多元统计方法用于太湖梅梁湾水质特征识别

旨在识别太湖梅梁湾水质特征，为水质保护、规划、管理、利用提供决策参考.研究利用太湖梅梁湾区域9个监测点数据，以主成分分析探讨主要污染来源；以聚类分析划分监测点类别并识别其空间相似性；以比对各类别监测点数据，讨论了污染物类别及浓度变化情况.结果显示梅梁湾水质主要受农业非点源、浮游植物生长、外源输入的有机悬浮物、含氮有机污染物及土壤土质5方面影响；梅梁湾区域9个监测点位划归为4类，即：河流入湖口、入湖口近岸、远离入湖口近岸及湖心点类；梅梁湾水质主要超标污染物为N、P，且各指标浓度变异不大.由此可见，太湖梅梁湾水质具有明确的空间分布与特征.

Multivariate statistical methods for recognition of water quality feature in Meiliang Bay of Taihu Lake

The aim is to identify feature of water quality in Meiliang Bay of Taihu Lake and provide scientific refer- ence for protection, planning, management and water utilization. According to the data of nine monitoring points in Meiliang Bay, the main sources of pollutants were firstly examined based on principal component analysis （PCA） ; secondly, the sampling points were categorized with cluster analysis （CA） and the spatial similarities and differences between sampling points were identified; at last, the types of pollutants and variations of concentration were ana- lyzed through comparison with the monitoring data of various types of sampling points. The results showed that： the water quality of Meiliang Bay of Taihu Lake is affected by agricultural non-point source, phytoplankton growth, organic suspended matter from external sources, nitrogenous organic pollutant and composition of soil; the nine mo- nitoring points in Meiliangwan Bay can be classified into four categories, namely, the points of river estuary, lake- shore near river estuary, lakeshore away from the river estuary and the lake central; Meiliang Bay was seriously pol- luted with N and P, and the concentration variation of each index was not obvious in each category of sampling. The results indicated that the water quality of Meiliang Bay has definite spatial distribution and feature.