基于PER-Kriging插值方法的降水空间展布

Spatial interpolation of precipitation using the PER-Kriging method

将PRISM插值对高程因素的计算方法引入普通Kriging插值中,提出了综合考虑观测点与插值点的位置、距离和高程关系的PER-Kriging插值方法。通过引入计算参考平面并进行数学变换,使该方法具有较好的可操作性。在对高程～降水量采用线性拟合的基础上,加入非线性拟合,进一步提高PER-Kriging插值的表达能力。在澜沧江流域分别采用PER-Kriging和普通Kriging方法进行降水插值对比,结果表明:前者能有效地消除观测点与周围区域地形差异太大造成的插值异常,并且能较好地消除插值运算的平滑效应,平均误差比后者减少了20 mm以上;不同曲线拟合插值的相对误差均小于12%。

Based on the parameter-elevation regressions on independent slopes model （PRISM）, a PER-Kriging interpolation method is developed to consider the effect of elevations on precipitation interpolation. The method can com- prehensively consider the complex relationships between grid points and observational sites in locations, horizontal dis- tances and elevations. The method uses a reference elevation in precipitation interpolation in order to practise easily. The performance of the method is enhanced through nonlinear refitting of the elevation-precipitation relationship ob- tained by linear fitting scheme. Both PER-Kriging and Ordinary Kriging methods are tested on the Lancangjiang basin for precipitation interpolation and results are compared with each other. The result shows that PER-Kriging can effec- tively eliminate the errors introduced by the large geographic differences and the smoothing processes. Compared to Ordinary Krijing, the PER-Kriging method can reduce the mean error of interpolated precipitation by 20 mm, and the relative errors in different types of curve fitting are less than 12%.