多元回归方法在芹山水库流域年降雨量系列插补的应用

Application of Multiple Regression Method in Annual Rainfall Series Interpolation in Qinshan Reservoir Basin

降雨量是水文预报和水资源配置所需要的重要基础资料之一,在实际工作中雨量站缺测降雨量需要进行插补。针对这一问题应用多元回归方法,建立了两个或者多个雨量站之间的回归方程,并利用最小二乘法求解回归方程系数。以福建穆阳溪流域芹山水库流域雨量站为研究对象,邻近流域水文站作为参证站,求解了多个雨量站回归方程,进行了显著性检验的比较,最后选择均方误差最小的回归方程,完成了降雨资料的插补,可为工程实际提供依据。

Precipitation,as one of the most important basic indication for hydrologic prognosis and water resources distribution,requires interpolation in measurement of precipitation.This paper,on the basis of multiple regression,forms two or more regression equations between two rainfall stations,and it will solve coefficients to the regression equations by least square method.The author studies Qinshan Reservoir rainfall station in Muyang Stream watershed of Fujian Province,by referring to reservoirs adjacent to it and provides resolves of regression equations for multiple rainfall stations.In compare of significance test between the stations,the author selects the the repression equation with the lease mean square error and applies it to interpolation of precipitation.The test result of this article offers reference to engineering practices.