Quality of a climate reconstruction for the CADSES regions

Abstract

We used the regional climate model CCLM 4 to reconstruct the climate over the CADSES regions for the period 1961-2000. This most recent model version is currently still in a test phase. The simulation quality for this part of the entire simulation area is exemplarily analyzed by using a small but representative subset of model variables. Mean sea level pressure is chosen as an indicator for the behavior of the near-surface dynamics. Furthermore, the two regional variables 2 m temperature and precipitation are evaluated, which are of relevance for hydrological applications. For additional analyses to identify sources of uncertainty, we used soil moisture and cloudiness. Altogether, the accuracy of the simulation results can be judged as being in good agreement with the results of similar simulations by other regional climate models over Europe. Area-averaged monthly mean 2 m temperatures during winter show a cold bias of about -2 K. The simulated mean sea level pressure is with a bias of about 2 hPa slightly too high compared to the used reference data set. Although cloud cover evaluation hints at insufficiencies in the corresponding parameterization, further analyses are required to unambiguously identify the causes for this cold bias. During summer, higher 2 m temperatures linked to reduced precipitation are identified in nearly identical regions over southern and southeastern Europe. Further analyses of soil moisture evolution and cloud cover give evidence for a source of uncertainty in the surface hydrological scheme. Altogether, the detected differences do not exceed the typical ranges of uncertainty in comparable regional climate simulations over Europe. For input in hydrological applications, however, the uncertainties produced by the CCLM simulation described here have to be reduced in the medium term.