Impact of high temporal frequency radar data assimilation on storm-scale NWP model simulations

Radial-velocity and reflectivity observations from Doppler radars can provide important information for initializing numerical storm-scale prediction models and in diagnosing the evolution of severe weather events like thunderstorms and tornadoes. Recent research indicates that the assimilation of Doppler radar data using the Ensemble Kalman Filter (EnKF) approach generates good estimates of storm structure. While the conventional Doppler radar takes 4-5 minutes to scan a thunderstorm, the new and emerging Phased Array Radar (PAR) rapid and adaptive scanning technology can scan the same storm in less than a minute and can enhance the scanning angles in real time to get a better depiction of the storm top. Thus, in an effort to explore the impact of high temporal frequency PAR observations in storm assimilation, Observing System Simulation Experiments (OSSEs) are designed using the EnKF as a method for initializing storm-scale numerical forecast models.

Several different OSSEs are conducted with radial-velocity and reflectivity observations constructed from simulated supercells in native radar coordinates using a realistic volume averaging technique. Two sets of experiment are run for each OSSE. One experiment assimilates the simulated Doppler radar observations while the other experiment assimilates the high temporal frequency PAR observations. Results obtained are compared to document the value of new PAR observations to the creation of improved storm analyses and short-range ensemble forecasts.