水环境模拟参数随机场的指示克里格研究

Indicator Kriging of random parameters fields in water environment simulation

在水环境模拟中,经常要对各种参数的空间分布数值做出可靠的估计。环境问题的复杂性往往使得这类估计存在不确定性。如何利用有限的监测数据来实现水环境中各种参数的精确估值具有重要的实际意义。以水环境模拟参数随机场中的离子质量浓度为例,介绍了指示克里格估值计算的理论方法及应用。通过对样本数据的统计分析得出,数据总体不服从正态分布,且表现出一定的非平稳性。在此基础上用指示克里格法对离子在模拟环境中分布的空间结构特征进行分析,并对估点进行估值计算。估值结果显示,指示克里格的估值效果优于普通克里格。这一实际应用表明,指示克里格法可以较好地描述水环境模拟参数随机场的空间分布规律。

The present article is to introduce the authors＇ improved indicator Kriging water environmental simulation model based on the traditional ordinary Kriging method. As is known, it is of key importance to estimate each kind of parameter accurately by utilizing the limited monitor data in the water environment assessment. The reason is that the multiplicity of enviromnental factors or the collectivity of data are not arranged as the normal school rules ask for but rather show non-stationary or regular features that. would cause the complication of the related factors in the whole estimation. Therefore, this paper prefers to give descriptions by using an improved method, the so-called indicator Kriging assessment method by taking the ion mass concentration of parameters in the random fields for our water environment simulation. However, the application of the improved indicator Kriging method allow us to describe better the spatial distribution laws of the parameters of random nature in ordinary water environmental simulation. Therefore, based on the indicator Kriging, three thresholds are chosen to analyse the ion＇s spatial structural characteristics in water environmental simulation in which the indicator variogram is taken as a main device. It is found that the parameters at the different thresholds can reveal the differences from one another among themselves. And the calculation process of indicator Kriging helps us to restrict the incidence of the outliers but still consider the outliers may sneak into the process through the way of probability when the structural characteristics of the spatial distribution of variables are brought into the water environmental simulation, Thus, the eventual results demonstrate the true estimated values of the indicator Kriging method.