Abstract
This study investigates cold-season characteristics and the occurrence of ice episodes for a target region in northern Germany. Additionally, it identifies the characteristic atmospheric patterns associated with ice episodes using reanalysis data. A pattern-matching method is developed, which uses the structural similarity index to determine whether regional climate models (RCMs) can reproduce the ice-episode specific atmospheric patterns identified from the reanalysis data. The results show that the frequency of ice days is overestimated, but not the frequency of frost days. This suggests a daytime cold bias in the RCMs in this region, as maximum temperatures are too low. Moreover, most of the analysed RCMs simulate too many ice episodes (>5 consecutive ice days). The developed pattern-matching demonstrates, based on reanalysis data, that ice episodes in the target region can be associated with a variety of different atmospheric patterns. Specifically, longer ice episodes (lasting 14–29 consecutive ice days) in the target region are associated with a blocking pattern over Iceland/the British Isles. This is the most frequent pattern observed during ice episodes in the target region. The least frequent patterns show a system of low GPH500 over the European continent and are associated with shorter ice episodes (6–8 consecutive ice days) in the target region. The RCMs can reproduce these patterns and their frequency well, regardless of their forcing, and these patterns are also associated with ice episodes in the RCMs. Furthermore, it can be concluded that the “typical” blocking pattern is not a reliable indicator for ice episodes in the target region.
DOI Link to Publication