Global Model Test Bed: Fostering Community Involvement in NOAA's Next-Generation Global Prediction System

The National Weather Service is engaged in an ambitious effort to substantially upgrade its model production suite with the goal of improving numerical prediction on scales ranging from nowcasting through subseasonal. This multi-year effort hinges on recommendations made by the NCEP advisory committee regarding reducing the complexity of the production suite and enhancing community engagement with operational modeling systems.

To support the goals above, the NWS Next-Generation Global Prediction System Program is sponsoring a Global Model Test Bed (GMTB) to create a software infrastructure and governance process for facilitating transition of research to operations specifically in the area of atmospheric physical parameterizations. To that end, GMTB is developing the Common Community Physics Package (CCPP), a collection of physical parameterizations with a well-described interface that lessens the complexity for community scientists to contribute innovations to be considered for operational implementation. Community involvement is further enhanced by the flexibility of using the CCPP in the GMTB Single Column Model (SCM). Future enhancements could include the use of the CCPP by other dynamic modeling systems.

In order to evaluate innovations proposed for inclusion in the supported CCPP and for advancement toward operational consideration, GMTB has put in place a hierarchical test “harness” that can be used both by its own staff and by community scientists with access to NOAA computational platforms. This harness contains a variety of tools of multiple levels of complexity, ranging from a parameterization simulator, the GMTB SCM, case studies for diagnostics, and an automated workflow for running, postprocessing, and evaluating physics for global models. This harness was tested by assessing an advanced cumulus parameterization option for the NCEP Global Forecast System, the Grell-Freitas scale-aware scheme, in both cold start and cycled data assimilation modes. In this presentation, we will review progress of the development of CCPP and the physics tests that have been conducted. We will also provide our perspective on the future role of GMTB in support of the continued development of NGGPS.