Assessment of timing and coverage of convection during the 2012 NOAA Hazardous Weather Testbed Spring Forecasting Experiment

Predicting the timing of convective initiation (CI) and convective coverage were key elements of experimental forecasting exercises during 2012 Spring Forecasting Experiment (SFE2012), conducted by the Experimental Forecast Program (EFP) of the NOAA Hazardous Weather Testbed from 7 May – 8 June 2012. Each morning, forecast teams used radar and satellite imagery to choose a subdomain of the CONUS in which no convective elements resided, but where CI was expected later in the day in association with an identifiable forcing mechanism. This process of isolating a limited regional domain (hereafter CI domain) allowed forecasters to focus on a specific CI episode each day and to automate diagnostic assessments of human and model forecasts of this episode after the event. The timing of the event was predicted to the nearest hour, and the uncertainty associated with that prediction based on the human and the convection-allowing model (CAM) ensemble from the University of Oklahoma Center for Analysis and Prediction of Storms (CAPS) was estimated quantitatively. Additionally, total convective coverage was forecast for 4-h time periods over a larger region that included the CI domain. A blend of observations, operational models, and experimental high-resolution CAM ensembles were used as guidance for these forecasts. The presentation will include working definitions of CI, methods for forecasting CI and convective coverage, and verification statistics for both the model guidance and human forecasts.