Impact of Assimilating Radar Observations on Storm-scale EnKF System within Realistic Mesoscale Environment

Recent studies of the assimilation of Doppler radar observations on storm-scale NWP model using Ensemble Kalman Filter (EnKF) in both idealized and real-data studies show reasonably good analyses after assimilating radar observations for about an hour, but the forecasts deteriorate quickly within tens of minutes. However, the development and evolution of severe thunderstorm events are strongly tied to the environment in which they develop, indicating the importance of incorporating the influence of horizontal environmental variability and mesoscale forcing on the storm scale flows. Therefore, a challenging but potentially rewarding avenue for improving severe thunderstorm analyses and forecasts is to assimilate radar observations into storm scale model using real environmental conditions. In an effort to explore the impact of a realistic mesoscale environment on storm-scale assimilation, a storm-scale ensemble is incorporated within the NSSL WRF/DART EnKF mesoscale ensemble system and evaluated using the 8 May 2003 Oklahoma City supercell tornado event. A 30-member mesoscale ensemble with a horizontal grid spacing of 20 km on a continental United States domain is initialized at 0900 UTC and routinely available observations from surface stations, rawinsondes, and wind profilers are assimilated on an hourly basis. After assimilating these observations over a 12-h period, a one-way nested 30-member storm-scale ensemble with 2-km horizontal grid spacing centered on the tornadic event is initialized. Radial velocity and reflectivity radar observations from the KOUN radar are assimilated into the storm-scale ensemble for about an hour. The analyses are then examined to evaluate the value of realistic mesoscale and near-storm environment on storm-scale assimilation.