基于协同克里金的饮用水硝酸盐浓度预测研究

Estimation Nitrate of Concentration in Drinking Water Based on Cooperative Kriging Interpolation

为了研究贵阳某机场机场附近居民饮用水源中硝酸盐浓度的空间分布情况,对24个居民饮用水点分别进行普通克里金和协同克里金插值,并对插值结果进行比较分析。结果表明:通过距离模型判断出镁离子同硝酸盐的相关性最高,相关系数达到0.726,但无法提供显著性P值;偏相关分析对两者相关性进行检验,相关系数为0.787,P=0.001表明两者相关性极强,可以将镁离子的浓度作为协变量进行协同克里金的插值;运用克里金插值时发现,协同克里金模型的标准平均值和标准均方根预测误差均优于普通克里金插值,表明将镁离子的浓度作为协变量,可以弥补主变量硝酸盐样本点少,变异函数欠稳定的缺点;通过协同克里金插值进而可以得到研究区内的硝酸盐浓度预测分布图,期待能给当地的水污染治理提供一定的指导价值。

To study the spatial distribution of Nitrate concentration in drinking water resources of residents near an airport,we analyze the 24 settlements. The results show that the correlation coefficient between magnesium and Nitrate is the highest,and the correlation coefficient is 0.726,but the significant P value can not be provided. The correlation between the two variables is tested by partial process,and the correlation coefficient is 0.787,P = 0.001. The correlation between the two is very strong,and the concentration of magnesium ions can be used as a covariate for the interpolation of collaborative Kriging. When Kriging interpolation is used,the standard mean and standard root mean square prediction error of the cooperative Kriging model are both excellent. By kriging interpolation,you can get the prediction of nitrate concentration in the study area distribution,expect to provide some guidance to the local water pollution control.