Estimated changes in flood quantiles of the river Meuse from resampling of regional climate model output

Precipitation and temperature data from three regional climate model (RCM) experiments were used to assess the effect of climatic change on the flood quantiles of the French-Belgian river Meuse. In two of these experiments the RCM was driven by the global atmospheric model HadAM3H of the Hadley Centre (HC), and in the other experiment the RCM was driven by the global coupled atmosphere-ocean model ECHAM4/OPYC3 of the Max-Planck Institute for Meteorology (MPI). RCM simulations for the control climate (1961-1990) and the SRES-scenario A2 (2071-2100) were available. The HBV rainfall-runoff model was used to simulate river discharges. Long synthetic sequences of precipitation and temperature were resampled from the RCM output using a nearest-neighbour technique to obtain the flood quantiles for long return periods. The maxima of 10-day precipitation and discharge for the winter half-year (flooding season) were analysed. It was found that the changes in the extreme quantiles of 10-day precipitation and discharge were highly sensitive to the driving GCM. In the runs driven by HC, there was little change in the most extreme quantiles, whereas the MPI-driven run projected a remarkable increase. It is shown that this difference between the HC- and MPI-driven runs is strongly related to the change in the coefficient of variation of the 10-day precipitation amounts, which decreases in the former and hardly changes in the latter. The relevance of bias correction of RCM output with regard to the estimated changes of flood quantiles is demonstrated.