Identification and Apportionment of Potential Pollution Sources Using Multivariate Statistical Techniques and APCS-MLR Model to Assess Surface Water Quality in Imjin River Watershed, South Korea

Abstract

Reliable water quality monitoring data, identifying potential pollution sources, and quantifying the corresponding potential pollution source apportionment are essential for future water resource management and pollution control. Here, we collected water quality data from seven monitoring sites to identify spatiotemporal changes in surface water in the Imjin River Watershed (IRW), South Korea, distinguish potential pollution sources, and quantify the source apportionment from 2018–2020. An analysis was performed based on multivariate statistical techniques (MST) and the absolute principal component score-multiple linear regression (APCS-MLR) model. Statistically significant groups were created based on spatiotemporally similar physicochemical water quality characteristics and anthropogenic activities: low-pollution (LP) and high-pollution (HP) regions, and dry season (DS) and wet season (WS). There were statistically significant mean differences in water quality parameters between spatial clusters, rather than between temporal clusters. We identified four and three potential factors that could explain 80.75% and 71.99% in the LP and HP regions, respectively. Identification and quantitative evaluation of potential pollution sources using MST and the APCS-MLR model for the IRW may be useful for policymakers to improve the water quality of target watersheds and establish future management policies.