The Development and Evaluation of Rapid Refresh Forecast System (RRFS) Multiphysics Ensemble Prototype

The ensemble forecast prototype for NOAA's next generation Rapid Refresh Forecast System (RRFS), a convective scale forecasting system based on the Unified Forecast System (UFS) is developed and evaluated in this study. The 9-member regional ensemble system prototype has a 3 km horizontal resolution. The ensemble members were initialized from a regional convective scale ensemble Kalman filter (EnKF) data assimilation system. The lateral boundary condition perturbations were created from the Global Ensemble Forecast System (GEFS). It combines a multiphysics approach and stochastic physics parameterizations (SPP). Various cloud microphysics schemes, planetary boundary layer (PBL) schemes and surface layer schemes are included into the ensemble. SPP is added to physics schemes to further increase the ensemble spread where the option is available, e.g. for Thompson cloud microphysics scheme, MYNN PBL and surface layer scheme, RUC land surface scheme, RRTMG radiation and gravity wave drag schemes. Single parameter perturbations are also applied for the NSSL cloud microphysics scheme. Stochastic physics perturbations are added to the ensemble, including stochastically perturbed parameterization tendencies (SPPT) and stochastic kinetic energy backscatter (SKEB). The ensemble was initialized from regional EnKF ensemble members and the lateral boundary conditions are from GEFS ensemble members. The ensemble system was run retrospectively for ten 2022 severe weather cases, and was verified against surface, upper air, precipitation and reflectivity observations. Ensemble spread and reliability with neighborhood probability were evaluated and compared against a single physics ensemble. The preliminary results show improvement of composite reflectivity reliability from the multiphysics ensemble with less overprediction and larger ensemble spread.