基于地统计学方法的泾河流域降水空间变异规律研究

Study on Spatial Variation of Precipitation in Jinghe River Basin Based on Geostatisitc Method

地统计学方法对具有结构和随机双重性质的变量的空间变异性研究 ,具有极强的科学性 ,它所提供的克立格法是一种最优的无偏内插估计方法。本文以泾河流域年降水量为研究对象 ,利用普通克立格法对 90年代 5个典型年份的年降水量的空间变异性进行分析。通过对几种模型估值误差指标的比较 ,认为用球状模型来拟合降水量的半变异函数的精确度最高 ;利用球状模型对典型年份的降水量的半变异函数进行拟合 ,对比分析各年份的主要参数 ,结果表明 1 996年降水量在空间变异的强度和影响范围均较大 ,而 1 995年的变异的强度和影响范围则较小 ;应用克立格插值生成各年份流域降水量的空间分布图 ,可直观分析降水量空间变异性 :总体趋势是流域的降水量从东南向西北逐步减少 ,呈平缓的梯状分布。而 1 995年较为特别 ,流域东南部降水量很小 ,主要集中在流域西部 。

Geostatistic method is scientific in the research of spatial variation of variable, which is both structural and stochastic. Kriging method, provided by Geostatistic , is a optimization method in spatial interpolation. In the paper, Kriging method was used to anal size the spatial variations of precipitation of Jinghe River Basin in 5 select years in 1990s. Through comparing the prediction errors of different models, spherical model was found to be the best one to fit the experimental semivariogram. Using spherical model to calculate the experimental semivariogram, main parameters in every select year were gotten, such as nugget, still, rang. Analyzing the parameters, The result showed that the spatial variation intensity and The range influenced of the precipitation in 1996 were the maximum, and those in 1995 were the minimum. Predict maps were generated according to Kriging interpolation method. From the maps, it could be concluded that the precipitations in Jinghe River Basin in 1990s were reduce from southeast to northwest, and the trend was gently. 1995 was a exception, the precipitation was mainly in the west of the basin. The trend of precipitation was changed to increase from the west to the east of the basin.