The Weather and Research Forecast Model (WRF) is a new mesoscale model under development by a number of agencies involved in weather forecasting and research. The model is fully compressible, nonhydrosatic, and can simulate dynamics from the cloud scale to the synoptic scale. At present, a number of microphysics, cumulus, and boundary layer parameterizations may be used to simulate sub-grid scale dynamics. WRF can be run with both idealized data or data from global forecast models as initial and boundary conditions. Further nesting is not yet available.

For operational hurricane forecasting, NOAA/NCEP is considering the replacement of the GFDL Hurricane Prediction System with WRF sometime in the next 5-10 years. We will present simulations of tropical cyclones in the WRF and GFDL models, for both a real data case and an idealized case. In these calculations, WRF will be compared with GFDL for identical initial and boundary conditions. The sensitivity of the WRF results to various physical parameterizations will also be presented. To aid in these comparisons, the WRF Standard Initialization has been augmented to allow initialization directly from GFDL model output. In the case of Hurricane Humberto (2001), WRF makes similar track and intensity predictions as the GFDL and global models. The idealized case is the simulation of a developing tropical cyclone on a constant-SST ocean with a homogenous environment.