Abstract : Q.1
Optimizing the COSMO-2 model for quantitative precipitation forecasts in the framework of MAP D-PHASE

Felix Ament, Marco Arpagaus
Federal Office of Meteorology and Climatology MeteoSwiss

The Swiss weather service MeteoSwiss is currently implementing a new high resolution, deep convection resolving version of the COSMO model (COSMO-2) with a grid spacing of 2.2 km (formerly known as LM). Model formulations of COSMO-2 are almost identical to those of the standard COMSO model with a resolution of 7 km. Exceptions are improved numerical schemes comprising more positive definite advection, enhanced microphysics by considering graupel and replacing the convection parameterization by a simple, shallow convection scheme. Participating in D-PHASE this summer will be the first practical test for COSMO-2. This presentation will address the questions what benefits from COSMO-2 can be expected with respect to quantitative precipitation forecasts (QPF) and whether additional optimization are needed at the convection resolving scale.

We will present hindcasts for precipitation events in the Alpine region, including e.g. MAP 1999 cases and the flooding event in August 2005. Traditional coarse scale verification against gridded rain gauges measurements reveals that both the standard 7km version of COSMO and the high resolution version COSMO-2 yield quite similar forecasts, which agree well to observations. However, a fine scale verification based on radar data using multi-scale fuzzy methods allows to identify the advantages of COSMO-2 at small scales. Additionally, the hindcasts are used as reference for sensitivity experiments which give some insight into the relevance of various factors, like e.g. initial/boundary conditions, numerical schemes or formulation of parameterizations, on the QPF quality. These findings will serve as a guideline for further optimization of COSMO-2.