Verifying low-level and midlevel rotation in convection-permitting ensemble forecasts

Improved computational resources have allowed for the advent of convective-scale numerical weather prediction. Convection-permitting forecasts are now generated from models that nominally resolve convective processes. Qualitative assessments of expected convective mode can be made from these forecasts, but operational models are not capable of explicitly resolving processes on the scale of severe weather phenomena. Therefore, severe storm proxies derived from model-simulated fields have been developed to aid in the prediction of severe weather hazards. For example, updraft helicity has been widely used as a proxy for the updraft rotation that occurs in supercell thunderstorms. This study aims to assess how effectively updraft helicity and other proxies for low-level rotation can act as predictors for occurrences of supercell thunderstorms and specific severe weather hazards including tornadoes. A five-member Weather Research and Forecasting (WRF) model ensemble was used to generate convection-permitting forecasts for a selection of severe weather events, including the high-impact events of 20 May and 31 May 2013. Forecasts of updraft helicity and low-level rotation were verified against observations of storm reports and radar-derived rotation tracks. Neighborhood probabilities of the forecast and observed fields were computed to quantify forecast skill. Preliminary results suggest there is some skill in predicting rotation when using the ensemble maximum forecast and rotation track observations.