WRF ensemble dynamical downscaling of precipitation over China using different cumulus convective schemes

Abstract

Ensemble dynamical downscaling of precipitation over China was conducted based on simulations of the Weather and Research Forecasting model using both Kain–Fritsch and Grell cumulus convective parameterization schemes. The simulations were driven by ERA-Interim reanalysis data with 25-km horizontal resolution for the period 1980–2015. Results indicated that superior performance was achieved by the ensemble based on the two different schemes because certain observed signals were captured complementarily by each scheme over distinct regions. For climatological mean precipitation, the ensemble improved the annual mean pattern, probability distribution, and seasonal evolution of precipitation. With regard to interannual variation, the ensemble showed the highest skill in representing the series of precipitation anomalies at regional scales for all subregions. Improvement was also evident in the spatial patterns of the dominant precipitation variability modes and the corresponding temporal variations. For extreme precipitation, several indices were selected, and the ensemble was better able to capture the main features with higher spatial pattern correlations and closer magnitudes. The advantages of the ensemble are due to its appropriate regime specific weights derived from the Kain–Fritsch and Grell schemes