High-Resolution Climate Predictions and Short-Range Forecasts to Improve the Process Understanding and the Representation of Land-Surface Interactions in the WRF Model in Southwest Germany (WRFCLIM)

摘要

The use of numerical modeling for climate projections is an important task in scien tific research since they are the most promising means to gain insight in possible cli mate changes. The quality of the prepared predictions has been constantly improved in recent years, enabled by more powerful supercomputers as well as advanced nu merical and physical schemes [e.g. 8, 14, 17]. The understanding of the interaction of anthropogenic climate change with natural climate variability on the global scale (grid boxes of 100 km and coarser) is steadily increasing. In the meantime, regional climate simulations with grid resolutions of 10-50 km became available in the cli mate modeling community [e.g. 3, 6, 10]. However, several effects severely limit the improvement of the skill of the simulations on the mesoscale. These include: a) Incorrect boundaries of global models resulting from incorrect physics and ini tial conditions, b) inconsistent physics between global and regional models, and c) poor consideration of orography and the heterogeneity of land-surface-vegetation properties. These deficiencies result in erroneous regional simulations of feedback processes between the land-surface and the atmospheric boundary layer as well as of the development of clouds and precipitation. The convection parameterization, which is required down to scales of the order of 4 km, has been identified as a key reason for significant systematic errors in QPF in regional climate simulations and mesoscale weather forecasting.