The Engineering Research Center for Collaborative Adaptive Sensing of the Atmosphere (CASA) recently deployed a network of four low-cost, short range Doppler radars in southwest Oklahoma. The high temporal and spatial resolution of CASA radar data make it ideal for assimilation into a storm-scale numerical weather prediction model. One branch of study within CASA focuses on the use of data in a semi-operational fashion to forecast convective events with 1 km model grid spacing. This study examines if the impact of CASA data is resolution dependent by decreasing grid spacing to 100 m. The Advanced regional Prediction System (ARPS) 3DVAR is used to combine CASA and NEXRAD data with more traditional forms of data such as ASOS, RAOBs, profiler data, and observations from the Oklahoma Mesonet. Experiments are conducted on two (nested) grids. The outer grid, relatively coarse with a 2 km horizontal resolution, is used to initialize the inner grid which features a very fine horizontal grid spacing of 100 m. CASA data are assimilated on the inner grid. All other data sources are assimilated on both grids with rapid (~5 min) assimilation cycles. To explore the benefit of CASA data, experiments conducted on the high resolution grid utilize conventional data with NEXRAD alone, CASA alone, and NEXRAD and CASA together. Our test case uses data of a mesoscale convective vortex that developed on 9 May 2007 in SW Oklahoma and moved through the CASA domain. This convective system, comprised of several multi-cell storms, produced a few small tornadoes including an EF-1 in the CASA domain near Minco, OK. The multi-cellular and small-scale nature of the 9 May 2007 event makes it well suited to be used as a test case for this high resolution study. Additionally, it is possible to gain insight into the dynamics behind the tornadogenesis of this unusual event. Most recent results from the study are presented.