Regional NWP Development at NOAA's Earth System Research Laboratory

The Global Systems Division (GSD) at NOAA’s Earth System Research Laboratory seeks to improve numerical prediction of severe weather; intense rainfall; convective storms, icing, and turbulence that affect aviation; wind and clouds for renewable-energy applications; and other small-scale phenomena such as smoke from wildfires. GSD’s partnership with the Federal Aviation Administration and NCEP’s Environmental Modeling Center led to the operational implementation of the Rapid Refresh and High Resolution Rapid Refresh (HRRR) Models. These hourly-updated analysis and forecasting systems will be upgraded in summer 2018 to add Alaska coverage to the existing CONUS coverage, and to provide 36-h forecasts every 3 h.

NWP systems developed at GSD are based on community modeling, data-assimilation, and post-processing software. GSD has contributed to the development of the Smirnova land-surface model, MYNN PBL parameterization, and Grell-Freitas-Olson (GFO) convective parameterization in WRF. This development is continuing with FV3, NOAA’s new global and regional NWP model. A goal is to provide a flexible physics package that performs well across model grid spacings from O(10 km) to O (100 m), and in complex terrain.

The future of convective-allowing NWP systems like the HRRR will be ensemble analyses and forecasts, to increase skill and provide uncertainty information for forecast applications. A prototype HRRR Ensemble (HRRRE) is under development and is running experimentally in real time. Ensemble diversity comes from initial-condition perturbations, boundary-condition perturbations, and stochastic physics. The HRRRE uses an ensemble Kalman filter to assimilate conventional and radar reflectivity observations hourly, and 9-member forecasts at regular intervals provide experimental probabilistic guidance. Parallel testing of other new applications will include initializing NSSL’s prototype Warn-on-Forecast system, developing RTMA-3D, providing QPF input to the National Water Model, developing advanced methods for assimilating satellite and radar observations, and testing components of the Joint Effort for Data assimilation Integration (JEDI) software. Experiences with HRRRE will guide development toward an operational implementation of an FV3 single-core convective-allowing ensemble analysis and forecasting system.