On May 22, 2008, an EF-3 tornado formed at 1727z near Platteville, CO and began traveling northwestward, eventually striking Windsor, CO, causing 1 fatality. High storm-relative helicity (SRH) supported the possibility of supercells with tornadoes, but the 18z Denver sounding indicated that the thermodynamic profile was questionable. Only immediately to the north of a warm frontal boundary was there sufficient boundary-layer moisture to support the development of strong storms. The capability (or lack thereof) of high-resolution models to accurately predict the mode of the convection on that day is a compelling research question.

Using this unusual northwest-moving EF-3 tornado event as a case study, radar reflectivity and radial velocity data from the Denver WSR-88D radar are assimilated into WRF-VAR. Simulations utilizing the ARW dynamical core are conducted using a small 3km explicit convection-resolving domain nested inside of a larger 15km domain for computational efficiency. By varying lead times with respect to the onset of tornadogenesis, the model's handling of the convection is judged based on accuracy of storm propagation, placement, intensity, and initiation. Experimentation is also performed with different microphysical schemes. Preliminary results presented at the 8th Annual AMS Student Conference will assess whether ARW can show skill at predicting an actual storm that occurs, as opposed to just adequately representing the convection in general.