Investigating radar data assimilation for winter cases using ICON-KENDA system

Abstract

Since 2017, the SINFONY (Seamless INtegrated FOrecastiNg sYstem) project has been under development at the Deutscher Wetterdienst (DWD). It is aimed to provide a seamless ensemble system for early predictions and warnings of severe weather events by combining the nowcasting based on extrapolating observed radar reflectivity and short-term forecasts initiated from the Rapid Update Cycle (RUC) of data assimilation for the convection-permitting ICON (ICOsahedral Nonhydtostatic) model. So far, the ICON-RUC setup has been extensively tested for convective summer cases. In this study, a series of sensitivity experiments have been conducted for the winter precipitation, including the choice of microphysics schemes and the Latent Heat Nudging (LHN). Results show that within data assimilation cycles the two-moment scheme outperforms the one-moment scheme, and the LHN has also positive impacts. For the 6-h reflectivity forecasts, the two-moment scheme is clearly better than the one-moment scheme and the added values by using the LHN persist almost 6 h. For the precipitation forecasts, the two-moment scheme also exhibits advantage for the light precipitation, however, for the moderate precipitation, the one-moment scheme prevails. Current results indicate that the two-moment has to be enhanced for the moderate precipitation in winter.