长江流域0.1°网格逐日降水数据集及其精度

利用气象部门在长江流域及其附近地区的国家基本站和15省市的一般观测站1971~2005年期间的逐日降水资料,根据降水资料特点,提出使用Barnes插值和最近台站降水频率相结合的混合插值方案,得到长江流域分辨率为0.1经纬度的网格化降水逐日资料集,并通过交叉检验确定了其误差估计。结果表明,该混合插值方案对降水的模拟不但绝对误差和均方根误差较小,相关系数大,而且对降水的方差和频率模拟与观测接近,为降水资料在气象和其它领域的应用提供了逼真、连续和格点化的数据。

A Daily Precipitation Grid Dataset with 0.1°Resolution in Changjiang River Valley and Its Precision

A spatial interpolation by combining the Barnes scheme and precipitation frequency is developed according to daily rainfall characteristics. With the interpolation, a daily grid precipitation dataset, which covers a domain ranging from 99° to 123°E and 24° to 36°N and has a 0.1 longitudinal and latitudinal resolution, is obtained by utilizing daily precipitation data of basic observation stations in China and ordinary observation stations in 15 provinces in the vicinity of the Changjiang valley during 1971 to 2005. The error estimation of the grid dataset is accessed through cross-validation statistics. The statistics show the combining scheme for daily precipitation interpolation is not only with small mean biased error, mean absolute error and mean square error as well as high correction coefficient, but also close to the observation variation and frequency. The primary analysis indicates that the datasets provide a finer precipitation distribution. And the years with the greatness (smallness) of the annual area precipitation, in particularly the summertime area precipitation, correspond to the distinguished flooding (drought) processes occurred in the Changjiang valley. So the daily precipitation datasets can be applied to meteorology and relative field with realistic, continuum grid data.