Configurationally analysis of relationships between land-cover characteristics and river water quality in a real scenario

Abstract

The composition and configuration of land uses/covers (LULC) have a key role in determining the quality of surface waters. This study investigated the relationships between 207 landscape metrics (LMs) at landscape and class levels and water quality parameters (WQPs) by analytically approach in Mazandaran sub-basins, north of Iran. For this purpose, at first the basic WQPs were identified by principal component analysis (PCA). Then, stepwise linear regression analysis was applied to recognize optimal LMs for estimating each of the WQPs individually. Finally, the effect of the spatial configuration of LULC classes on WQPs and their variability was analytically evaluated by real samples. According to the PCA results, SAR, TDS, pH, PO43−, and NO3− were identified as principal WQPs in Mazandaran Rivers. The results also showed that the interspersion and juxtaposition index of bare lands, related circumscribing circle of agriculture, percentage of the forest, connectivity index of residential, and percentage of agriculture were the optimal metrics for estimating SAR, pH, TDS, PO43−, and NO3− levels, respectively. The metrics at the class level also had more ability to describe the WQPs. In this study, a suitable model for future LULC establishment in Mazandaran Province with the aim of improving effects on surface water quality was proposed.