Lessons Learned from Three Years of Real-time Experiments with the NEWS-e

For the past three spring seasons (2016-2018), the NSSL Experimental Warn-on-Forecast System for ensembles (NEWS-e), has been run as a daily real-time system to provide rapidly-updating/high temporal resolution ensemble forecasts for experimental severe weather forecasts in NOAA’s Hazardous Weather and Hydrometeorological Testbeds. Since output from the NEWS-e is accessible via a web page (https://www.nssl.noaa.gov/projects/wof/news-e/realtime/), several NWS forecast offices have also used NEWS-e output as an additional tool to help inform their warning process during local severe weather events. With approximately 100 experimental forecast days over the three-year period, the NSSL Warn-on-Forecast (WoF) research project portfolio contains a unique set of cases having diverse characteristics in terms of convective mode, primary severe hazard, and geolocation across the CONUS.

Since the concept of operations for a WoF system will embed a regional and weather-adaptive “NEWS-e” type system into a CONUS-scale convection-allowing ensemble, in 2016 NOAA’s ESRL/GSD and NSSL decided to collaborate and test a combined real-time system where ESRL/GSD’s 2/3 CONUS-scale experimental HRRR ensemble would be used for the initial and boundary conditions for the NSSL’s NEWS-e system. This requires considerable coordination as the timing and reliability of running coupled experimental systems presents numerous challenges to both groups. Data latency/availability, management of and changes to code bases and software, and efficient post-processing of ensemble output are just a few of the issues that impact the ability to provide real-time ensemble forecast data to stakeholders.

This study will present how lessons learned from each experimental season have led to improvements in NEWS-e workflow and data management, observation pre-processing, and post-processing. Feedback from the experimental forecasts in the Hazardous Weather and Hydrometeorological Testbeds from operational forecasters, academics, students, and our NWSFO partners, have helped to improve our system from the original WoF paradigm, as well as help evaluate and create new ensemble-based output and verification metrics for the system. Finally, a vision for the near-term future of the NEWS-e will be presented.