最优插值法在对中国自动站降水量空间分析中的参数优化

Parameter Improvements of Hourly Automatic Weather Stations Precipitation Analysis by Optimal Interpolation over China

探讨了利用最优插值对中国3万多个自动站小时降水数据进行空间分析时核心参数(数据信噪比、相关随距离的变化关系等)的优化和确定方法,其中优化的数据信噪比约为0.3,相关随距离增加呈e指数递减关系。研究表明:当站网密度差异较大时,通过确定影响台站数来进行空间插值,一方面可以提高站点稠密区的插值效率,另一方面可以提高站点相对稀疏区的插值精度。另外,探讨了最优插值中初估场对插值结果地影响,认为采用气候标准值生成的初估场能显著降低插值结果的随机误差,其结果比采用同时刻的降水分析场作为初估场更稳定。

Optimal Interpolation method （short for OI） is one of common ways to precipitation analysis. In this pa- per, the key parameters in OI such as signal to noise ratio, changing relationship between precipitation correlation co- efficient and distance at any two observed points is determined. The research results indicate that 1） the optimized signal to noise ratio is about 0.3; 2） the correlation coefficient of precipitation at two points is a function of its dis- tance with exponential decrease; 3） the number of influencing stations is an effective index when the gauge density is relatively inhomogeneous as the distribution of Automatic Weather Stations over China, which can improve the inter- polation efficiency over the dense gauged areas and increase the interpolation accuracy over the sparse gauged areas; 4） taking the climatological standard analysis as the first-guess field can effectively reduce the random error in OI and tends to present more stable results than the simultaneous first-guess field.