中文摘要：

卫星反演降水产品作为降水资料获取的一种重要途径,其精度往往因时空尺度和雨强的不同而变化。本文以美国CMORPH卫星反演降水产品为研究对象,以2008年～2010年夏季中国东部的降水格点分析场为参考值,以均方根误差RMSE作为误差分析指标,定量化地探究了CMORPH产品的误差随时间分辨率、空间分辨率和雨强的变化规律。结果表明：卫星反演降水产品的误差随空间分辨率和时间分辨率的增大呈现递减的趋势,随雨强的增大而呈现出递增的趋势。进一步的分析发现：空间分辨率对误差的影响规律可以采用y=a＊e-b＊x函数进行拟合,时间分辨率和雨强对误差的影响规律可以采用y=a＊x-b函数进行拟合。在此基础上,构建了CMORPH产品随时空分辨率和雨强等3种因素而变化的误差模型σr=a＊e-b＊A＊T-c＊Rd。针对误差模型的适用性分析表明：误差公式在多数研究区均表现出良好的适用性。

英文摘要：

As the key way of acquiring precipitation data,the satellite-based precipitation product＇s uncertainty is largely dependent on the temporal and spatial scale and rain rate.This paper systematically examines the RMSE of CMORPH product at different temporal and spatial resolutions over 15 selected 5°×5° boxes.The study is based on the detailed analysis of precipitation in the summer of 2008~2010 in east China.The uncertainty estimated results show that the relative RMSE of CMORPH is to decrease by the degree of both temporal and spatial resolution,and to increase by the level of rain rate.Both the decrease by temporal resolution and increase by rain rate level could be simulated well by a power function,while the decrease by spatial resolution could be simulated well by e-exponential function.An uncertainty model has been built,which is integrated with the temporal spatial resolutions and rain rate.The analysis shows that the model fits most test areas very well.