We evaluate simulation of Central European heat waves in the recent climate and projected changes for the 21st century in an ensemble of regional climate model (RCM) simulations from the ENSEMBLES and CORDEX projects. Observed data are represented by the E-OBS gridded dataset.
Heat waves are defined based on excesses above the 90th percentile of summer daily maximum temperature (Tmax), and their temporal as well as spatial characteristics are considered. In historical runs, RCMs tend to simulate more heat waves than observed (although the percentiles are calculated for each RCM in order to remove Tmax biases), which is due to overestimated clustering tendency of days with high temperature in the RCMs.
The simulated heat waves have also too pronounced temperature peaks. We examine also driving mechanisms for major heat waves over Central Europe in the past, especially the most severe 1994 heat wave, and show that its reproduction in the RCMs is closely linked to simulated precipitation and soil moisture.
Projected changes of the heat wave characteristics are analysed for two time slices corresponding to near (2020-2049) and far (2070-2099) future. Although the RCMs generally agree on the increased heat wave severity, the multi-model spread is large, indicating considerable uncertainty in climate change scenarios of Central European heat waves.
Better understanding the RCMs' biases and uncertainties related to such factors as emission scenarios, driving global models, spatial resolution and natural variability, will be essential for impact studies as well as design of suitable adaptation and mitigation strategies.