On the Validity of ECHAM-Simulated Daily Extreme Temperatures

Three-dimensional global circulation models (GCMs) are state-of-the-art tools for projecting possible changes in climate. GCM simulations have frequently been validated against observations in terms of time-averaged variables while daily time series have not been studied extensively. In this paper, 30-year simulations of daily extreme temperatures are compared with the 30-year series recorded in Moravia. Attention is focused on the annual cycles, trimmed characteristics and average interdiurnal variability which are calculated for the average simulated series (4 gridpoints) and the average time series recorded in Moravia (5 stations). It is shown that daily extreme temperature variability is underestimated in the simulations, maximum (minimum) temperatures being underestimated (overestimated). Generally, the persistence of the simulated series is much higher, and small day-to-day changes are observed more frequenly than those in reality. The model is unable to reflect large changes between two consecutive days.