Stochastic Physics in the FV3-LAM/SRW App: Implementation, Testing, and Transition to Operations in the RRFS

Stochastic physics has now been successfully implemented within the Finite-Volume Cubed (FV3)-Limited Area Model (LAM) and Unified Forecast System (UFS) Short-Range Weather (SRW) App as part of a three-year, NOAA/Weather Program Office (WPO)-funded project. There are four different stochastic physics methods available for use by the community: Stochastically Perturbed Parameterizations (SPP), Stochastically Perturbed Physics Tendencies (SPPT), Stochastic HUMidity Perturbations (SHUM), and Stochastic Kinetic Energy Backscatter (SKEB). These schemes were extensively tested and evaluated within the SRW App configured to run a 3-km, CONUS domain to inform ensemble design of the next-generation, regional, convection-allowing Rapid-Refresh Forecast System (RRFS). This talk will describe the four different stochastic physics schemes, sensitivity testing performed, evaluations of retrospective forecasts conducted, and an assessment of the impacts on computational efficiency within the FV3-LAM. We will also describe the recommended stochastic physics configuration communicated to RRFS managers as well as those running the FV3-LAM in an ensemble configuration for other real-time purposes, such as for the Hazardous Weather Testbed-Spring Forecast Experiment (HWT-SFE). Transition-to-operations for the first version of the RRFS is currently planned for the end of FY24, and based in part on our findings, is likely to include a subset of the four stochastic physics currently available in the FV3-LAM/SRW App. This work represents an example of the effective R2O pathway that exists between universities, external laboratories, and NOAA, leading to the successful transition of innovative science into operations.