Estimating the contribution of environmental variables to water quality in the postrestoration littoral zones of Taihu Lake using the APCS-MLR model

Abstract

Water quality monitoring is one of the most important aspects of postrestoration assessments because it affects water pollution control and the development of sustainable management strategies. However, a comprehensive understanding of potential water pollution and source apportionment in restoration projects is still lacking. In this study, the water quality variables of three restored national wetland parks with different cofferdam systems (i.e., an eco-layered cofferdam, a fully enclosed cofferdam, and open water) in the littoral zone of Taihu Lake were monitored monthly for three years (2019–2021). Hydrochemical and meteorological variables were used as auxiliary parameters for multivariate statistics, including principal component analysis (PCA) and absolute principal component score-multiple linear regression (APCS-MLR), to accurately estimate the source apportionment of the potential factors influencing the water environment. PCA extracted three or four potential sources, accounting for 64.71 %, 65.40 %, and 63.85 % of the total variance. The APCS-MLR results showed that wind direction and volatile suspended solids were the primary sources affecting water quality in open water, with a sum of the mean source contributions of 40.7 %. In fully enclosed cofferdam systems, the dire state of endogenous pollution was the greatest potential source affecting water quality, with a mean contribution of 41.2 %. The eco-layered cofferdam alleviated the contributions of suspended solids (mean contribution of 23.7 %) and nutrients in the water column (mean contribution of 30.8 %); however, the contribution of organic matter in the cofferdam was relatively high (mean contribution of 13.4 %). Based on these results, eco-layered cofferdams play a positive role in eutrophication control and ecological restoration in the littoral zone of large shallow lakes. Meanwhile, adding meteorological variables to assist hydrochemical variables in multivariate statistics may improve the accuracy and certainty of pollution source apportionment and support decision-makers in developing water quality protection and management strategies for postrestoration projects in littoral zones.