Extended Abstract (396K)4.9
Evaluation of WRF's ability to predict the coverage of air-mass thunderstorms and applications to short-term forecasting
Cody L. Phillips, NCAR, Boulder, CO; and J. Pinto and R. M. Rasmussen
The ability of Weather Research and Forecasting (WRF) model to predict the coverage of air-mass thunderstorms across the eastern half of the United States is assessed. Correlations between thunderstorm coverage and large-scale environmental conditions are determined. In this study, the ARW version of WRF is run at 4 km resolution twice per day in real time (00 and 12 UTC initialization times) during the two-month period of June/July 2006. At this resolution (which is roughly the current limit in terms of what can be run over the CONUS operationally), the model is capable of explicitly resolving storm scale motions and microphysics without the need for convective parameterization. Periods of time are conditionally sampled to look for times and areas where the dominant mode of convection is unorganized cells. The model-predicted storm characteristics (including storm locations, storm densities and coverages) are then compared with the observations derived from national mosaic of WSR-88D reflectivity from WSI. The storm area coverage for these times/areas is determined in both the model predictions and observations by determining the fractional area covered by reflectivity or precipitation rate exceeding some predefined threshold (here we use 3 mm hr-1). The goal of this study is to determine whether high-resolution WRF runs can be used to guide operational, rapidly-updating nowcasting systems in predicting characteristics of air mass storms (including convective organization and coverage). As such, comparisons are also made with versions of the National Convective Weather Forecast (NCWF) running at NCAR and the Aviation Weather Center.
Session 4, Prediction (Use of Climate Statistics in Forecasting)
Wednesday, 17 January 2007, 8:30 AM-11:45 AM, 206A
Previous paper Next paper
Recorded presentation