The Sensitivity of WRF Daily Summertime Simulations over West Africa to Alternative Parameterizations. Part II: Precipitation

摘要

This paper evaluates the performance of the Weather Research and Forecasting (WRF) Model as a regional atmospheric model over West Africa. It tests WRF's sensitivity to 64 configurations of alternative parameterizations in a series of 104 twelve-day September simulations during 11 consecutive years, 2000-10. The 64 configurations combine WRF parameterizations of cumulus convection, radiation, surface hydrology, and the PBL. Simulated daily and total precipitation results are validated against Global Precipitation Climatology Project (GPCP) and Tropical Rainfall Measuring Mission (TRMM) data. Particular attention is given to westward-propagating precipitation maxima associated with African easterly waves (AEWs). A wide range of daily precipitation validation scores demonstrates the influence of alternative parameterizations. The best WRF performers achieve time-longitude correlations (against GPCP) of between 0.35 and 0.42 and spatiotemporal variability amplitudes only slightly higher than observed estimates. A parallel simulation by the benchmark Regional Model version 3 achieves a higher correlation (0.52) and realistic spatiotemporal variability amplitudes. The largest favorable impact on WRF precipitation validation is achieved by selecting the Grell-Devenyi convection scheme, resulting in higher correlations against observations than using the Kain-Fritch convection scheme. Other parameterizations have less obvious impacts. Validation statistics for optimized WRF configurations simulating the parallel period during 2000-10 are more favorable for 2005, 2006, and 2008 than for other years. The selection of some of the same WRF configurations as high scorers in both circulation and precipitation validations supports the notion that simulations of West African daily precipitation benefit from skillful simulations of associated AEW vorticity centers and that simulations of AEWs would benefit from skillful simulations of convective precipitation.