Water Quality Change of Dongting Lake Basin with the Influence of Landscape Metrics at Multiple Scales

Stream water quality deterioration has been a problem attracting global concern. Landscape metrics, including topography, land use pattern (landscape configuration and land use composition), can influence water quality by changing the processes of non-point source pollutants (NPS) into rivers. And the influence of each type of landscape metrics changed with spatial and temporal scales. Quantifying the unique influence of three types of landscape metrics can help improving water quality by managing landscape. With the monthly water quality data in 2016 of Dongting Lake China, we quantify the unique contribution of topography, landscape configuration and land use composition on water quality variation at spatial scale (buffer and sub-watershed scales) and seasonal scales (wet and dry seasons) by the statistical method of redundancy analysis (RDA) and partial redundancy analysis (pRDA). The results indicated that the explanation of the three types of landscape metrics on water quality was generally ordered as: topography (11.4-30.9%)> landscape configuration (9.4-23.0%) > land use composition (5.9-15.7%). And the influence of these landscape metrics on water quality changed with spatial and temporal scales, which explained water quality variation better at buffer scale and during the dry season. The influence of topography on water quality variation was most significant at spatial scales, while the interactive influence of the three environmental groups on water quality was more sensitive to seasonal difference. Besides, the concrete influence of each landscape parameters changed at multiple scales, with percentage of water area as the main index influencing water quality during the dry season and edge density highest explaining water quality condition during the wet season at both buffer and sub-watershed scales. The results of our study indicated that water quality management should considering the influence change of landscape metrics at spatial and temporal scales.