An Observational Study: Real-time Use of the NSSL Experimental Warn-on-Forecast System for Ensembles during NWS Severe Weather Operations (216)
Katie Wilson, OU CIMMS/NOAA NSSL, Norman, OK
Patrick Skinner, OU CIMMS/NOAA NSSL
Todd Lindley, NOAA NWS
Randy Bowers, NOAA NWS
Doug Speheger, NOAA NWS
Currently, National Weather Service (NWS) forecasters follow a warning decision process that uses real-time observations and reports to detect imminent or occurring weather threats. However, improvements in warning lead time have not occurred in this warn-on-detection paradigm in recent years. Therefore, a need for forecast information that enables earlier anticipation of weather threats has been recognized. Tasked with addressing this need, the NOAA Warn-on-Forecast project is developing and testing a convection-allowing ensemble analysis and prediction system that provides probabilistic guidance for a variety of weather threats within the watch-to-warning timeframe. This prototype system, termed the NSSL Experimental Warn-on-Forecast System for ensembles (NEWS-e), provides real-time short-term forecast guidance. These forecasts are provided on the regional-scale, with the position of the NEWS-e grid (750km^2 x 750km^2) being based on Day 1 convective outlook guidance from the Storm Predication Center and Weather Prediction Center.On days in spring 2017 when the NEWS-e grid was placed over the Norman Weather Forecast Office (WFO) domain, real-time NEWS-e guidance was provided to forecasters during operations. Briefings from a Cooperative Institute for Mesoscale Meteorological Studies (CIMMS) scientist on NEWS-e output proved useful to aiding forecasters' interpretation of this probabilistic guidance. Most notably, the NEWS-e output had a considerable impact on forecasters' decision making on 16 May 2017, when forecasters decided to issue an advisory expressing that there was &ldquoa high probability that tornado warnings will be issued&rdquo in western Oklahoma. Approximately two hours later, an EF-2 rated tornado touched down in Elk City, Oklahoma, resulting in one fatality and eight injuries.Given the success of this first introduction of NEWS-e output to the Norman WFO, this collaborative effort will continue in spring 2018. In addition to a CIMMS scientist providing briefings and interpretative support to forecasters, a second CIMMS scientist will engage in observational activities. These observations will take place within the naturalistic operational environment, and will be both direct (i.e., forecasters will be aware that observations are being made) and continuous (i.e., observations will made within the same operational setting as often as possible). Observations will be semi-structured, such that they will be designed to capture certain elements of the decision making process while also being largely driven by what activities emerge during real-time operations. These observations are expected to bring to light: 1) how forecasters draw meaningful information from the NEWS-e output, 2) the ways in which this information can be subsequently useful to their decision making within the watch-to-warning timeframe, and 3) what adjustments or modifications forecasters must make to their normal workflow to integrate this new information effectively. The qualitative data will be analyzed descriptively within context of their particular situations, and early findings from this analysis will be presented at this conference.