Short-term forecasting of summer and winter storms using a mesoscale model and radar data assimilation

A mesoscale modeling system, RTFDDA, based on a high-resolution MM5, is tested for short-term storm predictions during the winter and summer seasons. The system continuously assimilates conventional observations as well as high-resolution observations from Doppler radar network, and provides three-dimensional analyses and short-term forecasts in a cycling fashion. The radar data assimilation scheme in RTFDDA includes adjustment of the rainfall mixing ratio and heating on model grid points based on the observed radar reflectivity, and nudging the wind toward vectors derived from radar radial velocities using the volume-velocity processing (VVP) method. The system is currently being upgraded from MM5-based to WRF-based.

In this work, impact of radar data assimilation on the storm prediction using the RTFDDA system is evaluated through case studies as well as real-time tests. Tests for storms in summer and winter regimes are described. A triply nested grid with a fine mesh of 3.3 km resolution is used in the tests. Techniques that can effectively assimilate real-time radar data into MM5 are explored. Predictability of the summer and winter storms, and feasibility of using the MM5 and radar observations for storm predictions are investigated.