Recent advances in high-resolution atmospheric weather modeling include the introduction of the enhanced 13km Rapid Update Cycle (RUC) model which assimilates a wide range of data including radar reflectivity. In order to study the impact of these improvements on local modeling efforts at WFO Miami, an experiment is being conducted during South Florida's 2011 convective season (mid-May to mid-October) using the Weather Research and Forecasting (WRF) Environmental Modeling System. Two different configurations of the model are being evaluated: 1) a 9:3 km Advanced Research WRF nest run six times a day out to 18 hours using the RUC for both initial and boundary conditions; and 2) a 6:2 km Non-Hydrostatic Mesoscale Model nest run eight times per day out to 36 hours using the RUC for initial conditions and the NCEP Global Forecast System for boundary conditions. Both configurations use NASA's Short-term Prediction Research and Transition Center's Land Information System (LIS) for surface data, and MODIS/AMSR-E sea surface temperature (SST) composites that provide superior details of the surrounding ocean. The output from both local model configurations is being compared against that of the enhanced 13 km RUC model. The focus of the analysis is on short term convection, using grid scale threat scores composited at one-, three- and six-hourly intervals. Emphasis will also be placed on the effect of the NASA's experimental datasets on the model performance. Expectations are that the high resolution LIS and SST analyses would produce more skill convective forecasts. This paper will present the results of this threat score based verification effort.