Assessment Study Of Merged Precipitation And Multiple Satellite Precipitation Products

Based on the precipitation observation data from ground automatic meteorological stations of China, the quality of three high-resolution satellite precipitation products (CMORPH, TRMM3B42, FY2) and a merged precipitation product which developed by National Meteorological Information Center are assessed. In this paper, the merged precipitation product and the satellite precipitation products’ accuracy are compared, while spatial and temporal distribution of errors, and detestability on precipitation daily cycle and several precipitation events such as typhoon are also discussed. The purpose of the paper is to supply suggestions for users of merged and satellite precipitation products by these major conclusions:(1) CMORPH TRMM3B42、FY2and merged precipitation products have similar spatial and temporal distributions with ground observations on the average daily precipitation in many years, both of them can show a pattern that the precipitation in southern of China are more than in northern. But CMORPH underestimated precipitation all over the China, and FY2overestimated precipitation in western China.(2) CMORPH TRMM3B42and FY2satellite precipitation products’3hr precipitation correlation coefficients are less than0.7, biases are among±0.25mm, root-mean-square error(RMSE), mean absolute errors(MAE) and relative errors(MRE) of southeastern China are higher than northwestern China, and all of them show a seasonal period. Errors of merged precipitation data have reduced obviously while correlation coefficient of merged precipitation data can be above0.8.(3) The missing retrieval rates of CMORPH, TRMM3B42and FY2precipitation products have a more than12%high value area at southern China such as Szechwan Basin. Research reveals that both merged and satellite precipitation products show low intensity precipitation days are more than high intensity precipitation days through dividing precipitation intensity into11grades in China, and bias and RMSE of precipitation products increase as precipitation intensity increase.(4) CMORPH and TRMM3B42can not well monitor the rainfall "double peak" in some southeastern and northern area of China FY2and merged precipitation data have similar results with ground observation."Night Rain" distributing in Tibet Plateau and Szechwan Basin can be found in both three satellite and merged precipitation products.(5) CMORPH, TRMM3B42, FY2and merged precipitation data all underestimated short time strong precipitation caused by extreme weather, meanwhile, have errors in monitoring precipitation centre. As for long-lasting rainfall event in "Mei Yu", CMORPH still underestimated precipitation value and monitoring rainfall peak time incorrectly. TRMM3B42 and FY2have better performance because adding observation data information, and merged precipitation product has the best result in monitoring this long-lasting rainfall. In the contrastive analysis of typhoon rainfall spatial and temporal distributions, we found that FY2has some severe problems in monitoring sea surface typhoon pattern, CMORPH and TRMM3B42can well capture typhoon rainfall process to some extent, but there are still have some errors in monitoring rainfall centre and value. In contrast, the merged precipitation data has the most similar result with observation data in land area in monitoring typhoon, and sea surface information come from CMORPH data in major.