Linking regional stakeholder scenarios and shared socioeconomic pathways: Quantified West African food and climate futures in a global context

(1)

Linking regional stakeholder scenarios and shared socioeconomic pathways: Quanti ﬁ ed West African food and climate futures in a global context

Amanda Palazzo

^a,

*, Joost M. Vervoort

^b,c,d

, Daniel Mason-D ’ Croz

^e

, Lucas Rutting

^b,c

, Petr Havlík

^a

, Shahnila Islam

^e

, Jules Bayala

^f

, Hugo Valin

^a

, Hamé Abdou Kadi Kadi

^g

, Philip Thornton

^c

, Robert Zougmore

^c,h

aInternationalInstituteforAppliedSystemsAnalysis(IIASA),EcosystemsServicesandManagementProgram,Schlossplatz1,A-2361Laxenburg,Austria

bEnvironmentalChangeInstitute(ECI),UniversityofOxford,SouthParksRoad,OX13QYOxford,UnitedKingdom

cCGIARResearchProgramonClimateChange,AgricultureandFoodSecurity(CCAFS),UniversityofCopenhagen,FacultyofScience,DepartmentofPlantand EnvironmentalSciences,Rolighedsvej21,DK-1958FrederiksbergC,Denmark

dCopernicusInstituteofSustainableDevelopment,UtrechtUniversity,Heidelberglaan2,P.O.Box80.115,3508TCUtrecht,TheNetherlands

eInternationalFoodPolicyResearchInstitute(IFPRI),EnvironmentandProductionTechnologyDivision,2033KStreet,NW,Washington,DC20006-1002,USA

fWorldAgroforestryCentre(ICRAF),WestandCentralAfricaRegionalOfﬁce SahelNode,BPE5118,Bamako,Mali

gInstitutNationaldelaRechercheAgronomiqueduNiger(INRAN),BP429,Niamey,Niger

hInternationalCropsResearchInstitutefortheSemi-AridTropics(ICRISAT),BP320Bamako,Mali

ARTICLE INFO

Articlehistory:

Received3June2016

Receivedinrevisedform16November2016 Accepted4December2016

Availableonlinexxx

Keywords:

Agriculture Climatechange

Representativeagriculturalpathways Sharedsocioeconomicpathways Stakeholders

WestAfrica

ABSTRACT

Theclimatechangeresearchcommunity’ssharedsocioeconomicpathways(SSPs)areasetofalternative globaldevelopmentscenariosfocusedonmitigationofandadaptationtoclimatechange.Tousethese scenariosasaglobalcontextthatisrelevantforpolicyguidanceatregionalandnationallevels,theyhave tobeconnectedtoanexplorationofdriversandchallengesinformedbyregionalexpertise.

Inthispaper,wepresentscenariosforWestAfricadevelopedbyregionalstakeholdersandquantiﬁed usingtwoglobaleconomicmodels,GLOBIOMandIMPACT,ininteractionwithstakeholder-generated narrativesandscenariotrendsandSSPassumptions.Wepresentthisprocessasanexampleoflinking comparablescenariosacrosslevelstoincreasecoherencewithglobalcontexts,whilepresentinginsights aboutthefutureofagricultureandfoodsecurityunderarangeoffuturedriversincludingclimatechange.

In these scenarios, strong economic development increases food security and agricultural development.Thelatterincreasescropandlivestockproductivityleadingtoanexpansionofagricultural areawithintheregionwhilereducingthelandexpansionburdenelsewhere.Inthecontextofaglobal economy,WestAfricaremainsalargeconsumerandproducerofaselectionofcommodities.However, thegrowthinpopulationcoupledwithrisingincomesleadstoincreasesintheregion’simports.ForWest Africa,climatechangeisprojectedtohavenegativeeffectsonbothcropyieldsandgrasslandproductivity, andalackofinvestmentmayexacerbatetheseeffects.Linkingmulti-stakeholderregionalscenariosto theglobalSSPsensuresscenariosthatareregionallyappropriateandusefulforpolicydevelopmentas evidencedinthecasestudy,whileallowingforacriticallinktoglobalcontexts.

1.Introduction

Climatechangeisasigniﬁcantsourceofuncertaintyforthefood security,healthandlivelihoodofthepoorinmanyoftheworld’s vulnerable regions, interacting with and compounding other

sources of uncertainty such as socioeconomic development, political stability and the effects of widespread ecosystem degradation(IPCC,2014).Amongthemostvulnerableregionsis WestAfrica,where75%ofthepopulationoftheﬁfteencountries that aremembersoftheEconomicCommunityofWest African States(ECOWAS)liveonlessthan$2adayandmorethan35%of theregionalGDPisderivedfromagriculturalproduction(Hollinger andStaatz,2015;WorldBank,2011).Thoughtheregionishometo currentlylessthan5%oftheworld’speople,inthefutureitmaybe

* Correspondingauthor.

E-mailaddresses:palazzo@iiasa.ac.at,ampalazzo@gmail.com(A.Palazzo).

http://dx.doi.org/10.1016/j.gloenvcha.2016.12.002

xxx–xxx

Pleasecitethisarticleinpressas:A.Palazzo,etal.,Linkingregionalstakeholderscenariosandsharedsocioeconomicpathways:Quantiﬁed ContentslistsavailableatScienceDirect

Global Environmental Change

j o u r n al h o m ep a g e: w w w . el s e v i e r . c o m / l o c at e / g l o e n vc h a

(2)

thefastestgrowing(JiangandO’Neill,2015;KcandLutz,2014)and oneofthemostexposedtoclimatechangeduetoitsdependence on(rainfed)agriculture,and the estimatednegativeimpacts of climatechange(Leclèreetal.,2014;Mülleretal.,2011;Roudier etal.,2011).Thoseinvolvedingovernmentpolicy,privatesector investments,civilsocietyactionandotherstrategicprocessesmust considertheinteractinguncertaintiesofdevelopmentandclimate change in an integrated fashion when planning for the future (Vermeulenetal.,2013).

Scenario-guided planning allows decision-makers to engage with uncertain futures and assess and improve the feasibility, ﬂexibilityandconcretenessoftheirplans(Vervoortetal.,2014).

Theinternationalclimatechangecommunityisdevelopingasetof globalscenarios, consistingofvariouscombinationsofradiative forcingscenarios(RepresentativeConcentrationPathwaysorRCPs) andsocioeconomicand policyscenarios(Shared Socioeconomic Pathways;SSPs,andSharedPolicyAssumptions;SPAs)thatwhen combinedcanbeusedtoexaminetheimpactsofclimatechange.

Thesescenariosalsoprovideaglobalcontextand/ortemplatefor processesatlowergeographicallevelsthatseektousescenariosto guideregional,nationalor sub-nationalplanning(O’Neilletal., 2014). Conversely, there is scope for sub-global processes to complement the shared socioeconomic pathways (SSPs) with moreregionalcontextualizationofassumptionsandresults,even when using scenarios in the global setting. Regionally speciﬁc scenarios serve to assist policy makers in developing robust agricultureandclimateadaptationstrategies,whilealsoproviding thescientiﬁccommunity workingat theregional,national,and sub-nationallevelwithmultiplepathwaysfordevelopmentthat canbedisaggregatedorlinkedtoadaptationassessments(Antle etal.,2015;Kiharaetal.,2015;Valdiviaetal.,2015).

The frameworks to develop the global SSPs have been thoroughly documented (O’Neill et al., 2014; Schweizer and O’Neill,2013;van Ruijvenetal.,2014;vanVuuren etal.,2014), linkedtopreviousscenarioassessments(vanVuurenandCarter, 2013),andrecentlyintegratedwithclimatechangeandquantiﬁed (Riahi et al., 2016). They are just beginning to be scrutinized through regional and national (Absar and Preston, 2015), and humanimpact(Hasegawaetal.,2015)lenses.

Inthispaper,wepresentaprocessinwhichasetofstakeholder- generated,regionalscenariosforWestAfricawaslinkedquantita- tivelytotheSSPsbyusingtheregionalstakeholder scenariosto criticallyexamineandadaptSSPassumptionsmadefortheregion.

Thisway,asetofscenarioswascreatedthatfocusesprincipallyon regionalchallengesbut hasbeen madecoherentwiththeSSPs (ZurekandHenrichs,2007),allowingforaglobalsituatingofthe scenarios.Theresultingsetofscenarioswasdesignedtobeused for planning by policy makers (in the widest sense, including privatesectorandcivilsocietygroups)at nationalandregional levelsandhavebeenusedforthispurposeinanumberofplanning processes,amongwhicharenationalpolicyguidanceprocessesin Burkina Faso and Ghana. The process was led by the CGIAR Program on Climate Change, Agriculture and Food Security (CCAFS).

Wepresentthisprocessas1)anexampleofusingglobalmodels toquantifyregional scenarios to balancethe needfor regional perspectiveswiththeneedforconnectionstoglobalfutures;and 2)tomorespeciﬁcallyexaminetheimplicationsforagricultureand foodsecurityinWestAfricaunderfuture climateandsocioeco- nomicuncertainty.

Inthispaperwewillﬁrstdescribeourparticipatoryscenario development methodology, including how the scenarios were linked across levels and quantiﬁed. Then we will present the resultingregionalscenarios:thesocioeconomicdriversofchange andthequantitativemodelingresults,highlightingthelinktothe SSPs by their narratives, scenario drivers, and challenges to

adaptation.Finally,wewilldiscussthebeneﬁtsanddrawbacksof ourapproachoflinkingregionalandglobalscenariosandcompare ittoalternativeapproaches.Anoteonterminology:followingCash etal.(2006)weuse‘level’ratherthan‘scale’todescribelevelssuch as‘regional’and‘global’.

2.Methodology

2.1.Mainprocessobjectivesanddesignchoices

Scenarios are hypothetical futuresexpressed through narratives, numbers or other means (visual, interactive), to explore differentdirectionsofchange(vanNotten,2006;Vervoortetal., 2010).TheCCAFSWestAfricascenariosprovidegloballycontex- tualized meso-level futures for policy guidance at regional, national and sub-national levels acrossWest Africa. A number of policy guidance processes were co-developed between the projectresearchersand policymakers,and designedtodirectly examineagivenpolicyorplaninthecontextofmultiplescenarios, leading to an assessment and an improvement of the plan’s robustnessinthefaceoffutureuncertainty,basedonnewinsights comingfromtheexaminationoftheplanthrougheachdifferent futurescenario(Vervoortetal.,2014).

Thisstrongfocusonregionalandnationalpolicyguidancehas consequences for how the regional scenarios and global SSPs shouldbelinked.Toensurepolicyrelevance,driversconsideredto be the most important at the regional level should frame the scenarios,andpolicymakersshouldbeinvolvedintheidentiﬁca- tion of these drivers and the development of the scenarios (WilkinsonandEidinow,2008).Multi-levelscenarioprocessescan exhibitdifferentdegreesofintegrationofscenariosacrosslevels, though they are often conceived through a top-down process (Biggsetal.,2007;Koketal.,2007,2006a,2006b;Shawetal.,2009;

Kok et al. In Review).Zurek and Henrichs (2007) describe the differentpossibledegreesoflinkagebetweenscenariosorganized atdifferentgeographiclevels,from‘equivalent’(thescenariosare the same at different levels) to ‘independent’ (unconnected scenarios). We start with regional scenarios that Zurek and Henrichswouldcategorizeas‘comparable’totheglobalSSPs in thattheyhaveasimilarscopeofconcern,buttheframingdrivers and assumptions of the scenarios are not connected. This comparable regional set of scenarios was then quantiﬁed to provideinputsforglobalmodeling,inaprocessthatmappedthe regional scenarios to the global SSPs in terms of quantitative drivers. We will argue that this process moved the regional scenariostowardbeing‘coherent’withtheglobalSSPs–meaning thattheregionalscenariosandtheglobalSSPsmaptoeachotherin termsofcontentandassumptions.Havingtwodifferent,compa- rablestartingpointsforthescenariosateachlevelmeansthatthe regionalscenarios providean independent,regionally grounded perspectivefromwhichtheregionalassumptionsfortheSSPscan beexaminedandadapted.Atthesametime,movingthescenarios from comparable toward coherent through the quantiﬁcation processmeans thatthescenarios can besituated inglobal SSP contexts whichisessentialtounderstandingthedevelopmentof WestAfrica’sfutureinthefaceofglobaldriversofchange.

2.2.Scenariodevelopmentandframework

The CCAFS scenarios process in West Africa started by examining, with regional stakeholders, the impacts of future climateandsocioeconomicdriversonfoodsecurity,environment and rural livelihoods. Scenarios were developed over three separate workshops. Regional stakeholders took ownership of the process by offering information on the relevant drivers of changeastheyrelated toagriculture,foodsecurityandclimate

xxx–xxx

(3)

adaptation/mitigationinthefutureofWestAfrica(inWorkshop1), creatingnarrativesofeachscenario(inWorkshop2)andproviding semi-quantitative estimates for scenario variables and model inputs, in close collaboration with the modeling teams (in Workshop 3). Although Mali, Niger, BurkinaFaso, Senegal and GhanaarethefocuscountriesfortheCCAFSWestAfricaprogram, participantswerealsoincludedfromregionaland international organizations to provide a regional perspective. The regional participatory process depends on having the right balance of participantswithdiverseinterests.Tothisendweaimedforamix of participants across focus countries, sectors, disciplines, and gender (for more detail on stakeholder backgrounds see Appendix A in Supplementary materials). Additionally, it was important toselectparticipants who couldboth offer in-depth expertiseandinﬂuencestrategicprocessesintheirorganizations.

94participantsfromgovernments(agricultureandenvironment ministries, meteorological institutes), research organizations, nationalandregionalcivilsocietyorganizations(CSOs),interna- tionalnon-governmentalorganizations(INGOs),academiaandthe mediaparticipatedin theoriginaldevelopmentofthescenarios overthethreeworkshops(Palazzoetal.,2016).Theteaminvolved intheselectionofthestakeholdersincludedkeyprojectpartners from the West African Climate Change, Agriculture and Food Security program (CCAFS), from the Agriculture and Rural DevelopmentofﬁceoftheEconomicCommunityofWestAfrican States (ECOWAS), and from the West African Council for Agricultural Research and Development (CORAF/WECARD). To- gether,thisorganizingteamwasabletouseanextensiveregional networkfortheinvitationofparticipants.

The CCAFS scenarios were created to represent regional developments over time on the way to a 2050 time horizon.

Stakeholdersoutlinedfourscenarios,structuredalongtwoaxesof uncertainty,usingnarrativeflowcharts, conceptualmaps,storylines,and a range of trend indicatorsincluding information on governance,agriculture,foodsecurityandlivelihoods.Participants selectedtwoaxes froma broadsetof futureuncertainties that weredeemedmostrelevantanduncertainfortheregion.Theaxes identifiedwere1)whetherstateornon-stateactorsdominatethe regionaldevelopmentprocess;and2)whethershortorlong-term priorities dominate policy-making. Other drivers, such as eco- nomicandpopulationgrowth,wereconsideredimportant,butless uncertain which meant that they were selected to play a contributing role in each of the scenarios. Their impacts were differentacrossthescenariosand dependedontheothermain driver/axesstates,but thesedriversdidnotdefinethescenario.

Both drivers deﬁning the scenarios are essentially governance drivers–participantssawregionalgovernanceasthemostcritical anduncertaindimensionofWestAfrica’sfuture,that could,for instance, determinethedirection of investments into development, the useof resources,and other drivers. With ‘non-state actors’, participantsmeantboth private sectorand civilsociety actors – the resulting scenarios where non-state actors play a prominentroleresultedinadynamic interplaybetweenprivate sectorandcivilsocietyactors.Whileeachscenariodescribesthe future to 2050, in the scenarios where governance focuses on short-termpriorities,this doesnot meanthatthescenariosare themselvesshorter.Instead,throughoutthetimeperiodofthese scenarios,short-termconcernsaregivenpriority.Thisresultsina relativelackofinvestmentinlong-termprojects.Thusshortcycles of growth and investment make developments in the two scenarioswiththischaracteristicmoreunstable.

Self-Determinationis a scenariowhere stateactorsdominate developmentand agendas are focused onthe long-term. Cash, Control, Calories is a scenario where state actors dominate developmentandwithashort-sightedagendasetting.CivilSociety to the Rescue? is a scenariowith non-state actors, such as the

privatesectorandCSOs,dominatingregionaldevelopmentwitha long-termstrategicagenda.SaveYourselfisascenariowherenon- stateactorsdominatetheregionaldevelopmentandtheirfocusis ontheshort-term.CartoonrepresentationsoftheCCAFSscenarios (basedonthescenarionarratives)shownalongthetwoaxesare presented in Fig.1. Narrative summaries of the scenarios and detailsof thedevelopmentprocessarefoundin AppendixA in Supplementarymaterials.

2.3.QuantiﬁcationoftheCCAFSscenarios

Followingthedevelopmentofthequalitativescenarios,stake- holdersprovidedadetailedlookintothescenariosbysignalingthe logicofchangeandmagnitudeofchange(givenas+,=,and )fora setofindicatorsthatrepresentthescopeofinterestforthefuture of food security, livelihoods, and environments (A full list of indicators appears in Appendix Table B1 in Supplementary materials). To fully quantify the scenarios, a subset of these indicatorsweregivennumericalvaluesandusedasmodeldrivers in two global partial-equilibrium economic models of the agriculturesector–GLOBIOM(Havliketal.,2014), developedby theInternationalInstitutefor AppliedSystems Analysis(IIASA), andIMPACT(Robinsonetal.,2015),developedbytheInternational FoodPolicyResearchInstitute(IFPRI).

2.3.1.Quantiﬁcationofthescenariodrivers

The objectives ofthe CCAFSregionalscenarios development process are focused on policy engagement and planning and quantiﬁcationofthescenariosbyglobalmodelsbalancesregional priorities and perspectives within a global context. Quantiﬁed scenarios provide a tool to measure and examine the relative impactsofregionalsocioeconomicdevelopment,forinstancehow populationgrowthcanaffectecosystemsthroughtheexpansionof cropland.Weusethestakeholdergeneratedinformationfromthe trend indicators and narratives as the main link between the qualitative scenarios and the models. After establishing a quantitative link, werun themodels overthe time period and examinetheimpactsofthescenarioassumptions.Inthefollowing paragraphs we describe several steps taken to interpret the scenariotrendindicatorsfromsymbols(+and )intonumerical valuestobeusedasdriversinbothmodels.

2.3.1.1. Selection of indicators to quantify as drivers and ensure consistencyoftrendsamongscenarios. Ofthefullsetofscenario indicators,we selectedandquantiﬁedthosewhich signiﬁcantly impacttheagriculturesectorduetotheimportanceofthesectorin theregionandthedetailedrepresentationofthesectorwithinthe models: population, GDP, technology-driven improvements in cropand livestockyields,and farminputcosts.Becausesmaller groups of stakeholders directly gave the logic, direction, and magnitudeofchangeoverdifferenttimeperiodsforeachindicator, checks andadjustmentsweremadetothetrendstheusingthe logicandscenarionarrativestoensurethatthetrendindicators wereconsistentacrossthescenarios.Thefullsetofindicatorscan befoundinTableB1intheAppendixinSupplementarymaterials.

2.3.1.2.Comparescopeofinterestforbothsetsofindicators. Asa nextstep,wemappedsimilarindicatorsofboththeCCAFSandSSP scenarios,forexample,“grossdomesticproduct(percapita)”from the CCAFS indicators and “growth per capita and population growth”fromtheSSPs(O’Neilletal.,2015).InTable1(inSection3) wepresentthemappingoftheselectionoftheCCAFSscenarios indicators(column1)withindicatorsfortheSSPs(columns9and 11).Whileonlyasubsetofindicatorswereselectedtoquantifyand useasmodeldrivers,wehavemappedeachCCAFSindicatortothe SSPindicatorsinAppendixTableB2inSupplementarymaterials.

xxx–xxx

Pleasecitethisarticleinpressas:A.Palazzo,etal.,Linkingregionalstakeholderscenariosandsharedsocioeconomicpathways:Quantiﬁed

(4)

2.3.1.3.MappingtheCCAFSscenariosontotheSSPscenarios. There are some key differences between the regional West Africa scenariosandtheSSPs.TheSSPswerecreated bya community ofresearchers;theregionalWestAfricanscenarioswerecreatedby atransdisciplinarygroupof regionalstakeholders.However,for linking scenario sets, content matters most. Here, the main differenceis that theSSPs havebeen framed in terms of their consequencesforadaptationandmitigation,whiletheWestAfrica scenarios have been framed by their drivers in this case, dominantmodesofgovernance.TheSSPsaredeﬁnedbythelevel ofchallengetoclimateadaptationand thelevelofchallengeto climatemitigationandconstructedupontwoaxeswheretheend pointsofeachaxis,highandlowrespectively,combinetodeﬁne the“challengespace”of thescenario(O’Neilletal., 2014).This differenceinframingmakesthetwosetsofscenarioscomparable rather than coherent in Zurek and Henrichs (2007) terms, but coherence between the socio-economic assumptions in the scenarioscanbeestablished mappingtheregionalandglobal scenariostoeachotherthroughtheirnarratives.Weemployeda

“one-to-one”mappingsystem(ZurekandHenrichs,2007)guided bythenarrativesandtrendindicatorstomaptheCCAFSscenarios ontotheSSPswithinthecontextoftheSSPnarratives(O’Neilletal., 2015). We present the results of the mapping of each CCAFS scenariotoanSSPinSection3.1.

2.3.1.4.QuantifyingtheindicatorsinthecontextofSSPdrivers. The quantiﬁcation of the drivers of the SSPs that focus on the challengesassociatedwiththesocioeconomicdevelopmenthave beenwell-documentedandprovideinsightsintopopulationand urbanization (Jiang and O’Neill, 2015; Kc and Lutz, 2014) and

economic growth(CrespoCuaresma, 2015; Dellinket al., 2015;

Leimbach et al., 2015). Global integrated assessment models (IAMs)usedthemajordriversoftheSSPs(suchaspopulationand incomegrowth)toproducetheﬁrstsetoffullyquantiﬁedglobal SSPs(Calvinetal.,2016;Frickoetal.,2016;Fujimorietal.,2016;

Kriegleretal.,2016;O’Neilletal.,2015,2014;vanVuurenetal., 2014). The modeling teams offered interpretations of the key elementsof the narratives presentedin O’Neillet al. (2015) as model inputs (crop and livestock yields; (Fricko et al., 2016;

Herreroetal.,2014);energysector(Baueretal.,2016))andalsoas modeloutputs(agriculturallandusechange(Poppetal.,2016);air pollution(Raoetal.,2016)).Afterreview,wechosetousethese SSPsdriversasaboundaryconditionorenvelopeofpossiblevalues (van Ruijvenetal., 2014).Followingthe mappingofthe CCAFS scenarios ontotheSSPs,we usedthevalueofthedriverof the respective SSP as a starting point. Then, we used the trend indicatorstoguideandshiftthesevalueswhilemakingacritical comparison between both sets of narratives. The SSP indicator assumptionsaredeﬁnedfortheendofthecenturyratherthanthe CCAFS time period of 2050,this was taken intoaccount when adjustingthetrends.Adetailedlookintothescenariodriversis foundinSection3.2.

2.3.2.GLOBIOMandIMPACT

GLOBIOMand IMPACTare globalpartial equilibrium models with a detailed representation of the agricultural sector. The similarities and differences between modeling approaches, of GLOBIOMandIMPACTinparticular,havebeenexaminedthrough theintercomparisonactivitiesoftheAgMIPprojectwithfocused reviewsofthemodelingofagriculturalsystemstomeetfuturefood Fig.1.CartoonrepresentationsofthefourCCAFSWestAfricascenariosalongthetwoaxesofuncertainty.

Source:DrawingsbyartistAndreDanielTapsoba

xxx–xxx

(5)

Table1

CCAFSscenariostrendindicatorscomparedwithandmappedtosharedsocioeconomicpathwaysandindicators.

CCAFS

ind. Scenario

2010- 2020 2020- 2030 logic for change

2030- 2050 logic for change SSP SSP ind.

SSP qualitave informaon

SSP ind.

SSP qualitave informaon

Col 1 2 3 4 5 6 7 8 9 10 11 12

)atipac rep( tcudorP citsemoD ssorG

Cash, Control,

Calories ++ = Inial boosts are not sustained

as long-term growth + Periodical boosts and plateauing; reacve

SSP5

Growth (per capita)

High

Populaon growth

Relavely low

Self-

Determinaon + +

Some countries already involved in long-term transformaon, others make an eﬀort. Minerals exported/divide between countries, arﬁcial way of changing GDP through services

++

Transion into services and secondary industry, agricultural producon; processing

SSP 1

High in LICs, MICs;

Medium in HICs Relavely low

Civil Society to

the Rescue? + +

Increasing regional stability and strong civil sociees smulate investment, but governments are not able to facilitate investments well.

+

Populaon pressures increase; pressure on educaon; without governments it is diﬃcult to bridge the growing gap between poor, middle class and rich. Climate change makes things worse for the poorest.

SSP2 Medium, uneven Medium

Save Yourself ++ ++

Open market compeon with lile state interference, but also forming of cartels, society overall is worse oﬀ

++

Dynamic growth connues to build though resources have become a constraint; large informal economies

SSP

3 Slow High

secirp tupni remraF

Cash, Control,

Calories ++ +

Some buﬀering, polically movated by governments who need to stay in power, periodical

++ Same as before SSP

5

internaonal trade

High, with regional specializaon in producon

agriculture

Highly managed, resource- intensive; rapid increase in producvity

Self-

Determinaon = =

Fossil fuel prices increase, but government organizing price drops for ferlizer, seeds, subsidies, manpower

=

Trend connues, but dampened by big and smaller renewable energy projects in the Sahel

SSP1 Moderate

Improvements in ag producvity; rapid diﬀusion of best pracces

the Rescue? + + Rising input prices are

unmediated +

Fuel prices increase; technology has a cost; but renewable technologies become more and more available

SSP

2 Moderate

Medium pace of tech change in ag sector; entry barriers to ag markets reduced slowly Save Yourself +++ +++ No reformave change, some

tech improvements +++ Connues to follow global prices

SSP

3 Strongly constrained

Low technology development, restricted trade

egnahcdleiy kcotseviL

Cash, Control,

Calories + + New breeds, new inputs but no

structural investment = Veterinary services decrease SSP 5

agriculture

Highly managed, resource- intensive; rapid increase in producvity Self-

Determinaon ++ ++ Government planning to

support yield increase ++ Trend connues SSP

1

Improvements in ag producvity; rapid diﬀusion of best pracces Civil Society to

the Rescue? + ++

Demand for meat drives private sector investment.

Social entrepreneurs work with professionalized communies.

++ Trend connues as before – demand grows.

SSP 2

Medium pace of tech change in ag sector; entry barriers to ag markets reduced slowly Save Yourself = =

Very patchy investment, no government planning; no veterinary control; current trend= decrease

= No improvement SSP

3

Crop Yields

Cash, Control,

Calories + ++

More rainfed area expansion under low-input with lower concern for sustainability.

Eﬀects of land degradaon, climate change, low intensiﬁcaon; conservave esmate as irrigated yield gap may be lower

++ Lack of long-term thinking, hard to repair once degraded

SSP5

agriculture

Highly managed, resource- intensive; rapid increase in producvity

Self-

Determinaon + +++

Capital investment takes me; Strong government will but few resources; market signals drive yield investments supported by governments.

Backup irrigaon systems implemented; Water a liming factor but may be available at a connental level; technology a limited factor; new tech needs strong organizaon; beer management; culvars;

hybrids; beer water management

+++

+

Aer period of trial and error, instuons are now strong enough, and capacitated enough to cause real improvements; governments are invesng in research and development; Once irrigaon schemes are widely implemented, yields can increase strongly

SSP 1

Improvements in ag producvity; rapid diﬀusion of best pracces

the Rescue? + ++

The professionalizaon of farmers supported by social enterprises and CSO is combined with more eﬀecve communicaon tech ; though beneﬁng largely those who already have some capacity for yield increase; Conservave esmate as yield gap lower in irrigated crops

+

Adopon of new strategies through comms, agricultural technology etc. Big mulnaonal GMO - two responses: 1. resistance and alternave farming; 2. "home- grown" responsible GMO technologies.; small schemes can be set up by private sector, but no big government investment in irrigaon

SSP 2

Medium pace of tech change in ag sector; entry barriers to ag markets reduced slowly

Save Yourself + +

Private sector will focus on cash crops; farmers lobby for food producon; private sector some investment but not signiﬁcant

+

increases only for those rainfed crops that are economically viable on a large scale like biofuels; staple foods suﬀer

SSP 3

Note:TheﬁrstsevencolumnsrepresentthestakeholderproducedlogicanddirectionofchangeforindicatorsoftheCCAFSWestAfricascenariosprocessthatwereusedto developthemodeldrivers,withadjustmentsmadetoensureconsistencyamongthescenarios.TheCCAFSindicatorsappearingincolumn1fallwithinsamescopeastheSSP indicatorsfromcolumns9and11.ForacompletemappingofallCCAFSindicatorsandSSPindicatorsseeAppendixTableB2inSupplementarymaterials.Anadditional exampleofthemappingofindicatorsbetweenscenarioscanbefoundAppendixTableB3inSupplementarymaterils.

xxx–xxx

(6)

demand(Valinetal.,2014),theimpactsofclimatechange(Nelson et al., 2014b), and land usechange (Schmitz et al., 2014).The quantiﬁcationoftheCCAFSscenariosbeneﬁtsfromtheuseofboth modelsowingtotheirdifferencesmodelingapproaches.Outputs fromthescenariosmodeledbyGLOBIOMmayproveusefulasan inputfor modeling ofregionalimpactassessmentsbecausethe modelconsidersmultiplemanagementsystems,or technologies, thebiophysicalenvironmentof production,orclimates,andthe socioeconomiccontextoftheregion(Antleetal.,2015;Havlíketal., 2015;Leclèreetal.,2014).IMPACThasalonghistoryofscenario analysisofalternativefuturesintheglobalagriculturesystem,and recentmodelingimprovementshaveexpandedthecommodities andcountriesthatcanbedirectlyanalyzed(Nelsonetal.,2010;

Robinson et al., 2015). Appendix Table C1 in Supplementary materialspresentsthemainsimilaritiesanddifferencesbetween bothmodelsusedforquantifyingthescenarios.

2.3.3.Climatechangeimpacts

WestAfricaishighlydependentonagriculture,predominantly rainfedagriculture,whichmakestheregionparticularlyvulnera- bletoachangingclimate.Thestrictlybiophysicalimpactsoncrop production due to changes in climate have been examined extensivelywithin the model intercomparisonprojects, AgMIP and ISI-MIP,through globally-griddedcropmodels(Müller and Robertson,2014).ForWest Africa, analysesof impacts,through cropmodels,as well as throughempirical study, ﬁnd that the negativeimpactsofclimatechangeoncropyieldsareconsistently negativeacrosstheclimateandcropmodelingresults,thoughthe magnitudeofimpactsremainsuncertain(Jallohetal.,2013;Müller etal., 2011;Müller and Robertson,2014; Müller, 2011;Roudier etal.,2011;Sultanetal.,2013).

Thebiophysicaleffectsfromclimatechangeonagricultureare appliedhereconsistentlywiththeSSP/RCPframework,whichdoes notcontainanexplicitlinkbetweenthesocio-economicscenarios and climate change impacts, but rather suggests to test the differentclimatechangescenariosunderseveralsocio-economic scenarios. van Vuuren et al. (2014),Rothman et al. (2014) and O’Neilletal.(2014)suggestthatclimaterelatedbiophysicalfactors shouldnotbeelementsoftheSSPs,butusedincombinationwith SSPsandclimatepoliciestodeﬁneanintegratedscenario.Inthe CCAFS scenarios, climate impacts are not “better” or “worse” amongthescenarios, rather,climateimpacts areexaminedasa forceoutside thescenario.We considerthe projections offour GeneralCirculationmodels(GCMs)availablethroughtheISI-MIP project(Rosenzweigetal.,2014;Tayloretal.,2011; Warszawski et al., 2014)along witha constant 2000 climate. RCP 8.5 was selectedbecause together with the current climate scenario it allowsthescenariostoexplorethemostextremetrendenvisaged for climate futures averaged across multiple crop models and assumptionsoftheimpactsofCO2response.RCP8.5iscombined withSSP2forthesocioeconomicassumptionsfortherestofthe world.In principle RCPs and SSPsare completelyindependent dimensions (van Vuuren et al., 2014). However, more recent quantiﬁcationofemissionsunderSSP2inmodelensemblesuggest RCP8.5couldbepessimisticunderSSP2economicdevelopment, althoughnotimpossible(Collinsetal.,2013;Frickoetal.,2016;

Raupachetal.,2007;Sheehan,2008;vanVuurenandRiahi,2008).

Itisworthnotingthattherangeofclimateeffectsfortemperature increasesofRCP8.5overlapswiththerangeforRCP6.0in2050 (+0.8–1.8for6.0and+1.4–2.6for8.5in2050)(IPCC,2013;Riahi etal.,2016).

Using simulations of crop growth from two process-based globally-griddedcropmodelsthatconsidertheconditionsofthe futureclimates,weapplytherelativedifferencesofcropgrowth duetoclimatetothecropyields(Leclèreetal.,2014;Mosnieretal., 2014; Müller and Robertson, 2014; Nelson et al., 2014b). The

scientiﬁccommunityhasyettoreachanagreementonwhether thepotentialbeneﬁtsfromincreasesinCO2canbetakenupand usedbycrops,especiallyiftemperatureandprecipitationreduce cropyields, therefore we consider a multi-model approach by including two globally gridded crop models with different assumptionsonCO2fertilization.TheimpactsofCO2fertilization on crop yields is included in the EPIC (Environmental Policy IntegratedClimateModel)cropmodelingsimulationsusedwithin GLOBIOM,whileIMPACTsimulatesclimateimpactswithoutCO₂ fertilization using DSSAT (Decision Support System for Agro- technologyTransfer).TakentogethertheyieldsfromGLOBIOMand IMPACT can show the potential range of the biophysical and economicimpacts oncropyieldsfromclimatechange(morein AppendixF).

2.3.4.Representativeagriculturalpathways(RAPs)

Nationaland subnational impactassessmentsthat represent farm systems and households of small geographic units often requireagloballyconsistentmarketequilibriumforcommodities, which are produced by global or regional economic models (Valdiviaetal.,2015).TheintegratedscenariosoftheSSPs/RCPs provide modeling communities with the global, and to some extenttheregional,sector-specificstorylines.Thesestorylinesas they pertain to agriculture serve as global representative agricultural pathways (RAPs). At the same time, researchers focusingonadaptationchallengeshavedevisedsomelocalRAPs specific to particular contexts in Africa, although these are disconnected from a global consistent framing (Antle et al., 2015; Valdiviaet al., 2015).The CCAFS scenarios, quantified by GLOBIOMandIMPACT,examineregionalstakeholderdevelopment pathwayswithinthespaceoftheSSPsofferingthefirstglobally coherent,regionallyrelevantRAPs.Theperspectiveandexamina- tionofthepossibledevelopmentoftheregionthroughthelensof regionalstakeholderscanprovidefeedbacktotheglobalRAPsas wellasconsistencyfordownscaledscenarios(Fig.2).

3.Results

FirstwepresentthemappingoftheCCAFSscenariostotheSSP scenarios(Section3.1)andthequantifieddriversofchangeforthe CCAFSscenarios(Section3.2).Wethenhighlighttheimpactsofthe scenariosonimprovingfoodsecurity,theregionalsupplyofcrop andlivestockproductsandimpactsontheenvironmentincluding landusechange(Section3.3).Thoughtheregionalscenarioswere definedbyuncertaintyconcerningthemostactiveactorsandlong- termversusshort-termpriorities,wemakeanassessmentofthe vulnerabilityofthescenarios,completingthelinkoftheregional scenarios to the SSPs which are defined by the challenges to adaptation(Section3.4).

3.1.CCAFSscenariosinthecontextoftheSSPs

InSelf-Determination,wherestrongstateactorsfocusonlong- termissues,trendindicatorsaligncloselywithSSP1:Sustainability innearlyallqualitativeelementsdescribingtheSSPnarrative,such asinvestmentsinproductivityandextensionservices,increased educationandhealthandsanitationservices,regulationstoreduce deforestation, and effective social protection schemes. A key differenceisthat investments areestimatedtobelowerin the CCAFSscenarioduetoalackofﬁnancialsupportfromoutsidethe regionanda relianceonregionalresources.Additionally,within theCCAFSscenariothestruggleforinstitutionalchangemayopen uptheopportunityforcorruption,whichisinconsistentwithSSP1 where strong institutions are effective at the national and internationallevels.Inﬁgures,Self-Determinationwillappear as

“SelfDet”.

xxx–xxx

(7)

InSaveYourself,actionisnottakenbytheweakandunstable governments,butbyCSOsinanemergencyresponsemanner,and bytheprivatesectoractingwithshort-termprofitabilityinterests, which mirrors theglobal narrative of SSP3: RegionalRivalry, of weakinstitutions,lowtechnologydevelopmentfortheagriculture sectorandfoodsecurityissuesduetogrowinginequalityandhigh populationgrowth.TheCCAFSscenarioreflectssomeaspectsof thelow-incomecountrynarrativeofSSP4:Inequality,thoughthe keydifferenceisthattheWestAfricascenarioseesmorepolitical instabilityandineffectiveness of institutionswhich hinders the region’sdevelopmentandaccesstomarketswhile,SSP4represents aworldwheregrowinginequality(withinandbetweencountries) stems from limited access to education and consolidation of politicaland economicpowerbyelites.Thereforewe alignSave Yourself with SSP3, and the scenario in figures, appears as

“SaveYourself”.

Civil Society to the Rescue?, where weak governments are replacedwithstrongCSOstacklingfoodsecuritywithalong-term focus, together with strategic investments by a more socially conscious private sector, is most closely represented by SSP2:

Middleofthe Road,wheresomeactionsforprotectionleadtoa declineindeforestationrates,modestproductivityandcommer- cializationbeneﬁtsfalltothosewhoalreadyhavecapacityrather than inducing a transformation of smallholders, and moderate increasesineducationandhealthissuesarelargelytakenupby CSOswithprivatesectorsupport.Ultimately,inthisscenario,the lackof government supportand coordination meansthat non- stateambitionsareonlypartiallyachieved.Inﬁgures,CivilSociety totheRescue?willappearas“CivilSociety”.

Theshort-sightedprioritizationofgovernmentsinterestedin maintainingpowerintheCash,Control,Caloriesscenario,createsa highlyurbanized,higheconomicgrowthfocusedscenario,leading toreactiveinvestmentsineducationandhealthservices,similarto theSSP5:Fossil-fueledDevelopment.ThedifferencewithSSP5isthat in this scenario, investment cycles are short, creating unstable developmentthroughoutthescenarioperiod.Additionally,O’Neill etal.(2015)discussedthepossibilitythatactionstakenwithina

development pathway may change a pathway and alter the challengesforadaptationormitigation,whichcanbeseeninthe

“startand sputter” natureof Cash,Control,Calories,makingthe scenarioquitedifferentthanSSP5bytheendofthetimeperiod.In ﬁgures,Cash,Control,Calorieswillappearas“CCC”.

In our mapping of CCAFS scenarios onto the SSPs,we have lookedforoverlappinginthenarrativesandthestorylinesatthe regionallevel,becauseweassumetherestoftheworldfollowsthe SSP2storylineforallscenarios.AlthoughtheSSPsthemselvesare globalinnatureandscope,thisallowsustoexaminetheimpactsof theregionalassumptions.Fig.3illustratesthelinkagesbetween the stakeholder-deﬁned scenarios for West Africa and the narratives oftheSSPsfromO’Neilletal. (2015)withtheCCAFS scenariosappearinginitalics.

SSPs

GlobalRAPs

Globally Consistent RegionalScenarios (CCAFS scenarios)

Crop, Livestock, Economic, and other

Model Inputs and Parameters

Fig.2. GloballyconsistentregionalscenariosadaptedfromValdiviaetal.(2015).

Soci oeconomic ch allen ges for adapta on

Socio e co nomic challenges fo r migaon

SSP5: Fossil-fueled Development West Africa Scenario:

Cash, Control, Calories

SSP3: Regional Rivalry West Africa Scenario:

Save Yourself

SSP1: Sustainability West Africa Scenario:

Self-Determinaon

SSP4: Inequality SSP2: Middle of the Road

West Africa Scenario:

Civil Society to the Rescue?

Fig.3.Theﬁvesharedsocioeconomicpathways(SSPs)mappedtotheCCAFSWest Africascenarios.

Source:FigureadaptedfromO’Neilletal.(2015).

xxx–xxx

(8)

3.2.QuantiﬁedCCAFSscenariodrivers

Inthissection,wepresenttheresultsfromthemethodology presented in Section 2.3.1 analyzing the following drivers:

socioeconomic development, crop and livestock productivity, regional integration, expansion of cropland area, and the development of the rest of the world. Table 1 presents an overview of the four selected trend indicators that were translatedfromstakeholder information into numerical values tousewithinthemodelsasdrivers.Eachindicatorisgroupedinto four rows by scenarios with the direction and magnitude of change andlogic for change as provided by the stakeholders.

Columns 10 and 12 of Table 1, expand on the “one-to-one” scenariomappingpresentedinSection3.1,offeringinsightsinto howwelltheSSP-speciﬁcindicatorassumptionsmatchwithits correspondingCCAFS scenario, where greenindicates a “good match,”yellowindicatesa“neutralmatch”,andredindicatesa

“badmatch.”Whentheindicatorassumptionswere“badmatch” betweenSSP assumptions andthe CCAFS indicator trends, we adjusted the SSP drivers to better match the CCAFS narrative storylines.

3.2.1.Macroeconomicandsocioeconomicdevelopment

Understanding regional economic development, population growth,theroleofregionalintegrationonagriculturalinputs,and developmentoutside of theregionareessential toassessWest Africa’sfuturedevelopment.

Wecomparedtheeconomicanddemographicdevelopments, acriticalfactorindeterminingfooddemand(Valinetal.,2014), fortheregionupto2050fortheSSPs(Dellinketal.,2015;Kcand Lutz, 2014; O’Neill et al., 2015). Then, guided by the regional scenarionarrativesandthetrendindicationsofchangedeveloped bythestakeholdersduringthescenariodevelopmentworkshops (ﬁrst fourrows of Table1), we adjustedthese drivers for the regiontocapturetheuncertaintyaroundgovernanceandpolitical stability inherent in the regional scenarios as they pose a challengefordevelopmentinWesternAfrica(Palazzoetal.,2016, 2014).In WestAfrica,thepopulation oftheregiongrowsfrom 300 million in 2010 to almost 600 and 800 million in Self- Determination and Save Yourself respectively (Appendix D2 in Supplementary materials). GDP per capita increases across all scenarios,butby2050allscenariosremainlowerthantheregional SSPprojections(AppendixD1inSupplementarymaterials).Cash, Control, Calories initially sees the largest increase, but its GDP development is unstable, and it begins to slow and declines slightlyafter2040–reﬂectingtheshort-termismofthescenario.

PercapitaGDPisthehighestinSelf-Determinationby2050and CivilSociety tothe Rescue? experiences a steady and consistent increaseinpercapitaGDP.PercapitaGDPinSaveYourselfincreases theleastamongstthescenariosoverthetimeperiodandfollows cyclesofgrowthand recession,representingunstableeconomic development.

Theimpactsofthescenariosassumptionswithintheregionare isolatedtosomeextentbyassumingthattheglobalcontextineach of the scenarios follows the same trends for climate impacts, agricultural development, and socioeconomic development. In principle, underlying variables of demographic and economic developmentarecorrelatedacrossregions(fertilityrate,mortality, investment, technological adoption), however, some regional deviations are plausible, as political context also strongly inﬂuencestheevolutionofsuchvariables.Foreasiercomparability of the direct impacts of the regional drivers, we intentionally variedonlytheWestAfricaparameters,changingotherparameters overtimebutkeepingthemconstantacrossthescenariosforother regions of the world, similar to what was done in the CCAFS SoutheastAsianScenarios(Mason-D’Crozetal.,2016).Therestof

theworldfollowstheSSP2populationandeconomicdevelopment trajectorywhere,by2050,theglobalpopulationreaches9.2billion people (Kc and Lutz, 2014)and global averageGDP per capita doublestoaround16,000USD(Dellinketal.,2015).

Thedegreetowhichcurrentregionalintegrationeffortswithin AfricaandtheECOWAScommunityarestrugglingtoﬁndsuccessin agriculturehighlightsthechallengesfacingtheregion(UNCTAD, 2012; United Nations, 2009). The CCAFS scenario narratives considerthechallengestoregionalintegration,includingthelack of regulation, which have been brought into the quantitative modelingof GLOBIOMthroughimpacts inthefarm inputcosts (rows4–8inTable1andAppendixDinSupplementarymaterials).

Limitations in the tradeof both theinputs to and productsof agricultureandshocksintheagriculturalsupplychain,stemming fromconﬂictsorclimatechangecanhaveprofoundeffectsonfood security(BaldosandHertel,2015;Mosnieretal.,2014;Simsonand Tang,2013;vanDijk,2011).Conﬂictsarehighlightedineachofthe scenarios, although in Save Yourself the lack of strong state governmentscombinedwithshort-termprioritysettinggivesthis scenariothemostpotentialforfoodinsecurity.

3.2.2.Cropandlivestockproductivity

Technicalprogressincropproductionisrepresentedinboth models through increases in crop yields. To estimate crop productivityover thetime,weuseaneconometricestimateof the relationship between crop yields and GDP per capita assumptions of the SSPs (Fricko et al., 2016; Herrero et al., 2014).TheIAMsusedinthequantiﬁcationoftheintegratedSSPs projectchangesincropyields,withdifferentIAMsprovidingthe

“marker” for each SSP. For consistency, we have used the GLOBIOM yieldprojections for each SSP (as a starting point), andthenmadeadjustmentsbasedonthescenarionarrativesand trend indicators(foragricultural productivityandcrop-speciﬁc productivity).ThesetrendsappearinthelastfourrowsofTable1.

InIMPACT,thecropyieldtrendswerequantiﬁedbyapplyingthe scenario deviations from the GLOBIOM SSP2 baseline to the IMPACTSSP2baseline,whichwereestimatedbasedonhistorical yieldtrends,agriculturalresearchanddevelopment,andassump- tionsonhowthesecouldchangeovertime(Sulseretal.,2015).The gapbetweentheglobalaverageyieldsandyieldsinWestAfrica willremaina challengefor theagriculturalsystem evenin the scenariowiththehighestinvestmentinagriculture,Self-Determi- nation(Fig.4).

Thecontributionof thelivestock sectortothenational GDP rangesfrom10%to15%(Kamuangaetal.,2008).Recentlivestock foresightstudiesspotlighttheregion’spotential intransitioning from extensive land based systems to mixed crop-livestock systems(Herreroetal.,2014),intensifyingpastoralsystemswhile also protecting pastoralists and animal health, echoing assess- mentsmadebytheSahelWestAfricanClubSecretariatandOCED (Kamuangaetal.,2008).

Productivityof livestockcanbemeasuredbytheconversion efficiencyofthequantityoflivestockproductproducedperunitof feedconsumed.Weusedtheprojectionsofconversionefficiencies forlivestockfortheSSPs(Herreroetal.,2014)asastartingpoint and furtherdeveloped theprojections usingthenarratives and indicatorlogicsandtrends(rows9–12ofTable1).Theinvestments in ruminant production, due to the growing food demand as outlinedinthescenarionarratives,resultinyieldimprovementsin Self-Determination, while the focus on dairy production and monogastric production in the early decades of Cash, Control, Caloriesisconsidered.InCivilSocietytotheRescue?meatdemand drivestheinvestments fromprivatesectorand socialentrepre- neurs.Littleinvestmentismadeforlivestockorveterinaryservices in theSaveYourself scenario resultingin relatively insignificant yieldimprovements.

xxx–xxx

(9)

3.2.3.Croplandareaexpansion

Toharmonizethequantitativemodelingresults,croplandarea expansion as modeled byGLOBIOM was used as aninput into IMPACT,althoughthedistributionofcropareabycroptypeand managementsystem,inthiscaseirrigatedorrainfedcroplandarea, remainedendogenous(AppendixDinSupplementarymaterials).

Croplandintheregionexpandsnearly55%inSSP2by2050.InCash, Control, Calories and Self-Determination cropland increases less (51%and46%,respectively)andSaveYourselfandCivilSocietytothe Rescue?croplandincreasesslightly more(59% and57%,respectively).

3.3.QuantiﬁedCCAFSscenarios

Byapplyingthechangesinthescenariodriversfortheregion overtimewithinthemodelsweprovideaplausiblefutureofthe regional development of the agriculture sector, both in the demandand supplyofproductsas wellasthecompetitionfor land for agricultural production. Inthe following sections, we summarize the scenario results as they pertain to crop and livestockproduction(Section3.3.1),foodavailability,prices,and net trade (Section 3.3.2), andland usechange (Section 3.3.3).

While thispaper focuses onthe multiple, plausible futures of socioeconomicdevelopmentofWestAfrica,thedevelopmentof the restof theworld follows the trends of SSP2(Frickoet al., 2016). Economic growth improves food security spurring an increaseintheproductionofcropandlivestockproductsglobally.

Wehavehighlightedsomeofthewaysthedevelopmentof the regionaffecttherestoftheworldinAppendixHinSupplemen- tary materials. The changes in cropland area expansion from GLOBIOMwereusedasanexogenousdriverinIMPACT,therefore the scenario results presented in this section that explore the

expansion of cropland area regional land use change were modeledbyGLOBIOM.

3.3.1.Agriculturalproductionandclimateimpacts

Agriculturalproductioncurrentlyaccountsforaboutaquarter oftheregion’sGDP,butwasashighas35%inthe1980s(World BankDevelopmentIndicators,2015).WestAfrica,asaregion,isthe leader, or among the top global producers of cassava, millet, sorghum,andoilpalm(FAOSTAT,2015).Historically,increasesin productionwithintheregionhavecomefromexpandingcropland arearatherthanthroughsigniﬁcantyieldimprovements(Byerlee et al., 2014; Fischer et al., 2014; Hillocks, 2002). In the CCAFS scenarios,thehistoricaltrendcontinuesintheSaveYourselfand Civil Societytothe Rescue? scenariosfrom 2010to2050,where slightlylessthanhalfoftheaverageannualgrowthinproduction comesfromcropareaexpansion.Alternatively,almost66%ofthe increase inproductionin theSelf-Determinationscenariocomes fromyieldimprovements(Fig.5).

Inbothmodels,cropproductionintheregionincreasesfrom 2010to2050forallscenarios,withSelf-Determinationhavingthe highestlevelsofcropproductionandSaveYourselfhavingtheleast growthin cropproduction (AppendixFig.E2 inSupplementary materials). The developmentof crops in the regionremains of particularimportancetotheglobalproductionby2050,especially formillet,cassava,and sorghum(moredetailsinAppendixEin Supplementarymaterials).

Investmentsinlivestockproductionnearlyquadruplesthetotal livestockcaloriesproduced(fromdairy,ruminantandmonogastric meat)forCash,Control,CaloriesandSelf-DeterminationinGLOBIOM andtriplesinIMPACT.Althoughthereislittleinvestmentinthe livestocksector(asidefromthedairysector)inSaveYourselfand limitedinvestmentinCivilSocietytotheRescue?,thesescenarios 2

4 6 8 10 12 14 16

1965 1970 1980 1990 2000 2010 2020 2030 2040 2050

SSP1 SelfDet SSP5 CCC

SSP2 CivilSociety SSP3 SaveYourself

SSP4 FAO-West Africa (hist) FAO-Global Average (hist) SSP2- Global Average

Fig.4.Historicalandaggregateexogenouscropyields(gigacaloriesperha)forCCAFSWestAfricabyscenarioandforeachSSPandhistoricalandexogenousSSP2global average.

Source:FAOSTAT(2016)forhistorical;Frickoetal.(2016)fortheSSP2globalaverage;AuthorsforprojectionsforWestAfricascenarios.

xxx–xxx

(10)

stillseeanannualincreaseoftotallivestockproductionofbetween 2.3%and2.8%peryear,respectivelyinbothmodels.

Examining the impacts onthe mostimportant crops tothe regionshowsthat,onaverage,climatechangelowerscropyields (Fig.6).Thisisconsistentwithotherassessmentsundervaried climaticconditionsusingWestAfricaspeciﬁccropmodels,despite oneof the GCM climate models(MIROC) predictingconditions whereclimatechangeis morefavorable tocrops(Sultanet al., 2013).Aggregated cropyields providea rough estimateof the impactsofclimatechange,however,thesemayunderestimatethe impactstoindividualcrops(suchasmillet,sorghum,andcassava).

The impacts for individual crops can be found in AppendixTableF2inSupplementarymaterials.Forbothmodels,

thenegativeclimateimpactsonaggregatecropyieldsintheSelf- Determination scenario,which hasthe highest exogenousyield improvements,areinmostcases,stillgreaterthantheyieldsfor thethreeotherscenarioswithoutclimateimpacts,suggestingthat adaptationmeasuresand investments takenin thepresent can lessentheimpactsoffutureclimatechange.Theﬂexibilityofthe endogenous areareallocationresponse withinGLOBIOM makes themodelmoreresponsivetotheyieldeffectsofclimatechange thanIMPACT(Nelsonetal.,2014a,2014b).

3.3.2.Marketsandfooddemand

Kilocalorie availabilityper capitaperday, a commonly used indicator to measure food security, considers the total food

1.35 1.45 1.55 1.65 1.75 1.85 1.95 2.05 2.15

CCC CivilSociety SaveYourself SelfDet

NoCC GFDL HGEM IPSL MIROC

1.45 1.50 1.55 1.60 1.65 1.70 1.75 1.80

CCC CivilSociety SaveYourself SelfDet

NoCC GFDL HGEM IPSL MIROC

GLOBIOM IMPACT

Fig.6. Relativechangeinaveragecropyieldsin2050comparedto2010yieldsasmodeledbyGLOBIOMandIMPACTfortheCCAFSWestAfricaScenarioswithandwithoutthe climatechangeeffectsoncropgrowthincluded.Note:They-axisisnotthesameforbothmodels.

Source:GLOBIOMmodelresults(leftside);IMPACTmodelresults(rightside).

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

FAO 1980-1990

FAO 1990-2000

FAO 2000-2010

SSP2 CCC CivilSociety SaveYourself SelfDet Area Yield

Fig.5.ShareofthesourceofproductiongrowthbasedontherateofgrowthforWestAfricaoverhistoricaltrendsandscenarioprojectionsfortheCCAFSscenariosandSSP2.

Note:Areaiscroplandareaexpansionandyieldistheincreaseintheaggregatecropyieldintonsperhectare.

Source:FAOSTAT(2015),(leftside);GLOBIOMmodelresults(averageover2010–2050)(rightside).

xxx–xxx