Page - 129 - in Loss and Damage from Climate Change - Concepts, Methods and Policy Options

5 Attribution:HowIs ItRelevant forLoss… 129 Fig.5.3 Example of a simplified detection and attribution study for global temperature. Points showobservedglobal temperature anomalies relative to1880–1920 (shadedblue topink to repre- sent cooler towarmer temperatures). These are compared tomodel simulated temperatureswith natural forcings only (blue), anthropogenic forcing only (orange), and a combination of natural andanthropogenic forcings (black).Asshown, theobservationscanonlybe reproducedwithboth naturalandanthropogenic forcing. SourceBindoffet al. (2013) IPCCAR5WGI,Box10.1Fig.1, p. 876 extremeweather events are rare, and their occurrence is strongly influencedbynat- uralvariability, it isnotpossible tosay thataspecificeventwouldnothaveoccurred without anthropogenic interference. However, it is possible to investigatewhether andhowanthropogenic emissions influenced theprobability andmagnitude.There are several differentmethods for examining the influenceof anthropogenic climate change on extremeweather events, including observational andmodel-based stud- ies (Stott et al. 2016). All methods use either large ensembles of climatemodels or statisticalmodels to estimate the likelihood of an event occurring in the current climate as well as with the anthropogenic climate drivers removed. The resulting frequencydistributions canbeused toestimate thechange in theprobabilitydue to anthropogenic interference (as inFig.5.4). Extreme event attribution studies are increasingly being applied to understand contemporary extreme events, and for the past 6 years theBulletin of the Ameri- canMeteorological Society has published a summary of attribution studies refer- ring to the previous year (Peterson et al. 2012, 2013; Herring et al. 2014, 2015, 2016, 2018). The science is advancing rapidly, evidenced in the large growth in the number of studies published, and the ability to make attribution statements more quickly: scientists are investigating the possibility of operational event attri- butionwhich coulddeliver statements in theweeks andmonths followinganevent