Persistent uncertainties in the magnitude of future river floods

Air can hold more moisture as temperature increases, leading to more extreme precipitation events. Yet, in many locations, this does not result in larger river floods. Here we use global projections to explore differences in the response of the atmosphere and catchments to an increase in global mean temperature. We focus on changes in the amplitude of extreme precipitation events and river floods per °C above pre-industrial levels. We rely on global projections produced as part of the ISIMIP2b and ISIMIP3b projects based on CMIP5 and CMIP6 climate models, respectively. We compute changes in the median of annual maxima based on periods of 31 years on 0.5° global grids. 

We find that whilst extreme precipitation is projected to increase over a large majority of the land area, a much smaller fraction of the land area is projected to show an increase in extreme flow magnitude. Importantly, whilst there is high model agreement that extreme precipitation will increase, agreement that future flows will increase is significantly lower. Specifically, ensemble spread for fluvial changes is typically wider and more likely to encompass both increases and decreases than for pluvial changes. We connect these discrepancies to changes in land-surface processes projected by the global hydrological models, highlighting the importance of river flood drivers other than extreme precipitation and illustrating the limits of using a Clausius-Clapeyron narrative to predict future changes in river floods. We compare ISIMIP2b and ISIMIP3b projections to underline the persistence of uncertainties in the magnitude of future river floods and discuss their implications for adaptation.