On the Prediction of a Violent Tornado Outbreak in Central Oklahoma on 20–21 May 2019

During the evening and overnight hours of 20-21 May 2019, a line of severe thunderstorms moved east from west Texas through Oklahoma to Missouri producing numerous tornado, wind and hail reports. In all, 34 tornadoes were reported in the region, with a bulk of the reports in the Texas Panhandle on the afternoon of 20 May and in central and northeast Oklahoma during the overnight hours. Available EF-scale information included in the tornado reports suggests that all of the tornadoes were EF-0 and EF-1 intensity. This severe weather event occurred in an environment characterized by strong synoptic-scale forcing for ascent with high convective available potential energy and vertical wind shear. Convection-allowing model (CAM) guidance beginning on 19 May through about 2000 UTC 20 May indicated that numerous isolated supercells would develop in central Oklahoma on the afternoon of 20 May ahead of the main line of severe thunderstorms. All indications, including derived tornado parameters that reached historic levels, were that an outbreak of long-track tornadic supercells capable of producing violent tornadoes would occur in central Oklahoma on the afternoon of 20 May. With the operational guidance so consistent in this solution, the Storm Prediction Center issued a high risk for severe thunderstorms for much of Oklahoma for 20 May. In response to the impending tornado outbreak, all Norman Public Schools and the University of Oklahoma closed. In the real atmosphere, however, the tornadic supercells failed to develop in central Oklahoma ahead of the main line of severe thunderstorms. The aim of this presentation is to examine some of the CAM guidance available as part of the NOAA Hazardous Weather Testbed (HWT) to determine factors that contributed to incorrect prediction of a violent tornado outbreak in central Oklahoma.

Preliminary results suggest that a stout capping inversion at the base of an elevated mixed layer (EML) seen in special soundings at Norman, Oklahoma, and Dallas, Texas, suppressed convection that attempted to initiate over central Oklahoma on the afternoon of 20 May. Inspection of forecasts from the operational and experimental high-resolution rapid refresh (HRRR) and the experimental National Severe Storms Laboratory Warn-on-Forecast system (WoFS) initialized prior to 2000 UTC both showed evidence of the capping inversion, but with a “smoother” vertical profile and slightly weaker cap strength that allowed the development of supercell storms. Beginning with the 2000 UTC initializations, the CAM guidance backed off on the development of supercell thunderstorms ahead of the main line of severe thunderstorms. It appears that the EML and stout cap were better represented in the CAM guidance in these later forecasts. Diagnostic analysis of the EML and cap in the HRRR and WoFS guidance (and other CAMs used during the HWT) using air parcel trajectories and the lapse rate tendency equation will be presented.