaried
job
28.0%
Livestock
Charcoal
Production
24.6%
0.30
25.7%
Remittances
0.20
16.1%
28.6%
Tree
planting
0.10
0.00
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Incidence index (I)
Figure 54 Map representing the incidence/importance/contribution of income sources for the Massacre Watershed (n=159) The same three income-­‐generating activities stand out in the Pedernales watershed. However, despite the higher incidence of agriculture over commerce, the latter is ranked higher and contributes to a larger portion of total household income than farming (Figure 55). As in Massacre, livestock seems to be a common complementary activity and appears widespread among respondents. Charcoal production, salaried job, and fishing, have a lower incidence but represent important sources of income. Frontera Verde Final Report 59 1.00
Commerce
0.90
0.80
Agriculture
Fishing
47.3%
Importance index (P)
0.70
38.6%
0.60
47.1%
Charcoal
Production
0.50
Livestock
Salaried
job
0.40
30.5%
0.30
24.9%
41.7%
0.20
0.10
13.3%
Remittances
4.3% Tree planting
0.00
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Incidence index (I)
Figure 55 Map representing the incidence/importance/contribution of income sources for the Pedernales watershed (n=70). Figure 56 depicts the concentration of proportional income values derived from commercial activities. In Massacre, high concentration of values seems to be associated with population density and proximity to roads. On the other hand, low concentration values seem to be located in areas where transport accessibility is limited and population density is low. In Pedernales, the highest concentration values tend to be associated with the location of bi-­‐national markets and population density. Frontera Verde Final Report 60 Figure 56 Income from commerce density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Grouping the responses for fishing communities in Massacre and the three different eco-­‐
zones in Pedernales also revealed some interesting patterns (Figure 57). There were no important differences in the Massacre watershed in terms of other sorting criteria, such as distance from rivers/roads or intervention/non-­‐intervention villages (results not shown). On the other hand, for both fishing areas in Massacre and coastal areas in Pedernales, commerce and fishing are important sources of income often contributing 40-­‐50 percent of a household’s salary. In the mid-­‐altitude region of Pedernales charcoal production becomes a significant source of income, although commerce and agriculture are both important in terms of proportion of income generated and incidence among households, respectively. The mountainous area is the most unique with essentially all households relying on agriculture for some portion of their income and then augmenting that income with commerce and/or livestock activities. Charcoal production is not a significant source of income in this eco-­‐zone. Frontera Verde Final Report 61 <23=63>&/?&634/-5&@/A045@&?/0&B@C63>&D6EE2>5@F&G2@@2405&93H'';&
$"!!#
@$""*%)*+
J/--5045&
(#-(&
!",!#
JC204/2E&
1023@./01&
$(#)%&
L6@C63>&
,-./012345&63758&9:;&
!"*!#
P>064AE1A05&
!")!#
!"#$%&
$"#*%&
!'#*%&
!"(!#
!!#"%&
!"'!#
JC204/2E&
:0/7A4K/3&
M2E20657&
&N/O&
!"&!#
F,6A,(4+
(#*(&
-./0#
123456738#
!+#+%&
+'#$%&
(#,(&
!"#$%&'()*+,(-*.+/01+
!"+!#
2'(3,(4+$5+,()$"*+6$7%)*6+5$%+)$'6&'8+'%*'6+9+0*-*%('8*6+/(:;<1+
"#((&
I6D5@1/4=&
(#$(&
(#+(&
I%**+
#8'(B(4+
'(#(%&
@A'%)$'8+
0%$-7)B$(+
C'8'%,*-+
+G$H+
(#)(&
>,?*6&$)3+
(#!(&
'+#+%&
')#!%&
!(#(%&
'"#"%&
'(#!%&
(#'(&
!"%!#
!"$!#
'(#(%&
(#((&
(#((&
!+#+%&
!"$!#
!"%!#
"!#!%&
(#"(&
G6>0215&1/&Q<&
!"!!#
!"!!#
!"&!#
!"'!#
!"(!#
!")!#
!"*!#
!"+!#
!",!#
$"!!#
2*",J'()*6+
C'8*6+5%$"+
D%*E$$-+
(#"(&
(#'(&
(#!(&
(#)(&
(#+(&
(#$(&
(#*(&
(#,(&
2'(3,(4+$5+,()$"*+6$7%)*6+5$%+",-8'9:&7-*+'%*'68+0*-*%('9*6+/(;<=1+
)#&&%
+#)+&
&#'&%
+#$+&
&#-&%
@A'%)$'9+
0%$-7):$(+
>,?*6&$)3+
+#'+&
*(#'%&
!"#$%&'()*+,(-*.+/01+
!"#$%&'()*+,(-*.+/01+
F4%,)79&7%*+
&#(&%
()#"%&
+#-+&
C4%,)79&7%*+
!"#$%&
!'#$%&
+#(+&
+#*+&
!"#"$%
?$""*%)*+
&#,&%
&#!&%
C'9'%,*-+
+D$E+
&#+&%
"!#($%
"&#'$%
=,>*6&$)3+
"&#&$%
&#"&%
+#!+&
&#*&%
B%'(6#$%&*%+
+#,+&
&#)&%
(+#+%&
+#++&
+#++&
"#((&
2'(3,(4+$5+6$7%)*6+$5+,()$"*+5$%+"$7(&',($76+'%*'68+0*-*%('9*6+/(:;<1+
@$""*%)*+
+#"+&
(#-(&
!(),-*()*+,(-*.+/!1+
,34675345&63758&9,;&
,#++&
=4%,)78&7%*+
!"#$%&
)*#"%&
?@'%)$'9+
0%$-7)A$(+
"&#&$%
+#,+&
+#!+&
+#*+&
+#(+&
+#'+&
+#-+&
!(),-*()*+,(-*.+/!1+
+#$+&
+#)+&
+#"+&
,#++&
&#&&%
&#&&%
"#"$%
&#)&%
B%**+#9'(A(4+
&#*&%
&#"&%
&#+&%
&#!&%
&#,&%
&#-&%
&#'&%
&#(&%
)#&&%
!(),-*()*+,(-*.+/!1+
Figure 57 Maps representing the incidence/importance/contribution of income sources for the fishing villages in Massacre and each eco-­‐zone in the Pedernales watershed. Seasonal dependence on income sources For this baseline assessment we were interested in exploring how a household’s income and food or energy insecurity changed throughout the year. In terms of sources of income, farming, livestock and tree planting were the strategies that exhibited the greatest seasonality (Figure 58). This would be consistent with the specific times of year households are able to harvest production from their agricultural and tree planting activities, while in interviews respondents often cited the need to sell livestock in times of low income or when school fees were due. As a result, there appears to be a peak during the months of September to December for income from livestock. On the other hand, commerce, charcoal production, salaried jobs and fishing appear to be more constant throughout the year (results not pictured) or at least respondents did not provide specific months when these activities were most important. Frontera Verde Final Report 62 123".&*$4))
5%6789:)
"%#$
)"#$
)%#$
("#$
(%#$
!"#$
!%#$
&"#$
&%#$
"#$
%#$
')"#
)'#$
((#$
!'#$
!"#$
!"#$
!(#$
!(#$
!%#$
&%#$ &!#$
&&#$ &&#$
!#$ )#$ &#$
!"#$"%&'(")*+),*-.",*/0.)
!"#$"%&'(")*+),*-.",*/0.)
1'#23%())
4%56778)
&("#
&'"#
&)"#
!'"#
!)"#
$'"#
$%"#
$)"#
%)"#
'"#
%'"#
%!"#
%'"#
!"# $"#
!"# !"# $"#
%"# %"# %"#
%)"#
&"#
)"#
*+,$ -./$ 0+1$ 231$ 0+4$ *5,$ *56$ 257$ 8.3$ 9:;$ <=>$ ?.:$ 2,4$ ?@$$ <=,.$
*+,# -./# 0+1# 231# 0+4# *5,# *56# 257# 8.3# 9:;# <=># ?.:# 2,4# ?@## <=,.#
!"#$"%&'(")*+),*-.",*/0.)
1#"")!/'%2%())
3%4556)
&+"#
,&"#
,+"#
(&"#
(+"#
)&"#
)+"#
%&"#
%+"#
&"#
+"#
'"#
)*"#
%&"#
!"#
$"#
!"# !"#
%'"#
$"#
("#
!"#
$"# $"#
%)"#
%&"#
$"#
-./# 012# 3.4# 564# 3.7# -8/# -89# 58:# ;16# <=># ?@A# B1=# 5/7# BC## ?@/1#
Figure 58 Seasonal dependence of households to specific income strategies. Only those with a seasonal pattern are presented. Seasonal food insecurity, farming and energy vulnerability In both watersheds, food and energy insecurity are closely related. Months with higher dependence on farming, unsurprisingly, show minimums in these two forms of insecurity. In addition, the lower the percentages of monthly income coming from farming, the more likely households are to experience a lack of food. Both watersheds indicate two periods of high food and energy insecurity, although during slightly different months. In Massacre, the two harvest cycles, from November to January and from May to August, also correspond to lower food insecurity levels. Similarly, in Pedernales, December to April seems to be the main harvest period and corresponds to low food and energy insecurity. Figures 59 and 60 show the different patterns of this relationship for the two watersheds and indicates that the timing of interventions to augment household income would do well to be during the periods of high insecurity and will need to be adjusted according to which watershed is being targeted. Frontera Verde Final Report 63 1*-.",*/0.)#"2*#3%()4*%&,.)*+)+'#4)5%$*4")'%0)/'$6)*+)
+**0)'%0)"%"#(7)8)9'..'$#")
@,:=A.#B1=A#B+1AC,7#D,E$'&F#
G+:H#=B#B==I#D,E$'&F#
G+:H#=B#.,.174#D,E$'&F#
!"#$"%&'(")*+),*-.",*/0.)
)!"#
(!"#
'!"#
&!"#
%!"#
$!"#
!"#
*+,#
-./#
0+1#
231#
0+4#
*5,#
*56#
257#
8.3#
9:;#
<=>#
?.:#
Figure 59 Overlaying percentages of households reporting months with insufficient energy and food as well as months when farm income was the most important for Massacre. 1*-.",*/0.)#"2*#3%()4*%&,.)*+)+'#4)5%$*4")'%0)/'$6)*+)
+**0)'%0)"%"#(7)8)!"0"#%'/".)
@,:=A.#B1=A#B+1AC,7#D,E(FG#
H+:I#=B#B==J#D,E(FG#
H+:I#=B#.,.174#D,E(FG#
!"#$"%&'(")*+),*-.",*/0.)
)!"#
(!"#
'!"#
&!"#
%!"#
$!"#
!"#
*+,#
-./#
0+1#
231#
0+4#
*5,#
*56#
257#
8.3#
9:;#
<=>#
?.:#
Figure 60 Overlaying percentages of households reporting months with insufficient energy and food as well as months when farm income was the most important for Pedernales. Frontera Verde Final Report 64 Farming and livestock Agriculture as a main source of income Eighty-­‐seven percent of survey respondents reported relying on agriculture for some percentage of their income, 143 households in Massacre and 57 in Pedernales. Each respondent was asked to list the 5 main crops they grew. While this information is valuable for getting an idea of what main crops are being grown, it is not possible to present a detailed analysis of the diversity of a household’s diet or the importance of having a few productive cash crops rather than several smaller subsistence crops. A total of 31 crop varieties were reported by one or more households, with the top nine crops for each watershed listed in Table 11. There are a few clear differences, such as only farmers in Massacre produce peanuts and rice while only respondents in Pedernales reported planting coffee. Maize, manioc, plantain, potato and pois congo were highly important for both watersheds. Table 11 Top nine reported crops by incidence in each watershed. Massacre Pedernales Farmers in Farmers in watershed Crop Crop watershed (%) (%) Manioc 54.5 Black bean 36.8 Peanuts 53.8 Maize 35.1 Pois congo 48.3 Plantain 19.3 Maize 44.8 Carrot 15.8 Potato 33.6 Pois congo 14.0 Black bean 16.1 Pois chus 14.0 Plantain 18.9 Coffee 14.0 Rice 16.8 Sorghum 14.0 Sugar cane 9.1 Potato 12.3 We divided farmers into those who relied on more than 40 percent of their income from agriculture (57 percent of reported farmers) compared to those who would be receiving the majority of their income from off-­‐farm sources. We found majority farmers were significantly more likely to cultivate maize, manioc, pois congo and potato. This would indicate these are the most important cash crops across the two watersheds (aside from coffee in the mountainous region of Pedernales). This group also had significantly fewer assets, roughly half in value compared to non-­‐agriculture dependent households, although they tend to own larger parcels of land. In addition, agriculture dependent households were significantly less likely to buy food in the Dominican Republic. Finally, households dependent on agriculture are much more likely to be severely food insecure (Figure 61). Frontera Verde Final Report 65 8/+9:-0%/5%/%9/:-%514+,*%12%:-,19*%/,155%6/-7%/+*/%1;-*7%
<-=>""?%
1'#23%()'.)')2'3%).*-#$")*+)3%$*2")'$#*..)+**0)."$-#3&4)
$'&"(*#3".)5%67889)
-*$%
((")$%
'*$%
(*$%
@2>>%8A7=%,*$%
<=3/B2%;5/B%
;75B<=C%
!*$%
,*$%
+*$%
)'"!$%
#*$%
)*$%
*$%
6/52%8A7=%,*$%
<=3/B2%;5/B%
;75B<=C%%
#&"'$%
!"#$%
$'#(%&
$'%&
!'%&
!(%&
&'#($%
!"#"$%
$(%&
!"#$%&
!*#!%&
!$#"%&
)'#+%&
)'%&
4355&67/0&"(%&
;0<-=3&>:-=&
>/:=;0?&
)"#*%&
9-:3&67/0&"(%&
;0<-=3&>:-=&
>/:=;0?&&
)(%&
'%&
)"&$%
.//0%123452%
)*+,*-./0*%12%3145*31675%
!"#$"%&'(")*+),*-.",*/0.)
"(%&
&*"'$%
&*$%
(%&
6/0257829:%;//0%
<=>23452%
12?2529:%;//0%<=>23452%
,-&./01&-2031& 4355&67/0&(#'&
8/2-&
(#'&6-&)&8/2-&
9-:3&67/0&)&
8/2-&
Figure 61 Households dependent on agriculture (> 40 percent of income) across food security categories (Pr = 0.035) and area of land area owned (Pr = 0.017). Figure 62 Income from farming density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Figure 62 shows the concentration of proportional income values derived from agricultural activities. In Massacre, the areas with the highest values are consistently close to areas with high concentration of land used for agriculture. Inversely, households in areas around Lafleur and Dumas reported less reliability on agriculture as Frontera Verde Final Report 66 part of their annual income. For Pedernales, the highest concentration of households who derive most of their income from agriculture is close to the localities of Garreau and Plateau de Cedres. The latter is consistent with coffee production activities conducted n the area, as observed in the field during the data collection. Livestock as a livelihood strategy We collected minimal data on livestock holding as a livelihood strategy and were not able to see any clear socio-­‐economic or food security benefits to this activity from our analysis. However, this is probably more a function of the data we were able to collect than a definitive statement on the value of owning livestock. Although not significant, male-­‐headed households were slightly more likely to own livestock (see section on female-­‐headed households below). Livestock ownership is also highly correlated with farming as a source of income as well as somewhat correlated with access to larger parcels of land (Figure 63). These households also consistently report incidences of energy and water insecurity and to a lesser extent are victims of theft. Livestock theft has often been cited as a serious conflict between Haitians and Dominicans (Murray, 2010). 567'2+.)8%.9*'(206:%,*4%3,*4%,))'22%;*<==>?%
(,%&
&'()'*+,-'%./%0.12'0.342%
$(%&
!"#"$%
$,%&
+(%&
+,%&
'(%&
+!#'%&
'(#$%&
'$#)%&
''#$%&
'!#*%&
!"#$%&
',%&
7.689:;4<-.=/&.1643<&
>:;4<-.=/&.1643<&
!(%&
!!#"%&
!,%&
(%&
,%&
,&-.&,#(&/01.&
,#(&-.&!&/01.&
!&-.&'&/01.&
2.34&-506&'&&
/01.&
Figure 63 Livestock ownership over area of land access (Pr = 0.048). Charcoal Production Strong disparities across areas and communities On average in the two watersheds, 28 percent of the households are engaged in charcoal production. While the level is similar between the two watersheds, the percentage of people making charcoal production varies strongly across communities. Villages closer to transportation infrastructure, larger agglomerations, and markets seem less Frontera Verde Final Report 67 dependent on charcoal production. More isolated and rural communities, with easier woodland access and more limited job opportunities, appear more likely to engage in an activity that requires neither formal education nor large capital investment. Although time consuming and labor intensive, it represents an important cash crop for rural households. Figure 64 presents the incidence of charcoal production by community. !#*0-$1*%)*+)$,'#$*'/)23),*-.",*/0)'$#*..)$*44-%51".)6%7889:)
&"#$
'!#$
'"#$
!'#$
)"#$
!"#$
+"#$
%&#$
%%#$
%"#$
%"#$
("#$
("#$
("#$
O9
/P
$
0<
B0
=$
RBB0
-1
$
40
-1
0$
E
M9
/$
0$
B0
N1
G30
=$
O//0
$
/0
$<
K9
K9
,-
L3B
B3G
-
:3/
0$
"#$
/0
$<
H
0:
7F
$
I7J
10
=3 :
$
<0
$
0.
:30
./0
B$
$2 9
/G
10
5$
EF
,-
7$
0B
-/
03
-
C
D0
4$
0$
@
3/
-/
C
2-
$
01
B$
$A :
2-
=$
@
-=
3/
;<
67
1:
?1
91
"#$
0=
;>
345
$
67
-8
$2 -
$
"#$
0$
-1
./
*"#$
'#$
)#$
*"#$
,-
("#$
*%#$
"#$
%"#$
$<
+"#$
Q:
-
!"#$
!"#$
C
!"#$"%&'(")*+),*-.",*/0)
'"#$
Figure 64 Percentage of households reporting producing charcoal by community for both watersheds. In the Massacre watershed, only 7 percent of the households produce charcoal in areas close to roads and rivers, while about a third of them engage in charcoal production in more remote areas. In the southeast, people living in the costal area of Anse-­‐à-­‐Pitres, near the Pedernales bi-­‐national market, are significantly less likely to engage in charcoal production than villagers living in the more secluded middle altitude area. Plateau de Cedres and Garreau are notable exceptions to this trend (Figure 65). This can be partly explained by the need for shade over coffee plants, thereby requiring the maintenance of tree cover for this cash crop. In fact, most farmers in this mountainous region cite tree cover as an important protection for their coffee plants from weather damage. Frontera Verde Final Report 68 1*-.",*/0.)2#*0-$3%()$,'#$*'/)45)"$*6
7*%"6)!"0"#%'/".)8%9:;<)))
)'#'$%
!"#"$%
!'#'$%
('#'$%
('#'$%
'#'$%
+,-./-0%-12-%
345502%-06/752% 3,78/-48,7.%-12-%
-12-%
#(("($%
!*"($%
*("($%
)*"($%
&'"!$%
&&"&$%
01/234/%56.7389/%
:/;-7/%%
!"#$%
("($%
+,-./%
Figure 65 Incidence of charcoal production by eco-­‐zone in Pedernales (Pearson chi2(2) = 10.6283 Pr = 0.005) and for communities across proximity measures for Massacre (Pearson chi2(2) = 11.4816 Pr = 0.003). In the Massacre watershed, the level of charcoal production is significantly lower in villages selected for the Frontera Verde project than in the other villages. This means that the project is targeting communities that are currently less likely to produce charcoal and, by extension, drive further deforestation (Figure 66). 1*-.",*/0.)2#*0-$3%()$,'#$*'/)3%)
$*44-%35".)63&,)7#3('0".)*#)63&,*-&8)
9'..'$#"):%;<=>?)))))
&''#'$%
*)#'$%
!"#$"%&'(")*+),*-.",*/0.)
!"#$"%&'(")*+),*-.",*/0)
&'#'$%
!"#$"%&'(")*+),*-.",*/0.)
&&#'$%
*'#'$%
"'#'$%
1*-.",*/0.)2#*0-$3%()$,'#$*'/)45)
2#*6373&58)9'..'$#"):%;<=>?))))
)'#'$%
!"#!$%
&"#'$%
()#'$%
'#'$%
+,-./0123456%
70123456%
Figure 66 Incidence of charcoal production for communities with and without tree-­‐
planting brigades (Pearson chi2(1) = 7.6212 Pr = 0.006). A slightly more intense and recent activity in the southeast The intensity of charcoal production seems somewhat higher in the southeast (Figure 67). Indeed, the percentage of the total income resources coming from charcoal, as well as the number of bags produced during the last charcoal production cycle, are both slightly higher in the Pedernales watershed. Although only the difference in the volume of charcoal produced is significant. Frontera Verde Final Report 69 3*,4%5-2678%-.%/0*1/-*2%91-:6/8:%
)*+,%-.%/0*1/-*2%
&$"!#
(!#&%
&!"!#
!"#$%
%$"!#
!&#'%
%!"!#
$"!#
!"!#
'()*+#,-.$$/#
0*11*234#,-.56/#
74843-*+41#,-.%9/#
Figure 67 Estimated level of charcoal production by a household based on the previous volume of charcoal produced for both watersheds (F=4.65; Prob > F = 0.0356). Interpret with caution the volume of charcoal production: Last time they produced charcoal, charcoal producers made on average slightly more than 15 bags. However, the size of the charcoal bag varies strongly, from 6 to 15 marmite according to the respondents. Unfortunately, the units assumed were not rigorously assessed. On average, households interviewed have been producing charcoal for the last 14 years in both watersheds. Nearly half of the charcoal producers in the Pedernales area just started producing charcoal over the past 5 years (Figure 68), suggesting that this activity is relatively new for an important portion of the population living in the southeast. This finding is consistent with the land cover change mapping of Pedernales. Frontera Verde Final Report 70 1"'#.)*+)$,'#$*'/)2#*0-$3*%)
.4774:01%;5<'"=%
>1?1054@17%;5<!"=%
!,#&$%
9177%2345%(%61407%
*"#!$%
&+#"$%
(8!-%61407%
!!#)$%
&+#"$%
!!8&-%61407%
&'#($%
&*#'$%
./01%2345%&-%61407%
!"#"$%
-#-$%
(#-$%
!-#-$%
!(#-$%
&-#-$%
&(#-$%
'-#-$%
'(#-$%
*-#-$%
*(#-$%
(-#-$%
!"#$"%&'(")*+),*-.",*/0.)
Figure 68 Breakdown of years a household has been producing charcoal by watershed (Pearson chi2(3) = 6.2304 Pr = 0.101). This difference is only marginally significant. Frontera Verde Final Report 71 Figure 69 Income from charcoal density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Figure 69 depicts the concentration of proportional income values derived from charcoal production. In Massacre, areas with evident high values are close to Meiac, Lafleur and Acul du Pins. Jackzil shows no values for charcoal production at all. In the case of Pedernales, middle altitude areas close to Bony and Banane present the highest concentration values. This was consistent to what was observed during the fieldwork. Both datasets are depicted over forest land-­‐use density layers, derived from LandSat 2010. With the information available at this point, it is not possible to determine spatial causality between concentrations of forest land-­‐use and the proportion of income derived from charcoal (or charcoal producing households). However, a visual inspection –especially in Massacre-­‐-­‐ seems to confirm that households who rely mostly on charcoal for their annual income might have easier access to forested areas than households who do not. The exception to this assumption is the upper watershed area in Pedernales. During the fieldwork it was observed that, for farmers, cultivating coffee represents an important source of income. Due to the light/shade requirement for the coffee plants to grow, farmers are less inclined to cut down trees that provide shade. As a result, we observe relatively high concentration values of income derived from agricultural activities, and low concentration values of income derived from charcoal in the same area (see figure 69 above, and figure 62 for proportion of income derived from farming). Frontera Verde Final Report 72 A strong decrease in wood resources over the past 2 years Nearly eight in ten charcoal producers have noticed a sizeable decrease in the quantity of wood available in the past two years. Charcoal producers, especially in the northeast, deplore the overexploitation of wood resources. Oftentimes, charcoal producers are not planting trees to replace those cut for charcoal burning, therefore contributing to net forest loss in the region. Similarly, regeneration cycles are getting shorter and shorter, due to “the recent acceleration of the exploitation rhythm”. As an experienced male charcoal producer from Acul-­‐des-­‐pins mentioned, most charcoal producers are not allowing a minimum of 2 to 3 years to let the trees grow back to their normal size, cutting them before they reach maturity. This deteriorates the quality of the wood used for charcoal burning, and more trees are needed to produce similar amounts of charcoal. This vicious circle makes the whole activity less profitable, pushing charcoal producers to increase their volume of charcoal activity to maintain similar levels of income. “The amount of wood available for charcoal decreases from year to year. A while ago, we did not have to go very far to find wood. Nowadays, the landscape is cleared of trees, and you have to go far to find some.” Female, Savane longue “The amount of wood used for charcoal production decreases nowadays because everyone in the village use their fields to produce charcoal, and trees do not regenerate easily.” Female, Meiac “Since I have been producing charcoal, wood has decreased significantly in the area, especially the Bayahonde that I use.” Male, Meiac In the Pedernales region, almost every other charcoal producer has noticed an increase in the number of people depending on wood-­‐based charcoal production, putting more strain on forest resources. People in the southwest are more inclined to use charcoal as a primary energy for cooking, and therefore to buy charcoal on Haitian or Dominican markets. This might drive up the demand of charcoal in the area, as this woman from the middle-­‐altitude region suggests: “Today, we produce more charcoal than before, because charcoal has now much more value than before, and is much more in demand.” Female, Banane A small minority of charcoal producers has not reported witnessing a decrease in productivity or in wood availability. They claimed that charcoal-­‐producing performances mostly depends on supervision and care spent during its production. Others attribute the fluctuations in wood availability to weather conditions, the rainy season bringing unfavorable conditions for charcoal production. “The amount of wood used today to produce charcoal does not change. Everything depends on the time spent during the entire production process” Male, Dumas Frontera Verde Final Report 73 “My performance in production will depend on the supervision and time spent during the entire process of production. If I monitor well the process of production, the performance is always proportional to the amount of trees I use” Female, Dumas “[Wood availability] varies with the seasons. In the dry season, there is a lot of wood to make charcoal but when it rains a lot, it is difficult to find wood.” Female, Manket Species of trees and collection methods When possible, Haitians use trees that are easily available and can regenerate quickly. From Figure 70, it is evident that the majority of charcoal in both watersheds is being produced from bayahonde, which is a fast-­‐growing tree that is able to survive in harsh, particularly dry conditions. This is also a beneficial tree for grazing, as it offers animals nutritious pods (Timyan, 1996). The next most common tree used for charcoal is logwood, also known as “bloodwood” or Campeche. It has significant medicinal properties, can be used as a dye and was an important export during Haiti’s colonial period (Duke, 2008). It is a fast-­‐growing, ornamental tree that can grow in the direct sun and perhaps for this reason is an attractive tree for charcoal production. Both Caimito and Mango trees are important fruit producing trees and we assume that charcoal producers are preferentially cutting down older, less productive trees of these varieties. It is telling that the use of these fruit species for charcoal production is more common in the Massacre watershed, indicating there is a dearth of forest species available for conversion. This is consistent both with the land cover mapping of Massacre and the survey results concerning availability of trees and number of years households have been producing charcoal. 12"$3".)*+)&#"".)-."0)+*#)$,'#$*'/)2#*0-$4*%)
'$#*..)5'&"#$,"0.)
,-==->23#?.@)%A#
B3932.-C3=#?.@$+A#
*("#
:-;-<0.93#
'("#
''"#
70/8009#
$&"#
)&"#
4-56510#
$%"#
)*"#
,-./0#1233#
!"#
!"#
$!"#
&!"#
)!"#
*!"#
(!"#
'!"#
+!"#
!"#$"%&'(")*+),*-.",*/0.)
Figure 70 Percentage of households reporting the main species of tree they are using for charcoal production. Frontera Verde Final Report 74 Wood collection process Sixty-­‐four percent of the charcoal producers collect wood from their own land or garden. Most of the time, landowners use the wood they cut down for farming. When charcoal producers don’t own any land or cannot gather enough wood from their field, 38% of them buy wood directly from another charcoal producer. This also includes the possibility of paying somebody to let them cut branches or trees on a field that does not belong to them. Only a tenth of the charcoal producers admitted to cutting down trees illegally in the forest, bushes or thickets, risking a penalty. Some charcoal producers (approximately 12 percent of respondents), especially when gathering wood on their field, seem to carefully avoid cutting young trees. Most of the time, they collect wood when they actually need to trim trees or clear some land for farming. They often prefer cutting branches to trees, and choose mature over younger trees. “Generally, I gather wood when I prune or cut down trees on my land in order to farm” Male, Savane au lait “Usually, I gather pieces of wood in my garden to make charcoal, or cut some branches from the trees of my garden. I never cut down a tree to make charcoal.” Female, Savane longue “I only use fruit trees for charcoal if and only if they had been uprooted by strong winds.” Female, Dumas “I use the trees on my own land. Often, when these trees are not mature yet, I buy some more from other charcoal producers in the village.” Male, La Fleur A Haitian wood collection and charcoal production process Virtually all the wood collection and charcoal production is made on Haitian territory. The extraction of charcoal is illegal in the Dominican Republic, and above all, considered to be very risky. If a Haitian crosses the border to cut trees in the Dominican Republic, he can get severely punished if caught. Several Haitians from the southeast even described cases of mutilation or killing that took place over the past ten years. “Haitians can be killed if caught collecting wood in the Dominican Republic.” Female, Zone de refugies As a result, merely 9 percent of the charcoal producers interviewed collect or buy their wood in Dominican Republic, while only 5 percent actually produce charcoal on the other side of the border. When they do so, it is always under the supervision or in collaboration with Dominicans. Haitians can, for instance, produce charcoal directly on a Dominican field, under the supervision of the owner, therefore sharing profits together. They can also bribe the Dominican guards to be able to transport wood or charcoal across the border. With the exception of one producer from Manket, Haitians producing charcoal in the Dominican Republic come exclusively from the southeast region. Frontera Verde Final Report 75 “When farming on Dominican lands, if the owner wants me to clear his plot, I am authorized to cut down trees to make charcoal. But this can only be done if the owner gets along well with me. In this case, I produce charcoal on the border under the supervision of the Dominican owner. I then share the profits with him. For example, if I produce 30 bags, I take 20 for me and leave the 10 other bags to the Dominican owner.” Female, Manket “When I go to the Dominican Republic to make charcoal, the wood I bought is directly placed on the border between Haiti and the Dominican Republic for me. Then I carry it back home from there.” Female, Bony “I usually go to the Dominican Republic to buy wood to produce charcoal. Sometimes, I just pay a Dominican border guard to be able to gather some wood across the border. I also often go to the Dominican Republic to produce charcoal. In this case, some friends I have there give me some space to make charcoal for free, but I still have to pay the Dominican border guards to cross back the border at a rate of 1000 gourdes for fifty bags.” Female, Zone de refugies Evident rejection of a risky and demanding activity There is an unquestionable rejection of the charcoal activity among charcoal producers. When choosing between earning the same amount of money from charcoal and from some other income activity, every single charcoal producer would actually prefer to switch to another activity (Figure 71). Reasons why you would choose a different
activity than charcoal production (n=53)
45%
Dangerous for health
36%
Too physically demanding
Not profitable enough
15%
Bad for the environment
8%
Too time consuming
8%
0%
10%
20%
30%
40%
50%
Figure 71 Reasons cited across watersheds as to why charcoal producers would prefer to do another income generating activity. Several reasons can explain this consistent reticence toward charcoal making. First, it is perceived as being a real threat for a producer’s health. The heat generated by burning charcoal for several hours a day, in the sun and hot climate, is pointed out as extremely dangerous. It can provoke dramatic changes in body temperature, which can sometimes be fatal for charcoal producers. Wood smoke exposure is also associated with increased respiratory symptoms and diseases (cough, wheezing, lungs problems). Another Frontera Verde Final Report 76 common risk is to get severely injured while cutting trees. Simply put, producing charcoal “reduces your life expectancy”. Second, charcoal production is described as an extremely strenuous, physically demanding activity. A lot of older householders admit they are too old for such an exhausting activity. Third, producing charcoal is generally not profitable enough. It is very time consuming and labor intensive, charcoal producers spend up to 10 to 14 hours a day near burning kilns. Moreover, they often have to share their profit with somebody that cuts and transports the wood for them, especially women. Charcoal producers frequently have to team up with other charcoal producers, to exploit different sites and allow some time for trees to grow back, which generally considerably reduces their margin, sometimes up to a third. Last, several charcoal producers mentioned the harm that this was doing to the environment, largely contributing to deforestation. As a charcoal producers summarized, “Charcoal production is a dehumanizing job.” “This activity is too strenuous and dangerous…It’s easy to injure yourself when cutting down trees with an ax. Moreover, the heat generated by charcoal collection is very dangerous for the health.” Male, La Fleur “Charcoal production is very difficult, requires much patience and sacrifice.” Male, La Fleur “Charcoal production harms the environment so much, it is a strenuous activity, and it is dangerous for the health – especially the heat generated in collecting charcoal, that causes disease.” Female, Dumas “It is not easy to cut a tree, since when you cut a tree, you damage the earth. In addition, one needs someone else to cut trees, with whom one must split the benefits of the activity.” Male, Savane longue Fishing Basic demographics Only nine surveys were performed on individuals who identified themselves as dependent on income from fishing. As this is such a small sample, we will only be able to present some descriptive analysis of the information we collected. Table 12 provides a breakdown of the number of fishermen interviewed by community and watershed. We also performed a few qualitative interviews with fishing organizations in Pedernales, which will be presented as supporting information. Based on those we interviewed, fishermen in Pedernales have been fishing for a significantly longer period of time. This may be due to the age of the interviewees, as we did not ask for this information; however, this may be indicative of this livelihood being a longer-­‐term livelihood option off the Southeastern coast. Frontera Verde Final Report 77 Table 12 Number of fishing surveys performed by community and watershed. Average number Surveys Watershed of years reported Community completed being a fisherman Meiac 2 Massacre 15.6 Merande 1 Jackzil 2 Pedernales 28.8 Saline 4 !"#$"%&'(")*+),*-.",*/0.)
Only one fisherman surveyed claimed that he supported himself 100 percent on fishing alone. All other respondents listed at least one or two other activities they relied on to supplement their income. Eighty percent listed farming as an additional source of income, averaging 17 percent of the household’s earnings. Fifty-­‐six percent of fishermen also owned livestock, which contributed roughly 15 percent of their income. Other activities listed included charcoal production (2 respondents), salt production (1 respondent) and commerce (1 respondent). Those who mentioned charcoal production (both were in Massacre) listed the species of wood they used as black mangrove, logwood and bayahonde (mangrove destruction will be presented in another section). Eighty percent of fishermen surveyed listed food insecurity as one of their main household concerns. The average food insecurity score for all fishing households is 7.8, which falls between the average in both watersheds (Figure 21). The majority of the households surveyed fell into the two lowest wealth quartiles (Figure 72), although their average asset value was between the overall mean and median of the entire surveyed population. 12.,2%(),*-.",*/0.)34)5"'/&,)6-'#7/")8%9:;)
$!"#
($"#
(!"#
'$"#
'!"#
&$"#
&!"#
%$"#
%!"#
$"#
!"#
)*+,-.#+,/0.1# 6,7*89#+,/0.1# :1;49#+,/0.1# <;=1,-.#+,/0.1#
23/450,#
23/450,#
23/450,#
23/450,#
Figure 72 Fishing households across wealth quartiles for both watersheds. One hundred percent of respondents reported selling their catch, although their methods (e.g. local market, across the border, open air, etc.) varied considerably and seems to be dependent on the quality of their catch. There was not a clear seasonal dependence of households on fishing, with 56 percent of respondents reporting earning Frontera Verde Final Report 78 income for fishing during any or all months. Eighty percent of respondents reported receiving other forms of income throughout the year as well. For those that did report particular months for either fishing or other income, there appears to be an interesting overlap with reported months of energy and food insecurity (Figure 73). Food insecurity appears to peak during the months that fishermen report being the most active. While it is very difficult to make any inferences from such a small sample, these fishermen may be augmenting their income during times of food insecurity and energy insecurity. More information would be necessary to verify this. 12"#/'3)4"&5""%)6%$*7")+#*7)8.,6%(9)+**0)6%."$-#6&:)'%0)"%"#(:)
."$-#6&:);%<=>)
$!"#
!"#$"%&'(")*+),*-.",*/0.)
($"#
(!"#
,?@A?+6#B+9<C-#
'$"#
'!"#
8:A-0#?+9<C-#
&$"#
,<<D#?+@-940-#
&!"#
E+-063#B+@-940-#
%$"#
%!"#
$"#
!"#
)*+# ,-.# /*0# 120# /*3# )4+#
)45#
146# 7-2# 89:# ;<=# >-9#
Figure 73 Reported months of fishing or other income with reported months of food and energy insecurity. Otherwise, fishing appears to be a traditional livelihood option that is taught from parent to child. Two thirds of respondents stated they had come from a line of fishermen (Figure 74), conversely barely 55 percent reported they would encourage their children to become fishermen. Surprisingly, two thirds of surveyed fishermen in Massacre said they saw fishing as a viable livelihood for their children, compared to only a quarter of surveyed fishermen in Pedernales. Again a more thorough field campaign would be necessary to ascertain whether the number of fishermen have increased, stayed relatively stable or are declining in these communities. Frontera Verde Final Report 79 4"5*#&"3)1'6-)*+)2"'#%7%()&*),-.)8%9:;)
!"#$"%&'(")*+),-."#/"%)
,($%
#($%
!!"#$%
!($%
+($%
*($%
)($%
''"'$%
'($%
&&"&$%
&($%
($%
-./012%34.50678%
9:1;<3.=>?3%
0*1)2"'#%"3)&*),-.)
-47/%;40:85%
Figure 74 How fishermen said they had learned to fish across both watersheds. A reduction in the number of fish Across both watersheds, all respondents noticed that the overall number of fish has reduced, although the consensus was less unanimous concerning any changes to the size of fish (Figure 75). One of the two fishing organizations in Anse a Pitre, APEDA (Asosyasyon Pehè deside Anse a Pit), stated they have not noticed a change in size because they use the appropriate equipment that does not capture smaller fish. 23#4"56",$,)78"9$:$84:,;"$3,$<#4$#3="$3,$
>:#.$*$?":5#$@,ABC$
+),-.$
/,)0$
11'$
()$
**'$
!"#$
%&'$
Figure 75 Perception of fishermen whether the size of fish has reduced over the past two years. Table 13 lists the fish and marine species respondents mentioned as species they target. Again as the sample was so small, we are grouping the names by frequency of being mentioned (most, less and listed once). Only three fishermen reported having difficulty catching their preferred species in the past two years. One fisherman in Massacre said Frontera Verde Final Report 80 Fish and Marine species he could no longer catch suckerfish and two fishermen in Pedernales are less able to catch lobster, shark and jacks. Table 13 Fish species commonly caught by frequency of being listed by fishermen. Most frequently listed Less frequently listed Listed once Lobster Turtle Boramas Suckerfish Eel Ballyhoo Jacks Sea bass Conch Tilapia Manta ray Tuna Sardine Sturgeon Angelfish Squid Sole Dependence on traditional fishing equipment Haitian fishermen often lament the inferior equipment they have available for catching fish. This can also be a source of conflict with Dominican fishermen, as they blame Haitians for indiscriminately targeting smaller fish, which negatively impacts the long-­‐
term viability of these populations. Few fishermen reported having access to a motorized boat and rely predominately on nets and fishing line (Figure 76). Two thirds of the surveyed fishermen claimed to own their equipment and only one respondent described an equipment-­‐sharing scheme where he was able to keep 70 percent of his catch in exchange for access to equipment. 6"4*#&"5)1%$15"%$")*+),-.1%()"2314/"%&)3-"5)
!"#$"%&'(")*+),-."#/"%)
7%89:)
$!!"#
,!"#
+!"#
*!"#
)!"#
(!"#
'!"#
&!"#
%!"#
$!"#
!"#
-./#
0112#34-.#
516/#
71/1849.:#
516/#
01-.1%()"2314/"%&)3-"5)
068;11-#
Figure 76 Main types of fishing equipment used. Regardless of the current state of fish communities compared to how they were before, fishing as a livelihood strategy is extremely challenging and requires a significant commitment of time. Additionally, the depleted fishing stocks near the coast are driving fishermen to travel farther and, in some cases, putting them into dangerous situations. Frontera Verde Final Report 81 “Fishermen usually leave the port around 5 or 6 am, and sometimes take up to 5 or 6 hours on their small boat to get to a profitable fishing zone. Some fishermen don’t hesitate to spend 3 to 5 days on the sea to catch fish. But this type of fishing can be very dangerous due to changing weather conditions (boats can be turned over, often causing the death of the crew).” Male, Jackzil Fishermen appear to be better organized in the Southeast There are two fishing organizations in Saline/Anse à Pitre. The association APEDA is made up of 123 fishermen. A committee takes care of all the management issues (administrative, planning, inventory) and decides on the fishing days (that is days when a quota of the catch is given back to the association). The association only owns one yacht, while all the rest of the material belongs directly to the fishermen. Most of the time, members use their own fishing equipment and only occasionally use the yacht. The amount of money given back to the association depends on the species fished. This money is used to help families in case of illness or accidental death and/or support members repair or replace fishing equipment in case of damage There is also APPA (Asosyasyon pechè pou pwoteksyon ak anviwonman Anse a Pit), which consists of 100 fishermen. It owns 5 yachts, one tremail fishing net, as well as a generator and batteries in order to fish at night. To be able to use this equipment and benefit from support and assistance in case of equipment damage or illness/death in the family, each member has to give 5% of the income coming from their catch to the association. Incidence of fishing across the border differs by watershed Fishing in Dominican waters appears to be more common in Massacre where 60 percent of surveyed fishermen reported they had compared to 25 percent in Pedernales. Only one fisherman reported going across the border often in the past month, and this was in Merande where he would be fishing in the Massacre River. Almost 90 percent of fishermen reported Dominicans coming into Haitian waters to fish and an equally high 88 percent of those respondents said they were negatively impacted by this activity. However, only one respondent reported interacting directly with any Dominican fishermen and stated that his experiences have been overall positive. “Each time Haitian fishermen go to Dominican waters to fish, they get arrested. Their fishing equipment, and sometimes all the things they have with them (including food), is confiscated. This type of event occurs almost on a daily basis. That is the reason why most Haitians try to go to the DR secretly, in order to avoid being caught. On the other hand, Dominican fishermen can come and go freely in Haitian waters, without any type of restriction.” Male, Jackzil “Fishermen are used to go far away from the shore to fish, whether on the Haitian or Dominican side. When they go to the DR, they spend 6 to 8 hours on the ocean to get there, and usually spend 4 to 5 days fishing over there.” APEDA, Saline Frontera Verde Final Report 82 “When I go fishing in Dominican Republic, Dominicans take my fishing equipment. The only way to solve these problems would be to no longer travel to the Dominican Republic for fishing. Unfortunately, there is nothing in Haiti, we are forced to go into the DR. Dominicans have a marine head (“chef marin”) that can protect their interests, while we do not.” Male, Saline APEDA claims to have a good relationship with Dominicans because they use appropriate equipment, therefore they report being free to fish in the Dominican waters as they wish. The APPA stated that it also has very close ties with Dominicans, with its 5 boats being provided by a Dominican organization. They say they can fish in Dominican waters freely as well. Frontera Verde Brigade Membership As with the section on fishing, there were not enough surveys performed on brigade members for robust statistical analysis, however some qualitative data was gathered. In total, 12 brigade members were interviewed; 7 in Massacre and 5 in Pedernales. The questions asked covered how the respondent had been chosen to be a member of the brigade, whether there had been any internal conflict within the brigade or the community from where they were chosen and if they had any suggestions for how to improve the program. None of the respondents interviewed cited any problems within the brigade or that they had any personal problems with the capataz, the female brigade leader. Selection process for membership Most of the respondents explained the formalized procedure they underwent that eventually selected them to be a part of the program. Usually, the process involved the selection of possible members from existing work organizations within each community or input from the local CASEC or a committee of “civil protection”. A few of the brigade members in Pedernales did mention that members of their community had complained about not being chosen by the program. “Yes, some of the people that should have been selected [for the brigade] were not. Many people who were chosen are not going to take good enough care of the trees.” Male, Banane “The way the choice was made has led to some whispers, but nothing more.” Male, Bony Frontera Verde Final Report 83 Figure 77 Income from tree planting density map. Massacre watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Figure 77 shows the concentration of households who derive a part of their annual income from planting trees. This activity is associated with the activities that the Frontera Verde project promotes. In Massacre, the locality at Savane au Lait and Chevalier present the highest concentration of households under this parameter. Welch and Lamine also have active brigades, but we may not have captured a significant number of households that reported performing this activity. Pedernales has not started planting trees yet; therefore we did not obtain any data associated with this activity. Previous land use often forested already In addition, all brigade members in Massacre were also landowners, or individuals required to provide at least 1.25 kawo of land to the project. We asked them some qualitative questions regarding the previous use of the land and if they had suffered any losses previously due to theft or natural factors. Of the seven qualitative surveys performed, every respondent reported the previous land cover had consisted of some combination of fruit or timber trees (Table 14 lists the species provided). Few of the respondents stated they had cut down the trees to sell them, but instead mentioned using the wood to build their house or allowing cabinet-­‐ and furniture-­‐makers access to their wood for their wares. None of the landowners cited any problems with theft on their land and 6 of the 7 respondents mentioned significant tree loss during cyclone activity. Of the sites that were visited, all appeared to have a reasonable plan for keeping their seedlings watered or were a very short walk to the river for carrying up buckets of water. Frontera Verde Final Report 84 Table 14 Species of tree that had been planted on land intended or currently in use for the Frontera Verde project. Type of tree Common name Timber Oak, ironwood, pine and mahogany, eucalyptus Avocado, apple (cachiman), orange, mango, coconut, lemon, Fruit soursop, grapefruit In terms of recommendations for any changes to the project, few respondents offered concrete criticisms. Perhaps this is partially due to the project’s relative youth. However, payment did come up, particularly timing of salaries. It is understood by us that none of the brigades have received any payment as yet; however, few respondents complained of not having been paid. Project organizers only raised this concern informally. “As far as the payment method [salary], this should be done every fifteen days instead of every month. This would allow the household to ensure its survival, especially in difficult times.” Male, Welch Assessment of membership criteria In the introductory section we provided the criteria required of all Frontera Verde brigade members. Below we present some preliminary and descriptive analysis of households that meet these criteria, to see what characterizes the families being targeted by this project. Female-­‐headed households One of the requirements for membership is that at least 3 brigade members must be women as well as the head of the brigade, or capataz. In terms of main sources of income, female-­‐headed households are somewhat less likely than male-­‐headed households to earn income from farming, livestock or charcoal production; however, these households are significantly more likely to earn money from commerce activities. Female-­‐headed households are also significantly less likely to have access to any land, with 45 percent of these households being able to access 0.5 kawo or less (Figure 78). There is no significant difference in the amount of land either of these households own. In terms of livestock ownership, male-­‐headed households are significantly more likely to own larger or more animals than female-­‐headed households. Frontera Verde Final Report 85 1'%0)'#"')'$$"..)+*#)+"2'/"3)'%0)2'/"3,"'0"0),*-.",*/0.)
4%56789)
1"2*#&"0)3%$*4").*-#$".)+*#)+"4'/"5)'%0)4'/"5,"'0"0)
,*-.",*/0.)6%7)889:)
)*%&
&++$%
)*#$%&
!"#$%&
!*%&
'+$%
++#'%&
$'#"%&
'*%&
+*%&
$!#"%&
-+#*%&
$*%&
-*%&
",#!%&
((#"%&
"*%&
D/?7&
(*%&
,*%&
*%&
.71/?7&
!"#$"%&'(")*+),*-.",*/0.)
!"#$"%&'(")*+),*-.",*/0.)
,**%&
#+$%
,+$%
-+$%
(+$%
!!"#$%
!+$%
*+$%
)!"*$%
)+$%
8591:5%
)+",$%
&'"($%
526789:;<&
=>/0;:/?&
@0:AB;C:3&
=:1170;7&
+$%
)!"*$%
&+"*$%
&+$%
./01234&
31:5%
*+"-$%
)("*$%
+%./%+"(%012/%
+"(%./%&%012/%
&%./%)%012/%
3/45%.617%)%012/%
Figure 78 Breakdown of reported income sources (slightly significant) and land access (Pr < 0.05) for female-­‐ and male-­‐headed households. Land ownership From the previous section on female-­‐headed households we see they are less likely to have access to areas of land greater than a 0.5 kawo. This would put them at a distinct disadvantage to being chosen as a member for a Frontera Verde brigade, although we did not see such a stark difference in terms of ownership of land or at least not a significant difference from our relatively small sample. We did separate analyses of the characteristics of households owning greater than 1.5 kawo, to explore the type of households being selected with Massacre’s land ownership criteria. Consistent with the two Frontera Verde criteria, we found a strong relationship between those owning above 1.5 kawo and being dependent on farming as a main source of income. This was not significant with any other livelihood strategy we sampled. We also found that landowners were slightly more likely to have a larger number of children or at least reported a larger number of children who were attending school. As presented in a previous section, many of the wealth indicators are related and land ownership is no exception. Figure 79 shows that households owning the minimum amount of land to be considered a brigade member are consistently in the highest wealth quartiles or own high-­‐value livestock. This would indicate that the criteria are not targeting the households that are in the greatest need of intervention and perhaps this criterion, while convenient for identifying reforestation sites, is not adequately balancing the benefits of this project across the community. Frontera Verde Final Report 86 )(#(%&
!"#$"%&'(")*+),*-.",*/0.)
)(%&
),%&
',%&
'(%&
!"#$%&
*,%&
'(#)%&
!,#(%&
!,%&
!(%&
1'%0)*2%"#.,34)'$#*..)5'/-")*+)/35".&*$6)'.."&.)*2%"0)
7%89:;<)
1'%0)*2%"#.,34)'$#*..)2"'/&,)5-'#6/".)7%89:;<)
!)#*%&
!(#(%&
*+#(%&
*,#(%&
B.<C./<081&.8&4011&
253<&*#,&D3/.&
-3<=C./<081&EF*#,&
D3/.G&
*(%&
!"#$"%&'(")*+),*-.",*/0.)
),%&
+,%&
",%&
(%&
,%&
>5?8=&/03425&
6738940&
@?A5012&/03425&
6738940&
'$#"%&
',%&
(,%&
;84<8=4/:0&8:&
>/00&1234&(#+&?3=8&
!"#$%&
!,%&
,%&
-./012&/03425& :0;.<=&/03425&
6738940&
6738940&
-,#,%&
-,%&
('#+%&
''#+%&
()#"%&
(*#"%&
(+5,,,&18&
!,5,,,&670&
98:/&1234&
!,5,,,&670&
.347<8=4/:0&@A(#+&
?3=8B&
+#,%&
./00&1234&
(5+,,&670&
(5+,,&18&
(+5,,,&670&
Figure 79 Comparison of land owners of > 1.5 kawo compared to wealth quartiles (Pr < 0.05) and livestock asset value (Pr < 0.001). Cross-­‐purposes with charcoal production? The results we have presented thus far do not clearly elucidate what the Frontera Verde project is trying to accomplish through this intervention. The basic assumption is that by planting a large number of trees, the border region will benefit from increased forest cover; however, previous reforestation efforts have found that Haiti is far more complicated than that (Murray and Bannister, 2004). Certainly introducing more trees into both of these regions will have some benefits, particularly Massacre as it currently has such little forest cover. At the same time, it is important to recognize that there are pressures acting on existing standing trees, such as basic energy demand and charcoal production as a livelihood strategy. Therefore, efforts need to be made to target these activities as well in order to hope to encourage greater forest cover. What we have seen from the initial analysis of the Frontera Verde program is that it is hiring among the wealthiest households and, by extension, not the households currently practicing charcoal production, as these are among the poorest and most food insecure. Our qualitative analysis of charcoal production has shown that households that are currently producing charcoal would willingly perform almost any other livelihood strategy if it provided equivalent income. While the project may not want to “reward” households that have been involved in previously environmentally detrimental behavior, it appears that finding a way of hiring members of charcoal producing households may go a long way toward reducing pressure on dwindling forest resources. The focus on hiring women may not have clear environmental benefits, as these households are more likely to be reliant on commerce-­‐based activities and statistically less likely to be involved in charcoal production, they otherwise may not be chosen as they were found to have access to smaller areas of land if at all. However, we do not wish to make an argument against targeting female-­‐headed households, as they were roughly 50 percent of the households we surveyed and may not otherwise be selected for the brigade without this specific requirement. Frontera Verde Final Report 87 We would offer that the land ownership requirement does not appear to forward any of the Frontera Verde’s project goals, but instead seems to have been chosen to reduce the logistical burden of implementing this project. These households can still benefit from the project by being selected to receive seedlings planted on their land, but it would seem choosing brigade members should not disproportionately choose the communities’ wealthiest members or, at least, if the project wishes to have a positive impact on reducing food insecurity and poverty in the region. While we admit we only spoke to a small number of brigade members whose land had been donated to the project, we consistently found that these parcels of land previously already had trees growing on them. In fact, they were generally productive parcels of land that the household was already actively using for timber and fruit production. Again, this may be the intention of the project to augment existing household income, but if the idea is to reforest areas that were not previously forested the current model does not appear to be accomplishing that. However, this last conclusion would be better supported by the land cover change analysis. Unfortunately, as there was so little forest cover in the earliest image of the time series (1986) it was difficult to ascertain which areas had been subject to the highest rates of deforestation. Interaction and conflicts with Dominicans Border crossing Nearly 56 percent of Haitians declare having crossed the border at least once over the past month. In the two watersheds, the closer and the more accessible the border, the more inclined Haitians are to cross over to the Dominican Republic. In the Massacre watershed, 67 percent of the villagers living in communities close to roads and rivers went to the Dominican Republic, while 73 percent of those living in the coastal area of Pedernales went to the other side of the border at least once. On the other hand, none of the persons interviewed in the mountainous and isolated community of Plateau de Cedres had gone to the Dominican Republic in the past four weeks (Figure 80). Frontera Verde Final Report 88 ()#$
")#$
!)#$
%)#$
')#$
&)#$
+)#$
*)#$
)#$
!"#$
%&#$
,-./0$
1203450$67/849:0$
'(#$
1#*..)&,")2*#0"#)/'.&)3*%&,)'$#*..)
"$*45*%".)6%789:)
!"#$"%&'(")*+),*-.",*/0.)
!"#$"%&'(")*+),*-.",*/0.)
1#*..)&,")2*#0"#)/'.&)3*%&,)4)
5'..'$#")6%789:;)
;0<.80$$
+&#$
!&#$
%&#$
*&#$
)&#$
"&#$
(&#$
'&#$
&#$
!"#$
%&#$
"&#$
,-./0.1$.23.$
456613$.170863$ 4-890.59-8/$.23.$
Figure 80 Households reporting crossing the border by proximity measure in Massacre (Pearson chi2(2)= 3.3155 Pr = 0.191) and eco-­‐zone in Pedernales (Pearson chi2(2) = 9.2944 Pr = 0.010). Main reasons for crossing the border The first and foremost reason for crossing the border is buying food supplies (Figure 81). Sixty-­‐five percent of Haitians in the northeast and 83 percent in the southeast go to the Dominican Republic to purchase food, mostly at bi-­‐national markets held each Monday and Friday in Dajabon and Pedernales. Weekly transboundary migrations seem, therefore, driven by necessity and food insecurity, due to the recurrent inability of Haitian farmers to maintain a steady food supply through the seasons, and the absence of an appropriate market on the Haitian side. This food is then directly consumed by the households or sold at local community markets to make a profit. Haitians who can afford to cross the border to purchase food are significantly less likely to be severely food insecure and score, on average, significantly lower on the food insecurity scale. In fact, 68 percent of the people crossing over to the Dominican Republic to buy food supplies are in one of the top two wealth quartiles of our sample. Another reason to go to the Dominican Republic is the lack of available jobs in Haiti. Nearly one in four Haitians crossing the border in the past month did so to find a job. Usually, Haitians either work in the Dominican Republic a few days a week, or relocate there temporarily when they are unable to produce sufficient food and income to meet basic household needs in their own country. Frontera Verde Final Report 89 H2<<203.#I?J%%K#
C.9.3?2/.<#I?J+$K#
)&"#
CD3012<.#=449#
%!"#
'%"#
5.6#2#F4G#
'!"#
()"#
-.//#461.3#:349D06<#
'!"#
($"#
CD3012<.#461.3#E449<#
'!"#
*"#
-.//#/8A.<640B#034:<#
!"#
)"#
@/8?80#
&"#
$"#
;8<86#=278/>#=38.?9<#
%"#
'"#
5.6#:289#
$"#
("#
5.6#72338.9#
$"#
$"#
-.//#0123042/#
!"#
$"#
($"#
'$"#
!$"#
+$"#
&$"#
)$"#
*$"#
%$"#
,$"#
!"#$"%&'(")*+),*-.",*/0.)
Figure 81 Breakdown of mains reasons members of a household are crossing the border into the DR by watershed. A third reason to cross the border is to trade non-­‐perishable goods with Dominicans -­‐ that is purchasing or selling other types of goods, crops, or livestock. Because of the lack of infrastructure and markets on the Haitian side of the border, bi-­‐national markets represent a great opportunity for merchants to sell and buy products. The other motives for crossing the border, such as going to a clinic or visiting friends or family, remain marginal. 1"'.*%.)+*#)$#*..2%()&,")3*#0"#)2%)&,")4'.&)5*%&,)'$#*..)
6'&"#.,"0.)7%899:;)
A strong asymmetrical interdependence, source of potential conflicts There are profound differences in terms of wealth and socio-­‐economic development between the Dominican Republic and Haiti. The Dominican Republic is a rapidly developing democracy, with productive agriculture and manufacturing sectors, and one of the most flourishing tourism industries in the Caribbean. Whereas, Haiti is the poorest country in the Western Hemisphere, handicapped by natural disasters, political corruption, a much higher population density, and an obvious lack of resources compared to its neighbor. Because of this substantial economic gap between the two countries on the island, the productive capacity of the wealthier Dominican Republic is only matched with the raw materials and the cheap labor on the Haitian side. If there are proportional benefits for both parties, this situation also creates what C. Schofield called a strong “asymmetrical interdependence”(p 4, Schofield, 1994), where the Dominican Republic is the stronger nation and accordingly plays the dominant role in the region. Frontera Verde Final Report 90 Issues and conflicts in the borderland Despite the socio-­‐economic, cultural, linguistic, and color differences, Haitians and Dominicans at the borderland carry on friendly and cooperative relationships most of the time. Overall, more than 7 out of 10 Haitians living in the two watersheds do not report any kind of problems with Dominicans. Similarly, 64 percent of the Haitians who went to the Dominican Republic in the past month have had cordial relations with their counterparts. However, a lot of complaints were also reported by the Haitian population, most of them referring to the border crossing and to the bi-­‐national markets. They can be summarized in the following four categories: • Seizing/theft of products • Physical and verbal mistreatment • Issues at the border crossing (including arbitrary customs and military fees) • Random arrests/expulsions • Exploitation Seizing/theft of products One of the most recurring complaints by Haitians is money or merchandise theft. Since Haitians cross the border to either work or buy and sell products, they often carry a lot of money or merchandise, hence being an easy target for thieves, soldiers, and customs authorities. Haitian merchants regularly get their products confiscated by custom officers at the border, or by the military at checkpoints located further inside the Dominican Republic. Haitian people who work and get paid in Dominican Republic are often victims of theft, sometimes losing the totality of the cash they have just earned. Similar complaints are made by Haitians who contract credits from “fonkoze” (micro-­‐
finance institutions) in the Dominican Republic. “After having worked hard to earn a small amount of money, it sometimes get stolen on my payday by Dominicans.” Male, Lamine “Most problems I have with the Dominicans take place at the border, when I want to come back to Haiti, Dominicans generally steal my money. Dominicans don't like working, and when they see a poor Haitian that has been working hard to get some money, they steal it from him.” Male, Welch “During the month of February, when my cousin was coming back from the Dominican Republic after having bought some products, border guards threw her goods away.” Female, Bony Physical and verbal mistreatments Haitians regularly deplore the use of physical and/or verbal violence by Dominicans. Most of the time, Haitians complain that they get shoved, pushed, and provoked by their Frontera Verde Final Report 91 counterparts at the border or in crowded bi-­‐national markets, though it sometimes gets much worse. Several respondents complained about beating, whipping, or rape. Dominicans also throw projectiles or shoot at them. In addition, verbal mistreatment is frequent, Haitians being injured and compared to “wild animals,” “devils,” or “thieves.” Dominicans from the military, the custom, as well as tax collectors on the markets are generally the most inclined to mistreat them. During the recent cholera outbreak, tensions between the two communities peaked, accentuating the physical and verbal violence against Haitians. “My wife has repeatedly been the victim of unfair treatments and atrocities on the border when she goes to the Dominican Republic on market days. Dominicans steal her goods, shove and beat her.” Male, Savane longue “Dominicans at the border have thrown rocks and bottles at us. We were also hustled by Dominican guards that were dispersing the crowd that wanted to cross the border. My wife and daughter are often humiliated, beaten with whips by Dominicans.” Male, Dumas “Dominicans always compare us to dogs, pointing out the fact that killing a Haitian is like killing a dog.” Male, Lamine “When I go sell my product in Dominican Republic, Dominicans sometimes throw some water at me or provoke me. They often shove me at the entrance of the market.” Female, Acul des pins “Me and my eldest son have been beaten at least three times at the border by the Dominican police during the cholera outbreak. Dominicans were whipping Haitians, telling them not to bring the disease into their country.” Female, Manket Issues at the border crossing: arbitrary customs and military fees The border is a critical source of tension. On market days, Haitians are generally allowed to cross into the border towns to buy and/or sell merchandise without passport or visa requirements. While individuals with smaller goods can normally access the market for free, merchants with larger amount of goods have to pay custom duties to access bi-­‐
national markets. On top of that, vendors who cross the border and are not able to sell their products during the market have to pay another custom duty to bring their merchandise back across the border. These regulations are not standardized at all, and can vary considerably from one day to another. If Haitians don’t give the amount requested by border guards, they have no choice but to turn back and go home. Therefore, Haitians denounce the amount and the quantity of taxes they have to pay to trade or simply buy goods in the Dominican Republic. “When I go to Dominican Republic, border guards often get me into trouble. My husband was arrested several times already. [...] Sometimes, if you do not give the amount requested by border guards, you are forced to turn back. Dominicans might also beat you with their guns.” Female, Zone de refugies Frontera Verde Final Report 92 “Dominicans have a habit of mistreating them, putting them in prison, or even shooting at them when they cross the border to go to work.” Male, Merande “The Dominicans do not want Haitians living in their country. We have great difficulty to cross the border.” Male, La Fleur Random arrests and expulsions Dominican border guards sometimes arbitrarily arrest Haitians near or at the Dominican border, without apparent reasons. The fact that almost no Haitian in the two watersheds has a passport or meets the visa requirement makes it very easy for Dominican custom officers or for the military to arrest them randomly. “Dominican border guards sometimes place Haitians in an area known as “frontales” for a whole day, without food. Sometimes they take someone and force him to plow a field. If Haitians cannot explain the reason why they are visiting Dominican Republic, they can be thrown in the “frontales too.” Female, Zone de refugies “I went to prison, and spent the whole night and following morning there because I came back late from work and wanted to cross the border at 11pm.” Male, Lamine “One day, several Dominicans arrested one of us and put him in custody for a day. The person was just going to the Dominican Republic get a transfer of money from his friend.” Female, Lamine “When some members of the household travel to the Dominican Republic, Dominicans can put him or her in jail for no reason.” Male, Merande Exploitation When Haitians find a job on the other side of the border, they are often paid little compared to Dominicans. They are not treated equally with their counterparts and are more likely to be mistreated by their coworkers (“no respect”, “inhumane treatments”, “filthy remarks”, etc.). Most importantly, they often have trouble getting paid once their job is done. The most vulnerable Haitians sometimes happen to be trapped into forced labor, with Dominicans guards blackmailing them and forcing them to work for free (in their fields for instance). “When I go to the Dominican Republic to work or buy products, the Dominican captain that monitors the border often forces me to work on his own land (hoeing) near the barracks. Sometimes I get whipped.” Male, Meiac “Dominicans have a habit of mistreating Haitians, to beat them, humiliate them, and throw down their goods. They also treat Haitians working in Dominican fields as their slaves.” Female, Meiac Frontera Verde Final Report 93 “I was treated like a slave when I worked in the cane fields, and shipped back to Haiti whenever they didn't need me anymore.” Male, Acul des pins “When working for Dominicans, we are paid too little. Dominicans make us work very hard to earn money.” Male, Lamine “My husband went to the Dominican Republic to work for half-­‐days. He sometimes had trouble to get paid at the end of the day, Dominicans being reluctant to pay him.” Female, Garreau Other issues Among other issues, Haitians selling merchandise on credit to Dominicans often have trouble getting their money back. They are unable to get any authorities to act on their behalf against those who have defaulted on payment. Tensions and frustrations on the Haitians side These mistreatments and various predatory practices give rise to strong resentments on the part of some Haitians. The most frequent are: • Lack of respect • Humiliation • Hate • Fear • Xenophobia/racism from Dominicans Lack of respect Haitians living in the borderlands often deplore being treated with a lack of respect by Dominicans. They don’t have the feeling that they are treated as equals, having the feeling that Dominicans perpetually “look down on them”. “I am revolted by the way Dominicans mistreat Haitians on the border.” Female, Manket “…humiliation, lack of respect for Haitians…Dominicans say that Haitians are devils, and they treat them as thieves.” Male, Lamine “Dominicans inflict inhumane treatments to Haitians” Female, Welch “Dominicans are pretentious and are too at ease in Haiti in comparison with Haitians when they go to the Dominican Republic.” Female, Acul des pins Frontera Verde Final Report 94 Humiliation Some Haitians tend to feel deeply humiliated by the Dominican’s behavior and mistreatments. “The wife and daughter of the household are often humiliated, beaten with whips by the Dominicans.” Male, Dumas “Dominicans refuse to let them work at the border, while they are free to enter Haiti, which is humiliating.” Male, Manket “Dominicans often look down at Haitians and humiliate them.” Female, Welch Hate A minority of Haitians expresses a profound hatred for Dominicans. The atavistic weight of history, as well as the succession of tensions and conflicts between the two countries, has left a trail in the memory of Haitians living in the border area. A few respondents interviewed even mentioned the Parsley Massacre of 1937, when the Dominican dictator Rafael Trujillo ordered the execution of the Haitian population living in the borderlands with Haiti, killing more than 15,000 Haitians. However, it is mostly the daily injustices, discrimination and mistreatment that feed the aversion some Haitians have to their counterparts. “I don’t like Dominicans. They killed a lot of Haitians in 1937. Male, Acul des pins “I have an incalculable hatred towards Dominicans. That's why I never go there. If I were to go to the Dominican Republic and a Dominican was daring to touch me, I would just kill him.” Male, Manket “When I go to the Dominican Republic to buy goods, Dominicans often keep me from crossing the border, seize my goods, and sometimes just beat me with pleasure. I hate Dominicans but work with them because it contributes to improve my business.” Male, Zone de refugies “I don't like when Dominicans humiliate Haitians, inflicting them with bad treatment. I deeply hate Dominicans and I only go there because there is no other alternative in Haiti.” Female, Acul des pins Fear Some Haitians are afraid of Dominicans, representing them as “extremely violent and hostile” to their compatriots and pointing out their “cruelty.” There is a real fear of being killed on the other side of the border (“if I went there, I could die,” “fear of death,” “Haitians sometimes disappear over there,” etc.), which is often given as a reason why Frontera Verde Final Report 95 they avoid crossing the border at all cost. When one of their family members has no choice but to cross the border, they get very anxious (“[I get] really worried when a member of my family goes to the Dominican Republic.”) “I know that Dominicans kill Haitians with machetes and dogs. So for me, going to Dominican Republic is like tossing a coin: you can either return alive or simply get killed.” Female, Acul des pins Xenophobia/racism from Dominicans A few Haitians mentioned that they had been victim of xenophobic views from Dominicans, some of them attributing that to a “color prejudice.” Haitians from the borderland clearly perceive some hostility from their neighbors, especially when going there to work or sell their own products. “Dominicans hate Haitians. They often say there are too many Haitians in their country.” Female, Manket “Dominicans don't want Haitians to live in the Dominican Republic. Dominicans don't understand why Haitians would not stay in their own country” Male, Acul des pins “The Dominicans do not want Haitians living in their country. […] They always ask us what we are looking for in Dominican Republic, and why we do not stay in Haiti. Dominicans think we are going to make their country poorer.” Male, La Fleur Conflicts and feelings across watersheds Types of conflict and their frequency are similar across watersheds. However, resentments towards Dominicans are significantly higher in the Massacre area than they are in Pedernales. In fact, Haitians from the northeast are four times more likely to spontaneously report being treated with a lack of respect by Dominicans. They are also twice as likely to report feelings of humiliation, hatred, or fear (Figure 82). Frontera Verde Final Report 96 1*%23$&.)4)5""/3%(.).6*%&'%"*-./7)8"%9*%"0):%;<<=>)
1*%23$&.)4)5..-".)
.6*%&'%"*-./7)8"%9*%"0):%;<<=>)
N/;;/J82%O3P!,'Q%
!-"#$%
!#",$%
!&"-$%
!#"*$%
!#"*$%
'"(#$
,"*#$
96:/$
%"&#$
'"!#$
7/68$
%"&#$
,")$%
("($%
./01231456$
#"#$%
E882;4F2AB:6;>C3%
./01234%526/0%
++"*#$
9;<5=56>10$
'"#$%
IC8528%>;;:2%
7289/6%/9:;2%
)"*#$
#"*$%
.G0;>J/6%/9:;2%
<=>66>3?%
+("!#$
?6@A$1B$8/C2/@:$
!&"'$%
E9:;2F1>;482/41234%%
@AB6C>4/DC3%
H/I/806=/C$E0F(!G$
.25283/62;%O3P)(Q%
KG2LF;2>M>3?%
.:;G25F;GCH25%
D6CC6@8/$E0F+'&G$
+"&#$
!"!#$
)"!$%
*"+$%
("($%
!"!#$
%"!#$
)"!#$
,"!#$ -"!#$ +!"!#$ +%"!#$ +)"!#$ +,"!#$ +-"!#$
!"#$"%&'(")*+),*-.",*/0.)
*"!$%
&"'$%
*"!$%
!"#$%
!"*$%
!"#$%
("($% &"($% #"($% -"($% +"($% !("($% !&"($% !#"($% !-"($% !+"($%
!"#$"%&'(")*+),*-.",*/0.)
Figure 82 Households in both watersheds reporting specific conflicts and feelings towards Dominicans. Frontera Verde Final Report 97 Figure 83 Conflict score density map. Left: Massacre watershed; Right: Pedernales watershed. Source: Center for International Earth Science Information Network, CIESIN, 2012. Based on the responses obtained from the qualitative assessment, a conflict score was created. The score takes the most important factors spontaneously mentioned, shown in figure 82, and assigns an equal weight to each factor. The score goes from value 0 to 134. Figure 83 shows the concentration of conflict score values in both watersheds. In Massacre, the highest concentration appears to be close to Lamine, Meiac and Acul du Pins. These areas might be associated with respondents declaring having crossed the border to look for any job related activity. The area close to Meiac is also showing relatively high incidence; the research team suspects this might be associated with fishing practices from Dominican fishermen. Areas with less concentration of conflict values appear to be close to Lafleur, Dumas, Jackzil and Savane au lait. Further research is needed to understand factors associated with this spatial distribution. In Pedernales, the area with the highest concentration of conflict values are close to the coastal communities; and visibly less concentrated in middle or high elevation areas. This pattern is expected, as the border interaction in Anse a Pitre is considerable. 4 Factors included in the calculation of the conflict score are: humiliation, fear, hate, lack of respect, xenophobia, theft, physical abuse, verbal abuse, issues at the border, exploitation, arrest/ deportation, delays in payments. Frontera Verde Final Report 98 A strong sentiment of feeling abandoned and resignation Most Haitians don’t have any authority to turn to. The overwhelming majority of the respondents mentioning having had a problem with Dominicans complain about the absence of a Haitian body able to defend their own interests. The Haitian state is considered as “not responsive,” “powerless,” and “not able to take responsibility.” The lack of presence of the Haitian authority is felt at the border and in the Dominican territory, whether Haitians go to the Dominican Republic to work or trade at a bi-­‐
national market. There is a real feeling of resignation in the face of daily issues with Dominicans because of the lack of protection by their government. Most Haitians are just giving up and accepting their condition, putting themselves “into the hands of God” or simply making sure to avoid interactions with Dominicans (Figure 84). 1"'/2%()32&,)$*%42$&.)5%6789)))
:86B=%C8%.6?=>564%DC0E7810%=>0%7.</%82%@.5A.6%
)*"#$%
.3=>815=?%
900%.%@.5A.6%.3=>815=?%
&#"'$%
&'"($%
900%.%:8;565<.6%.3=>815=?%
-./0%102340%56%10754586%
!"#$%
'"'$%
+,"'$%
,'"'$%
),"'$%
&''"'$%
!"#$"%&'(")*+),*-.",*/0.)
Figure 84 Reported ways Haitians deal with conflicts. “There is nobody you can contact for Haitians, who always end up being the victim because there are no officials from the state in charge of solving these problems.” Male, Manket “Nothing is done to solve these problems. Dominicans have a marine head that can protect their interests, while we do not.” Male fisherman, Saline “When there is a family member who gets arrested, usually, they won't see anyone, because Dominicans are the only ones who decide whether to release them or not, whenever they want to.” Male, Merande A few Haitians are still trying, however, to resolve the conflicts or issues encountered with Dominicans. First, a minority of them has tried to contact the Haitian authorities. Judges, magistrates, or local consuls are sometimes reached out to, especially concerning more important issues. Often, however, the results of these efforts are pretty minor, and several people that used to contact the Haitian authorities do not bother Frontera Verde Final Report 99 anymore. This may also be due to their usual contact having left the borderland region. Second, in the absence of a Haitian authority, some Haitians deal directly with their Dominicans counterparts when they have a problem, like Dominicans police chiefs for instance. “If there is something more important, like a fellow in prison, we will see the consul.” Male, Meiac “The magistrate Mr. Lorenseau […] provides very little or almost no solution.” Female, Banane “When there is conflict with Dominicans, I used to go see a judge named Benjamin. Despite the power of the magistrate, he did not really try to help. Instead, he often told me that it was not him who sent me in the Dominican Republic and, therefore, that it was not his problem.” Female, Zone de refugies Recommendations made by Haitians for avoiding conflicts The most pessimistic respondents believe that Haitians should simply stay in their country and, if possible, avoid going to the Dominican Republic. However, many Haitians made suggestions for avoiding conflicts with Dominicans. One answer would be to increase the Haitian presence in the border area and to have an authority dedicated to manage issues with Dominicans. This would reinforce security and prevent any kind of abuse, especially for merchants and individuals traveling frequently between the Dominican Republic and Haiti. “Have a state official who will protect Haitians interests, especially those who cross the border all the time” Male, Zone de refugies “The Haitian authorities should enforce the same rules and rights for Haitians and Dominicans.” Male, Jackzil However, most respondents think that the problem has deeper roots and cannot be reduced to a question of security or state presence. The scarcity of resources on the Haitian side (food, jobs, other products), even if not clearly stated, is a major issue. The Haitian government should support local communities along the border. This implies developing and diversifying production to meet the various demands of the local population, as well as boosting the regional economy to create more jobs. Similarly, Haitians frequently complain about the absence of a market on the Haitian side of the border. One of their recommendations is, therefore, to build up better infrastructure and markets in Haiti. This includes having banks or institutions able to provide some credit. Another recommendation for conflict prevention is to support farming activities, through credits, subsidies, and the development of irrigation structures in order to reduce food insecurity and enable Haitians to be more independent. Supporting local communities, developing infrastructures and markets, and farming activities would help to restore the balance in relations between Haitians and Frontera Verde Final Report 100 Dominicans. Haitians would not need to go to the Dominican Republic out of necessity hence reducing the tensions between the two populations. “Produce enough in the country and diversify the production, so that Haitians don't have to go to Dominican Republic anymore to buy products. Promote job creation in the country to allow Haitians to increase their purchasing power.” Male, Savane longue “To solve these problems, a large market should be created in Haiti to sell harvested products. They should also have a credit bank in Haiti in order to have the means available to trade.” Male, Welch “To solve these problems it is necessary to develop structures adapted to their situation, such as creating income-­‐generating activities.” Male, Manket “Produce enough within the country and diversify the production, so that Haitians don't need to go to the Dominican Republic anymore. Promote job creation in the country.” Female, Manket Frontera Verde Final Report 101 Conclusions While we must be careful of making overly strong pronouncements based on our limited data, there are some relatively clear patterns emerging from these two watersheds and some initial recommendations we can make for the Frontera Verde project. Above all from our time in the field and the brigade members we interviewed, this project is quite popular and many are optimistic about its potential for success. The results of this report are in no way meant as a critique of the overall approach or worthiness of this kind of intervention, instead we hope to offer some suggestions as to how interventions like it can be improved and more tailored to the needs of the Massacre and Pedernales watersheds. We would also welcome the opportunity to do a similar study on the Dominican side of the border in order to get a more complete picture of these border dynamics. Interventions should focus on reducing food insecurity These results indicate that households that are the most food insecure are those dependent on farming and also tend to rely on charcoal production as an income supplement. A reforestation effort needs to tackle some of the drivers of deforestation as well as encourage the planting of new trees. While it is difficult to conclude that it is periods of food insecurity specifically driving an individual household to produce charcoal, we have found that households with lower food insecurity scores and greater wealth assets are significantly less likely to be involved in environmentally damaging livelihood strategies. Currently, the project happens to be targeting communities that are significantly less likely to be producing charcoal5. This could be accomplished through targeting brigade membership to households that are among the poorest in the community as well as helping to improve farm productivity through subsidized fertilizer programs or facilitating greater access to irrigation. Seasonal and spatial variability in vulnerability should be considered Both watersheds exhibit clear overlaps between drops in income and increases in food and energy insecurity. A household may not be exhibiting the same consumption behavior throughout the year due to these changes in insecurity; therefore, development interventions should be timed to augment household income during times of necessity. We were not able to identify the seasonality of activities like charcoal production, but it is likely these are occurring opportunistically and could be reduced if households have income alternatives. In addition, the Pedernales watershed in particular showed clear differences in socio-­‐economic conditions across its three eco-­‐
zones. As implementation of the Frontera Verde project is still underway in this watershed, these differences should be considered, especially as the mountainous region appears to suffer the most from food and energy insecurity while exhibiting a very low incidence of charcoal production. Finally, while we were not asking specific 5 It seems unlikely this is a result of the project’s influence on the community as it had only been operating for 4 months at the time of data collection and had not actually paid any brigade members. Frontera Verde Final Report 102 questions regarding the timing of paying school fees, we did see a seasonal spike in food and energy insecurity during months when these payments are traditionally due. This would suggest that this expenditure is putting undue burdens on households and perhaps a scheme of supplementing or subsidizing these school fees would also reduce pressures on local forest resources. Improving livelihoods in Haiti will reduce border tensions In the interest of reducing tensions between Haiti and the Dominican Republic, respondents consistently stated they felt they had no choice but to cross the border, often illegally, to support their families. However, if given the option to stay they would happily remain in Haiti. First, this is much easier said than done, as there is a significant investment in transport, water and building infrastructure needed to really bring this region onto a productive path. Secondly, the border region in the Dominican Republic appears to have heavily, intensive agriculture, which will continue to need affordable labor to manage it. This leads to the following conclusion section. Greater certainty in border governance would benefit Haitians Several respondents bemoaned the burdensome taxes they had to pay for bi-­‐national markets, the lack of presence of Haitian authorities to defend their interests (e.g. the case of livestock theft or payment defaults) and the prevalence of theft when a Haitian returns from working on a Dominican farm. Again, these are not straightforward problems to solve, as they can be indicative of either a weak state, such as Haiti, and/or a centrally governed state with little oversight of military and border guards, such as the Dominican government. However, introducing some formalized methods for Haitians to pursue legal recourse against Dominicans that have stolen or physically abused them may go some way to reducing resentment and tensions. Also, slight modifications to how these bi-­‐national markets are managed would be beneficial, which could reduce the cost of Haitian participation or, at least, not require a Haitian to pay hefty import and export duties multiple times on the same goods. Suggested Follow-­‐up Research The data collected for this research project was under a very short timeframe and with limited means; therefore, it was never meant to provide a statistically significant representation of either of these watersheds. However, we feel that this initial data collection has provided some useful insights as well as pointed to some possible future research directions that may be fruitful to pursue. Certainly, for citing areas for reforestation, a more robust on the ground mapping effort would be a huge improvement. We understand that this will be undertaken in Pedernales by CNIGS, where land ownership along the border is relatively non-­‐existent; however, Massacre would also benefit from this greatly. The most straightforward would be to identify areas of steep slopes and relatively bare soils, which are contributing to significant soil erosion that has silted the Massacre River and negatively impacted water quality and fisheries down stream. Then identifying the Frontera Verde Final Report 103 landowners of these parcels, while perhaps complicated, may go further toward accomplishing some of the main goals of the Frontera Verde project. Similarly, there is not currently a coherent plan regarding how best to monitor the progress of these seedlings or enforce their management by the parcel’s landowner. This was not a priority of this research project, as the oldest seedlings were only 4 months old; however, some thinking and project design could be instructive. Also, there are complicated dynamics occurring in terms of timing and scale of charcoal production for sale and gathering of firewood for energy use, which this project has only been able to touch on. We were not collecting data regarding the volumes of production or consumption and certainly could not rely on heads of households to remember the seasonality of these activities; however, in order to design effective interventions to minimize these behaviors a better understanding of the trade-­‐offs would be needed. This would be a data intensive field campaign that would, ideally, occur over a number of months or in different seasons to attempt to capture these dynamics. In addition, a concerted effort to understand the income garnered from commerce activities as well as ways of improving income from selling agricultural goods would be extremely helpful in project design. For instance, are farmers forced to sell their goods at the same time as all other farmers because they do not have the option of storing their products and wait for a better market price? Or are farmers suffering from severely degraded soil fertility and would benefit most from simply improving their own subsistence and market productivity through greater access to fertilizers? While labor intensive data to collect, these regions could be useful test-­‐beds for designing meaningful interventions that, if effective, could greatly reduce tensions along the border as well as help support the success of reforestation efforts like Frontera Verde. Finally, the most straightforward follow up research, would be to perform a similar study on the Dominican side of the border to compare what we have found. Certainly the results we have presented regarding border conflicts have to be read with the understanding that they are only from the perspective of one side of the border and would be greatly enhanced by the Dominican perspective. Frontera Verde Final Report 104 Appendix 1-­‐1 Conceptual mapping of sources of income and a household’s main concerns We used participatory ranking and scoring of the major activities and major concerns and conceptual mapping to highlight the main income-­‐generating activities/worries for each community, and to better understand how important these sources of income/worries were compared to one another. Respondents were first asked to free-­‐list the various sources of income/worries they had experienced over the past year. Then, they had to rank them by order of importance, the most important being ranked first. Finally, they were asked to indicate the importance or severity of each income-­‐generating activities/worries. Pebbles were used to indicate the importance of each activity/stressor, ranging from 1 (barely noticeable) to 5 (very important). We followed a method used by Tscharkert (2007) to analyze the responses from the participatory ranking and scoring activity. For each source of income/worry mentioned by the respondents, we first calculated an incidence index (I) that is the frequency of each source of income/worry. For instance, a score of 0.7 for farming means that 70% of the respondents identified farming as a source of income. Second, we calculated an importance index (P) based on the order in which each participant ranked income-­‐generating activities/worries. This index also ranges from 0 to 1, with Pj=[r-­‐
1)/(n-­‐1)*(-­‐1)+1], r representing the rank and n is the total number of sources of income/worries identified by that respondent. For instance, a respondent mentioning three income-­‐generating activities (n=3), with farming being his second source of income (r=2) would receive a score of 0.5 (Pj=[2-­‐1)/(3-­‐1)*(-­‐1)+1]). The average value for Pj was then calculated for the subset of participants who identified the particular income-­‐generating activity/worry. Last, a severity index (S) was created based on the number of pebbles respondents assigned to each activity/worry, ranging from 1 to 5. Food insecurity scores and categories Five questions were used to measure the level of household food insecurity access. •
•
•
•
•
Worry about food (q401) Eat smaller meal (q402) Eat fewer meals in a day (q403) No food of any kind in the household (q404) Go to sleep hungry (q405) These questions were ranked from the least to the most severe condition experienced by the participants over the four previous weeks: uncertainty about acquiring food (q401), reduction of the size and the frequency of the meals due to a lack of resources (q402 and q403), absence of food at home and feeling hungry at bedtime due to lack of food (q404 and q405). Each of these questions was followed by a frequency-­‐of-­‐
occurrence question, asking how often a reported condition occurred during the previous four weeks (1 = rarely, 2 = sometimes, 3 = often). Following the Household Food Insecurity Access Scale Indicator Guide, two indicators were created to assess food insecurity levels: the Household Food Insecurity Access Scale (HFIAS) Score, or Food Insecurity Score and the Food Insecurity Status (Coates et al., 2007). Food Insecurity Score: The Food Insecurity Score is a continuous measure of the degree of food insecurity access in the household in the past four weeks. It is the sum of the frequency-­‐of-­‐occurrence during the past four weeks for the 5 food insecurity-­‐
related conditions mentioned before, ranging from 0 to 15. The minimum score is 0 if the household did not experienced any of the food insecurity conditions mentioned above in the past month and therefore responded “no” to all occurrence questions (q401 to q405); the maximum score is 15 when the household response to all 5 frequency-­‐of-­‐occurrence questions was “often”. The higher the food insecurity score, the more food insecure a household is. Food Insecurity Status: The Food Insecurity Status indicator categorizes households into four levels of household food insecurity (access), described below. The more frequent they respond affirmatively to more severe conditions and/or experience, the more likely they are to fall into the more severe food insecure categories. Food secure: any household that experiences none of the food insecurity (access) conditions, or just experiences worry, but rarely. Mildly food insecure: any household that worries about not having enough food sometimes or often, but does not cut back on quantity nor experience any of two most severe conditions (running out of food, going to bed hungry). Moderately food insecure: any household that cuts back on quantity by reducing the size of meals or number of meals, rarely or sometimes, but does not experience any of the two most severe conditions (running out of food, going to bed hungry). Frontera Verde Final Report 106 Severely food insecure household: any household that experiences any of the two most severe conditions (running out of food, going to bed hungry) over the past 30 days, even as infrequently as rarely. Wealth Quartiles The Wealth Quartiles was calculated based on a list of assets that the household did or did not possess, partly following the DHS Wealth Index methodology (Rutstein and Johnson, 2004). Here is the list of the assets that were included in the survey: Assets used to generate the Wealth Quartiles Battery torch Iron Solar lamp Animal drawn cart Radio Bicycle Tape recorder Motorcycle CD player Other motor vehicle Cell phone Solar panels Wall clock Gas cook stove Television Gas cylinder To be included in the index, the assets needed to be currently functioning. If a household had a total of three radios, but only one currently functioning, only the value of the later would be taken into account to generate the wealth quartiles. Then, each functioning asset was multiplied by its overall market value. In order to determine the market value of an asset, interviews were conducted in each community of the Massacre and Pedernales watersheds with people that were knowledgeable about these market prices. The sum of the value of each functioning asset possessed by the household was added to obtain the total assets owned. It ranges from 0 to 101,217 gourdes. Finally, the sample population was broken down into quartiles, from the lowest to the highest wealth quartile. Frontera Verde Final Report 107 Appendix 1-­‐2 1"'/&,)2-'#3/".)45)$*66-%7&5)8)9'..'$#"):%;<=>?)
%"",#
%"#
!"#$"%&'(")*+),*-.",*/0.)
&",#
$"#
'",#
+",#
!)#
$%#
("#
("#
)",#
(",#
("#
$",#
+#
$$#
!)#
!"#
)"#
$'#
$"#
)+#
)"#
(!#
%$#
%"#
'#
%+#
)"#
%"#
*",#
+#
+#
+#
%(#
%"#
!)#
$'#
$'#
!)#
("#
$%#
-1790<#L1.:5F#J2.BM:1#
%&#
!",#
%",#
!"#
$%#
$"#
$"#
*#
$"#
39L1=5#L1.:5F#J2.BM:1#
$'#
$$#
!&#
!&#
!"#
%$#
K4GF1=5#L1.:5F#J2.BM:1#
NF4B<#L1.:5F#J2.BM:1#
%&#
#
1=
4:
#
I.
7J
2
H
1:
7F
1#
39
0G
21
#
41
B#
/.
0
-.
EF
1/
.:
<1
#
1B
.0
C
C
C
14
.
7#
D1
5#
.0
1#
@
40
3.
#
12
B#
#A :
3.
@
.=
=#
40
1=
;>
;<
67
2:
?2
#
92
67
.8
-.
/.
0
1#
.2
#3 .
45#
####
",#
1"'/&,)2-'#3/".)45)$*66-%7&5)8)!"0"#%'/".)9%:;<=))
$""(#
,"(#
!"#
!"#
&"#
&"#
+"(#
&"#
%"#
!"#$"%&'(")*+),*-.",*/0.)
'"#
$"#
*"(#
)"(#
%"#
C08D2<@#E2./@D#F;.79/2#
&"#
'"(#
'"#
%"(#
&"(#
'"#
&"#
!"#
-2H415#E2./@D#F;.79/2#
I4E2<@#E2./@D#F;.79/2#
!"#
!"(#
!"#
!"#
&"#
$"(#
GD075#E2./@D#F;.79/2#
%"#
%"#
$"#
!"#
!"#
$"#
$"#
$"#
=.1.12#
=41>#
"(#
-./012#
3412#5607708.941#3412#52#72:;802<#
?/.@2.;#52#
A2572<#
B.772.;#
Figure A.1 Wealth quartiles by community in Massacre (above) and Pedernales (below). Frontera Verde Final Report 108 References Coates J, Swindale A, Bilinsky P. Household Food Insecurity Access Scale (HFIAS) for Measurement of Household Food Access: Indicator Guide (v. 3). (2007) Washington, DC: Food and Nutrition Technical Assistance Project, Academy for Educational Development. Duke JA. Chapter 8: H. In: Duke's Handbook of Medicinal Plants of Latin America (2008): CRC Press. Murray GF. Sources of Conflict along and across the Haitian – Dominican border (2010) Santo Domingo, Dominican Republic: Pan American Development Foundation (PADF). 1-­‐42. Murray GF, Bannister ME. Peasants, agroforesters, and anthropologists: A 20-­‐
year venture in income-­‐generating trees and hedgerows in Haiti. Agroforestry Systems (2004) 61-­‐62:383-­‐397. Rutstein SO, Johnson K. The DHS Wealth Index (2004) Calverton, Maryland: ORC Macro. Schofield CH. Global Boundaries: World boundars volume I (1994) New York, NY: Routledge. Timyan J. Bwa Yo: Important trees of Haiti. (1996) Washington, DC: South-­‐East Consortium for International Development (SECID). Tschakert P. Views from the vulnerable: Understanding climatic and other stressors in the Sahel (2007): Department of Geography/Alliance for Earth Sciences, Engineering, and Development in Africa (AESEDA), Pennsylvania State University. Frontera Verde Final Report 109