REAL-TIME MESOSCALE PREDICTION ON WORKSTATIONS

Experience in performing real-time mesoscale numerical prediction fore casts using the Regional Atmospheric Modeling System (RAMS) over Color ado for a winter season on high-performance workstations is summarized . Performance evaluation is done for specific case studies and, statis tically, for the entire winter season. RAMS forecasts are also compare d with nested grid model forecasts. In addition, RAMS precipitation fo recasts with a simple ''dump bucket'' scheme are compared with explici t, bulk microphysics parameterization schemes. The potential applicati ons and political/social problems of having a readily accessible, real -time mesoscale forecasting capability on low-cost, high-performance w orkstations is discussed.