Intercomparison of spatiotemporal variability in severe weather environmental proxies and tornado activity over the United States

Tornadoes cause numerous deaths and significant property damage each year, yet how tornado activity varies across climate states, particularly under global warming, remains poorly understood. Importantly, severe weather events arise during transient periods of extreme thermodynamic environments whose variability may differ from that of the environmental mean state. This study analyzes the climatological relationships between commonly-used severe weather environmental proxies (the product of convective available potential energy and bulk vertical wind shear, energy-helicity index, and the significant tornado parameter) and tornado density on three dominant timescales of climate forcing: diurnal, seasonal, and interannual. We utilize reanalysis data to calculate the spatial distributions of the mean, median, and a range of extreme percentiles of these proxies across each timescale as well as for the full climatology. We then test the extent to which each measure captures the spatiotemporal variability of tornado density over the continental United States. Results indicate that the mean is a suitable statistic when used with the full climatology of the energy-helicity index and the significant tornado parameter without using convective inhibition in calculations, the diurnal cycle for convective available potential energy and the product of convective available potential energy and bulk vertical wind shear, and the interannual variations for all proxies except convective available potential energy. The mean is outperformed by extreme percentiles otherwise. This understanding of climatological relationships between tornadoes and the large scale environments can improve prediction of tornado frequency and provides a foundation for understanding how changes in the statistics of large-scale environments may affect tornado activity in a future warmer climate state.