基于聚类和多重评价法的河流质量评价研究

Water Quality Assessment of a Large River Based on Multiple Assessment Method and Cluster Analysis

采用多重评价方法与聚类分析法对研究很少的大型江河——沅江下游的水质情况开展了调查、分析和质量评价。主成分分析法评价该江段的水质均属于Ⅰ类～Ⅱ类；加权的Topsis法、综合指数评价法、模糊识别评价法和判别分析法计算也均属于Ⅰ类～Ⅱ类，且方法问重复性较高，相互补充。UPGMA聚类分析发现，沅江桃源江段的水质可分为三个亚区；从修正TSI来看，所调查样点多为中营养水体；Ross法评判发现江段下游水质有变差趋势，除下游靠近县城水体外，多数仅为轻度污染水。总之，采用多重评价法综合对比才能弥补单一方法存在的漏洞，采用主成分分析法＋加权的Topsis法＋判别分析的水质评估以及TSIm河流富营养化评估和Ross法评估的多重评价和聚类分析等的评价体系能综合水质指标的不同侧重点，能更真实开展水质评价工作。

Water quality of lower reaches of Yuanjiang River, a rarely studied large fiver, was sampled and assessed through multiple assessment method and cluster analysis. Through principle component analysis （PCA）, the water quality of 7 sampling plots satisfied Grade Ⅰ-Ⅱof National Standard of Surface Water Quality. Other assessment methods such as Topsis method, comprehensive evaluation index method, fuzzy identification model and discriminating method were also used to evaluate the water quality, with calculated results similar to that of PCA. UPGMA cluster analysis was used to divide the section of the River to three parts according to clustering of water quality. Results showed that most of water samples were in the eutrophic state, with a tendency of decrease at lower reaches of the River. Multiple assessment methods such as PCA, Topsis method and discriminating method and cluster analysis were integrated into the study, which were much closer to the reality of water body.