A coupled atmosphere/ice/ocean model for the North Sea and the Baltic Sea

A hindcast experiment with a regional coupled atmosphere/ice/ocean model for the Baltic and North seas has been carried out. The experiment was performed over a full seasonal cycle and verified by comparisons with independent data. Further un-coupled model runs with the atmospheric and oceanic sub-models have been made and analyzed to evaluate the sensitivity of different coupling and regionalization strategies of atmospheric global climate variability to the regional system. Overall it could be shown that the regional coupled atmosphere/ice/ocean model is stable over a full seasonal cycle. Furthermore, the coupling on the regional scale turned out to be a clear improvement compared to the un-coupled run with the atmospheric model. The regional model results in the climate mode (without re-initialization) are of similar quality compared to atmospheric re-analysis for the North and Baltic seas, due to the stabilizing effect of the coupling. In the forecast mode, i.e. when no observational data are available to improve model estimates by assimilation, it can be expected that the usage of a regional coupled model system will improve the quality of predictions strongly. Nevertheless, the response of the oceanic sub-system on the different regionalization strategies showed some differences important for local applications of the coupled system. Three different sensitive surface-parameters have been identified by the present study: the sea surface temperature and connected the sea ice, showing considerable sensitivity on the regionalization as well as on the coupling, and the sea surface elevation, showing a high sensitivity for the short term variability on the regionalization