Extremes and their trends from the ENSEMBLES regional model integrations driven by ERA-40

The EU project ENSEMBLES is about is to produce probabilistic climate scenarios on the centennial time scale at high horizontal resolution, based on a multi-model ensembles of GCMs downscaled by an independent set of RCMs over Europe. Probabilistic scenario will be based on ability of these models to reproduce the present climate. For this purpose, the skills of 16 RCMs have been assessed from their ERA-40 driven integrations. In this work, daily extremes from the ERA-40 simulations have been compared with the ENSEMBLES 25km daily observational dataset of precipitation, and minimum and maximum temperature. Extreme indices have been estimated by Extreme Value Analysis. The main outcome of this study is to confirm the ability of RCMs to reproduce daily extreme events with good accuracy, although there are regions where the models are showing biases with common spatial patterns. The relation with some important covariates has also been studied, in particular the strong relationship between both extreme precipitation and minimum temperature and the NAO index over Europe are well reproduced by all the models.