Impacts of horizontal resolution on simulated climate statistics in ECHAM4

The sensitivity of a general circulation model to changes in resolution is studied using the Max Planck Institute for Meteorology (MPI) 19-level model, ECHAM4. Simulations extending over a period between 10 and 15 years, with observed sea surface temperatures as lower boundary conditions from 1979 onward, have been performed using four different horizontal resolutions, T21, T30, T42 and T106. The atmospheric time-mean state and the intraseasonal variability are compared to the European Centre for Medium Range Weather Forecasts (ECMWF) reanalyses and a few other observational datasets. The T30, T42 and T106 simulations are similar in many respects, whereas the T21 simula- tion is qualitatively different. Several effects related to model resolution could be identi- fied, such as increasing tropical upper tropospheric warming with increasing resolution. This is due to more vigorous tropical convection, larger ice water content and, hence, increasing cirrus cloud greenhouse effect. Associated with this increasing warming at higher resolution is a poleward expansion of the zonally averaged circulation regime. On the other hand, the zonally asymmetric component of the circulation, i.e., the stationary waves, improve with higher resolution. Also, higher resolution has a positive impact on regional precipitation patterns which are affected by orography such as the summer mon- soon precipitation over India. Intraseasonal variability has been analyzed only for the higher resolution models, T42 and T106. Compared to the ECMWF reanalyses, both models are able to simulate the intrasea- sonal geopotential height variability, eddy fluxes of heat and momentum, and eddy kinetic energy with reasonable accuracy. This applies to transient eddies in both the bandpass and lowpass regime and to the stationary eddies as well. Some biases can be identified which are more or less independent of resolution. These include the mislocation of the Azores high and the overestimation of its intensity, a cold bias in the polar upper troposphere and lower stratosphere and the poleward and upward displacement of the maxima of geopotential height variability, momentum fluxes and eddy kinetic energy. An important finding is that the operational ECMWF analyses, which have been widely used for model validation, considerably overestimate the lowpass variability, as compared to the reanalyses, due to frequent changes of the forecast model and data assimilation scheme. This implies that the results from our investigations are not directly comparable to previous investigations that used operational analyses for validation.