Linking Weather Regimes to the Variability of Warm-Season Tornado Activity over the United States

Abstract. The contiguous United States (CONUS) experiences considerable interannual variability in tornado activity. The high impacts of tornadoes motivate the need to better understand the link between seasonal tornado activity and large-scale atmospheric circulation, which may contribute to better seasonal prediction. We employed K-means clustering analysis of 500 hPa geopotential height (500H) daily anomalies from the ERA-5 reanalysis and identified five warm-season weather regimes (WRs). Certain WRs are shown to strongly affect tornado activity, especially outbreaks, due to their relationship with environmental parameters including convective available potential energy (CAPE) and vertical wind shear (VWS). In particular, WR-B, which is characterized by a three-cell wave-like pattern with an anomalous low over the central-CONUS, is associated with enhanced CAPE and VWS in tornado-prone regions and represents a tornado-favorable environment. Persistent WRs, those lasting for ≥5 consecutive days, are associated with 76 % of all tornado outbreaks (days with >10 EF-1+ tornadoes) since 1960, with a persistent WR-B accounting for about 30 % of all tornado outbreaks. The impacts of WR persistence on tornado activity anomalies, however, are found to be asymmetric: compared to non-persistent WRs, persistent WRs amplify positive tornado activity anomalies but may not further enhance negative tornado activity anomalies. An empirical model using WR frequency and persistence captures the year-to-year variability of warm-season tornado days and outbreaks reasonably well, including some years with high-impact outbreaks. Our study highlights the potential application of WRs for better seasonal prediction of tornado activity.