Hybrid Prediction of Severe Storm Activity: Utilizing Weather Regimes

Severe storms (tornadoes, strong wind, and hail storms) are one of the high-impact weather disasters that can induce severe life loss and property damage. Extended range prediction of severe storms is challenging as the complex dynamics and physics of severe storms are not fully understood. In addition, the coarse model resolution of extended-range models do not fully resolve individual storms explicitly. The connection between severe weather and synoptic scale weather systems are well known and it is likely that synoptic-scale events strongly modulate tornado activity, even in seemingly unfavorable environments. Here, we investigate the relationship between large-scale weather regimes and severe weather variables during May. Results show the probability of occurrence of a tornado in the United States during a weather regime 1 (ridge in southeast) day to be greater than 70%. Analysis suggests that the modulation of weather regimes on the regional tornado activity can be explained by the changes in vertical wind shear and CAPE. Due to the connection between regimes and tornado occurrence, we successfully predict above/below normal tornado activity using model predicted weather regimes as predictors. The model has skill better than climatology out to week 3. This work aims to contribute to real-time forecasting and aide in mitigation of severe weather events.