A High-Resolution Simulation of the Year 2003 for Germany Using the Regional Model COSMO

Abstract

n this article, the authors examine the effect of a high-resolution grid (grid resolution lower than 3 km) in the context of a realistic climate simulation. For this purpose global simulation results of the German Weather Service were dynamically downscaled in a one-way nesting approach to 2.8 km using the regional forecast model of the Consortium for Small-Scale Modeling (COSMO) of the German Weather Service. The simulations were performed for the region of central Europe in 2003. In particular, the authors investigate whether COSMO adequately handles extreme events such as the persistent drought and heat of the summer of 2003. Comparisons of the simulated atmospheric conditions in terms of 2-m temperature, mean sea level pressure, and precipitation demonstrated a good correspondence to their associated observational data. Simulation results and observed data are both given as time series at locations such as grid cells or station locations. By cluster analysis a representation of the spatial structure for observation data and simulation results is found. The kappa statistic evaluates how well the two spatial structures correspond to each other. The different kappa variants are helpful to diagnose where shortcomings of the simulation results are located.