Dynamics of sulfate and nitrate in a temperate stream network: combining multi-isotope analysis and generalized linear mixed-effects models for precautionary conservation

Abstract Environmental pollution in stream networks is a major concern. In this study, the effects of land-use and geological characteristics on basin-scale sulfate and nitrate dynamics in the Chikusa River basin, Hyogo, Japan, where substantial environmental pollution has not yet been officially reported, were evaluated from the perspective of an overview. Using sulfate and nitrate concentrations as well as their isotopic signatures (δ15N-NO3−, δ18O-NO3−, δ34S-SO42−, δ18O-SO42− values) as response variables, we employed generalized linear mixed-effects models (GLMMs) to clarify the effects of land coverage, geological rock type, and season on the loading of these solutes. The results obtained showed an increase in sulfate concentration with an increase in the proportional area of exploited land regardless of the season. Isotopic signatures (δ34S-SO42−, δ18O-SO42− values) and the GLMMs suggested that sulfate primarily originated from the soil owing to rock weathering processes, and its distribution was mainly driven by the confluence of river branches along with altitude. In contrast, nitrate concentration varied with season, decreasing with an increase in the proportional area of exploited land in summer and showing an opposite trend in winter. The δ15N vs. δ18O plot showed that the impacts of direct nitrate input from precipitation and chemical fertilizer were negligible, while topically applied manure and septic waste played important roles. The GLMMs also indicated that the proportion of fertilized area affects the spatial distribution of δ15N-NO3− values but not nitrate concentration, implying the existence of nitrate loading from fertilized areas that could not be detected via concentration-only measurements. Therefore, the combined use of isotopic signatures and GLMMs is expected to provide valuable information for discerning potential sources of pollution in areas without substantial environmental issues. Taken together, this approach could be applied as a “proactive indicator” particularly for areas without apparent pollution, to ensure effective river management.