Chemometric modeling of main contamination sources in surface waters of Portugal.

Various chemometric data analysis methods, such as principal components analysis, multivariate curve resolution-alternating least squares, parallel factor analysis, and Tucker3, are proposed and compared for the resolution and modeling of main contamination sources in a large environmental data array obtained in an exhaustive environmental monitoring program that examined the quality of surface waters of Portugal. The study covered the analysis of 19 priority semivolatile organic compounds (SVOCs) frequently found in a total number of 644 surface water samples, including 46 different sites from throughout Portugal and corresponding to a period of 14 months, from April 1999 to May 2000. Main contamination sources of the analyzed SVOCs were identified and interpreted according to their chemical composition and according to their resolved geographical and temporal distribution profiles.