6种卫星降水产品在中国区域的精度特征评估

Validation of Six Satellite-Derived Rainfall Estimates over China

采用2012—2013年国家气象信息中心2419个国家级地面气象站格点化数据和TRMM(3B40RT、3B41RT、3B4 2RT)、CMORPH、GSMaP、HYDRO共6种0.25°×0.25°卫星反演降水数据,从中国区域、中国东部区域及中国西部区域3种空间尺度综合验证和比较了日尺度上卫星反演降水产品的精度。结果表明:春、夏、秋季卫星反演降水产品对降水描述的准确性高于冬季。相关性、均方根误差等指标随季节变化明显,且差异较大,夏季每种降水产品的评估指标均优于冬季。不同区域之间定量评估指标差异明显,具有月尺度变化特征。卫星反演降水产品对冬季小降水及固态降水的反演能力有限。对于中国区域的大范围降水事件,每种降水产品可以较好进行再现,但均以多传感器联合后的产品质量较高,单一微波、红外或可见光的效果较差。TRMM的3B42RT在夏季对降水反演的落区较好,但极值偏大,CMORPH的综合评价较好。卫星降水产品精度的空间分布与地形有一定的关系。

Using grided data from 2419 stations of National Meteorological Information Centre during 2012-2013 and the TRMM（3B40RT,3B41 RT,3B42RT）,CMORPH,GSMaP,HYDRO satellite-derived rainfall data,this paper evaluated and discussed the validation of rainfall estimates from the six data with0.25° ×0.25° spatial resolution and the accuracy of the retrieved daily rainfall products over the whole China,Eastern China and Western China.Results show that the accuracy of satellite-derived data is higher in spring,summer and autumn than in winter.The linear correlation coefficient,mean absolute error（MAE）,root mean square error（RMSE）,probability of detection（POD）,and the false alarm ratio（FAR） change with seasons.Meanwhile,the index of each data is better in summer than in winter.The quantitative statistical metrics are obviously different,changing in monthly scale.The estimate ability of satellite-derived rainfall is poor when estimating solid and small-amount precipitation.For the large range of precipitation events over China,all the 6 rainfall products have better result.The better quality is realized by merging multi-sources,and the worse quality results from single-source microwave,infrared or visible light.The 3B42 RT retrieves better falling area of rainfall,but extreme value is larger.The comprehensive evaluation of the CMORPH is better than others.The spatial distribution of accuracy of satellite-derived rainfall is related to terrains of China.