Short Range Mesoscale Ensemble Forecasts of Precipitation over the Southwest United States during the Summer Convective Season

The summertime Southwest United States experiences a distinct convective season that leads to frequent microburst winds, heavy precipitation and flash flooding. Accurate, operational prediction of these convective events remains very difficult despite the availability of high-resolution (~10 km grids), deterministic numerical model forecasts. One obstacle in obtaining accurate quantitative precipitation forecasts (QPF) is the inability of current convective parameterizations to accurately simulate convection forced by rugged terrain under conditions of weak synoptic flow. Since mesoscale models do capture important aspects of the convective environment over the Southwest, such as the diurnal cycle and convective initiation over higher terrain, but fail in terms of high resolution deterministic details, an ensemble approach might improve forecasts of convection and precipitation. Our results show that ensemble forecasts on a 36-km grid, with perturbations to the initial conditions, can improve probabilistic QPFs. The inclusion of stochastic perturbations to the convective scheme to account for subgrid scale variability increases the ensemble spread and can lead to further improvements in performance.