Deriving Precipitation Type Probabilities in the National Blend of Models

To support the National Weather Service goal of building a Weather-Ready Nation, the National Blend of Models (NBM) project was started in 2014 to create a skillful and consistent suite of calibrated guidance to leverage in the forecast process (Gilbert et al. 2015). Recently, NWS Central Region showed that model blending can provide a skillful first guess forecast for many National Digital Forecast Database (NDFD) weather elements. NDFD Weather is a more challenging grid to provide guidance for as it is typically derived by NWS forecasters using a variety of competing technical solutions (smart tools).

Central Region has tested a top-down approach for forecasting precipitation type. Analysis of techniques in the literature and in operational use, have led to a probabilistic approach, derived via forecast sounding information from a consensus blend of numerical model output. Specific environmental inputs being leveraged in this top-down approach include the elevated warm layer aloft, the surface-based cold layer, and the presence of cloud ice in the precipitating cloud layer (Reeves et al., 2014; Bourgouin 2000; Baumgardt, after Rauber et al. 2001). Further, statistical bias correction and calibration will be employed to achieve the type probabilities which will feed the production of NDFD predominant weather. This presentation will review the NBM development approach used to achieve the precipitation type probabilities.