Source apportionment of water pollutants in Poyang Lake Basin in China using absolute principal component score–multiple linear regression model combined with land-use parameters

Abstract

Source apportionment of surface water is essential for effective pollution control and sustainable water management. Physical mechanism models usually need so much data and parameters for calibration that their application for complex hydrologic condition watershed becomes difficult. However, reverse source tracing methods only based on water quality parameters present a certain subjectivity and uncertainty. In this research, additional land-use parameters were applied as an auxiliary in principal component analysis (PCA) for accurate identification of pollution sources. Thirteen water quality parameters and two meteorology parameters were used in the PCA and absolute principal component score–multiple linear regression (APCS–MLR) model to quantitatively identify potential pollution sources and their contributions to surface water pollution of the Poyang Lake Basin, in which frequent flow and sediment flux exchange with Yangtze River make the river–lake relationship complex. The results showed that urban wastewater with 34% contribution and agricultural non-point sources with 16% contribution, were the major sources of pollution in water quality. TP and NH 3 –N, the most serious pollutants, causing agricultural non-point source pollutions with 40% contributions and urban wastewater with 21% contributions were the major sources in the Poyang Lake Basin. Urban wastewater with 60% contributions was the major source of organic contamination. It can be concluded that with associated land-use parameters, the GIS approach with the APCS–MLR model can improve the accuracy and certainty of source apportionment, providing aid decision information for managers on protection of surface water quality.