An Ensemble Perspective of Changes to Severe Thunderstorms and Hail

Several recent studies have pointed to the prospects of changes to severe thunderstorms in a future climate, including hail, tornadoes and damaging winds. The warming climate is expected to influence these storms and their attributes, primarily through increasing thermodynamic instability driven by increased atmospheric moisture as warming occurs to near-surface temperatures. Most of these studies have focused on either regional projections, or high-resolution dynamic downscaling to project our expectations for the future on a continental scale. This presents a challenge to our understanding as many parts of the world that currently regularly experience severe thunderstorms are found in different climate regimes or latitudes, and which exhibit non-linear responses to the warming climate. Hence what is known for some latitudes and warming scenarios may not reflect a complete picture of the expected changes to hazards on a global level.

Despite the apparent confidence that would be implied by thermodynamic scaling towards increasingly favorable conditions for severe thunderstorms, there are considerable differences in the response rate of severe weather outcomes, and changes to variability. To address this uncertainty, we quantify how severe weather environments will evolve in response to moderately and strongly warmed scenarios in terms of both climate variability and change as projected using the latest generation of Coupled-Model Intercomparison Project Version 6 (CMIP6) data. We use both physical model proxies and newly developed hierarchical machine learning approaches to environmental proxies that synoptic-scale clustering and profile characteristics to explore how changes manifest across time and the 12 ensemble members. Results highlight that increases over the Northern and Western hemispheres are not necessarily reflective of changes in the subtropics and Southern hemisphere, and the rate of response can be dependent on the environmental proxy used. Variability across the CMIP6 ensemble in terms of synoptic processes is much larger than might be expected from prior studies. While the driving mechanism of projected changes are strong increases in CAPE, these changes can be modulated by local processes, and are often accompanied by factors that either resist convection (CIN), or seasonally may coincide with high likelihood of storm organization (S06, SRH) suggesting projections may have lower confidence than previously thought.