Development and Status of NOAA’s Rapid Refresh Forecast System

The Rapid Refresh Forecast System (RRFS) is a high-resolution (3-km), hourly updated ensemble prediction system covering North America. The data assimilation system uses a hybrid 3DEnVar algorithm together with an ensemble Kalman filter (EnKF), the latter of which is used to evolve the 3-km ensemble covariances derived from 30 members. The ensemble of analyses from the EnKF also provides the initial condition perturbations for the forecast ensemble. The physics configuration is derived from the RAP/HRRR framework and features updates across the suite, including the boundary layer, land surface, and microphysics schemes. Development of the system has been a collaborative endeavor encompassing NWS/EMC, NOAA Labs, and the wider academic research community all under the auspices of the Unified Forecast System (UFS).

In this talk we will provide a broad overview of the configuration of the RRFS and will focus on recent developments in four of our priority areas. Those areas are: (1) implementing a multiscale data assimilation algorithm, (2) reducing a high bias associated with the 0-3 hour forecast of convective storms, (3) reducing biases associated with convective-storm intensity (i.e., precipitation) and coverage, and (4) improving the forecast ensemble.