ABSTRACT
ABSTRACT: Multivariate statistical techniques, such as cluster analysis (CA) and principal component analysis (PCA) were applied to investigate the spatial variation and source of coastal water pollution in Xiamen Bay. According to CA, the sampling sites were significantly classified into two clusters, Inner bay (1st cluster) and outer bay (2nd cluster). The 1st cluster was distributed in Jiulong River Estuary (JRE), Eastern Sea (ES) and Dadeng Sea (DS), and the other one was distributed in Western Sea (WS) and Tongan Bay (TB). Nutrients of 1st cluster were higher than the other. based on PCA, 88.65% and 90.97% of the total variance can be explained for 1st and 2nd cluster respectively. Three and two Varifactors (VFs) were intentified for the two clusters respectively. VFs of 1st cluster represent river input and agricultural wastewater, livestock and poultry and organic pollutant of domestic sewage, while VFs of 2nd cluster can interpreted by industrial wastewater, sewage and agricultural wastewater discharge. The focus of coastal pollution control in Xiamen Bay was land-based pollution, and how to make the watershed comprehensive management.
