Telescoping model approaches to evaluate severe convective storms under future climates

Questions regarding changes in frequency, location, and intensity of severe convective storms in future climates remain unanswered owing to the inability of typical global climate models (GCM) and regional climate models (RCM) to explicitly resolve such events. One could circumvent this problem by equating storm occurrence to ranges of storm environmental parameters, such as CAPE and vertical wind shear, evaluated using output from these larger-scale models. Alternatively, one could employ a more direct approach that culminates in convection-permitting simulations. We will describe our experiences and results thus far using the latter.

Our scale-spanning, multiple-model system begins with a GCM (or global dataset), which drives long-term RCM simulations, which in turn drive mesoscale and nested cloud-resolving model simulations using Weather Research and Forecasting (WRF) model. Historical simulations have been used to test whether this model strategy can, for different convective events, accurately represent: the antecedent conditions on the synoptic and mesoscale, the initiation of deep convection, and then the resultant type or mode of the convective storms. The 3 April 1974 Tornado Super Outbreak, as well as the relatively more benign sequence of tornado and severe weather events during 3-9 May 2001, have been simulated. These results will be discussed, as will our use of model-based proxies of severe-storm reports for the generation of climate statistics.