Reduction Biases in Regional Climate Downscaling: Application of Bayesian Model Averaging on the Large-scale Forcing

Simulations of the 1998 East Asian summer monsoon were carried out using the Weather Research and Forecast (WRF) model forced by four reanalysis datasets (NCEP-R2, ERA-40, ERA-IN, and JRA-25), their equally weighted ensemble mean, and their ensemble mean based on Bayesian model averaging (BMA). Large discrepancies were found among experiments forced by individual reanalysis datasets. The uncertainties in the moisture field of large-scale forcing over ocean were responsible for the discrepancies. The control experiment forced by the equally weighted ensemble forcing reduced the biases in simulated circulation to a large extent, but only reduced the biases in simulated precipitation in some cases. The experiment forced by the BMA ensemble forcing outperformed not only the experiments forced by individual reanalysis datasets but also the control experiment. The results suggest that the uncertainties in lateral boundary forcing can be reduced through the BMA ensemble method based on the satellite data, and more three-dimensional satellite data is urgently demanded.