Impact of Doppler radar and mesoscale surface observations on the storm-scale analysis and prediction of a mesoscale convective system

Dowell et al. (2004) demonstrated the positive impact of assimilating data from a single 10-cm Doppler radar using an Ensemble Kalman Filter (EnKF) technique for a supercell thunderstorm event. They found that the locations of the main updraft and mesocyclone were determined rather accurately by assimilating 10 scans of radar reflectivity and velocity data over a 47-minute period. However, they also found that assimilation of these Doppler observations did little to improve the representation of the cold pool, partially because of poor spatial resolution in the radar data near the ground. This brings into question the usefulness of assimilating radar data alone for the analysis and subsequent prediction of convective systems, in which the cold pool plays a primary role in determining its strength and longevity.

We will report on preliminary results from a study that explores the impact of Doppler radar and mesoscale surface data assimilation using an EnKF technique on the storm-scale analysis and forecast of a convective system that occurred on 11-12 June 2003 in Oklahoma. This event was well sampled by three operational Doppler radars and one research Doppler radar, and produced heavy rainfall and multiple reports of severe wind gusts over a 6-hour period. We use the NSSL Collaborative Model for Multiscale Atmospheric Simulation (NCOMMAS) to produce a 50-member ensemble of NCOMMAS simulations of the 11-12 June 2003 event using the EnKF technique. Each simulation is performed on a 400 by 400 km grid at 2 km horizontal resolution. Reflectivity and velocity radar data provided by the Frederick, OK operational Doppler radar and surface data provided by the Oklahoma Mesonet is assimilated. For each of the assimilation experiments, all 50 ensemble members will be used to produce an ensemble forecast of the convective event and verified against all available information out to 2 h beyond the last assimilation time.