Spatio-temporal patterns and source apportionment of coastal water pollution in eastern Hong Kong.

The comprehensive application of different multivariate methods and geographic information systems (GIS) was used to evaluate the spatio-temporal patterns and source apportionment of coastal water pollution in eastern Hong Kong. Fourteen variables were surveyed at 27 sites monthly from 2000 to 2004. After data pretreatment, cluster analysis grouped the 12 months into two groups, June-September and the remaining months, and divided the entire area into two parts, representing different pollution levels. Discriminant analysis determined that NO3- -N, DO, and temperature and TN, SD, PO4(3-)-P, and VSS were significant variables affecting temporal and spatial variations with 84% and 90% correct assignments, respectively. Five potential pollution sources were identified for each part by rotated principal component analysis, explaining 71% and 68% of the total variances, respectively. Receptor-based source apportionment revealed that most of the variables were primarily influenced by soil weathering and organic pollution, nutrient pollution (or agricultural runoff), and mineral pollution. Furthermore, GIS further facilitated and supported multivariate analysis results.