 |
Center for Nonlinear Studies |
High resolution climate information is needed for adaptation to and mitigation of the impacts of climate change and variability at a range of time scales. For both regional and global climate models (RCMs and GCMs), there is a general perception that increased horizontal resolution can improve the simulation of regional climate features. However, there have been relatively few systematic investigations of the sensitivity of climate model simulations to horizontal resolution. Here we make use of a unique multi-model ensemble produced by the European ENSEMBLES project. Participating institutions performed two reanalysis-driven RCM simulations at 50 km and 25 km grid spacings over a large European domain. In all other respects (e.g., lateral boundary conditions, common minimum domain, parameterization settings) the model simulations are the same at the two resolutions, thus providing an opportunity to assess resolution effects on RCM simulations. We focus on precipitation, likely the most sensitive variable to model resolution. A comparison of the seasonal spatial and temporal mean and variability of the modeled precipitation with high-resolution gridded climate data shows improvements at higher resolution for the summer season in most e regions, while the opposite is true in winter. Application of the Kolmogorov-Smirnov test indicates that the distributions of the 25 km and 50 km simulated daily precipitation are significantly different over most land areas, especially over areas of high topography. This difference is due to a better simulation of rainy day frequency and intensity at 25 km. Overall the results indicate that higher resolution yields improvements in some areas and in some seasons, but improved performance is not universal, indicating that increased resolution alone does not necessarily result in better simulations.