Water quality assessment and the influence of landscape metrics at multiple scales in Poyang Lake basin
Jinying Xu,Yang Bai,Hailin You,Xiaowei Wang,Zhifei Ma,Hongwei Zhang
DOI: https://doi.org/10.1016/j.ecolind.2022.109096
IF: 6.9
2022-08-01
Ecological Indicators
Abstract:Stream water quality has been increasingly deteriorated in recent decades because of anthropogenic activities. Assessing water quality via key environmental parameters and quantifying the respective influence of landscape metrics like landscape configuration, land use composition and topography, which can overall represent anthropogenic activities, can facilitate the development of improved water quality management strategies. However, few studies concentrate on the quantitative effect of these landscape metrics groups on the change of key water quality parameters. In this study, with Poyang Lake basin, we selected key factors of the 14 environmental variables for the cost-effective water quality evaluation model of the minimum water quality index (WQImin) by the random forest method and quantified the contribution of different groups of landscape metrics on key water quality parameters by redundancy analysis and variance partitioning analysis. The results indicated that six water quality parameters of ammonium (NH4+-N), total nitrogen (TN), permanganate index (CODMn), dissolved oxygen (DO), fecal coliforms (F.coli) and total phosphorus (TP) can significantly represent the overall water quality of the Poyang Lake basin. Based on WQImin, the water quality of Fuhe River and Xiushui River was significantly better than that of Ganjiang River, Raohe River, and Xinjiang River in Poyang Lake basin. The contribution of landscape metrics to water quality variation followed the order of landscape configuration (19.0–22.5%) > land use composition (5.2–20.2%) > topography (5.7–8.0%). Here, Build-up land was the environmental factor contributing most significantly to the variation in water quality at both spatial and seasonal scales (5.7–21%). Build-up land, Agricultural land, Contagion index and Cohesion index were positively correlated with TN, TP, CODMn, and NH4+-N and negatively with DO at different scales; Grassland, Forest land, Hypsometric Integral index, Slope, and Mean Shape index showed the opposite relationship. The influence of these landscape types changed with spatial and seasonal scales, more significantly explaining variations in water quality during the dry season and at buffer scale. Changes in water quality at the spatial scale were most sensitive to land use composition. Our results represent a basis for the improvement of water quality management.
environmental sciences