Evaluation Study of Mesoscale Models MM5 and BFM Over the Model Domains of Utah Using Surface Meteorological Data by Mesowest

摘要

The forecast skills of the Penn State/National Center for Atmospheric Research Mesoscale Model version 5 (MM5), and the Battlescale Forecast Model (BFM) were statistically evaluated by comparing the model forecasting data with surface observation data. The study was done for the periods of January-March 2002, and April-June 2002. The MM5 used quadruple-nests of 67.5, 7.5, and 2.5 km grid increments, all of which have 51x51 grid points centered near Salt Lake City, UT. The forecast data of domains 3 and 4 were compared with the observation data. The BFM calculations were made separately over two model domains of grid increments of 5 and 2.5 km, both centered at Salt Lake City, UT. Findings included: No significant statistical differences in forecast skills between the two MM5 domains, and the two BFM grid increments are similar. Both models produce better results for surface temperature than for dew point temperature, and both models had difficulties in producing good forecast results for wind vectors. The MM5 produced statistically better results. Both models showed better results for April-June than for January-March.