Multi-model Ensemble QPF for Southeastern China Using Bayesian Model Averaging

摘要

TIGGE database contains 10-16 day ensemble quantitative precipitation forecasts (QPFs) produced by numerical weather prediction (NWP) models from about 10 NWP centers around the world. How to make use of the different ensemble QPFs has been a subject of intense research interest. This paper discusses how to use Bayesian Model Averaging (BMA) to combine ensemble QPFs from different NWP models. Ensemble QPFs from three NWP models are used in this study: China Meteorological Administration (CMA), Japanese Meteorological Administration (JMA) and US National Center for Environmental Predictions (NCEP). The study area is the Southeastem China. Preliminary results are presented in this paper.