Spa$al analysis of environmental degrada$on and household
insecurity along the Hai$an and Dominican Border
Alexandra Morel EI Fellow, CIESIN
Frontera Verde
• Bi-‐na$onal reforesta$on project
Massacre Pedernales
Methodology / Fieldwork • Combina$on land cover mapping and household surveys
• Survey: – 229 household surveys across 20 communi$es
– October-‐ November 2011 – Team of 10 Hai$an enumerators – Use of android tablets & ODK to collect the data
• Land cover mapping: – Landsat 5 – 3 $me periods (1986, 2001 & 2010)
Ini$al results of survey
Livelihoods targeted – Farming (87%) – Livestock (62%) – Fishing (4%) – Frontera Verde tree plan$ng
(5%) – Charcoal produc$on (28%)
Basic Demographics – HH size, gender of HH head, no
of HH members a`ending school, etc
Measures of insecurity – Food, energy & water – Seasonal repor$ng
Wealth measures – Assets, livestock & land
Border conflicts
!"#$
%#$
!"#$
&#$
'!#$
(#$
%#$
"(#$
"%#$
!(#$
!%#$
'(#$
'%#$
)*+$ ,-.$ /*0$ 120$ /*3$ )4+$ )45$ 146$ 7-2$ 89:$ ;<=$ >-9$
!"#$%&'()$%')#&*+,)-#$'-#-./0'1,."&&'(1$-.&%-2&'?<:*5$@+A"%BC$ /*DD*90-$@+A"(BC$ E-F-0+*5-D$@+A%BC$
3#0'4"#$%&5'/*DD*90-G$'!#$E-F-0+*5-DG$%(#$$
!"#$
%"#$!&#$
%'#$
()#$
!)#$
)#$
")#$
!)#$
%)#$
&)#$
')#$
()#$
*)#$
+,-$ ./0$ 1,2$ 342$ 1,5$ +6-$ +67$ 368$ 9/4$ :;<$ =>?$ @/;$
!"#$"%
&'(")*+),
*-.",*
/0.)
1*%&,.)23&,)/'$4)*+)+**0)'$#*..)"$*56*%".)5)!"0"#%'/".)
A>,B<,7$,2/,$C-D!EF$ 1GHH7/$,7I<6H/$,2/,$C-D"*F$ 1>-<,G->6B$,2/,$C-D!)F$
Interac$on with income by watershed
!"#
$!"#
%!"#
&!"#
'!"#
(!"#
)!"#
*+,# -./# 0+1# 231# 0+4# *5,# *56# 257# 8.3# 9:;# <=># ?.:#
!"#$"%&'(")*+),%$*-")+#*-)+'#-,%()'%.)/'$0)*+)+**.)'%.)"%"#(1)2)3'44'$#")
@,:=A.#B1=A#B+1AC,7#D,E$'&F# G+:H#=B#B==I#D,E$'&F# G+:H#=B#.,.174#D,E$'&F#
!"#
$!"#
%!"#
&!"#
'!"#
(!"#
)!"#
*+,# -./# 0+1# 231# 0+4# *5,# *56# 257# 8.3# 9:;# <=># ?.:#
!"#$"%&'(")*+),%$*-")+#*-)+'#-,%()'%.)/'$0)*+)+**.)'%.)"%"#(1)2)!"."#%'/"3)
@,:=A.#B1=A#B+1AC,7#D,E(FG# H+:I#=B#B==J#D,E(FG# H+:I#=B#.,.174#D,E(FG#
Farmers Less than 40%
income from farming
More than 40% income from
farming Total Statistical testbase 115 114 229
Household compositionHousehold size 7.5 6.8 7.2 Pr(|T| > |t|) = 0.0937 Nb of children 3.5 3.0 3.3 Pr(|T| > |t|) = 0.0899
Land ownershipNo land owned 24.3 15.9 20.0 Pr = 0.017Less than 0.5 kawo 35.0 23.4 29.10.5 to 1 kawo 14.6 29.9 22.4More than 1 kawo 26.2 30.8 28.6
Food insecurity Food Secure 5.2 1.8 3.5 Pr = 0.035Moderately food insecure 28.7 17.5 23.1Severely food insecure 66.1 80.7 73.4
Food insecurity score 7.1 8.2 7.7 Pr(|T| > |t|) = 0.0050 Buy food in DR 50.4 28.1 39.3 Pr = 0.001
Main crops base on farmers only 85 114 199Maize 17.7 60.5 42.2 Pr = 0.000Manioc 17.7 58.8 41.2 Pr = 0.000Pois congo 10.6 58.8 38.2 Pr = 0.000Potato 7.1 42.1 27.1 Pr = 0.000
Land Cover Change Analysis
Landsat 5 Mosaic for 2010
“Vicarious valida$on” with 30 cm aerial photographs.
Maximum Likelihood Classifica$on (MLC) overall accuracy = 64.4 %, κ = 0.53
Pedernales Massacre
Change in Mangrove Cover
Incidence and Prevalence of Charcoal Use/Produc$on Spa$ally
Charcoal producers Non-
charcoal producers
Charcoal producer Total Statistical test
base 166 63 229
GenderMale 51.8 63.5 55.0Female 48.2 36.5 45.0
Sources of incomeFarming (>40%) 54 43.3 50.95 t = 1.40; p=0.16Livestock 59.0 71.4 62.5 t =-2.02; p=0.04Remittances 6.0 0.0 4.4 t = 1.71; p=0.09
Food insecurityFood insecurity score (0-15) 7.4 8.4 7.7 t =-2.15; p=0.02
Energy insecurityDays with insufficient energy 65.1 76.2 68.1 t =-1.60; p=0.11
Wealth levelLack of money concern 54.2 63.5 56.8 t =-1.59; p=0.21
Lowest wealth quartile 22.3 25.4 23.1 X2(3)=1.67; p=0.64Second wealth quartile 25.9 30.2 27.1Third wealth quartile 25.3 23.8 24.9Highest wealth quartile 26.5 20.6 24.9
Assets value (in Gds) 8,602 6,569 8,043 t = 1.15; p= 0.37
Livestock ownershipLess than 1,500 Gds 27.6 17.7 24.9 X2(3)= 8.52; p= 0.041,500 to 15,000 Gds 29.5 17.7 26.215,000 to 30,000 Gds 16.6 22.6 18.2More than 30,000 Gds 26.4 41.9 30.7
Livestock value (in Gds) 22,147 34,191 25,466 t =-2.02; p=0.04
X2(1)=2.52; p=0.11
Indicates sta$s$cally significant differences at a 0.1 level
Future plans • Spa$al analysis of
environmental variables and household poverty/insecurity indicators
• Include MODIS imagery to be`er inform land cover classifica$on
• Generate community maps for future reforesta$on sites
• Facilitate bi-‐na$onal mee$ng between Dominican and Hai$an GIS agencies
Acknowledgements • My collaborators:
– Paola Kim-‐Blanco, Kevin Tschirhart, Kim Peng, Joseph Muhlhausen, Jen Mulvey
– Frontera Verde Staff – State University of Hai$ students
• Incredibly helpful advice from: – Hope Michelson
• Funding: – UNEP DEWA