Impact of radar data assimilation on storm predictions using a mesoscale model

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 latent heating on model grid points based on the observed radar reflectivity, and nudging the wind toward the vectors derived from radar radial velocities using the volume-velocity processing (VVP) method.

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. 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.