Long Term Evaluation of NCEP, NSSL, and a New NSSL WRF Microphysical Parameterizations Ensemble

Severe weather forecasting is crucial to protecting life and property in the U.S. Currently, the Storm Prediction Center (SPC) uses a variety of convection permitting models and ensembles of these models to aid in generating their forecasts. However, model performance may often vary greatly from day to day making it difficult to determine which models should be trusted for a particular forecast. In this study, NCEP WRF, NSSL WRF, and a microphysical parameterizations based ensemble of NSSL WRF will be evaluated over a long time period (three years of simulations) including hundreds of convective events. This thorough evaluation will utilize ground based radar data, cloud property retrievals from GOES data, and the University of North Dakota Hybrid Classification Product (Feng et al. 2011). Additionally, Self-Organizing Maps (SOMs) will be used to evaluate model performance during a variety of synoptic conditions. This will aid in determining whether certain models, or microphysical parameterizations perform better during specific synoptic regimes. The performance of the model simulations will be determined by numerous metrics. Of particular focus will be precipitation amounts and intensity (evaluated using Stage IV and NMQ Q2), storm intensity and coverage (evaluated using the UND Hybrid Classification Product), and cloud properties of simulated convection (evaluated using NASA LaRC GOES Retrieved cloud properties). A description of the evaluation datasets and methodologies along with preliminary results will be presented