Ensemble and Probabilistic Prediction of the 20 May 2013 Newcastle-Moore EF5 Tornado

On the afternoon of 20 May 2013, a supercell thunderstorm passed over the Oklahoma City metropolitan area in central Oklahoma, producing an EF5 tornado that resulted in 24 fatalities and 212 injuries, and extensive damage in the cities of Newcastle, Moore, and Oklahoma City.  In this study, an ensemble of six very-high-resolution forecasts is produced for the Newcastle-Moore EF5 tornado and its parent supercell to investigate the ability of such an ensemble to produce skillful ensemble forecasts of the tornado.  The ensemble members use 50 m horizontal grid spacing on a grid of 2003 × 1433 × 63 grid points, and are nested within a storm-scale ensemble with 500 m grid spacing which was initialized via cycled EnKF data assimilation of radar reflectivity and radial velocity, as well as conventional weather observations. 

The ensemble of six forecasts is initialized at 1930 UTC, approximately 26 minutes prior to the start of the observed Newcastle-Moore tornado, and is run for 90 minutes.  All ensemble members predict tornadic vortices with surface winds of 30 m s^-1 or greater, and three members predict intense tornadoes with wind speeds meeting EF5 criteria (90 m s^-1).  While the storm in all ensemble members moves to the northeast somewhat faster than the observed Newcastle-Moore supercell, there is substantial variation among members in the duration and track of predicted tornadic vortices.

Probabilistic and ensemble forecast products are produced for the tornado, using several combinations of criteria, including near-surface wind speed and vertical vorticity.  These probabilistic forecasts are verified against the observed tornado, and compared against probabilistic forecasts produced for commonly-used proxy variables, including updraft helicity.  While the tornado track indicated in the probabilistic forecasts extends well to the northeast of the observed tornado track, and despite the often substantial disagreement among ensemble members, the probabilistic and ensemble forecast products indicate that the forecast ensemble exhibits marginal to moderate skill in predicting the track of the tornado.  Despite the track error, the ensemble indicates a substantial probability of an intense tornado within approximately 10 km of the observed tornado.