The NCEP's WRF NMM and hazardous weather prediction

The NCEP Nonhydrostatic Mesoscale Model (NMM) has been developed building on NWP experience within the WRF effort. The dynamical core of the model was discretized following the so called mimetic approach first introduced by Arakawa. With this approach important properties of the continuous equations and differential operators are preserved in the discrete system. The conservation of energy and enstrophy improves the accuracy of the model's nonlinear dynamics. Despite the complexity of the formulation, the computational efficiency of the model has been significantly higher than the computational efficiency of most nonhydrostatic models.

The NMM has been run operationally in NCEP for several years. Since June 2006, the new end-to-end system based on the WRF NMM with 12km resolution has become the main operational regional forecasting system for North America (NAM). Efforts are under way to implement the NMM operationally as the Hurricane WRF. Finally, commencement of operational runs with 4km resolution in support of the Storm Prediction Center (SPC) operations is planned for the near future.

During the first winter in operations, the new regional forecasting system showed noticeable skill in several major winter storms in different parts of the country and in different synoptic conditions. In addition to operational forecasting the model has been tested in many case studies and several validation campaigns. Further evidence has been gathered about the WRF NMM ability to predict tropical storms during the exceptionally active tropical storm season of 2005. For two consecutive years the WRF NMM participated in a carefully controlled springtime experiment in which the model was run at near-cloud resolving horizontal resolution of 4.5 km without parameterized convection. The model demonstrated ability to spin-up severe convective systems on the 24 hour time scale more frequently, and with a stronger signal, than if this were happening only by chance. This indicates that further improvements in deterministic forecasting of severe weather phenomena may be possible with increased resolution.