Abstract: In this paper, we adopted a stepwise regression algorithm based on regional monthly division to establish precipitation estimation model based on the data from TRMM and FY-2C. This model also combined with observation data of weather stations and DEM data. By employing the estimation model, we obtained the spatial distribution of precipitation of the Yangtze River in January, April, July and October 2007, and tested and analyzed the results. The simulation results showed that the model could revise TRMM and FY-2C effectively. Further analysis and calculation showed that the averaged relative errors of TRMM precipitation in the four months were 37.7%, 47.3%, 44.2% and 41.9%, respectively,while those of FY-2C precipitation were 46.3%, 50.9%, 39.8% and 48.8%, respectively. From the perspective of the whole year, the correlation coefficient of simulated TRMM precipitation was 0.838, while the correlation coefficient of simulated FY-2C precipitation was 0.811. The simulation result showed that TRMM was more accurate than FY-2C. Moreover, the distribution of precipitation remained almost the same with the original data, and the results of the present model reflected the distribution pattern of precipitation.

Key words: Yangtze River basin, monthly precipitation, TRMM, FY-2C, grid estimation