Using EEM-PARAFAC to identify and trace the pollution sources of surface water with receptor models in Taihu Lake Basin, China.

The identification and apportionment of the multiple pollution sources are essential and crucial for improving the effectiveness of surface water resources management. In this study, the surface water samples were collected from Taihu Lake Basin, and the optimal water quality parameters for the receptor models were selected firstly with multivariate statistical analyses. In order to identify the potential pollution sources in surface water, dissolved organic matter (DOM) was analyzed with the excitation-emission matrix coupled with parallel factor analysis (EEM-PARAFAC). Through the Pearson correlation analysis of water quality parameters and DOM components, the pollution sources were further verified, i.e., agricultural activities, domestic sewage, phytoplankton growth/terrestrial input and industrial sources. In addition, principal component analysis (PCA) combined with the absolute principal component score-multiple linear regression (APCS-MLR) and positive matrix factorization (PMF) models were employed to quantify pollution sources. Compared with PCA-APCS-MLR model, PMF model resulted in higher performance on evaluation statistics and lower proportion of unexplained variability, thus showed more realistic and robust representation. The results of PMF showed that agricultural activities (42.08%) and domestic sewage (21.16%) were identified as the dominant pollution sources of surface water in the study area. This study highlights the effectiveness of EEM-PARAFAC in identifying the pollution sources, and the applicability of PMF in apportioning the contributions of each potential pollution source in surface water.