Rapid Urban Impact Appraisal

Abstract

Bridging the global‐regional divide inclimate impact research for urban areas means to establish a comprehensive picture which covers all urban agglomeration of the world. This is different from the case of, e.g., hydrological impact modeling where coarse‐scaled (spatial and functional) global models and detailed regional studies have to be brought together. Therefore we suggest a structured approach towards a full spatial and functional coverage of urban impact analyses: (1) Filtering ‐ all urban agglomerations are identified where a specific Climate Change impact path is probably relevant or even the dominant one and (2) a targeted, fast quantitative impact assessment of the respective impact path is performed for these urban areas. Step (1) starts with the existing knowledge on potential urban impact paths and extracts through different natural, social and economic filtering steps the urban agglomerations where these impact paths have to be studied quantitatively. In step (2) this is done by applying a set of tools which are mainly based on urban remote sensing to overcome the data scarcity bottleneck. It occurs that single filtering steps and tools can be reused for different impact paths. To illustrate the approach we presenta filtering example, resulting in a global map which shows the urban agglomerations where the following impact path is relevant: pluvial flooding of slum settlements under increasing frequency of heavy rain events To exemplify step (2) we present a remote sensing based toolset for quantitative assessment.

Full text

ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
1

RapidUrbanImpactAppraisal
MatthiasK.B.Lüdeke&OleksandrKit

Abstract
Bridgingtheglobal‐regionaldivideinclimateimpactresearchforurbanareasmeansto
establishacomprehensivepicturewhichcoversallurbanagglomerationoftheworld.This
isdifferentfromthecaseof,e.g.,hydrologicalimpactmodelingwherecoarse‐scaled
(spatialandfunctional)globalmodelsanddetailedregionalstudieshavetobebrought
together.Thereforewesuggestastructuredapproachtowardsafullspatialand
functionalcoverageofurbanimpactanalyses:(1)Filtering‐allurbanagglomerationsare
identifiedwhereaspecificClimateChangeimpactpathisprobablyrelevantoreventhe
dominantoneand(2)atargeted,fastquantitativeimpactassessmentoftherespective
impactpathisperformedfortheseurbanareas.Step(1)startswiththeexisting
knowledgeonpotentialurbanimpactpathsandextractsthroughdifferentnatural,social
andeconomicfilteringstepstheurbanagglomerationswheretheseimpactpathshaveto
bestudiedquantitatively.Instep(2)thisisdonebyapplyingasetoftoolswhichare
mainlybasedonurbanremotesensingtoovercomethedatascarcitybottleneck.Itoccurs
thatsinglefilteringstepsandtoolscanbereusedfordifferentimpactpaths.To illustrate
theapproachwepresentafilteringexample,resultinginaglobalmapwhichshowsthe
urbanagglomerationswherethefollowingimpactpathisrelevant:pluvialfloodingof
slumsettlementsunderincreasingfrequencyofheavyraineventsToexemplifystep(2)
wepresentaremotesensingbasedtoolsetforquantitativeassessment.
IndexTerms—climateimpactassessment,urbanagglomerations,remotesensing,data
scarcity
————————————————————
1 Introduction
Severalsinglestudiesonclimatechangeimpactsonurbanagglomerationsareavailablewhilea
globalimpactmodelfortheurbanagglomerationsoftheworlddoesnotexist.Soinurbanimpact
researchthemethodologicalchallengeofbridgingtheglobal‐regionaldivideisdifferentfromthe
caseof,e.g.,hydrologicalimpactmodelingwherecoarse‐scaled(spatialandfunctional)globalmod‐
elsanddetailedregionalstudieshavetobebroughttogether.However,globalcoverageofurbanim‐
pactassessmentsisnecessarybecause(1)eachurbanareashouldhaveatleastaroughestimateof
climatechangeimpactstheywillencounterasafirstorientationforlocaladaptationdecisions,(2)
thesumofalllocalurbanadaptationcosts/effortshastobeincludedintotheglobalbalancebe‐
tweenadaptationandmitigationand(3)international(EU,UN)policiesthatneedtostrikeabalance
betweenthecostsandbenefitsforindividualmemberstatesneednationalquantitativeestimateds
ofimpactsonurbanareas.
Inthispaperwesuggestanapproachtowardsamorecomprehensiveandsystematicglobalurban
impactassessmentwhichidentifiessubsetsofcitiesbeingsensitivitetospecificclimatechangeim‐
pactsandprovidestoolsforquantitativeimpactassessmentalongthesespecifities.Inparticular
thesequantitativeassessmentsareratherdifficultinlargeurbanagglomerationsindevelopingand

ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
2

newlyindustrializedcountries.Mostoffutureurbanizationwillhappenherebutduetoinformality
andrapidnessofdevelopmentthedatabasisisforquantitativeimpactassessmentisoftenunsuffi‐
cient.Theassessmenttoolshavetoreflecttheseconditionsby,e.g.,usingurbanremotesensing
techniquesfordataacquisitiontoovercomethedatabottleneck.Startingfromexperiencesgainedin
acomprehensiveimpactassessmentforHyderabd/Indiaweproposeasystematicandfeasiblewayto
obtainaglobalandquantitativeoverviewonclimatechangeimpactsoncities.Wefurthermoreshow
aspecificexamplewherewealreadyappliedthisapproach.Inthefollowingsectionwesketchthe
basicstructureoftheapproach,insection3wegiveanexamplefortheidentificationofcitysubsets
withsimilarimpactsensitivitiesandinsection4anexampleforaquantitativeimpactassessment
tool.
2 BasicIdea:atwo‐stepprocedure
Wesuggestastructuredapproachtowardsafullspatialandfunctionalcoverageofurbanimpact
analyses:
(1)Filtering‐allurbanagglomerationsareidentifiedwhereaspecificClimateChangeimpactpathis
probablyrelevantoreventhedominantoneand
(2)Atargeted,fastquantitativeimpactassessmentoftherespectiveimpactpathisperformedfor
theseurbanareas.









Figure1:Subsetofurbanclimateimpactpaths.Theredpathwillbeexemplarilyanalyszedusingthe
suggestedrapidurbanimpactappraisalapproach(hereimpactpathsweretakenfromReckienetal.
2011)

Step(1)startswiththeexistingknowledgeonpotentialurbanimpactpathsandextractsthroughdif‐
ferentnatural,socialandeconomicfilteringstepstheurbanagglomerationswheretheseimpact
pathshavetobestudiedquantitatively.Theimpactpathsarecharacterizedbyaspecificclimatic
stimulus(e.g.aflood,heatwaveorstormevent),anexposureunit(e.g.thetrafficsystem,settle‐
ments,thewatersupplysystem)andthetypeofimpact(e.g.structuraldamage,operationaldeterio‐
rationorhealthimpacts)–seeFig.1.Sourcesfortheseimpactpathsarethenumerousdetailedcase
studiesforsinglecities(forourexampleweusedtheHyderabadcaseasastartingpoint).Oncean
impactpathischosen,filterscanbeconstructedwhichexcludeurbanareaswheretherespective
climaticstimulusortheexposureunitareirrelevant.Thesefiltersarebasedonglobaldatasetschar‐

ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
3

acterizingclimatological,physicalandsocio‐economicpropertiesoftheurbanareasfromdifferent
sources.Theclimaticstimulus“Pluvialflooding”forinstancewillbeonlyrelevantforcitiesinclimatic
zoneswithstrongrainevntsandahillyurbanorography.Ontheotherhand,“fluvialflooding”re‐
quiresacitywithalargeupstreambasin.Thisstimulusisnottobeexpectedforlocationsnearwater‐
sheds.Thebenefitofthisfilteringstepforaspecificcasestudyisthepriorisationoftheimpactpaths
tobestudied.Regardingtheglobaloverviewalreadythisfirststepresultsinaninterestingmapof
urbanagglomerationsbeingsensitivetowardsthesamespecificimpactpath.Forstep(2)anurban
remotesensingorientedtoolboxwasdevelopedtoquantifyimpactsalongthechosenrelevantim‐
pactpath.InFigure1differenturbanimpactpathsaredisplayedexemplarily(see,e.g.,Reckienetal.,
2011).Theredimpactpathasksforthenumberofslumdwellersseverelyaffectedbypluvialflooding
andhowthiswouldchangeunderclimatechange.
3 Anexampleforthefilteringstep
InthefollowingwewilldemonstratethefilteringstepsfortheredimpactpathinFig.1,dealingwith
theclimaticstimulusofpluvialflooding.
Figure2illustratesthefilteringstepsnecessarytoidentifyurbanareaswhicharesusceptibletothe
choosenimpactpath.Thefirstfilteringstepexcludescitiesinclimaticzoneswhichtypicallydonot
experiencehighintensityrainfalleventsasgivenbytheKoeppen‐Geigerclimaticzones.Thesecond
stepidentifiesurbanagglomerationswhicharenotsensitivetofluvialfloodingbecausetheyare
closetoawatershed(i.e.veryupstreamintheriverbasin,withinabufferzonearoundthewatershed
of100km)andfarfromcoasts(noestuary,atleast50kmdistancefromcoast).Step3excludescities
whichdonotshowahillyurbanlandscape(smallmeanabsolutcurvature)andatleasturbanareas
withalowprobabliltyofslumoccurrence(lessthen3%urbanslumpopulationaccordingtoUNsta‐
tistics)arefilteredout.ThereddotsinFig.2ddenotetheremainingurbanareaswhicharesuscepti‐
bletowardsthechosenimpactpath.Fig.3zoomsintotheglobalresultandshowsthecitieswhere
theslumpopulationispotentiallyendangeredbypluvialflooding.Insection4wewillshowforone
ofthesecitieshowtodoafastquantitativeimpactassessmentalongthisimpactpath.



ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
4


A

B

C

D

Figure2:Largeurbanagglomerations(>1000km2)filteredforthefollowingcharacteristics:
a)experiencinghighintensityrainfall,b)additionallyclosetowatershedsanddistanttocoasts,c)
additionallyhillyurbanlandscaped)additionallyhighprobabilityofurbanslumsettlements.Red:
urbanagglomerationsremainingaftertherespectiveconsecutivefilteringsteps.Blackandgrey:ag‐
glomerationsexcluded.





ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
5


Figure3:LargeurbanagglomerationsinIndiawhicharesusceptibleforpluvialfloodingofslumset‐
tlements(detailofFig.2d)
4 Fastquantitativeimpactassessment
Inthissectionwepresentanexampleforthesecondstep.Wechoosetheimpactpathofpluvial
floodingofslumsettlementsforwhichweintroducedtheglobalfilteringinsection3.Theidentified
urbanagglomerationsareaffectedbythisprocessbutthequantitativeimpacthasstilltobedeter‐
mined.InFigure4weshowallstepstobeperformedforobtainingthequantitativeimpactandits
uncertaintyfortheexampleofHyderabad/India.Fig.4ashowsurbanlocationswhichareseverely
floodedunderdifferentprojectionsofthe“onceintwoyearpercentile”ofexpecteddailyprecipation
dependingondifferentglobalemissionscenarios(B1,A2).ForthepresentHyderabadclimatethis
percentileamountsto80mm/dayandwaschosenduetohistoricalevidenceofsevere,city‐wideim‐
pacts.Ifpossible,forothercitiesaffectedbythisimpactpaththisthresholdhastobeempiricallyver‐
fified.Therangeoftheprojectionsoftheconsideredclimatevariableisdenotedbythehatchedrec‐
tanglesinFig.4a,top.Halfoftheconsideredglobalclimatemodels(AOGCMsfromtheIPCCAR4
modelensemble)projectvalueswithinthisrangeaftertheywerestatisticallydownscaledtotheHy‐
derabadregion(Lüdekeetal.,2012).Toidentifywhichadditionalareaswillbeaffectedbysevere
floodinginthefutureaflow‐accumulationanalysiswasperformed(DEMtakenfromSRTMremote
sensing,seeKitetal.,2011).To identifytheexposureunit,aremotesensing(QuickBirdsatellite)
basedidentificationofslumareaswasdeveloped.Hereweusetherelationoftheurbantexture
(measuredbylacunarity)withtheprobabilityofslumoccurrencebecauseslumareasshowatypical
settlementstructure(Kitetal.,2012).AppliedtodifferentQuickBirdtimeslicesitallowstoidentify
spatiallyexplicittrendsinslumdevelopmentduring2003to2010(Kitetal.,2013)asshowninFig.
4b.Thiscurrenttrend(roughly:reducedslumpopulationinthecentralpartofthecity,mostlydueto
slumupgradeandnewlyoccurringslumareasatthefringeoftheinnercity)wasusedtogetherwith

ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
6

projectionsofthetotalpopulationtoproduceplausiblescenariosoffutureslumdevelopmentupto
2050.InFig4ctheimpactonslumdwellersisquantified.Itshowstheward‐wiseevaluationofaddi‐
tionalslumdwellersseverlyaffectedbyfuturepluvialfloodingin2050undertheA2scenario,the
extrapolatedcurrentslumdevelopmentandtheassumptionofexponentialpopulationgrowth
withinthecity.Clearspatialhotspotscanbeidentifiedwhichimplyprioritizationofe.g.stormdrain‐
ageimprovementactivities.Thetotalnumberamountstoabout78000dwellersadditionallyaf‐
fected,theuncertaintyrangeof[20000,193000]takesintoaccountthewholerangeofclimatepro‐
jectionsbytheensembleoftheAOGCMs,includingtheoutliers.Assumingtheaverageclimatepro‐
jectionandchangingbetweenexponentialandlinearpopulationgrowthgeneratesanuncertainty
rangeofthesameorderofmagnitude.


Fig.4:FastquantitativeclimatechangeimpactassementforHyderabad/Indiawithregardtotheex‐
pectednumberofslumdwellersseverelyaffectedbypluvialfloodingunderclimatechange.a)Drive‐
r:onceintwoyearpercentileofexpecteddailyprecipationunderdifferentglobalemissionscenarios
(B1,A2,fordetailsseetext).Flow‐accumulationbasedidentificationofareasseverelyaffectedbythe
resultingpluvialflooding(Kitetal.,2011).b)Remotesensingbasedidentificationofslumareas(Kit
etal.,2012).c)Ward‐wiseevaluationofthenumberofslumdwellersadditionallyseverlyaffected
underfuturepluvialflooding(fordetailsseetext)undertheA2scenarioandtheassumptionofex‐
ponentialpopulationgrowthwithinthecity.

ImpactsWorl d2013,InternationalConferenceonClimateChangeEffects,
Potsdam,May27‐30
7

5 Conclusions
Thepresentedexamplesforthefilteringofcitiesaffectedbyspecificimpactpathsshowedhowcom‐
parablesubsetsofcitiescanbeidentifiedandthen,inasecondstep,befurtherinvestigatedwith
similaranalysistoolstoobtainquantitativeimpacts.Theexamplefromsection4forsuchatoolset
mailydependsonremotelysensedandgloballyavailableinputdatasets,i.e.globaldataavailability
wouldallowtoapplyittoallfilteredcitiesresultinginaworldwidequantitativeevaluationofthe“se‐
verepluvialfloodingofslumdwellers”impactpath,relyingonaminimumofgroundbaseddata,in‐
cludingsomecalibrationdatafortheslumidentificationalgorithm,atleastexemplarilyforlarger
worldregionslikeIndia,Souith‐America,Africa.
Slightmodificationsandrecombinationofthefilteringstepsinsection3yielddifferentbutalsovery
relevantpathssothatanincreasingcollectionofsuchpartialfilterswillcoveraverylargenumberof
relevantclimateimpactpaths.Thesefiltersrelyonaggregated,structuralindicatorsforurbanareas
whicharerelatedtothesensitivitytowardsclimatechange.Furtherresearchtodiscoversuchrela‐
tionsisaprerequisiteforachievingamorecomprehensiveoverviewonclimateimpactsoncities.
Theproposedapproachprovidesaframeworktointegratethiskindofpartialknowledgeinasys‐
tematicmanner‐possiblyleadingtoawellfoundedglobalpictureofurbanclimatechangeimpacts.
6 References

Kit,O.;Lüdeke,M.K.B.;Reckien,D.,2013.Definingthebull'seye:satelliteimagery‐assistedslum
populationassessmentinHyderabad/India.UrbanGeography,onlinefirst
Kit,O.;Lüdeke,M.K.B.;Reckien,D.,2012.Text u r e ‐basedidentificationofurbanslumsinHydera‐
bad,Indiausingremotesensingdata.AppliedGeography32,660‐667p.
Kit,O.;Lüdeke,M.K.B.;Reckien,D.2011.Assessmentofclimatechange‐inducedvulnerabilityto
floodsinHyderabad/Indiausingremotesensingdata.In:ResilientCities‐CitiesandAdaptationto
ClimateChangeEd.:Otto‐Zimmermann,K.Dordrecht:Springer35‐44p.
Lüdeke,M.K.B.;Budde,M.;Kit,O.;Reckien,D.2012.ClimateChangeScenariosforHyderabad:in‐
tegratinguncertaintiesandconsolidation.EmergingmegacitiesV1/2010,ISSN2193‐6927,pp3‐37
ReckienD,LüdekeM,ReusswigF, KitO,Meyer‐OhlendorfL,BuddeM,2011.Hyderabad,India,in‐
frastructureadaptationplanning.InRosenzweigC,SoleckiWD,HammerSA,MehrotraS:Climate
ChangeandCities–FirstAssessmentReportoftheUrbanClimateChangeResearchNetwork,Cam‐
bridgeUniversityPress,pp152‐154


