The Penn State/NCAR MM5 mesoscale model was running real-time at the St. Louis Forecast Office at the time of the event. This model receives its initialization fields from the 0000 UTC Eta model each day and is run out to 30 hours. It is used to assist in daily forecasts at the office. It has been shown that the MM5 output at 27 km resolution produced a viable forecast alternative to the Eta forecast with the same initialization time (O'Sullivan et al. 1999).
The MM5 model is run with 3 domains of 27, 9, and 3 km grid spacing, respectively. The domains are two-way nested, with coarser domains providing the boundary conditions for finer grids, and these inner domains then updating the outer domains with better-resolved information. The Grell cumulus parameterization scheme has been shown to be the best approximation of convection for the simulations using the 27 and 9 km domains. Convection is calculated explicitly in the smallest domain with 3 km resolution. There are 29 vertical levels.
Output from the finer grid(s) will be shown to examine the observational accuracy of the model's performance. The information from these domains will be compared to WSR-88D data from the 11 February 1999 event from WFO St. Louis (KLSX). Various kinematic and dynamic fields will be shown which correspond to fields shown to have good correlation with the synoptic observations. Among these are 850 mb winds, surface moisture convergence, and 0-3 km vertical wind shear. Also, output will be shown to determine the characteristics of the squall line, such as momentum transport and the perturbation pressure differences within the linear system. Whereas previous studies attempted to determine the utility of the model in an operational setting, this presentation will describe the scientific validity of the simulation by examining the fine details of the predicted linear MCS.