An Empirical Model for Assessing the Severe Weather Potential of Developing Convection

Abstract The formation and maintenance of thunderstorms that produce large hail, strong winds, and tornadoes are often difficult to forecast due to their rapid evolution and complex interactions with environmental features that are challenging to observe. Given inherent uncertainties in storm development, it is intuitive to predict severe storms in a probabilistic manner. This paper presents such an approach to forecasting severe thunderstorms and their associated hazards, fusing together data from several sources as input into a statistical model. Mesoscale numerical weather prediction (NWP) models have been developed in part to forecast environments favorable to severe storm development. Geostationary satellites, such as the Geostationary Operational Environmental Satellite (GOES) series, maintain a frequently updating view of growing cumulus clouds over the contiguous United States to provide temporal trends in developing convection to forecasters. The Next Generation Weather Radar (NEXRAD) network delivers repeated scans of hydrometeors inside storms, monitoring the intensification of hydrometeor size and extent, as well as hydrometeor motion. Forecasters utilize NWP models, and GOES and NEXRAD data, at different stages of the forecast of severe storms, and the model described in this paper exploits data from each in an attempt to predict severe hazards in a more accurate and timely manner while providing uncertainty information to the forecaster. A preliminary evaluation of the model demonstrates good skill in the forecast of storms, and also displays the potential to increase lead time on severe hazards, as measured relative to the issuance times of National Weather Service (NWS) severe thunderstorm and tornado warnings and occurrence times of severe events in local storm reports.