Assessment of water quality in Baiyangdian Lake using multivariate statistical techniques

Abstract

Abstract Water quality can be considered a key contributor to both health and disease for humans. This study involved the evaluation and interpretation of complex water quality data and the sources of pollution in Baiyangdian Lake (China). It also allowed us to obtain more advanced information about water quality, and to design a monitoring network for this study area. Multivariate statistical techniques, including principal component analysis (PCA) and hierarchical cluster analysis (CA) were applied to evaluate water quality of the lake. The 21 physicochemical parameters focused on in the study were analyzed in water samples collected monthly over a two-year period from 13 different sites located in and around the lake. Exploratory analysis of experimental data involved use of PCA and CA in an attempt to discriminate sources of variation measured in the samples. PCA was used to identify a reduced number of five principle components, demonstrating up to 92% of both temporal and spatial changes. CA classified similar water quality stations into 5 clusters based on the PCA scores. The results showed that cluster 5 (site 2) was characterized as the most heavily polluted site, a result that can be attributed to the pollution from the nearby Fuhe River (a upstream river that receiving almost all of the domestic sewage and some industrial wastewater from Baoding City). Cluster 1 (sites 3, 4, 5, 6 and 7) and cluster 4 (site 1) were identified as moderately polluted in association with the both domestic and agricultural sewage, as well as fishery-related pollution in the lake. Cluster 2 (sites 11, 12 and 13) and cluster 3 (sites 8, 9 and 10) were less polluted, which suggests that the water quality was better in the eastern and central portions of the lake.