Seite - 143 - in Loss and Damage from Climate Change - Concepts, Methods and Policy Options

5 Attribution:HowIs ItRelevant forLoss… 143 events, and how those drivers are changing across time and spatial scales, in order to informadaptationandriskmanagement,hasbeenrecognisedasakeypriority for the next IPCC report (RC/RCCC2017).Herewegive a fewexamples to illustrate the importanceof this approach. InEastAfrica,manyclimatemodels suggest that the regionwill becomewetter in future (Shongwe et al. 2011; James et al. 2014b), whichmight imply the need to adapt towetter conditions, and potential for losses and damages from extreme precipitationorflooding (Shongweetal. 2011;Tayloret al. 2013).However,obser- vationssuggest that therehasbeenarecent increase indrought (Rowelletal.2015), andanalysisof theclimateprocessesassociatedwithprecipitationdeclinesuggest it couldbe causedbywarmingof the IndianOcean,which is expectedunder anthro- pogenic climate change (Copsey et al. 2006;Williams and Funk 2011;Williams etal.2012).Anattributionstudyof thedriversof theclimatehazards that resultedin theextremedrought in2010–11 (Lott et al. 2013) showed that itwas influencedby both climate change andnatural variability (with an important role for theElNino SouthernOscillation).Bycombiningevidencefromattributionresearchwithobser- vationalevidence,physicalunderstanding,andfutureprojections, it seemsclear that adaptationandL&DplanningforEastAfricashouldnotassumewetterfutures(Funk 2011),andshouldstrengthenmeasures to respondtodrought,whichcouldcontinue tooccurdue tonaturalvariability, andmaybeamplifiedbyclimatechange. Another reason that attribution research canprovide important evidence to help address losses anddamages, is that it offers an assessment of how risk is changing now. Formany decision-makers, information about how climatemight change in 30 years is not relevant because their planning horizons are much shorter (Jones etal. 2017).This is true for severalof theapproacheswhichhavebeensuggested to address losses and damages. For example, there has been a great deal of emphasis on risk pooling schemes and (re)insurance. These systems rely on estimates of the probabilityofextremeweathereventsbasedonhistoricaldata,whichmaynolonger be relevant in a changing climate. Attribution studies can provide an estimate of the current probability of extremeweather events. Finally, the above has focused on the hazard component of climate riskmanagement but attribution science can alsobe extended toprovide relevant informationon thevulnerability and exposure components. This principle is demonstrated by considering the impact of different responses to two category 4 tropical cyclones inMozambique (Benessene 2007; UNISDR2010).This showedsignificantly less lossof life inamore recentevent in 2007compared to2000,asa resultofbetterearlywarningsystemsreducinghuman exposure to ahazardof similarmagnitude.Another examplebyOtto et al. (2015b) showedthatclimatechangehadnotaltered the likelihoodof theprecipitationdeficit associated with the 2014–15 droughts in the Sao Paolo area. Thus higher losses in this case compared to earlier events could not be attributed to a change in the hazardand sowereattributable tohighervulnerability andexposure resulting from socio-economic changes. These examples demonstrate that attribution science can beuseful toguide thedesignof improved future responses toclimate-related risks.