Beyond probability: Providing information to warning forecasters using the NSSL Experimental Warn-on-forecast system for ensembles (NEWS-e)

Ensemble frequency/probability is often used in Warn-on-Forecast studies to highlight the potential for mesocyclones and tornadoes, and it is hypothesized that providing such information to warning forecasters will help increase warning performance (e.g., warning lead-time). Before we get to that point, however, we have to acknowledge that tornadoes are rare events and may not always be represented as simply large probability values. As we provide forecast data down to the minute scale, we assume data assimilation and model accuracy may not match all the scales at which these phenomena form or are present. Here we provide object-based tracking of storm rotation and low-level vorticity maxima to warning forecasters via raw model data representations that do not depend on ensemble probability alone. We provided a situational awareness version of the NEWS-e data set that could be used to query the underlying objects. Warning forecasters could provide feedback on real-time forecasts from the NEWS-e, as performed during the 2016 NOAA Hazardous Weather Testbed Probabilistic Hazard Information tool experiment. We will present our post-processing methodology and possible applications along with forecasters perspectives on using NWP at the warning desk. Challenges and opportunities that have emerged will be discussed.