基于主成分聚类分析的地下水化学分类方法

Groundwater Chemical Classification Method Based on Principal Component Cluster Analysis

根据新疆巴里坤县汉水泉地区20组承压水水样的K^+、Na^+、Ca^(2+)、Mg^(2+)、SO_4^(2-)、C^l-和HCO_3^-等7个化学指标检测数据,采用舒卡列夫分类法进行地下水水化学类型分类。结果显示,部分同类型之间组分含量差异较大而不同类之间组分含量差异较小,分析认为是舒卡列夫分类法划分界限人为性太强导致的。因此,我们采用主成分聚类分析法代替舒卡列夫分类法,对水化学类型进行分类。本文借助SPSS软件进行主成分分析,提取2个主成分,将主成分得分作为指标进行聚类分析,最终将20组水样划分为6类。主成分聚类法通过消除变量之间的相关性使结果更加客观合理。

According to the determination result of seven chemical indexes including K~＋,Na~＋,Ca^（2＋）,Mg^（2＋）,SO_4^（2-）,Cl^- and HCO_3^-of 20 groups of confined groundwater samples in Hanshuiquan region of Balikun County in Xinjiang,Shkalev classification is applied to classification of hydrochemical types. Results show that the difference among the same type of samples is relative greater and the difference among different type of samples is relative smaller which is probably due to certain artificial factors of Shkalev classification which can lead to the inaccuracy of classification results. Therefore,principal component cluster analysis is applied to classification of hydrochemical types instead of Shkalev classification. Using principal component analysis with SPSS software,two principal components were extracted in our study. The 20 groups of samples were divided into six categories according to the principal components scores which are regarded as the metric of cluster analysis. The principal component cluster analysis eliminates the correlation between variables and makes the results more objective and reasonable.