User-focused Research to Build a Better Warn-on-Forecast System

Owing to advancements in high-resolution modeling and increased computational power, the frequency of NWP forecasts has been increasing. The HRRR, for example, currently runs on an hourly cycle. However, due to delay in spin-up of deep convection there is limited real-time usability of the first 3–6-hour forecasts from operational convection-allowing models. Given the dynamic nature, inherent uncertainty, and societal impacts of near-term hazardous weather events (e.g., extreme rainfall, tornadoes, hail and damaging wind), an ensemble forecast system with frequent data assimilation cycling (every 15 min) and frequent forecast launches (e.g., every 30 min) is needed. Toward this end, researchers at the NOAA National Severe Storms Lab, OU CIMMS, NOAA ESRL/GSD, and other partner institutions have jointly developed the experimental Warn-on-Forecast System (WOFS).

WOFS produces real-time, short-term (0-6 h) probabilistic forecast guidance for the occurrence of weather hazards associated with individual storms. This system is envisioned to enable greater continuity of forecast information and greater specificity of the location and timing of weather events within the Watch and Warning spatiotemporal scales. The concept of operations is for NWS operational centers, such as the Storm Prediction Center (SPC) and Weather Prediction Center (WPC), to decide when and where a WOFS sub-domain is required to enable better weather forecasts and services. The performance of this science-based system is measured by forecast quality and its usability and value to decision makers. Hence, WOFS development is aligned with the FACETs (Forecasting a Continuum of Environmental Threats) framework and is envisioned, in the longer term (e.g., 2030), to be a part of the NWS Unified Forecast System.

An important aspect of the WOFS development process is forecaster-in-the-loop studies. Because the concept-of-operations is driven by NWS operational centers, initial experiments have focued on usefulness and usability of WOFS guidance WOFS guidance for 1) severe weather forecasting at the SPC, and 2) flash-flood forecasting at the WPC. The longevity of these experiments enables forecasters to identify strengths and limitations of WOFS for a broad spectrum of real-time weather events, which in 2019 included the summer monsoon in the southwest U.S. for the first time (owing to the CONUS-wide HRRRE runs provided by GSD). NSSL is also partnering with a few NWS Weather Forecast Offices to explore potential applications of WOFS to flash flood and severe weather forecasting. Of particular interest are potential improvements to hydrologic forecasts of flash floods using WOFS ensemble-based precipitation forecasts as the driver.

These R2O2R interactions are motivating and guiding model and data assimilation enhancements, post-processing, and visualizations needed to develop and eventually transition a WOFS whose forecast guidance is useful to forecasters and has a positive impact on protective weather-focused decision making. This presentation will summarize the spectrum of user-focused research conducted using WOFS within the last three years, progress with respect to these forecaster-in-the-loop studies, and next steps planned in the R2O2R process.