Modelling regional cropping patterns under scenarios of climate and socio-economic change in Hungary

2018·2.7 MB·English

by  Sen Li

ScienceoftheTotalEnvironment622–623(2018)1611–1620
ContentslistsavailableatScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv
Modelling regional cropping patterns under scenarios of climate and
socio-economic change in Hungary
SenLia,⁎,LindaJuhász-Horváthb,LászlóPintérb,c,MarkD.A.Rounsevelld,e,PaulaA.Harrisonf
aEnvironmentalChangeInstitute,UniversityofOxford,SouthParksRoad,OxfordOX13QY,UK
bDepartmentofEnvironmentalSciencesandPolicy,CentralEuropeanUniversity,Nádoru.9,Budapest1051,Hungary
cInternationalInstituteforSustainableDevelopment,325-111LombardAvenue,Winnipeg,MBR3B0T4,Canada
dInstituteofMeteorologyandClimateResearch(IMK-IFU),KarlsruheInstituteofTechnology,Kreuzeckbahnstrasse19,Garmisch-Partenkirchen82467,Germany
eSchoolofGeoSciences,UniversityofEdinburgh,DrummondStreet,EdinburghEH89XP,UK
fCentreforEcology&Hydrology,LibraryAvenue,LancasterLA14AP,UK
H I G H L I G H T S G R A P H I C A L A B S T R A C T
• Weemphasisethecriticalroleofregion-
allevelactorsindevelopingeffective
responsestoarablelandusechanges.
• We describe the development of an
empirically-grounded agent-based
model for projecting future cropping
patterns.
• We apply the model to stakeholder-
driven scenarios of plausible future
socio-economicandclimatechange.
• Themodelprojectsstrongdifferencesin
future land use change between two
Hungarianregions.
• Theresultssupporttheneedtoimple-
mentfocusedadaptationpolicyatthe
regionallevel.
a r t i c l e i n f o a b s t r a c t
Articlehistory: Impactsofsocio-economic,politicalandclimaticchangeonagriculturallandsystemsareinherentlyuncertain.
Received24July2017 Theroleofregionalandlocal-levelactorsiscriticalindevelopingeffectivepolicyresponsesthataccommodate
Receivedinrevisedform25September2017 suchuncertaintyinaflexibleandinformedwayacrossgovernancelevels.Thisstudyidentifiedpotentialregional
Accepted5October2017
challengesinarablelandusesystems,whichmayarisefromclimateandsocio-economicchangefortwocounties
Availableonline18October2017
inwesternHungary:VeszprémandTolna.Anempirically-grounded,agent-basedmodelwasdevelopedfroman
extensivefarmerhouseholdsurveyaboutlocallandusepractices.Themodelwasusedtoprojectfuturepatterns
Editor:D.Barcelo
ofarablelanduseunderfourlocalised,stakeholder-drivenscenariosofplausiblefuturesocio-economicandcli-
Keywords: matechange.Theresultsshowstrongdifferencesinfarmers'behaviourandcurrentagriculturallandusepatterns
Agriculturallanduse betweenthetworegions,highlightingtheneedtoimplementfocusedpolicyattheregionallevel.Forinstance,
Croprotation policythatencourageslocalfoodsecuritymayneedtosupportimprovementsinthecapacityoffarmersto
Empirically-groundedagent-basedmodel adapttophysicalconstraintsinVeszprémandfarmeraccesstosocialcapitalandenvironmentalawarenessin
Environmentalchangeimpactassessment Tolna.Itisfurthersuggestedthatthetworegionswillexperiencedifferentchallengestoadaptationunderpos-
Farmerhouseholdsurvey siblefutureconditions(upto2100).Forexample,Veszprémwasprojectedtohaveincreasedfallowland
Stakeholder-drivenscenarios
underascenariowithhighinequality,ineffectiveinstitutionsandhigher-endclimatechange,implyingrisksof
landabandonment.Bycontrast,Tolnawasprojectedtohaveaconsiderabledeclineinmajorcerealsunderasce-
narioassumingade-globalisingfuturewithmoderateclimatechange,inferringchallengestolocalfoodself-
⁎ Correspondingauthor.
E-mailaddress:[email protected](S.Li).
https://doi.org/10.1016/j.scitotenv.2017.10.038
0048-9697/©2017ElsevierB.V.Allrightsreserved.
1612 S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620
sufficiency.Thestudyprovidesinsightintohowsocio-economicandphysicalfactorsinfluencetheselectionof
croprotationplansbyfarmersinwesternHungaryandhowfarmerbehaviourmayaffectfutureriskstoagricul-
turallandsystemsunderenvironmentalchange.
©2017ElsevierB.V.Allrightsreserved.
1.Introduction Thisstudybuildsonrecentresearchonapplyinganempirically-
grounded ABM for scenario-based future LUCC projections, e.g.
Socio-economicandpoliticalchangesarecriticaldriversofagricul- Castellaetal.(2007),Guillemetal.(2015),andMurray-Rustetal.
tural land use in Europe (Ewert et al., 2005; Holman et al., 2016; (2013).Theobjectiveswere(i)toprojectpossiblefuturepatternsofar-
LambinandMeyfroidt,2011;RounsevellandReay,2009).Development ablelanduseattheregionallevelfortwocounties,VeszprémandTolna,
ofnewfarmingtechnologiesandcrops,increasingpopulationdensity, inwesternHungary,and(ii)toidentifypotentialcounty-specificchal-
mobilityandchangingconsumptionpatterns,anddiversifiedinterna- lenges for crop production, which may arise from environmental
tionaltradingpatternshaveallaffectedtheprofitabilityofdomesticag- change.Toachievetheseobjectives,anextensivefarmerhouseholdsur-
riculturalproductiontovaryingextentsandinfluencedfarmerdecision- veywasdesignedandimplementedtoinvestigatethecurrentpatterns
making.Climatechangealsoplaysanimportantrolethroughshiftsin ofarablelanduseinwesternHungary.Thiswasusedtoexploretheen-
weatherpatterns(e.g.precipitationandtemperature)andextreme vironmentalfactorsinfluencingfarmerdecisionsinselectingcroprota-
weatherevents(e.g.floods,droughtsandstorms)affectingcropsuit- tion plans. The model's structure was designed within the Aporia
abilityandyieldvariabilityinagriculturalsystems(OlesenandBindi, framework,anABMofagriculturalLUCC(Murray-Rustetal.,2014),
2002;Olesenetal.,2011).Suchsocio-economic,politicalandclimatic andfunctionswerecalibratedbasedonthesurveydataaswellascensus
changesareexpectedtocontinue,andmayevenaccelerateinthefuture datafromvarioussources.Ascenarioanalysiswasperformedusingfour
resultinginpotentiallysevere,buthighlyuncertain,impactsonagricul- integratedclimateandsocio-economicscenariosco-createdwithstake-
turallandsystems(Holmanetal.,2017). holdersfromthetwocounties.Themodellingexperimentssoughtto
Landuse/coverchange(LUCC)modellingtechniqueshaveadvanced identifypotentialchangesinfuturearablelandusepatternsduetoen-
rapidly(c.f.reviewsbyLambinetal.,2000,Matthewsetal.,2007and vironmentalandrelatedsocio-economicchangeandchallengesfor
Verburgetal.,2004).Thishasledtomodel-basedprojectionsoffuture adaptingtothesechangeswithinthetwocounties.
LUCCbeingmoreoftenusedfordecisionsupportinurban/ruralplan-
ningpolicy(Harrisonetal.,2016;Presteleetal.,2016).Severalauthors 2.Materialsandmethods
havesuggestedthatformodellingstudiesoffutureLUCCtobeuseful
andrelevantforlandusedecision-makers,theyshould:(i)represent 2.1.Studysites
real-worldprocessesandsectoralinteractionsofimportancewithin
thecontextofaspecificsocio-ecologicalsystem;(ii)developparticipa- Farmers'landusedecisionsandhowtheyaffectarablelandusedy-
toryscenarioswithstakeholderstogainqualitativeinsightintopossible namicswereexploredfortwopredominantlyruralcountiesinwestern
futurechanges;and(iii)recogniseuncertaintyanderrorpropagationin Hungary,VeszprémandTolna(Fig.1).Thesecountieswerechosenbe-
modeloutputsexplicitlyratherthanfocusingonprecisepredictions causeoftheinterestofkeystakeholdersinthecommunitiesinanas-
(Audsleyetal.,2006;Harrisonetal.,2015;Millaretal.,2007).Moreover, sessmentoftheimpactsofenvironmentalchange,andparticularly
effectivepolicyresponsestoenvironmentalchangeneedtocoordinate extremeorhigh-endclimatechange,onvulnerabilityintheirregions
acrossgovernancelevelsandacrosssectors(Adgeretal.,2005;Ciscar andthedegreetowhichadaptationoptionsmayreducetheirvulnera-
etal.,2011;HurlimannandMarch,2012).Inparticular,theroleof bility.Thetwocountieshaveasimilarculturalprofile,andthegeo-
lowerlevel(regionalandlocal)actorsindeterminingresponsesto graphicaldistancebetweentheircentresisapproximately100km.
LUCCareessentialinunderstandingtheeffectivenessofmanagement Hydro-climaticconditionsinthetwocountiesareparticularlyimpor-
interventionsunderuncertainfuturescenarios(Antonsonetal.,2016; tantforregionalagriculture,withVeszprémlocatedatthenortheast
EikelboomandJanssen,2013;KumarandGeneletti,2015).Todate, shoreoftheLakeBalaton,thelargestlakeinCentralEurope,andTolna
mostscenario-basedLUCCstudieshavefocusedonthecontinental- locatedonthewestbankoftheriverDanube,Europe'ssecond-longest
andglobal-scale(Ewertetal.,2005;Holmanetal.,2017;Hurttetal., river. Individual farmers from Veszprém andTolna manage 75,782
2011;Rounsevelletal.,2005).Regional-andlocal-levelLUCCstudies and119,485haagriculturalland,respectively,representingbelow-
havebeenrare,withafewrecentexceptions(Guillemetal.,2015; averageandaveragelevelsinthecountry(KSH,2011).Eventhough
Houetetal.,2016;Lietal.,2017b;Murray-Rustetal.,2013). theextentoftotalarablelandatthenationallevelhasonlyslightlyin-
Farmersmakeagriculturallandusedecisions.Whilefarmersmaybe creased by 0.5% between 2000 and 2010, the proportion of major
operatingwithinsimilarsocio-economiccontextsattheregionalscale, crops has changed considerably: by −9.4% for cereals, −33.8% for
landusedecisionscanbehighlydiverse,owingtodifferencesinindivid- pulses,−54.8%forpotatoes,−70.6%forsugarbeet,−18.9%forfresh
ualobjectives,preferencesandexperiences,aswellasfarm-levelphys- vegetables, melons, strawberries and +87.1% for industrial crops
icalconstrainssuchassoiltypeandaccessibility.Thissuggeststhat (EUFSS,2012).Basedonthesecensusdata,thestudyfocusedonthedy-
modelsofagriculturalLUCCneedtoincorporateexplicitlyrepresenta- namicsoflanduseproportionsratherthantheirextents.
tionofthemechanismsofhowindividualand/orfarm-levelattributes
affectfarmermanagementstrategies(Rounsevelletal.,2003).From 2.2.Modeldevelopment
thisperspective,agent-basedmodelling(ABM)isanapproachthatof-
fersaflexible,bottom-upwayofrepresentingtheprocessesofindivid- Anempirically-groundedmodelwasdevelopedforthecasestudies,
uallandusedecision-makingandcross-scaleinteractions(Murray-Rust usingtheAporiaframeworkforagent-basedmodellingofagricultural
etal.,2014;Valbuenaetal.,2010).ABMshavebeenshowntoprovide landusechange(Guillemetal.,2015;Murray-Rustetal.,2014).The
greaterexplanatorypowerthanmanytop-downapproaches(e.g.linear modeldevelopedinthisstudy,termedAporia-Lite,adoptedtheAporia
programmingmodelling)(Filatovaetal.,2013;Kellyetal.,2013)and conceptsandwasprogrammedinareducedformtoonlyincludethe
thusofferthepotentialforinterpretingthereasonsbehindprojected important components for the case study (see Fig. 2). As an
patternsoflandusechange. empirically-grounded model, the database and key functions of
S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620 1613
Fig.1.ThetwostudysitesinHungary:VeszprémandTolnacounties.
Aporia-Litemodelarehighlyspecialisedforthecasestudy.Readersin- agriculturallandmanagerofspecificsocio-economicandattitudinalat-
terested in the technical details in developing agent-based models tributes.Theirfarmsarerepresentedasparcelagentsdepictedbyasetof
usingtheAporiaframeworkarereferredtoMurray-Rustetal.(2014). physicalattributes.Afarmeragentmanagesaparcelagentthroughare-
ThefreesoftwareandguidanceforthefullversionoftheAporiamodel gime,oracroprotationplan,whichhasspecificattributesdescribingits
areavailableathttp://www.wiki.ed.ac.uk/display/Aporia. social,economicandenvironmentalvalues.Thus,theregimedatabase
TheAporia-Litemodelconsistsoftwotypesofagents(farmerand servesasalocalknowledgebaseofcroprotationplansforfarmerselec-
parcel)andaregimedatabase.Afarmeragentrepresentsanindividual tion.Theregimeselectionprocessisassumedtobedrivenbyajoint
Fig.2.ThetheoreticalframeworkoftheAporia-Litemodel.
1614 S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620
considerationofallthefarmerandparcelattributesandtheavailable isthecommonrangefoundinthesurvey.Avariabledescribingthegen-
knowledgeonadoptableregimes.Afarmertypologyisembeddedin eraladoptabilityoftheregimewasgeneratedtoreflecttheeasewith
themodeltobetterunderstandandcategorisethedifferencesinregime whichacroprotationplancouldbeadoptedbasedonexistinglocal
selection.Themodelgeneratespatternsoffarmertypesandarableland knowledge.Thiswasestimatedfromthelikelihoodofdifferentcrop-
useonanannualbasis. to-crop chains as found in the local survey and termed the ‘crop
matchpossibility’(Fig.3).Forexample,thetotalcropmatchpossibility
2.2.1.Localfarmerhouseholdsurvey ofathree-yearregime‘Maize→Sunflower→Wheat’wasestimatedas
Anextensive,multi-purposefarmerhouseholdsurveywasconduct- the‘Maize→Sunflower’matchpossibility∗the‘Sunflower→Wheat’
edbetweenJuneandAugust2015forthetwostudyareas.Thesurvey matchpossibility.Allpossiblecroprotationsoflength2–8yearswere
wasdesignedtoservetwokeypurposes.Firstly,toinvestigatecurrent generated.Inordertoreducethecomputationalburden,noN1000
agriculturallandusepatternsandcollectinformationonmajorlocal croprotationswereretainedforeachlength,byselectingthe1000
cropsandsamplecroprotationplans.Secondly,tocollectattitudinal, withthegreatestcrop-matchpossibility.Finally,4855and5526poten-
socio-economicattributesandphysicalattributesthatmayinfluence tiallyadoptablecroprotationplanswereincludedintheregimedata-
farmers'landusedecision-making.Thesurveystudywasdesigned basesforVeszprémandTolna,respectively.
withassistancefromtheHungarianCentralStatisticalOffice(KSH), Thedatabaseswerethencompletedbycompilingfurthereconomic,
andwasconductedjointlywiththeKSH'sregularagriculturalsurvey socialandenvironmentalattributesforeachregime.Theattributesused
in2015bytrainedandexperiencedKSHsurveyors.Foreachcounty, inthisstudyfollowedpreviousAporiamodellingstudies(Guillemetal.,
theKSHdetermined110surveyparticipantsbyrandomlyselecting 2015;Murray-Rustetal.,2014),includingregimelength,regimediver-
farmerswhomanageprivateholdingsfromitsannualsamplingpools sity,grossmargin(averageyield∗price+subsidies−costs),costs
which contain 687 farmers (out of 13,529) in Veszprém and 1253 (seed,labour,machinery,fertiliser,pesticideandothercosts),soilstruc-
farmers(outof18,092)inTolna.Thesesamplingpoolsweredeter- ture,nitrogendemand,croplandcoverage,andscoresfortradition,
minedbytheKSH'sinternalexpertstoadequatelyreflectthelocaldis- sceneryandrecreation.The2013agro-economicstatisticsprovidedby
tribution of farm size and farming objective. For the 220 survey the Hungarian Research Institute of Agricultural Economics (AKI)
participantsselectedfromthetwocountiesandvisited,allhelpedto wereusedtocalculategrossmarginandcosts.Thequalitativescores
completethequestionnaire.Finally,172farmerswhoprovidedinfor- ofsoilstructure,nitrogendemandandcropcoverwerecalculatedfol-
mationontheirregimeswereincludedinthisstudy.Furtherdetails lowingthemethodintheDEFRAguidebookof‘ArableCropping&the
aboutthesurveystudycanbefoundinLietal.(2017a).Theattributes Environment’(DEFRA,2002).Thetraditionscorewasbasedonthe
of farmers and regimes collected for this study are summarised in crops'contributiontosownareain1990,basedontheFAOcountrypro-
TableS1,theAppendixA. filesforHungary.Therankingsofsceneryandrecreationweregenerat-
ed with the help of local experts during stakeholder meetings
2.2.2.Regimedatabase (Section2.4).
Inthefarmerhouseholdsurvey,participantswereaskedtoprovide
detailedinformationon(i)thelengthofcurrentcroprotationplan(in 2.2.3.Farmer'sexecutivetypology
years),(ii)therotationschemelistingthecropsplantedforeachyear, Typologiesareoftenusedtosimplifythediversityoffarmersand
and(iii)whichyearofthiscroprotationschemetheyarecurrentlyin. their land use strategies, by defining different groups statistically
Thesurveycollected172regimerecordsintotal,164ofwhichweredif- basedonspecificcriteria(McKinney,1950).Thesecriteriareflectnot
ferent,suggestingahighdiversityinfarmers'regimeselectionprocess. onlytheresearchobjective,butalsodefinehowthediversityofland
Toreflectthisdiversity,aregimedatabasewasdevelopedfromthesur- usedecisionsissimplifiedandincludedintheABMofLUCC(Valbuena
veyresultstocoverasmanyadoptableregimesaspossible.Thedata- etal.,2008).Inthisstudy,the“executivetypology”isdefinedasa
basewasconstructedtoincluderegimesof2–8yearduration,which categorisationofthefarmersbasedonthedifferencesinthestylesof
Fig.3.Matchpossibilityofmajorcropsinthetwocasestudyareas:(A)Veszprémand(B)Tolna.Scalesofthecoloursegmentsindicatethetotalnumberoftimesthecropinquestionwas
mentionedinaregimebyarespondent,eitherasa‘from’(greatergaptothecircumference)orasa‘to’(smallergap)ina‘from→to’cropmatchpossibility.
S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620 1615
croprotationplanactuallyapplied.Aclassificationanalysiswasper- regimeattributesandregimeselectionwererepeated.Finally,there-
formed following the approach used in previous typology studies gime'sadoptability(orthe‘cropmatchpossibility’,Section2.2.3)was
(Guillemetal.,2012;Karalietal.,2013).First,aprincipalcomponents consideredtorepresentthepotentialtoacquirelocalknowledgeand
analysiswasconductedtoextracttheprincipalcomponentsdescribing supportforimplementingtheregime.Arandomselectionschemewas
the variation of regime attributes. Attributes with anti-image appliedtoallowregimeswithahigheradoptabilitytohavegreater
correlationb0.3wereexcluded,leadingtosixattributesbeingkept,in- chanceofbeingselectedfromthecandidatepool.
cludinglengthofregimeadoption,diversityofregime,grossmargin,
costs,soilstructureandtradition.Second,alltheregimerecordswere 2.3.Modelimplementation,workflowandevaluation
classifiedbasedontheprincipalcomponentsusingaK-meanscluster
analysis(withWard'smethod).Fourgroupsofregimes,ortheexecutive TheAporia-LitemodelwasscriptedintheJava-basedRepasttoolkit
typesoffarmers,weredetermined: version2.1foragent-basedmodelling(Northetal.,2013).Allthedata
analysisandempiricalfunctionsmentionedpreviouslywereperformed
(i) Business-orientedfamerswhoadoptregimeswithhighmargin/
anddevelopedusingtheIBMSPSSstatisticsversion22(Field,2013),be-
costs,whichcontainnon-traditionalcrops,butresultindegraded
forebeingintegratedintothemodel.
soilstructure;
Wheninitialisingthemodel,farmeragentsandtheirparcelagents
(ii) Traditional/non-diversifiedfamerswhoadoptlong-term,low-
wereco-created.Eachfarmer-parcelpairwasassignedrandomlygener-
diversityregimeswithtraditionalcrops;
atedattributesaccordingtothesummarystatistics(meanandvariance)
(iii) Traditional/diversified famers who adopt short-term, high-
fromtheagriculturalcensusdataorthehouseholdsurveydata.Theag-
diversityregimeswithtraditionalcropsandrelativelyhighmar- riculturalcensusfromtheHungarianCentralStatisticalOffice(KSH)
gin/costs,and;
(retrievedfortheyearof2013)wasusedtogenerate:farmingobjective
(iv) Supplementaryfamerswho adoptshort-termregimesoflow
(commercial,semi-subsistence,orsubsistence)andfarmtype(arable,
margin/costs,andcreategoodsoilstructure.
mixed,orlivestock)forfarmeragents;andfarmsizes(distributionof
sizesavailableinthecensus)forparcelagents.Asthisstage,1000farm-
eragentsweregeneratedforbothstudysites,andbecauseofthere-
Theselectedregimeattributeswerecomparedbetweentheexecu- searchfocusonlythosewhomanage‘arable’or‘mixed’farmtypes
tivetypesofthefarmersandthesummarystatisticsareprovidedin werekeptinthemodel.Thisresultedinapproximately590and570
TableS2,theAppendixA.Multinomiallogisticregressionswereper- farmeragentsbeingretainedforVeszprémandTolna,respectively.
formedforthetwositesseparately,toidentifywhichfarmerandparcel Then,thesummarystatisticsbasedonthehouseholdsurveyonfarmers'
attributesaffectthecategorisationoffarmers'executivetypes.Attri- landuse(TableS2,theAppendixA)wereusedtogeneratetheremain-
buteswereenteredintotheregressionmodelandselectedbasedon ingattributesfoundtobeimportantinclassifyingtheagentexecutive
theirstatisticalsignificance(Pb0.1),correlationswithotherattributes types(Section2.2.3).Aftertheagentswerecreated,theregimeselec-
(−0.5bSpearmancorrelationcoefficientb0.5)andcontributionto tionprocedure(Section2.2.4)wasapplied.Allthefarmeragentswere
thefunctionspredictivepower(usingthelikelihood-ratiostatistics). assignedaninitialregimeandarandomnumbertoindicatetheyear
Usingtheseempiricalfunctions,itwaspossibletoprojecthowsocio- oftheregimethefameragentwasin.Then,foreachyear,farmeragents
economicandenvironmentalchangescouldshiftfarmertypeand,con- managedtheirfarmsandmadelandusedecisionssimultaneously.They
sequently,theirregimeselection.Thetwofinalcounty-specificfunc- wouldfirstcheckifaregimewasfinished:ifno,thelanduseplannedfor
tionshadsatisfactorypredictivepowerwiththeNagelkerke'sPseudo thenextyearwouldbeapplied;ifyes,theregimeselectionprocedure
R-SquareN0.6andanoverallrateofcorrectclassificationhigherthan (Section2.2.4)wasexecutedandnewregimesapplied.
67%.Thefollowingfourattributeswereincludedinbothfunctions: Themodelwasevaluatedthrough(i)stakeholders'qualitativeas-
sizeoftheparcel,multipleparcelmanagement(binary),attitudeto- sessment and acceptance of the model structure and usefulness of
wardstheimportanceoftheenvironment,andaccessibilitytoinforma- model projections,and(ii) aquantitativecomparisonbetweenthe
tionviasocialgroups(binary)ofthefarmer.Inaddition,thefunctionfor projectedarablelandusepatternsandtheobservedpatternsrecorded
Veszprémincludedfarmers'incomeandcommercialobjective(binary), inthelocalagriculturalcensus.Foreachcounty,themodelwasrun
whilethatforTolnaalsoincludedfarmers'education,attitudestowards for100yearswithbaselinesettings(for2010)andtheproportionof
theimportanceofrevenueandeffort,soilqualityandarabletypeofthe landusesformajorcropswassummarisedandcomparedwithcensus
parcel.DetailedparameterestimatesarepresentedinTablesS3andS4, data(retrievedfromtheKSHfor2013).
theAppendixA,forVeszprémandTolna,respectively.
2.4.Modellingexperiments:projectingfuturearablelandusepattern
2.2.4.Regimeselection
Inthemodel,thedecisionstrategyofafarmeragentfollowsautility 2.4.1.Integratedscenariosforlocalenvironmentalchanges
maximisationprocedure.Theattributesofthefarmeragentanditspar- Understandinghowenvironmentalconditionsarelikelytochange
celagentdepictthelimitsofsocio-economicandphysicalresources. inthefutureisanimportantfirststepinscenarioanalysis,asitprovides
Consideringtheselimitations,anexecutivetypewasassignedtoafarm- theconditionsforwhichthemodel'sparametersneedtobemodified.
er agent by applying the multinomial logistic regressive functions Throughworkingcloselywithlocalstakeholdersfromthetwostudy
(Section2.2.3).Apoolofcandidateregimeswasthenselectedusing sites,asetofintegratedscenariosweredevelopedtodescribeplausible
threesteps.First,arandomnumberwasgeneratedforeachregimeat- alternativefutureclimateandsocio-economictrajectories.Thesocio-
tributefollowinganormaldistributionwiththesurveyedmeanand economic scenarios were based on downscaling four of the global
variancevalues(TableS2,theAppendixA).Theserandomlygenerated Shared Socio-economic Pathways (SSPs) (O'Neill et al., 2015) for
attributeswereassumedtobethetargetattributethatreflectsthemax- Hungary.PKD(Pinkdream–SSP1)describesasustainablefuturewith
imum utility the farmer would seek. Secondly, the regimes which ahealthyeconomyandimprovedenvironment.RRV(RegionalRivalry
metallofthefollowingrequirementsfromthedatabasewereselected –SSP3)representsaninternationallyfragmentedfuturewithhighlyde-
intothecandidatepool:(i)candidate'slength=targetlength;(ii)can- graded environmental conditions. IEQ (Inequality – SSP4) is
didate'sgrossmargin≥targetgrossmargin;(iii)candidate'scosts≤tar- characterisedbyapoliticalandbusinesselitewithincreaseddisparities
getcosts,and;(iv)candidate'diversity,soilstructureandtraditionϵ[0.8 betweentheeliteandmassesinsocial,economicandpoliticaldimen-
∗targetvalues,1.2∗targetvalues].Thirdly,ifthepoolremainedempty sions,andadegradedenvironment.PPU(PatóPálÚr–SSP5)portrays
afterloopingthroughthedatabase,thentherandomisationoftarget afossil-fuel-basedandhighlyindustrialisedfuturewithrapideconomic
1616 S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620
growthandhighlyoverexploitedenvironment.ThePKDandIEQsce- improvementsduetoimprovedagronomyand/orcropbreeding.Pro-
narioswerecoupledwithintermediateclimatechangebasedonthe jectionsforseveralmajorcropswereextractedforHungary(seeFig.
RCP4.5emissionsscenario,whiletheRRVandPPUwerelinkedwith S1,theAppendixA).Ingeneral,cropyieldswereprojectedtoimprove
high-endclimatechangebasedontheRCP8.5emissionsscenario.De- dramaticallyunderPPU(SSP5)andIEQ(SSP4),butdecreaseslightly
tailsoftheparticipatorymethodologyfordevelopingthescenarios underPKD(SSP1)andRRV(SSP3).Sinceminorcerealsandlegumes
through a series of stakeholder workshops are given in Li et al. werenotmodelledbytheIAP,theircropyieldswereassumedtoremain
(2017b).Afulldescriptionofthedevelopmentofthefourlocalscenar- unchangedunderthedifferentscenariosinthemodellingexperiments.
ioscanbefoundinKokandPedde(2016).Therationalebehindthese-
lectionandintegrationofthelocalscenariosarediscussedinKoketal.
(2015)andMadsenetal.(2016). 3.Results
2.4.2.Scenariodriverselectionandquantification 3.1.Modelperformance
Foreachscenario,futurepatternsofarablelandusewereprojected
by(i)tuningtheagents'attributesbasedonaquantificationexciseand Qualitativeevaluationsbythestakeholders(throughdiscussions)
(ii)adjustingtheyieldsofmajorlocalcropsaccordingtoprojections suggestgeneralacceptanceofthemodelstructure.Inapost-workshop
fromcropmodels.TheAporia-Litemodelwasrunfor100timesteps surveywithstakeholders,20outofthe23respondentsfoundmodel
withscenario-specificsettings,throughwhichresultsattheannual projections useful or very useful for discussing possible policy re-
levelweresummarisedandcompared. sponses. The comparison between the baseline model projections
Thefourattributesfoundtobeimportantintheclassificationof (2010)withthecensusresultssuggestsasatisfactorymodelperfor-
farmers'executivetypesinbothcasestudyareas(Section2.2.3)were mance(Fig.4):theprojectedrankoftheproportionformajorcropsin
selectedforthequantificationexcise.‘Farmsize’wasassumedtobeas- totalarablelandusewasinlinewiththerankfoundincensusdata;
sociatedwiththetotalagriculturallandusechange,i.e.,farmsizein forthemajorcrops(9outof11inVeszprémand7outof11inTolna),
generaldecreasedastotalagriculturallandusedecreases.Whethera therecordedproportionofthecropintotalarablelanduseinthecensus
farmermanages‘multipleparcels’wasassumedtocorrespondtothe datafellwithintheprojectedmin-maxrange.
sizeoftheruralpopulation(i.e.,asmallerruralpopulationleadsto
morepeoplemanagingmultipleparcels)andtechnology(i.e.,greater
developmentof technology means less labouris needed and, thus, 3.2.Projectedchangesinfarmer'sexecutivetypes
moremultipleparcels).‘Accesstosocialnetwork’wasassumedtobe
dependentonchangesinsocialcapital,i.e.,increasedsocialcapitalcre- Projectedpatternsofchangeinfarmertypesaresimilarinbothcase
atesmorechancesforfarmerstoaccessknowledge.‘Attitudetowards studyareasforsomescenarios,butdifferentforothers(seeFig.5A,for
(theimportanceof)theenvironment’wasassumedtobeinfluenced theextentsofchangesby2100;seeFig.S2,theAppendixA,fortrends
byageneralawarenessofenvironmentalprotectionandsustainablede- overtime).Ingeneral,projectedchangesinfarmertypesweredriven
velopment,i.e.,enhancedawarenessleadsfarmerstohaveastronger bythedifferencesinthebaselinefarmertypeproportionsandassump-
feelingoftheimportanceofenvironment.Thepotentialchangesin tionsmadeaboutthescenariodrivers(Table1).Themostdramatic
theseattributeswereinterpretedqualitativelyforeachscenario,based changeswereprojectedunderRRVandPPUinVeszprémbutunder
ontherationaleandstorylinesunderpinningthelocalisedscenarios PKDinTolna.TheIEQscenariowasprojectedtoleadtoarelatively
(Table1).The‘low’and‘mediumtohigh’changeextentswereassumed lowerlevelofchangeinbothareas.
tobe15%and35%forallattributesforexperimentalpurposes.These UnderPKD,farmerswereexpectedtobemoreenvironmentally
valueswereselectedastheyarenottooextremetobeimpossiblein friendlyandself-sustainedforbothcasestudyareasby2100:anin-
thefuturebutstillabletocauseconsiderablechangesinprojectedfu- creasein thesupplementary (+10.6%in Veszprém and+14.8%in
turecroppingsystems. Tolna) and traditional/non-diversified (+25.5% in Veszprém and
PotentialchangesincropyieldswerederivedfromtheIntegrated +17.8%inTolna)farmertypeswereprojected,aswellasadecrease
AssessmentPlatform(IAP)atthepan-Europeanlevel(Harrisonetal., inbusiness-oriented(−3.0%inVeszprémand−11.4%inTolna)and
2016;Holmanetal.,2015).IntheIAP,cropyieldprojectionsareinflu- traditional/diversified (−33.1% Veszprém and −21.1% in Tolna)
encedbyclimate,CO levelsandSSP-specificassumptionsaboutyield farmers.
2
UnderRRV,farmersweremoreinclinedtoadopthighergrossmar-
ginregimesandtobelessresponsibleinprotectingsoilstructure:de-
creaseswereprojectedforthesupplementary(−9.8%Veszprémand
Table1 −7.2%inTolna)andtraditional/non-diversified(−11.0%Veszprém
Settingsofselectedagents'attributesforfuturelanduseprojectiona. and −39.5% in Tolna) farmers. Owing to a strong increase in the
business-orientedfarmersinTolna(+39.5%by2100),traditional/di-
Scenario Timeslice Farm Multiple Accessto Attitudetowards(the
size parcels social importanceof) versifiedfarmerswereprojectedtodecreasein2070–2100.Incontrast,
network environment inVeszprém,thebusiness-orientedfarmerswereprojectedtobestable
PKD 2010–2040 − − ↗ ↗ (+1.4%by2100)andthetraditional/diversifiedfarmerswereprojected
2040–2070 ↘ ↘ ↗ ↗ toincreasecontinually(+7.2%by2100).
2070–2100 ↘ ↘ ↗↗ ↗ UnderIEG,theproportionsoffarmertypeswereprojectedtoberel-
RRV 2010–2040 ↗ ↘ ↘ ↘
ativelystableby2100:Veszprémhadamoderateincreaseinsupple-
2040–2070 ↗↗ ↘↘ ↘ ↘
2070–2100 ↗↗ ↘↘ ↘↘ ↘↘ mentaryfarmers(+17.7%)anddecreases(absolutechangesb10%)in
IEQ 2010–2040 ↗ − ↘↘ − alloftheotherfarmertypes;Tolna,however,hadanincreaseinthetra-
2040–2070 ↘↘ ↘ ↗ ↗ ditional/diversifiedfarmers(+11.3%)andslightdecreasesintheother
2070–2100 ↘ ↘ − − types.
PPU 22001400––22004700 ↘↘↘↘ ↗↗↗↗ ↘− ↘−↘ UnderPPU,farmerswereexpectedtoadoptmoreflexiblelanduse
2070–2100 ↘ ↗ ↗ ↗ strategies (with a shorter regime length) and look for an optimal
trade-offbetweenenvironmentsustainabilityandprofit,i.e.,ashiftto
a Changesbetweenthestartandendofthetimeslice:‘–’indicatesamarginalchange;
‘↗’and‘↘’meanlowincreaseanddecrease;‘↗↗’and‘↘↘’meanmediumtohighin- the traditional/diversified type was projected by 2100 (+21.6% in
creaseanddecrease. Veszprémand+33.0%inTolna)forallotherfarmertypes.
S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620 1617
Fig.4.Comparisonbetweenthepredictedproportionofarablelanduse(2010)andcensusdata(KSH2013)inVeszprém(A)andTolna(B).
3.3.Projectedchangesinarablelanduse Veszprémwheatdecreasedby18.2%andmaizeincreasedby7.5%,and
inTolnawheatdecreasedby11.1%andmaizeincreasedby9.2%.Such
Inthemodel,apartfromthetypeoffarmer,cropyieldalsoinflu- ashiftwasassociatedwithadecreasedgrossmarginofwheat(asare-
encedthepatternsofarableLUCCbyaffectingthepotentialgrossmar- sultofdecreasedcropyield).Forenergycrops,sunflowerwasprojected
ginofacroprotationplan,thekeyeconomicdriveroffarmerlanduse toincreasemoderatelyinbothVeszprém(+11.7%)andTolna(8.4%).
decisionmaking.Thisfurtheramplifiedthedifferenceintheprojected Under RRV, a small change in arable land use was projected in
regionalpatternsofarablelandusebetweencasestudies(seeFig.5B, Veszprém,witha3.9%increaseinmajorcerealsanda1.3%decreasein
fortheextentsofchangesby2100;seeFig.S3,theAppendixA,for energycrops.However,arableLUCCinTolnawasprojectedtochange
trendsovertime). rapidly,witha13.1%decreaseinmajorcereals(−8.9%formaize)and
UnderPKD,thetwoareaswereprojectedtohaveasimilararable a 12.9% increase in energy crops (+13.4% for soybean, +3.6% for
LUCCpatternby2100,bothwithaslightdecreaseincerealcroplands sunflower and −4.1% for rape). This difference was mainly due to
(−5.8%inVeszprémand−1.7%inTolna)andanincreaseinmajoren- theprojectedincreaseinbusiness-orientedfarmersinTolna,making
ergycrops(+8.2%inVeszprémand2.6%inTolna).Forcereals,both moreprofitablecrops(e.g.soybeanandsunflower)morefrequently
areaswereprojectedtoexperienceashiftfromwheattomaize:in selected.
Fig.5.Projectedchangesby2100intheexecutivetypesoffarmer(A)andarablelandusepattern(B)inVeszprémandTolna,underthefourintegratedscenario.ns:notselectedbythe
farmeragents.
1618 S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620
UnderIEQ,projectedarableLUCCwasatalowlevelforbothareas,as objectives(Guillemetal.,2012;Karalietal.,2013).Previousstudies
lesssignificantchangesinfarmertypeswereprojectedthanforthe alsoincludedbothcropandmeatproduction,whereasthisstudyfo-
otherscenarios.InVeszprém,wheresupplementaryfarmersincreased, cusedsolelyoncrops,owingtodifferentresearchobjectives,dataavail-
regimeswithlowergrossmargin,butwhichcreatebettersoilstructure, abilityandlocalcharacteristics.
weremorefrequentlyselected.Asaresult,landuseformaizewas Inthisstudy,differentdriversofarablelandusechangedetermining
projectedtoreduceby7.5%.InTolna,theincreaseintraditional/diversi- currentandprojectedfuturecroppingpatternswerefoundinthetwo
fiedfarmersandimprovementsincropyieldsledtoincreasesinwheat counties,suggestingthatpolicyoptionsfordealingwith,oradapting
(by6.6%)andsunflower(+4.8%),anddecreasesinmaize(−8.6%)and to,thesedriversareneededthatarecounty-specificandimplemented
rape(−3.8%). bylowerlevelactors.Thefarmertypologyshowsaclearlydistinctbase-
UnderPPU,similarpatternsofchangewereprojectedforthetwo linepatternoffarmers'actualselectionoflandusepracticesbetween
areas:increasesinthemajorcereals(+3.4%inVeszprémand+1.8% thetwoareas.Thismayseemcounterintuitive,asonemightnotexpect
in Tolna) and decreases in energy plants (−7.9% in Veszprém and suchlargedifferencesbetweenthetwocountiesfromaculturaland
−4.2%inTolna).Wheatwasprojectedtobethemostpopularcropbe- technologicalperspective.Inparticular,themodelestimatedthepro-
causeforbothareasunderIEQandPPU,owingtoanincreaseinitspo- portionof‘supplementary’farmers,tobe50.5%inVeszprémand7.4%
tentialgrossmarginresultingfromalargeyieldincrease.Comparedto inTolna,whilethe2013censusdatafromKSHshowsanearlyidentical
maize,wheatalsohadlowerbaselinecosts,andtheregimeswhichin- proportionofsuchfarmers,withVeszprémhaving54.6%andTolna,
cludedwheatweremoreaccessiblebytraditionalandsupplementary 54.4%. It should be noted that the ‘supplementary’ farmers in this
farmerswhodominatedregionalfarmerpopulationsinbothareas. studywereassumedtorepresentfarmerswhonotonlyproducefood
fortheirownconsumption,butalsoadoptsustainablelandusepractices
4.Discussion toavoidsoiloverexploitation.Aprojectedreductioninsustainableuse
practiceswouldposefuturechallengesforTolnainachievingitslocal
TheAporia-LiteABMwasdevelopedasanexplanatorytooltobetter foodself-sufficiencyandsecuritygoals.Theresultspresentedheresug-
understandtherangeofarablelandusepatternsunderdifferentplausi- gestthatfuturepoliciestoaddressthischallengemayneedtoincrease
blefutures,ratherthantomakeprecisepredictions.Thevalidityofthe farmeraccesstosocialcapital(e.g.creatingagriculturalorganisations
modelprojectionsaresupportedinseveralways.Firstly,thescenarios andsocieties)andenhancetheirawarenessoftheimportancetofarm-
describingthepossiblefutureclimateandsocio-economicconditions ingoftheenvironmentandmaintainingnaturalcapital.Conversely,itis
weredevelopedbyworkingcloselywithstakeholderstoensurethesce- likelythatphysicalconstraints(asindicatedinthesurvey)havecontrib-
narioswerecrediblewithinthelocalcontextandrelevantforpolicyde- utedtoahigherproportionof‘supplementary’farmersinVeszprém.For
cisions(seeLietal.,2017bforthetimelineofscenariodevelopment). example,poorersoilquality,atopographylessfavourableforlarge-
Secondly,theconceptualframeworkofthemodelfollowedanexisting scale farming and thus smaller farm sizes may indicate that some
approachwhoseflexibility,extensibility,verification,andtransparency farmershavenochoicebuttoutiliseless-exploitativepractices.Thisis
hasbeenextensivelytestedagainstreal-worldissues(Guillemetal., supportedbyarelatedstudythatfoundthattheoverallrateofclimate
2015;Murray-Rustetal.,2014).Thirdly,theaccuracyofthebaselinear- change adaptation was lower among farmers from Veszprém than
ablelanduseprojectionswasfoundtobesatisfactoryforthetwocase thosefromTolna(Lietal.,2017a).Thus,futurepoliciesinVeszprém
studyareas(Section3.1).Finally,localstakeholdersfoundthefuture targetingsustainablelocalfoodproductionmayneedtosupportim-
projectionsappropriateinthecontextofthemodelandscenarios,and provementsinfarmers'abilitytoadapttovariousphysicalconstraints.
usefulintermsofsupportingdiscussionsonpolicyresponsestoenvi- Thefutureprojectionsforthefarmertypologyandarablelanduse
ronmentalchanges. alsoshowedcleardifferencesbetweenthetwocounties,underthe
Themodelisempirically-groundedandcalibratedondatacollected sameintegratedclimateandsocio-economicscenarios.ForthePKD
fromahouseholdsurveyoffarmers.Empiricaldatahavevarioususesin (Pinkdream)andPPU(PatóPálÚr)scenarios,theextentofchangedif-
thedevelopmentofABMs(SmajglandBarreteau,2017)andinthis feredsomewhatinspiteoftheprojectedtrendsbeingsimilarinboth
study the survey data contributed in two ways. Firstly, the survey areas.However,undertheRRV(RegionalRivalry)andIEQ(Inequality)
addsimportantempiricalevidencetosupportunderstandingofthe scenarios,boththetrendsandextentsofchangewereprojectedtobe
characteristicsoflocalcroprotationplansandfactorsaffectingfarmers' distinctlydifferent.Thesedifferencesattheregionallevelareconsistent
decision-making.InHungary,thedata(i)improvesunderstandingof withthenationallevelsocio-economicstorylinesunderpinningeach
what motivates individual-level decision-making which previously scenario,whichimpliesdifferentmagnitudesofchallengesforadapta-
was very limited, and (ii) encapsulates existing evidence about tionpolicies.PKD(SSP1;environmentalsustainabilityorientedscenar-
landscape-leveldeterminantsofarablelandusechange(Munteanu io)andPPU(SSP5;highlyindustrialisedinfrastructurescenario)are
etal.,2014;Verhulstetal.,2004),thusprovidingaricherpictureof bothassociatedwithrapideconomicdevelopment,reducedinequality
theprocessesunderpinningHungary'slandusesystemdynamics.Sec- andlowchallengestoadaptation.Bycontrast,RRV(SSP3;fragmented
ondly,theobservedpatternsplayedanimportantroleinguidingthede- withineffectiveinstitutionsscenario)andIEQ(SSP4;highlydividedso-
signoftheAporia-Litemodelstructure.Fromthetheoryofthepattern- cietyscenario)bothhavesloweconomicdevelopment,increasedin-
orientedapproach(Grimmetal.,2005),wesoughttobuildrigorous equalityandhighchallengestoadaptationwithfutureenvironmental
modelsbasedonrelevantpatternsobservedintherealsystem,rather change(O'Neilletal.,2015).Suchchallengesmaybefurtheramplified
thandevelopanoverly-complexmodeltoapproximatethesystemin forTolnaunderRRVandVeszprémunderIEQ.Forexample,ahugede-
asmuchdetailaspossible.Assuch,severaldifferencescanbenoted cline(−13.1%)inmajorcerealswasprojectedforTolna,whichmay
whencomparingthisstudywithtwopreviousstudiesforLunnainScot- threatentheregion'scapacitytobeself-sufficient.Moreover,securing
landandAarauinSwitzerlandwhichappliedthesameAporiamodelling foodfromotherregionsmaybedifficultasRRV(SSP3)ischaracterised
framework(Murray-Rustetal.,2014).Theempiricalanalysispresented byvariousbarrierstotrade,especiallyforagriculturalproducts(O'Neill
heresuggeststhatcroprotationplansusedinwesternHungaryare etal.,2015).Veszprémwasprojectedtohavedeclinesinlandforboth
highlydiverse.Hence,acomprehensiveregimedatabasewasconstruct- majorcereals(−4.7%)andenergycrops(−3.5%),butincreasesinfal-
edforfarmers'cropselection,insteadofusingacharacteristicregime lowland(+5.8%),whichwasthehighestinallsimulations.Thisimplies
databasewithonlyafewkeycroprotationplans.Thesurveydataalso thatsomeindividualfarmers,especiallysmallholders,mayfacesevere
enabledanexplorationofactuallanduse,throughwhichanexecutive difficultiesinsustainingregularfoodproductionandthusdecidenot
typologyoffarmertypeswasdeveloped.Bycontrast,thetypologiesde- to use their land. This situation is likely to be even worse in IEQ
velopedinpreviousstudiesonlycategorisedfarmers'attitudesand (SSP4),asresourcesarecontrolledbyanelitepartofsociety,leaving
S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620 1619
small-scalefarmersstrugglingwithlowproductivity(O'Neilletal., betweenthetwocountieswerefurtherexacerbatedinthefutureeven
2015).Fallowlandsare,therefore,atriskofabandonment. underthesameclimateandsocio-economicscenario,suggestingdiffer-
Thefindingsofthisstudyareofsubstantialregionalimportance entfociforlocalpoliciestoadapttolong-termenvironmentalchanges.
withinVeszprémandTolnaandhavepotentialrelevancetootherre- Forexample,underanunequalfuturewithincreasedsocialdisparities
gions.Inthenextstep,theprojectedimpactsonagriculturewillbe andhigher-endclimate change,Veszprém wasprojectedtohavea
usedtotesttheabilityofexistingdevelopment/sectorstrategiesandad- growthinfallowland,whichimpliesdifficultiesforsmallholderfarmers
aptationplansforthetwocountiestoreducevulnerabilitybylowering insustainingregularfoodproductionand,therefore,risksoflandaban-
exposureandincreasingresilience.Asagriculturallanduseinthetwo donment.Alternatively,underade-globalisingfuturewithineffective
countiesisbroadlysimilartootherpartsofthecountry,ourfindings institutions and moderate climate change, Tolna was projected to
aregenerallyrelevanttootherareasofHungary,aswellasareasofEast- haveaconsiderabledeclineinlandformajorcereals,whichmayinflu-
ernEuropethathavesimilarfarmownershipstructuresandhistoric encelocalfoodself-sufficiency.Thefutureprojectionsprovidefunda-
landusechanges.Inparticular,ourimpactassessmentsarerelevantto mentalinformationtosupportariskassessmentforenvironmental
theareasaroundtheLakeBalatonandwithintheDanuberiverbasin change.Theagent-basedmodeldevelopedcanbeusedinfuturestudies
thathavesimilarhydro-climaticconditions.Moreover,themodelling totesttheeffectivenessofpolicyresponsesinadaptingfarmer'sbehav-
methodhasmuchwiderrelevanceasitcanbeimplementedindifferent iourstovarioussocio-economicandclimatechanges.
cropping environments provided a survey is undertaken to
parameterisetheAporiamodel.Whenapplyingthemethodforotherre- Acknowledgements
gions,researchersneedtoreflectonwhethersimplifyingorextending
themodelcomponents/functionsusingdifferentapproaches(e.g.the Theresearchleadingtotheseresultshasreceivedfundingfromthe
pattern-orientedtheoryusedinthisstudy)isnecessaryinordertobet- EuropeanCommunity'sSeventhFrameworkProgramme(FP7/2007–
termeetthecase-specificresearchobjectivesanddataavailability.Good 2013)undergrantagreementnumber603416(TheIMPRESSIONSPro-
practicesrelatedforevaluatingmodelcomplexityforacasestudypur- ject-ImpactsandRisksfromHigh-EndScenarios:StrategiesforInnova-
posearedescribedinSunetal.(2016). tiveSolutions;www.impressions-project.eu).
Fromamethodologicalperspectiveseveralfutureimprovements
anddirectionsinthisresearcharepossible.Firstly,tobetterrepresent AppendixA.Supplementarydata
individual-leveldecision-makingprocesses,theregimedatabasecould
bereplacedbyanalgorithmtoallowfarmeragentstobuildtheirown Supplementarydatatothisarticlecanbefoundonlineathttps://doi.
crop rotation plans automatically, such as approaches reviewed in org/10.1016/j.scitotenv.2017.10.038.
Dury et al. (2012). In this study, due to data constraints, only the
within-regimecropmatchpossibilitieswereconsidered.Futureinvesti- References
gationsareneededtoevaluatetheconnectivitybetweenregimes,espe-
ciallyonthefactorsinfluencingthecropmatchesthatconnecttheend Adger,W.N.,Arnell,N.W.,Tompkins,E.L.,2005.Adaptingtoclimatechange:perspectives
across scales. Glob. Environ. Chang. 15:75–76. https://doi.org/10.1016/
ofoneregimetothebeginningofanother.Secondly,theimpactofre-
j.gloenvcha.2005.03.001.
gional/localclimatechangecouldbeexploredfurther.Thisincludes Alexander,P.,Moran,D.,Rounsevell,M.D.A.,Smith,P.,2013.Modellingtheperennialen-
modellingtheconsequencesofclimaticvariabilityandextremeweather ergycropmarket:theroleofspatialdiffusion.J.R.Soc.Interface10.https://doi.org/
10.1098/rsif.2013.0656.
events on crop yields, water availability and pests and diseases
Antonson,H.,Isaksson,K.,Storbjörk,S.,Hjerpe,M.,2016.Negotiatingclimatechangere-
(BachingerandZander,2007;DelplaandRodriguez,2014;Steffens sponses:regionalandlocalperspectivesontransportandcoastalzoneplanningin
etal.,2015;Whiteetal.,2011).Thirdly,itwouldbeusefultobetterin- South Sweden. Land Use Policy 52:297–305. https://doi.org/10.1016/
j.landusepol.2015.12.033.
tegratespatialheterogeneitywithinthemodel.Here,wewereunableto
Audsley,E.,Pearn,K.R.,Simota,C.,Cojocaru,G.,Koutsidou,E.,Rounsevell,M.D.A.,Trnka,
explorethespatialpatternoflandusechange,duetodataconstraints, M., Alexandrov, V., 2006. What can scenario modelling tell us about future
i.e.,alackofobservedagriculturalparcelsforcasestudyareasandinsuf- Europeanscaleagriculturallanduse,andwhatnot?Environ.Sci.Pol.9:148–162.
ficientevidenceforthespatialeffectsoffarmerdecision-making.Empir- https://doi.org/10.1016/j.envsci.2005.11.008.
Bachinger,J.,Zander,P.,2007.ROTOR,atoolforgeneratingandevaluatingcroprotations
ical studies are needed to address these shortcomings and further fororganicfarmingsystems.Eur.J.Agron.26:130–143.https://doi.org/10.1016/
understandhowlandusemaybepromotedorconstrainedbytheex- j.eja.2006.09.002.
change of knowledge between adjacent farmers (Alexander et al., Castella,J.-C.,PhengKam,S.,DinhQuang,D.,Verburg,P.H.,ThaiHoanh,C.,2007.
Combiningtop-downandbottom-upmodellingapproachesoflanduse/coverchange
2013; dos Santoset al.,2011). Finally, while this study focusedon
tosupportpublicpolicies:applicationtosustainablemanagementofnaturalre-
projectingfuturescenarios,themodelcouldalsobeusedtotesttheef- sourcesinnorthernVietnam.LandUsePolicy24:531–545.https://doi.org/10.1016/
fectivenessofdifferentpolicies.Forexample,themodelcouldbeusedto j.landusepol.2005.09.009.
Ciscar,J.-C.,Iglesias,A.,Feyen,L.,Szabó,L.,VanRegemorter,D.,Amelung,B.,Nicholls,R.,
explorehowdifferentsubsidystrategieschangelandusepatterns,or
Watkiss,P.,Christensen,O.B.,Dankers,R.,Garrote,L.,Goodess,C.M.,Hunt,A.,
whennovelcropsaretaken-up. Moreno,A.,Richards,J.,Soria,A.,2011.Physicalandeconomicconsequencesofcli-
matechangein Europe. Proc. Natl.Acad. Sci.108:2678–2683.https://doi.org/
10.1073/pnas.1011612108.
5.Conclusions
DEFRA,2002.ArableCropping&theEnvironment-AGuide.DepartmentforEnviron-
ment,Food&RuralAffairs(DEFRA),London.
Wedemonstratethedevelopmentandapplicationofanempirically- Delpla,I.,Rodriguez,M.J.,2014.Effectsoffutureclimateandlandusescenariosonriverine
sourcewaterquality.Sci.TotalEnviron.493:1014–1024.https://doi.org/10.1016/
groundedABMtoexploredifferencesinarablelanduseunderfuture
j.scitotenv.2014.06.087.
scenariosfortwocasestudyareasinwesternHungary,Veszprémand dosSantos,L.M.R.,Michelon,P.,Arenales,M.N.,Santos,R.H.S.,2011.Croprotationsched-
Tolnacounties.Despitethefactthatthetwoareasshareasimilarcultur- ulingwithadjacencyconstraints.Ann.Oper.Res.190:165–180.https://doi.org/
alandtechnologicalbackgroundandlargelycomparableagroclimatic 10.1007/s10479-008-0478-z.
Dury,J.,Schaller,N.,Garcia,F.,Reynaud,A.,Bergez,J.E.,2012.Modelstosupportcropping
conditions,ourmodelprojectedadistinctlydifferentpatterninthe planandcroprotationdecisions.Areview.Agron.Sustain.Dev.32:567–580.https://
typesoffarmersfromthetwoareas,inparticularforthe‘supplementa- doi.org/10.1007/s13593-011-0037-x.
ry’farmerswhoareself-sustainedandadoptsustainablelanduseprac- Eikelboom,T.,Janssen,R.,2013.Interactivespatialtoolsforthedesignofregionaladapta-
tion strategies. J. Environ. Manag. 127:S6–S14 (Supplement). https://doi.org/
tices, and, hence, different patterns of arable land use. Policy that
10.1016/j.jenvman.2012.09.019.
encourageslocalfoodself-sufficiencyandsecuritymayneedtotarget EUFSS,2012.AgriculturalCensusinHungary.Eurostat.
improvingthecapacityoffarmerstoadapttophysicalconstraintsin Ewert,F.,Rounsevell,M.D.A.,Reginster,I.,Metzger,M.J.,Leemans,R.,2005.Futurescenar-
iosofEuropeanagriculturallanduse:I.Estimatingchangesincropproductivity.
Veszprémandenhancingfarmeraccesstosocialcapitalandenviron- Agric.Ecosyst.Environ.107:101–116.https://doi.org/10.1016/j.agee.2004.12.003.
mental awareness in Tolna. The projected differences in land use Field,A.,2013.DiscoveringStatisticsUsingIBMSPSSStatistics.Sage.
1620 S.Lietal./ScienceoftheTotalEnvironment622–623(2018)1611–1620
Filatova,T.,Verburg,P.H.,Parker,D.C.,Stannard,C.A.,2013.Spatialagent-basedmodelsfor Matthews,R.B.,Gilbert,N.G.,Roach,A.,Polhill,J.G.,Gotts,N.M.,2007.Agent-basedland-
socio-ecologicalsystems:challengesandprospects.Environ.Model.Softw.45:1–7. usemodels:areviewofapplications.Landsc.Ecol.22:1447–1459.https://doi.org/
https://doi.org/10.1016/j.envsoft.2013.03.017. 10.1007/s10980-007-9135-1.
Grimm,V.,Revilla,E.,Berger,U.,Jeltsch,F.,Mooij,W.M.,Railsback,S.F.,Thulke,H.-H., McKinney,J.C.,1950.Theroleofconstructivetypologyinscientificsociologicalanalysis.
Weiner,J.,Wiegand,T.,DeAngelis,D.L.,2005.Pattern-orientedmodelingofagent- Soc.Forces28:235–240.https://doi.org/10.2307/2572006.
based complex systems: lessons from ecology. Science 310:987–991. https:// Millar,C.I.,Stephenson,N.L.,Stephens,S.L.,2007.Climatechangeandforestsofthefuture:
doi.org/10.1126/science.1116681. managinginthefaceofuncertainty.Ecol.Appl.17:2145–2151.https://doi.org/
Guillem,E.E.,Barnes,A.P.,Rounsevell,M.D.A.,Renwick,A.,2012.Refiningperception- 10.1890/06-1715.1.
basedfarmertypologieswiththeanalysisofpastcensusdata.J.Environ.Manag. Munteanu,C.,Kuemmerle,T.,Boltiziar,M.,Butsic,V.,Gimmi,U.,Lúboš,H.,Kaim,D.,Király,
110:226–235.https://doi.org/10.1016/j.jenvman.2012.06.020. G.,Konkoly-Gyuró,É.,Kozak,J.,Lieskovský,J.,Mojses,M.,Müller,D.,Ostafin,K.,
Guillem,E.E.,Murray-Rust,D.,Robinson,D.T.,Barnes,A.,Rounsevell,M.D.A.,2015.Model- Ostapowicz,K.,Shandra,O.,Štych,P.,Walker,S.,Radeloff,V.C.,2014.Forestandagri-
lingfarmerdecision-makingtoanticipatetradeoffsbetweenprovisioningecosystem culturallandchangeintheCarpathianregion—ameta-analysisoflong-termpatterns
services and biodiversity. Agric. Syst. 137:12–23. https://doi.org/10.1016/ and drivers of change. Land Use Policy 38:685–697. https://doi.org/10.1016/
j.agsy.2015.03.006. j.landusepol.2014.01.012.
Harrison,P.A.,Holman,I.P.,Berry,P.M.,2015.Assessingcross-sectoralclimatechangeim- Murray-Rust,D.,Rieser,V.,Robinson,D.T.,Miličič,V.,Rounsevell,M.,2013.Agent-based
pacts,vulnerabilityandadaptation:anintroductiontotheCLIMSAVEproject.Clim. modellingoflandusedynamicsandresidentialqualityoflifeforfuturescenarios.En-
Chang.128:153–167.https://doi.org/10.1007/s10584-015-1324-3. viron.Model.Softw.46:75–89.https://doi.org/10.1016/j.envsoft.2013.02.011.
Harrison,P.A.,Dunford,R.W.,Holman,I.P.,Rounsevell,M.D.A.,2016.Climatechangeim- Murray-Rust,D.,Robinson,D.T.,Guillem,E.,Karali,E.,Rounsevell,M.,2014.Anopen
pactmodellingneedstoincludecross-sectoralinteractions.Nat.Clim.Chang.6: frameworkforagentbasedmodellingofagriculturallandusechange.Environ.
885–890.https://doi.org/10.1038/nclimate3039. Model.Softw.61:19–38.https://doi.org/10.1016/j.envsoft.2014.06.027.
Holman,I.,Audsley,E.,Berry,P.,Brown,C.,Bugmann,H.,Clarke,L.,Cojocaru,G.,Dunford, North,M.,Collier,N.,Ozik,J.,Tatara,E.,Macal,C.,Bragen,M.,Sydelko,P.,2013.Complex
R.,Harrison,P.A.,Janes,V.,Kovats,S.,Lafond,V.,Li,S.,Lobanova,A.,Mokrech,M., adaptivesystemsmodelingwithRepastSimphony.Compl.Adapt.Syst.Model.1:
Rounsevell,M.,Sandars,D.,Savin,C.,Wimmer,F.,2015.SpecificationforEuropean 1–26.https://doi.org/10.1186/2194-3206-1-3.
modelimprovementanddevelopment.IMPRESSIONSDeliverableD3B.1. Olesen,J.E.,Bindi,M.,2002.ConsequencesofclimatechangeforEuropeanagricultural
Holman,I.P.,Harrison,P.A.,Metzger,M.J.,2016.Cross-sectoralimpactsofclimateand productivity, land use and policy. Eur. J. Agron. 16:239–262. https://doi.org/
socio-economicchangeinScotland:implicationsforadaptationpolicy.Reg.Environ. 10.1016/s1161-0301(02)00004-7.
Chang.16:97–109.https://doi.org/10.1007/s10113-014-0679-8. Olesen,J.E.,Trnka,M.,Kersebaum,K.C.,Skjelvåg,A.O.,Seguin,B.,Peltonen-Sainio,P.,Rossi,
Holman,I.P.,Brown,C.,Janes,V.,Sandars,D.,2017.Canwebecertainaboutfutureland F.,Kozyra,J.,Micale,F.,2011.ImpactsandadaptationofEuropeancropproduction
usechangeinEurope?Amulti-scenario,integrated-assessmentanalysis.Agric.Syst. systems to climate change. Eur. J. Agron. 34:96–112. https://doi.org/10.1016/
151:126–135.https://doi.org/10.1016/j.agsy.2016.12.001. j.eja.2010.11.003.
Houet,T.,Marchadier,C.,Bretagne,G.,Moine,M.P.,Aguejdad,R.,Viguié,V.,Bonhomme, O'Neill,B.C.,Kriegler,E.,Ebi,K.L.,Kemp-Benedict,E.,Riahi,K.,Rothman,D.S.,vanRuijven,
M.,Lemonsu,A.,Avner,P.,Hidalgo,J.,Masson,V.,2016.Combiningnarrativesand B.J.,vanVuuren,D.P.,Birkmann,J.,Kok,K.,Levy,M.,Solecki,W.,2015.Theroads
modellingapproachestosimulatefinescaleandlong-termurbangrowthscenarios ahead:narrativesforsharedsocioeconomicpathwaysdescribingworldfuturesin
forclimateadaptation.Environ.Model.Softw.86:1–13.https://doi.org/10.1016/ the 21st century. Glob. Environ. Chang. https://doi.org/10.1016/
j.envsoft.2016.09.010. j.gloenvcha.2015.01.004.
Hurlimann,A.C.,March,A.P.,2012.Theroleofspatialplanninginadaptingtoclimate Prestele,R.,Alexander,P.,Rounsevell,M.D.A.,Arneth,A.,Calvin,K.,Doelman,J.,Eitelberg,
change.WileyInterdiscip.Rev.Clim.Chang.3:477–488.https://doi.org/10.1002/ D.A.,Engström,K.,Fujimori,S.,Hasegawa,T.,Havlik,P.,Humpenöder,F.,Jain,A.K.,
wcc.183. Krisztin,T.,Kyle,P.,Meiyappan,P.,Popp,A.,Sands,R.D.,Schaldach,R.,Schüngel,J.,
Hurtt,G.C.,Chini,L.P.,Frolking,S.,Betts,R.A.,Feddema,J.,Fischer,G.,Fisk,J.P.,Hibbard,K., Stehfest,E.,Tabeau,A.,VanMeijl,H.,VanVliet,J.,Verburg,P.H.,2016.Hotspotsofun-
Houghton,R.A.,Janetos,A.,Jones,C.D.,Kindermann,G.,Kinoshita,T.,KleinGoldewijk, certaintyinland-useandland-coverchangeprojections:aglobal-scalemodelcom-
K.,Riahi,K.,Shevliakova,E.,Smith,S.,Stehfest,E.,Thomson,A.,Thornton,P.,van parison.Glob.Chang.Biol.22:3967–3983.https://doi.org/10.1111/gcb.13337.
Vuuren,D.P.,Wang,Y.P.,2011.Harmonizationofland-usescenariosfortheperiod Rounsevell,M.D.A.,Reay,D.S.,2009.LanduseandclimatechangeintheUK.LandUsePol-
1500–2100:600yearsofglobalgriddedannualland-usetransitions,woodharvest, icy26(Supplement1):S160–S169.https://doi.org/10.1016/j.landusepol.2009.09.007.
and resulting secondary lands. Clim. Chang. 109:117. https://doi.org/10.1007/ Rounsevell,M.D.A.,Annetts,J.E.,Audsley,E.,Mayr,T.,Reginster,I.,2003.Modellingthe
s10584-011-0153-2. spatialdistributionofagriculturallanduseattheregionalscale.Agric.Ecosyst.Envi-
Karali,E.,Brunner,B.,Doherty,R.,Hersperger,A.,Rounsevell,M.A.,2013.Theeffectof ron.95:465–479.https://doi.org/10.1016/S0167-8809(02)00217-7.
farmerattitudesandobjectivesontheheterogeneityoffarmattributesandmanage- Rounsevell,M.D.A.,Ewert,F.,Reginster,I.,Leemans,R.,Carter,T.R.,2005.Futurescenarios
mentinSwitzerland.Hum.Ecol.41:915–926.https://doi.org/10.1007/s10745-013- ofEuropeanagriculturallanduse:II.Projectingchangesincroplandandgrassland.
9612-x. Agric.Ecosyst.Environ.107:117–135.https://doi.org/10.1016/j.agee.2004.12.002.
Kelly, R.A., Jakeman, A.J., Barreteau, O., Borsuk, M.E., ElSawah, S., Hamilton, S.H., Smajgl,A.,Barreteau,O.,2017.Framingoptionsforcharacterisingandparameterising
Henriksen,H.J.,Kuikka,S.,Maier,H.R.,Rizzoli,A.E.,vanDelden,H.,Voinov,A.A., humanagentsinempiricalABM.Environ.Model.Softw.93:29–41.https://doi.org/
2013.Selectingamongfivecommonmodellingapproachesforintegratedenviron- 10.1016/j.envsoft.2017.02.011.
mentalassessmentandmanagement.Environ.Model.Softw.47:159–181.https:// Steffens,K.,Jarvis,N.,Lewan,E.,Lindström,B.,Kreuger,J.,Kjellström,E.,Moeys,J.,2015.
doi.org/10.1016/j.envsoft.2013.05.005. Directandindirecteffectsofclimatechangeonherbicideleaching—aregional
Kok,K.,Pedde,S.,2016.IMPRESSIONSSocio-economicScenariosEUFP7IMPRESSIONS scale assessment in Sweden. Sci. Total Environ. 514:239–249. https://doi.org/
ProjectDeliverableD2.2. 10.1016/j.scitotenv.2014.12.049.
Kok,K.,Christensen,J.H.,Sloth,M.M.,Pedde,S.,Gramberger,M.,Jäger,J.,Carter,T.,2015. Sun,Z.,Lorscheid,I.,Millington,J.D.,Lauf,S.,Magliocca,N.R.,Groeneveld,J.,Balbi,S.,
Evaluationofexistingclimateandsocio-economicscenariosincludingadetailedde- Nolzen,H.,Müller,B.,Schulze,J.,Buchmann,C.M.,2016.Simpleorcomplicated
scriptionofthefinalselection.EUFP7IMPRESSIONSProjectDeliverableD2.1. agent-based models? A complicated issue. Environ. Model. Softw. 86:56–67.
KSH,2011.Censusof2011.HungarianCentralStatisticalOffice(KSH),Budapest. https://doi.org/10.1016/j.envsoft.2016.09.006.
Kumar,P.,Geneletti,D.,2015.Howareclimatechangeconcernsaddressedbyspatial Valbuena,D.,Verburg,P.H.,Bregt,A.K.,2008.Amethodtodefineatypologyforagent-
plans?Anevaluationframework,andanapplicationtoIndiancities.LandUsePolicy basedanalysisinregionalland-useresearch.Agric.Ecosyst.Environ.128:27–36.
42:210–226.https://doi.org/10.1016/j.landusepol.2014.07.016. https://doi.org/10.1016/j.agee.2008.04.015.
Lambin,E.F.,Meyfroidt,P.,2011.Globallandusechange,economicglobalization,andthe Valbuena,D.,Verburg,P.H.,Bregt,A.K.,Ligtenberg,A.,2010.Anagent-basedapproachto
loominglandscarcity.Proc.Natl.Acad.Sci.108:3465–3472.https://doi.org/10.1073/ modelland-usechangeataregionalscale.Landsc.Ecol.25:185–199.https://doi.org/
pnas.1100480108. 10.1007/s10980-009-9380-6.
Lambin,E.F.,Rounsevell,M.D.A.,Geist,H.J.,2000.Areagriculturalland-usemodelsableto Verburg,P.H.,Schot,P.P.,Dijst,M.J.,Veldkamp,A.,2004.Landusechangemodelling:cur-
predictchangesinland-useintensity?Agric.Ecosyst.Environ.82:321–331.https:// rent practice and research priorities. GeoJournal 61:309–324. https://doi.org/
doi.org/10.1016/S0167-8809(00)00235-8. 10.1007/s10708-004-4946-y.
Li,S.,Juhász-Horváth,L.,Harrison,P.A.,Pintér,L.,Rounsevell,M.D.A.,2017a.Relating Verhulst,J.,Báldi,A.,Kleijn,D.,2004.Relationshipbetweenland-useintensityandspecies
farmer'sperceptionsofclimatechangerisktoadaptationbehaviourinHungary. richnessandabundanceofbirdsinHungary.Agric.Ecosyst.Environ.104:465–473.
J.Environ.Manag.185:21–30.https://doi.org/10.1016/j.jenvman.2016.10.051. https://doi.org/10.1016/j.agee.2004.01.043.
Li,S.,Juhász-Horváth,L.,Pedde,S.,Pintér,L.,Rounsevell,M.D.A.,Harrison,P.A.,2017b.In- White,J.W.,Hoogenboom,G.,Kimball,B.A.,Wall,G.W.,2011.Methodologiesforsimulat-
tegratedmodellingofurbanspatialdevelopmentunderuncertainclimatefutures:a ingimpactsofclimatechangeoncropproduction.FieldCropRes.124:357–368.
casestudyinHungary.Environ.Model.Softw.96:251–264.https://doi.org/10.1016/ https://doi.org/10.1016/j.fcr.2011.07.001.
j.envsoft.2017.07.005.
Madsen,M.S.,Maule,C.F.,Christensen,J.H.,Fronzek,S.,Carter,T.,2016.IMPRESSIONScli-
matescenarios.EUFP7IMPRESSIONSProjectDeliverableD2.3.