Development and refinement of the RRFS regional model – the R2O2R loop

NOAA’s Global Systems Laboratory (GSL) and Environmental Modeling Center (EMC) are working together toward refinement and deployment of a new deterministic / ensemble regional prediction system. Known as the Rapid Refresh Forecast System (RRFS), this system will replace the High-Resolution Rapid Refresh (HRRR) and other current NCEP operational regional prediction systems. Initial work has focused on use of the UFS-based FV3 model dynamic core coupled with a customized version of the GSI analysis package. Extensive testing has resulted in a system that has performed well for many weather aspects, including winter weather, upper-air wind, near surface moisture and clouds, and Alaskan weather.

Real-time testing during the Hazardous Weather Testbed (HWT) 2024 Springtime Forecast Experiment (SFE) revealed cases where intense convective elements were missed, leading to lower probability of detection (POD) for higher reflectivity thresholds. Based on this, the National Weather Service (NWS) is considering different configuration options for RRFS operational implementation.

GSL is pursuing research and development work toward one option, proceeding with a planned transition toward a new model core (MPAS – Model Prediction Across Scales system) coupled with a new analysis package (JEDI – Joint Effort for Data Assimilation Integration). Work to date has included installing and refining physics components within the MPAS model and testing and refining JEDI-based data assimilation capabilities for many different observations (including refining the background error specification). Additional work has included running and evaluating multiple cold-start MPAS runs and coupling JEDI to MPAS toward creating a real-time hourly cycled regional JEDI/MPAS run.

This RRFS work provides an excellent example of the R2O2R development cycle, with feedback from users and stakeholders informing RRFS development and refinement work. Throughout the RFFS development / refinement process, RRFS developers have actively coordinated with regional model users and stakeholders, including especially users and providers of aviation hazard specific guidance products (such as FIP, GTG, cloud and convection hazard guidance providers). This coordination has taken the form of regular stakeholder meetings to provide status updates, providing RRFS information to groups developing aviation hazard guidance products, completing adjustments and upgrade to the RRFS model post-processing, and providing RRFS retrospective gridded data for use by hazard guidance product developers.

In the presentation, we will provide a status update on the RRFS development / refinement work, including NWS implementation plans and timelines, system configuration, and retention / potential retirement of other forecast systems. We will also provide the latest information on GSL work on the JEDI-MPAS based RRFS system, including updates on the configuration, real-time and retrospective testing, objective verification, and subjective case-study evaluation.