Climatological studies indicate that climate change lead to an increase in the mean global temperature of around 0.5 °C until the end of the twentieth century. This warming impacts the atmospheric humidity, wind, radiation, and precipitation. However, the magnitude of changes is not equally distributed over the globe but differs markedly with regions, making a regionalization of the global information essential. The GLOWA-Danube project follows such a downscaling approach with the focus on the drainage basin of the Upper Danube River.
Since the horizontal resolution of global climate simulations (approx. 200 km) is still too coarse for studies on regional or even local scales, the results of the global projections are post-processed using the regional climate model MM5 to downscale the information to 45 km. The empirical-statistical method AtmoMM5 subsequently downscaled the MM5 results to the desired resolution of 1 km.
In order to test the quality of the meteorological data used in DANUBIA, the time series of different proxels within the period from 1991 to 2000 were compared with corresponding periods from the station measurements. The station measurements are interpolated with respect to distance and direction on a regular grid. These gridded fields in general follow the pattern of the observed precipitation and 2 m air temperature. The scaling of the regional model MM5 with AtmoMM5 causes a distinct reduction of the root mean square error between the observations and MM5 or AtmoMM5, respectively.