Structural Vulnerability Map of Burkina Faso - Feb. 2015 Definition of Structural Vulnerability:

Structural vulnerability is a tendency to be in a state of high-risk to negative well-being outcomes(ie. undernutrition, anemia) on account of persistent exposure to various potential shocks (ie. climatic, price) in combination with a chronic resilience deficit (ie. lack of absorptive, adaptive and transformative capacities).

GOURMA

Vulnerability Estimate

Linear Regression Analysis: Scatterplot shows grouping around trendline

A linear regression was run to determine how well the Vulnerability Estimate composite model predicts (correlates to) the Well-Being Outcome composite. The assumption is that if Resilience Capacities and Exposure to Shocks and Stresses are combined to form a Vulnerability Estimate , the vulnerable zones identified therein should correlate to geographic zones of negative well-being outcomes (ie. undernutrition). Modeled grid cell values from both geographic models were used for the regression analysis.

R-Squared = 0.39

TAPOA

Resilience Capacity

--

Absorptive, Adaptive, & Transformative Capacities

Service Access

Literacy Rates

Poverty

Food Security/ Ag. Productivity

Anemia Undernutrition

cm.aoe Well-Being Outcome

Vulnerability Estimate (left) gets "corrected" here by averaging it with real well-being outcome measurements through the well-being composite above.

The result of this correction is the Final Vulnerability Map to the right.

Exposure (Shocks and Stresses)

Malaria Prevalence

Recurrent Price Shocks

Historic Sites of Conflict

Recurrent Climate Shocks

Purpose of Map: The purpose of the final vulnerability map is to better identify "hotspots" of structural vulnerability USAID in order to better geographically target longer term resilience investments (development adapted to vulnerable contexts) where the need is greatest. Improved geographic targetting of the most vulnerable is critical to decision making regarding strategic investments for resilience and redressal of social justice related grievances which may lead to conflict.

FROM THE AMERICAN PEOPLE

Basic Methodology: Big data analytics were leveraged to identify structurally vulnerable zones. These zones of development need were calculated by averaging together all relevant and available sub-national development indicators across a broad spectrum. In all, 36 datasets, many of which were historical, were aggregated into composites, which were then aggregrated into higher level composites. Geographic areas where most development indicators were negative are more red and areas where indicators were relatively better are more blue.

Basic data processing steps are listed below:

1) Identify most relevant sub-national indicators available for the analysis.

2) Convert each geographic dataset to raster format. Use Kriging interpolation in the case of point data.

3) Winsorize data where appropriate based on histogram analysis. This prevents the data from being skewed by outlier data and amplifies geographic variation .

4) Rescale all datasets to a common 0-100 scale so that they are comparable for averaging to create composites.

5) Average datasets using weighting based on consensual subject matter expert judgement to create composites.

Mali

SOUROU

NAYALA

BA NWA

ZIRO

Ghana

--C'AS€-JAli>ES

COMOE

SUDOUEST PONI

Cote d'Ivoire N

Niger

ES7' GOURMA

TAPOA

KOULP ELOGO KOMPIENGA

Togo

Final Vulnerability Map

Legend: how to read this infographic

1) Red Zones on all maps are most vulnerable and the blue zones represent least vulnerable relatively.

2) Boxes with blue border contain composite index maps. The + sign between composite maps signifies that they are averaged together to form yet another composite map. Follow the arrow symbol with = sign to see which composite they create when averaged together.

3) Boxes with brown border contain sub-composite index maps. All subcomposite maps in brown bordered boxes are averaged together to form the composite maps they feed into (see the tip of brown box with = sign in it). It should be noted here that subcomposite maps in this infographic are based on other component maps not visualized on this infographic. The table at right gives the specifics about datasets used and the nature of the components maps that were aggregated using a weighted average to produce the subcomposite maps you see in this infographic.

6) Numbers in red circles represent relative weighting of each composite or subcomposite as it contributes to the average. Weightings are represented as a percentage out of 100%. The reason that some of the weightings are odd numbers is because they were derived from an initial intuitive weighting based on double-weighting, half-weighting, one and a half weighting etc., based on a group consensus of subject matter experts. These informal weightings were then converted to a percentage of 100 to calculating weighted averages to form composites.

Top 50 Most Vulnerable Communes "'"' Oe parlmenl

' TIN-AKOH

' DfOU

3 O URSI

4 GOROM-GOROM

' MARKOYE

6 KOUTOUGOU

' TITAEIE

' FALAGOUNTOU

' TANKOUGCUNADIE

'" ARl!I NDA

11 SEEIBA

" BOUNDORE

" GORGI\DJI

" DORI

" SEYTEi%A

" BOURCUM

" SOLHAN

" NAGEIING OU

" SAMPELGA

w TONGOMAY~L

H NASSOUMBOU

" BANI

'" YA LGO

" kEl..80

" M ANSILA

" PENS,q.

" DJIBO

" DABLO

" DIGUEL

" TOUGOURI

n SOLLE

" BARMOULE

" BANH

" COALLA

" lIMNI

" toUIU/l.NGA

" BARSALOG HO

" OUINDICiUI

" POR E· MENGAO

" TITAO

" NAM ISSIG UIMA

" THlmJ

" BOTOU

'' FOUTOURI

" 8ARTIE80UGOU

" TOEGHI N

" 80GANOE

" LI PTOUGOU

" NIDU

so ZEGUEDHOUIN

Province

Oud~IEO

Oudolon

Oudalao

Ouda lao

Oudalan

Soum

Y.igha

Seno

Y.igha

Saum

Y.igha

Yagha

Seno

Seno

Seno

Nom~nkn~

Yagha

Nemenl en~

Seno

Saum

Soum

Sen• NamenIen~

Soum

Yagha

Sanmatenga

Saum

Sanmat~nga

Saum

Namentenga

LOfoum

Soum

Lo.-oum

Gnagna

G, .. ,~nJ

B,m Sanmat en&,!

LOfoum

Saum

L0<oum

Sa11mdl C:!1 f,'l

G<1ogna

Tapoa

Komancljai i

Ko m~·ncljari

Kourweogo

Gr1<gna

G~sna

Kcurweogc

Namentenga

Rq;ion Vulnerabi lity §core POP [2006 Ce n..,11

SAHEL %3 21,013

SAHEL 95.~ 2'.i,321

SAHEL 91_9 15,.806

S/l.HEL 88.8 106,346 SAHE L 8B.S 27,478

SAHE L 88 .4 18,655

SAHEL 87.S 20.63Sl SA HEL :-l 'i.~ 18,18()

SAHEL 85.2 16.453 SAHE L ~4.5 91.(J20

SAHEL 84.1 32.374

SI\HEL 133-9 22,773

SI\HEL 83-7 29,913 SI\HEL 82,6 106,808

SAHEL 82.2 31.S85

QNTRE NORD 81 .6 46,232

SAHE L 81.1 25,108

QNTRE NORD 80.9 16,004

SAHEL 80.6 19,127

SI\HfL 80."l 70,3n

SAHEL 80.1 20,165

SAHEL 79 .6 S9,27B CENTRE NORD 78.7 31,541

SAHEL 78.3 24,1~7

SI\HEL 78.2. 4l,30S

CENTRE NORD 77.3 36.158

SAHEL 76.3 GD,042

CENTRE NOIIO 75.2, 20,707

SAHEL 75.0 8,989

CENTRE NORD 74.4 76.824

NORD 72.8 17,526

SAHE L 72.0 29,383 NORD 70.8 30,332

m 70.8 42,652

m 70 .7 68,4'18

CENTRE NOI\O 69.8 47,75 1

QNTRE NORD 69 .7 78,'}19

NORD 69.4 Z8,l78

SI\HEL 69 .3 Z4,~2

rJO RD 692 66,717

QNTRE NORD 69.0 9,752

m 68.3 23,025

m 68.0 46,"59

m 67.7 H,583

m 67 .S 16,067

PlATEAU CENTRA L 67 .4 16,500

m 66.7 84,&38

m 66.S 41,823

PLATEAU CENTRAL 66.2 26,998

UNTRE NORD 66.0 21,904

For more information: Jeremy Chevrier USAID/Senegal

Sahel Regional Office JChevrier@usaid.gov

or Siaka Millogo

USAID/Burkina Faso Millogos@state.gov

Title

fEWSNH

"' ~olnor,-s,,H sufli<ienl farm

household,

:ioilCri1an1< Carbon Con,;ty

[ducatlonal l.evt!I

Ltter•cv R•les

Pov@r-\'

Reminonce>

Wodhlnd~

L-a<k of Acee" to Health s.,-,1,., on A<<ount of

Financial Constrain IS

t"mpl,a l li•esto,k UnilS

lmm~rali<m Rates

La<kct ~cce,no Healch s,,..,1,., M ....,oun, of

c;.t.in<a

• 1sran,o 10 Heolrh c,,nw

~cf People i,,,rUnlt•<e•

Acee., t<> improsed

Sanitation Ac,~,, to lmpru"1l0 Orin kins

wo,orS<><Jru

> 10 minu:es wai. ,o """"" Ctlnklngw;,t.,>Q<Jn;•

~verage «alnfalt Variability

Ave••£• Teml"'roture ~"""' l!ainl!iea><>n

A,ior,ot,, LOO£\~ Of R•ir,y f,oa5on

AveragoTN•IAnnutl

Preclplliftlo~

Hi>l6rloar C..allkt

Rolo8•••

M8lori• P-•l•nc~

•v•r,,g~ niillet prw;e duri"l!I

'"""••awn

Averngr, '(l' ll <>w earn ,..1ce duri"it 1,an _..,n

Avera1~whl le <om pri<e

d"rinc k,;,,,=on As~rage whlle<0rghum ~"

~Ur1"'8 1Nll-,<on •s•••~~ rod sorghum pr Tee

~ur!ngleal,..,,.,~n

Provaten«r~S-k•on Anemia

An<!fflQ Prevak>nu! jWtrmen]

Ae~rago GAA1 ~ale>ISMUll

Aver~il~ GAM Rates (OHS)

Avcr.ttte Stunllng Rates (SMART\

,o,,-~•Sluolio~ Rat .. (DHS)

All datasets used for composites (datasets averaged based on weightings listed): ~ datasets averaged based o n weight ings Usted

"' historical datasets used when ailallable in orde r to map structu ral vulnerability \IS. conjunctural * all t ime series datasets have been averaged over entire period to map tendency (struct ural issues)

Admin SubCom

SubCampoake Fina l Source Cate Range

C@""I Methodotog v noie.. po,lte

lltle Woitht tomPo•lk!Tltl<i

Wei:ht

FCWS~' ET food ••curi1•1 Prcr,ic<f'./

O\Jt lookdala 1000-2014 LivehMod ,,., erai,,d 11'C ..:ore rer ,one o,~r ent~e t i""' pe-riod . ,~ - F.ood See"rfty =-Sl"'ISme d'Alerte Pr€coce

Corcm,ne ,cote gene r~•d ,V tctalling rar,,t, r of time, (SA/') \IUlnera', le 1009.2014 Cnmmrn<

commu11••·;d•n~~od • • , ulr.orable <luring lime pef;,,d, 5r.t:

commu ne, Faad MirMslr.ful ')oriculture - imlf! • .l(fl.q ROgiM ~""""£"d % per r,,ginn n,,., hnth '/"-'" 4.l'II, Secur lt¥/Ag . Bu-kin, Faso

Th~ snll aganlc ,_..,rboo predlaed mes n furl he l;t Prm:luct iv it.,.

lr,temational Soll si:anCard depth (·l-Scmt in,1 ,rn,..,ard dep1h ($,- !Semi

A!lril:uMural -Ref~r~l'C1' an,J Prod u,tivity lnlormatlon Cei tre -

2013 Ras:.tr .a nd >rd •landard '" world >Ool lnf<:rmation

depth (15- >0<m] were summd fur- anappm,cimation <lTth~ s<ill org,, n l: carhan In l"fl soil, which i, n-lD<m.

Am.1, ir,, Statisti~ue d, 2a!D-201J Provinco

The _,ir.~ rotes for ~••~e, 1 lhru 5 were "'"'r•~~d ,nd w. l'ed"cad on nation, le then IC.e,e averages were avern.ged ~·,e, the 4 years.

llrnlng th<> ~00~ ~"""~ e,pry,,n~ r,,,,r lh~ •l'JI nl 3 Liter acy Ra t es

cen,os Data 1006 comm....., were as,ed whecher or not t~e ·responden, oould read -,nci wr~• in anv lae~uag,o. 8urldnobe HO\Jseho ld

lworagod P""•rtv rate< 11r region 1oappm,c;m~te IJUin~ (Onditicn, 5u(',"@y 1003. 1009 R•i•on '" ti:C3~MI

gc r.era l tenden c·1 Poverw

6a nqtJP,. C,,nt rak> ~ !'tats tad)..,tedl 15¼

A per eai:itol ,mount .,,, " lcui,1e<11o, remitta.-.:e, de l" Al~que de l'Oc,e;I 2011 Re Bi On

~•rre{:ion. '" IBcrAO) Oemo;ira phi< and Health 2[1,)3. 2010

CluSler ~nt .. interpo'.ateQ to Raster usin~ Xrlg in~ Method (both JJ¼

SuNe ., IOIISI e•= ,,.,..,r_, fflr each eara,era "' Demographic and Ilea Ith Clusicr Points lnrorpD'ate<l to Raster ..-, Ing ~rtg1n~ McthOO (Dnth

SuNe.\« (DH<;I 1003, 2010 ~oinr, "'""'' fn, oach year a,orag,,<tl

.. Poverty

Mini,:,yct Live,toi:k • 2012

6urkN ~a,o Prtll'ince Proje cted Livesluck figur~s mnverted lu TW ,~

Du,111~ t ile 1006 censu:;, """'Y fan,ily was asbs:111 lhey had mOl'ed la The lasayear, and n so, tmm where ro

whoro. lmmigr,tion ~•••• v1ere u,ee "8 pro,~ for Census Data coo; Commu-re vul aerabiliC'1 bawd oo 1he a,sumption that generally ,~

,nne, rhat.arc less •mlncr.1ble .a re ,r.o.'1! ,atlractlw l<Iffcr

mo•~ o pportu nltleS1 and thus h'"'" higher rates of imm'8["~.on .

Domo~,optic ,n, Ho, lth C1u,1or f'oirits 1nterpo!,te,;l to ~utor w ing ,;,ig,n~ Method \ooth Survev, [DHSI

l003, lOlO Point,. rasters for each ·1ear a,e ragedl '"' Ellmncuo

Yea~y repon from The Mlnlsl,v of Health that ~••ll~Ser,icw, -MCniSln,' of Heatt~ • ,Oil r,o\llnce. calcu!at<es how many people In each """""'e a<e JO ,m -Bu,kiru Fa,o

er m~r• awav from , health ,..,,~,

AFRIPCP W14estimote Raster _.., . .._., ot low,.,- po;wlati <>1 at• oor.,ido,;,,d a,;, po,,•1 to vs lade oJ acc~ss to seMces l=note nessl Service Access

Demo!)rophinnd H,.ltt, l.00:1, 2010

Clu,1er f'oints ;n te,- pQ'oted to Ra >ter u, lnR ( ri~ :n; Met liod (both

"" $"""'~' [DH>I ,.~ ro,i,,,. lot • • <h ·1 .. , o,• ro,;edl Demo;;raphic and Health

1tm • .01u Cluster f'oint> interpolated to Raster <£ ins Krigin o Method (Doth

l.3½ SU s IO!l5i e•~ "''"'" loreach ear ••••• · ••

oemoarapllk an~ Health l 003, 2010

(1(,,1,r F'Oints 'interp~ ,te<i to Ra ;t,r llSiOg ~~g in~ Method (bOlh -5ur,,ev<[CHSI eo~ raster< fur e•<h y_ear a , ~raged1

CO<ffic,on t oi variation of r,, inf; II Oata w..s ca l<ulatod acroz entire ,ime periccl for the mon:h cl May

[plant"'1: t ime) and !he monrh cf Ottoberi har,,etJ. The rnlRPI .... ,e,, C. funk ot Ian, 1%1 ·

Rascer variat ion m ,,.;nf•II du,;o-'l •h.,• tv,:o m'""1\h, 1, -" Sept. 2011 consi, ere d crlicat The lwo ra,ters ""'" lhen

averai,-ed to hi~lilii:1,1 t he mu, t vu l,.,,able n,re; i., ,eg,CO, .to ,oinfa'.l ,,.,;,bi! ~v and i t, affep on ag

amduc1 ion.

11ve••ll' ,~mi,erature duri,g eaeh reio1· si,a;,,., VJ-.5) Recur rent

Uni,eroily -0/ <i6\ k ·,gr ;, •, Z000-2011 <>ve r OOlirtr time i;erio~ wa>~vecago~ :OIIOU li"Mr.l l a :rna ti , l\eser•r<+, Unit (JJt.S-rdilly R"\er rainy ,~ae<n av<,ra~e \ernperaluro. ~l\eniver•~• ~- Cli mate 5hock

(UEA,11'.AU) ...,,,on) t»mperat,.-e durlfll: !9"'Y season can be ,:,insi<leted a i:=1• 1, plan! ,1,.,, at highor tomporoture,.

famine ~• rl v War.sin~ l<lne> With s ~orter ,air.~ ;eosons a re con,ide/ M more

s1·srem, Nmwo-1: 1.JOl-2010 Rasrer vui,-.,,a ble . ,~

(ft'WSNHi

CHIRP> ,;a,o,e~ C. funk •t Ian, '1961 · Calc l-i•te,J .,.,., wth ;;....., period.. Zor.e, d l""'er

Rascer .Ne rage total precip1:.otion are consioered rn:,re -" Sept. 2014 vu1nr;ra b1e ,

41.of incident, per location p!u , ~umbe r cffa,alitle.

~,med ( <inflict Cc,,;ati,a !'< 1/1/1~,o

multipiod t,1, tv,o wa, u;ed lo g,,ee,ate, "<o"fti<t Ev;,,~ Dala (AClECI

7/10/2014 Po1r,1001a l e, .. ~ .. pe, PPinl l"'a l;,,11, All !we, ofWo\ft iCl fr"1, ,,.

Oatahase ~•taMse •.11e"' lnilurled (le- prnt,,.;r,, , mien £'0'Jp;, Historic Siles of

police, e<mic militie->. etc.). Canfl kl

I W0tlo Food Froi;ramme Total ~• count was used per locat;,,n as• prcxy to

(WFl'I 2014 Poir,t Da ta cu,,nid b•c•-~f pu,ulaUon m;l r~><...-e< pres.u,., li¼

crEated by- refuse• preserace.

Demo,rophic and H.,lth 2010 Point Dot•

Points inl-<roolaled to R"ter us ing ~ri8 iag Method (both

I - Heal th Shock Sur,e,,. IDl!SI rasle,s fu, each ·1ear a,orago,;1]

Puinl Udla repre sent< a ll markets l urveied rnur,lhly !o< ~r ice,. A·,·erage marl<et prices"'"" c, 1,u!ated for all

m•rl.:.t, ov•-"tim• du~nel••n '"'"'"- LI!•n ~••on i, when t.i~h pri<es have the biggest ~ti"" impact on

SIM}SO~'AGESS 2~04-2014 Pelot Daia house hold fuod se curity. 1'<:trn d:ara was ln:erpolated = 10 ~aster u,li,g Krlglng Me,hod fbnth r;;:ers lor each

v.,.r aWtra;i•d). ib>latiWI w•ightieg lor-•ach commodil'/ was ca lc ulated prol>C'1 iooa ll¥ to each comrROO~ie,

Recurrent P r ice production level.

Shocks SIM,(SO~AG.SS 2U 04•l 014 Point o at•

SAME MfTHODOWGV NOTES FOR ALL PlllCE CATA IN COMP051H(SEENOlE>FOR MILlET PRl([S]

5~"1 Carn Prlaes ~,

SIM/50t1'1GE55 2004·2014 Point D&t• SAME MfTHDOOLOGf NOTES fOR ALl ~ ICE DATIi IN

'"' COMPOSITE jSEE NOTES FOR MILLET PPJCES]

SIM}SO~' /\GESS lll04-2014 Puir,l 0~1< SAME MITHOOOLOGV NOTES fOR ALL PlllCE OAT~ IN

O:,MFOSI H 15EE NOH:i FOR Mill.ET PRICES] n,

SIM/SO~•oi:;c,5 ;w04-:l014 t'<linlCO>U SAM< METHDOOLOGt NOTES fOR ALL P!l lO: CATA IN ,,.

COMPOSITE ]SEE NOUS fOR MILLET PI\ICES]

Derno,raphic and Healtti 1cm, 20,o

Clu,lOr Point,. interoo:,teQ to Raster uoiag ~ri8"1g Method (l>oth = Surveys [Dl!SI Points r•srer, for each ·1ear auerai:ed) ,t.~emla Demo~raphio an~ Heal th Cluster F'oi,m ioterpd awi to Raster u;in~ Kr.eJae Methcd (OOlh Prevaleoce l.OO'l, WI• li¼

5ur,e'/$ (OHSI Po ints rast<ero fur each ·1ear a, e,aw,dl

5'8nd~di,.d Monitoring Each re gion was polled bf SMARr every oiC.er vear, all

and Asse;sme 11t ol Re :;e f l009·2013 Re&;oa inroi-matlim was averaged 1cgether "' AvorageGAM

3n0 Transil i>ns ISM~RT I ~7 ?> ~-"· Demop;raphic en~ Health Clu.,•r F'Oiots ·interpo' ated to Ra,ter u;inp; ~•l~ ing Method (Oorh

100_3, 2010 '" 5"""''1S IDHSI Points raster, 10, each eara,ora!IOdl Undernutrl t lor,

St>ndodi>ed MITTtonng Each region was pol leC every othorycoc, all ln'o.'fflation

and A,,;,.,,srn~nl ci1 R~:i~f 2~09-2013 Re~iou ,.,., •v~r~J!<d toAAther '" "'~""'II<' andTra oSltlCHIS !SMART]

srun11ng !late< ~¼

Derno.iraphic an:! Health 10)3. WJ.0

Clus1or Points ir,ter po'aleQ.to Raster us ing Kr1!! "1S Ml!lhod (bnth ~0'6

SUM:V, 101151 Points rasters fur each year auc raf:C(fJ

Composite Top lnde• ~., -~In Compooit> W@ij:ht

31%

23%

Reslllence

'"' Capacity

31%

15%

44%

Exposure

)1% !Shock ~ So 21% stresses )

11%

33%

20%

WelHlelng Ouloome

~0%

BO%