Changes in European drought characteristics projected by the PRUDENCE regional climate models

Abstract

One of the key features of global climate change will be perturbations to the hydrological regime across Europe. To date, assessments of the impacts of future change have generally used results from only one climate model, thus underestimating the range of possible change projected by different climate models. Here, the skill of six regional climate models (RCMs) in reproducing the mean precipitation for the 1961-1990 period for six catchments across Europe is compared and their projections of changes in future precipitation are assessed. A simple drought index based on monthly precipitation anomalies is also described and used to assess the models. Considerable variation in model skill in reproducing monthly mean precipitation and drought statistics is observed, with model errors in the reproduction of drought events independent of those for the mean, suggesting that the models have difficulties in reproducing the observed persistence of low monthly rainfall totals. In broad terms, the models indicate decreases in summer and increases in winter precipitation across Europe. On the regional scales required for impacts analysis, considerable model uncertainty is demonstrated for future projections, particularly for drought frequency. Although increases in the frequency of long-duration droughts are identified for catchments in southern Europe, the magnitude of this change is not certain. In contrast, for a catchment in northern England, such events are likely to become less frequent. For shorter-duration droughts, future changes encompass the direction of change. For stakeholders in each of the regions, these changes and uncertainties pose different challenges for the management of water resources. For the scientific community, the challenge raised is how to incorporate this uncertainty in climate change projections in a way that allows those groups to make informed decisions based on model projections. It is suggested that probabilistic scenarios for specific hydrological impacts offer considerable potential to achieve this. Copyright © 2007 Royal Meteorological Society.