Comparison of three-dimensional winds derived from assimilated phased array radar data with mobile dual-doppler analyses from a tornadic storm

On 29 May 2004 a tornadic supercell traversed central Oklahoma during a several hour period. This storm caused considerable damage beginning near the town of Geary, OK (Payne et al. 2008) and continuing eastward through northern sections of the Oklahoma City metropolitan area. The Thunderstorm Electrification and Lightning EXperiment (TELEX, Rust et al. 2004, MacGorman et al. 2007) collected a wide range of observations from this storm using mobile platforms as well as other unique instrumentation. These special platforms included a three-dimensional lightning mapping array, the KOUN 10 cm dual polarimetric radar, proximity soundings using the NCAR M-CLASS systems, and Doppler radar observations from two mobile C-band SMART radars (Biggerstaff et al. 2005). This study uses data gathered from the 10 cm Phased Array Radar (hereafter, PAR, Forsyth et al. 2001) in conjunction with the other observations in order to investigate the following hypothesis: Can the rapid-scan capability of the phased array radar generate, through the use of storm-scale Ensemble Kalman data assimilation (Snyder and Zhang 2004, Dowell et al. 2004), accurate three-dimensional wind analyses for the convective storm?

The PAR gathered a unique data set from this storm using 90 degree sector volumes containing 7 individual tilt levels which where collected every 20-30 seconds during a several hour period when the storm was less than 80 km from the radar. Using the EnKF methodology from Dowell and Wicker (2008) the velocity and reflectivity data are assimilated every 60-120 seconds to generate storm-scale analyses representing the thermodynamic and kinematic structure of the supercell. We then compare the three-dimensional wind fields from the assimilation output with the dual-Doppler analyses from the SMART radars (Biggerstaff et al. 2007). During the period between 0100 and 0230 UTC the storm is located in the northern dual-Doppler lobe and within 40 km of both SMART radars. During this period volumetric data are available every 2-3 minutes from both mobile radars, providing high-resolution space/time data to generate a set of wind analyses to examine the retrieved winds from the PAR radar.

Results will show the sensitivity of the assimilation analyses to the frequency of data insertion. Comparisons are made between the vertical velocity characteristics generated by the two analysis systems. The results will be used to guide the deployment of the SMART radar systems in the upcoming VORTEX-2 field program in 2009-2010.

References available upon request.