Effects of forest structure from graph theory connectivity indicators on river water quality in the Caspian Sea Basin

Abstract

Abstract Landscape structure is one of the most important factors affecting the sources of food and organic matter in rivers. The spatial patterns of landscape patches including dimension variability and gap sizes between patches contribute in the amount and type of materials discharged to waterbodies and watercourses. We studied the effects of forest-patch connectivity and landscape corridors on water quality in the Greater Caspian Sea Basin. We used 10 landscape metrics and 11 water-quality indicators developed from graph theory to examine if contiguous/unfractured landscapes enhance water quality. We used independent calculations of Pearson’s and Spearman’s correlation coefficients to explore the association of forest-patch connectivity and water quality metrics. Stepwise regression was also used to generate allometry-based power, exponential, and logarithmic models. The results of correlation between forest-patch connectivity indicators and water quality parameters showed that several forest-patch metrics indicating connectivity including dLCP (Landscape coincidence probability) and dIIC (Integral Index of Connectivity) had a significant negative correlation with water quality metrics indicating pollution. This means that increasing forest connectivity is associated with improvement in water quality. The modeling results also showed that almost all selected models with acceptable AIC coefficients were nonlinear models. As connectivity of forest patches decreases and more fragmentation occurs in a watershed, the parameters of water pollution increase and the quality of water decreases. Models showed high R2 values for water quality metrics including CO3 (0.82), water discharge (0.73), Ca (0.77), and TDS (Total Dissolve Solids) (0.70).