An Evaluation of Changes to the NSSL Experimental Warn-on-Forecast System for Ensembles (NEWS-e) in Spring 2017

The NOAA Warn-on-Forecast (WoF) Project seeks to increase tornado, severe thunderstorm, and flash flood warning lead times using probabilistic hazard guidance from a storm-scale ensemble prediction system. The NSSL Experimental Warn-on-Forecast System for ensembles (NEWS-e) assimilates radar, satellite, and conventional data and generates new 0-3 h probabilistic forecasts two times an hour. This system was run in real-time during the NOAA Hazardous Weather Testbed 2017 Spring Experiment with two changes from the previous year: 1) use of the NSSL 2-moment microphysics scheme (as a replacement for the partial two-moment Thompson scheme); and 2) additional assimilation of METAR surface data.

For three severe weather days during spring 2017, the NEWS-e system will be rerun using Thompson microphysics to better understand the cost-benefit of its replacement with a full two-moment scheme. Also for each day, a second data-denial experiment will be run in which the METAR surface observations are removed from the data assimilation procedure (while preserving all other aspects of the 2017 NEWS-e configuration). Probabilistic forecasts from the three NEWS-e configurations will be compared based on the ability of model-derived measures of storm rotation (i.e., updraft helicity) to anticipate supercell thunderstorms and mesoscale convective systems. These forecasts will be further evaluated to better understand any differences in simulated storm structures (reflectivity fields, updraft profiles, etc.) and near-storm environments.