The representation of a synoptic-scale weather system in a thermodynamically adjusted version of the ECHAM4 general circulation model

Summary In this work, a strong cyclone event is simulated by the general circulation model (GCM) ECHAM4 for studying the representation of weather systems in a climate model. The system developed along the East Coast of the U.S.A. between the 12th and 14th of March 1993. The GCM simulation was started from climatological conditions and was continuously forced to the analyzed state by a thermodynamical adjustment based on the Newtonian relaxation technique (nudging). Relaxation terms for vorticity, divergence, temperature, and the logarithm of surface pressure were added at each model level and time step. The necessary forcing files were calculated from the ECMWF re-analysis (ERA15). No nudging terms were added for the components of the water cycle. Using this forcing, the model was able to reproduce the synoptic-scale features and its temporal development realistically after a spin-up period. This is true even for quantities that are not adjusted to the analysis (e.g., humidity). Detailed comparisons of the model simulations with available observations and the forcing ERA15 were performed for the cyclone case. Systematic errors were detected in the simulation of the thermodynamic state of the atmosphere, which can be traced back to deficiencies in model parametrizations. Differences in the representation of the surface fluxes lead to systematic deviations in near-surface temperature and wind fields. The general situation is very similar in both model representations. Errors were detected in the simulation of the convective boundary layer behind the cold front. The observed strong convective activity is missed both by the adjusted ECHAM4 simulation and ERA15. This is most likely caused by weaknesses in the cloud and convection schemes or by a too strong downdraft compensating the frontal lifting and suppressing the vertical transport of moisture from the boundary layer to higher levels. This work demonstrates for the investigated case the value of simulating single weather events in climate models for validating model physics.