Flood event attribution and damage estimation using national-scale grid-based modelling: Winter 2013/2014 in Great Britain

摘要

A sequence of major flood events in Britain over the last two decades has prompted questions about the influence of anthropogenic greenhouse gas emissions on flood risk. Such questions are difficult to answer definitively, as a range of other factors are involved, but modelling techniques allow an assessment of how much the chance of occurrence of an event could have been altered by emissions. Here the floods of winter 2013/2014 in Great Britain are assessed by combining ensembles of climate model data with a national-scale hydrological model and, for one severely impacted river basin (the Thames), a detailed analysis of flood inundation and the increased number of residential properties placed at risk. One climate model ensemble represents the range of possible weather under the current climate, while 11 alternative ensembles represent the weather as it could have been had past emissions not occurred. The pooled ensemble results show that emissions are likely to have increased the chance of occurrence of these floods across much of the country, with a stronger influence on longer duration peaks (similar to 10 days or more) than for shorter durations (consistent with observations). However, there is substantial variation in results between alternative ensembles, with some suggesting likely decreases in the chance of flood occurrence, at least in some regions of the country. The influence on flows and property flooding varies spatially, due to both spatial variation in the influence on precipitation and variation in physical properties that affect the transformation of precipitation to river flow and flood impacts, including flood defences. This complexity highlights the importance of using hydrological modelling to attribute hydrological impacts from meteorological changes. Changes in snow occurrence in a warming climate are also shown to be important, with effects varying spatially.