Analysis of CCLM model biases in relation to intra-ensemble model variability

Abstract

We present an analysis of the climatologies and the interannual variations of 2 m temperature and precipitation for ensemble simulations performed with the CCLM for Europe. Special focus is put on the analysis of the intra-ensemble variabilities. CCLM predicts a pronounced cold bias in spring and summer over large parts of Europe north of approximately 40N. In contrast, the model suffers from a distinct warm bias in Northern Europe in winter. Precipitation is clearly overestimated in Europe during all seasons, with the exception of a dry summer bias extending from Italy to the Black Sea. CCLM distinctly overestimates the precipitation variations over most of Europe while temperature variations tend to be underestimated in Northern Europe and overestimated in the Mediterranean. The study shows that temperature and precipitation biases for both the mean and interannual variability are critically dependent on the ensemble member that is selected for the evaluation. It is therefore essential to use ensemble simulations for model validation in order to avoid random model biases for both the mean and interannual variability of temperature and precipitation. The results suggest that the intra-ensemble variability relative to the model bias is higher for interannual variability than for the mean. It can thus be concluded that the model's performance in predicting climate extremes cannot be properly evaluated using only one model simulation. Finally, we also compare several extreme indices (largest number of consecutive frost days, summer days, wet days, and dry days). The study shows that the cold summer bias leads to an underestimation of the largest number of consecutive summer days over major parts of the model domain. The wet model bias in Central and Northern Europe leads to a distinct underestimation of the largest number of consecutive dry days.