科研之友
微信
新浪微博
Facebook
分享链接
摘要
A new merging scheme (referred to as HL-OI) was developed to combine daily precipitation data from high-resolution gauge (HRG) observations, The Climate Prediction Center morphing technique (CMORPH) satellite estimates, and National Centers for Environmental Prediction (NCEP) numerical simulations over China to perform reliable high-resolution daily precipitation analyses. The scheme is designed using a three-step strategy of removing systemic biases, reducing random errors, quantitatively estimating error variances, and combining useful information from each data source. First, a cumulative distribution function matching procedure is adopted to reduce biases and provide unbiased background fields for the following merging processes. Second, the developed error estimation algorithm is implemented to quantify both the background and observation errors from the background departures. Third, the bias-corrected NCEP and CMORPH data are combined with the HRG data using the optimal interpolation (OI) objective analysis technique. The magnitudes and spatial structures of both observation errors and background errors can be estimated successfully. Results of cross-validation experiments show that the HL-OI scheme effectively removes most of systemic biases and random errors in the background fields compared to the independent gauge observations and is robust even with imperfect background fields. The HL-OI merging scheme significantly improves the temporal variations, agreements between the spatial patterns, frequency, and locations of daily precipitation occurrences. When information from gauge observations, satellite estimates, and model simulations are combined simultaneously, the merged multisource analyses perform better than dual-source analyses. These results indicate that each independent information source of daily precipitation contributes to improving the quality of the final merged analyses under the framework of HL-OI scheme.
