Verification of probablistic forecasts for severe weather parameters in the 2010 Storm-Scale Ensemble Forecast System and the operational SREF system

As advances in computing capabilities facilitate the increasing use of convection-allowing modeling (CAM) systems, one of the important scientific issues to address is the value of CAMs relative to coarser resolution operational modeling systems that parameterize convection. Several recent studies reveal advantages in convection-allowing relative to convection-parameterizing simulations in rainfall prediction, diurnal precipitation cycle depiction, and forecasts of MCS frequency and mode. However, it is not clear whether convection-allowing simulations also provide improved forecasts of pre-convective environmental parameters such as low-level temperature and moisture, wind shear, and instability. In fact, one recent study found larger magnitude errors for pre-convective environmental parameters in deterministic convection-allowing forecasts relative to forecasts that parameterize convection.

Using traditional metrics for evaluating probabilistic forecasts (e.g., ROC area and Brier Score), this study examines forecast exceedance probabilities for selected fields in pre-convective environments derived from the Storm-Scale Ensemble Forecast (SSEF) system run by the Center for Analysis and Prediction of Storms (CAPS) at the University of Oklahoma and used for the 2010 NOAA/Hazardous Weather Testbed Spring Experiment. Comparisons are made to probabilistic forecasts derived from the NCEP operational Short-Range Ensemble Forecast (SREF) system. Fields examined include 0-6 km wind shear and surface-based CAPE, as well as the composite indices Significant Tornado Parameter (STP) and Supercell Composite Parameter (SCP).