Understanding the spatiotemporal variability in nonpoint source nutrient loads and its effect on water quality in the upper Xin’an river basin, Eastern China

Abstract

Understanding the spatiotemporal characteristics of major nonpoint sources (NPSs) are crucial for protecting river water quality in drinking water source areas. This study employed a geomorphology-based nonpoint source pollution (GBNP) model to simulate hydrological processes together with NPS pollutant transport and transformation processes from 1980 to 2019 in the upper Xin’an River basin, an important drinking water headwater. The simulations revealed the spatiotemporal characteristics of nitrogen and phosphorus loads and their influences on river water quality. The results showed that the long-term average total nitrogen (TN) and total phosphorus (TP) loads were 3219 kg∙km−2∙yr−1 and 293 kg∙km−2∙yr−1, respectively. The proportional coefficients for TN and TP entering the rivers from the watershed were 0.40 and 0.37. The TN and TP loads during the April to July flood season accounted for 65.0 % and 63.2 % of the annual total, respectively. Annual TN load was significantly correlated with annual runoff depth, soil erosion and TN input in the interannual variation, while the interannual variation of TP load was significantly correlated with annual runoff depth and soil erosion. The long-term average TN and TP loads showed spatial patterns similar to those of land use and runoff depth, respectively. The daily TN concentrations showed a clockwise hysteresis relationship with discharge during flood events, while the daily TP concentrations had a single positive correlation with discharge throughout a whole year. Without atmospheric N deposition, the average TN load would decrease from 3122 kg∙km−2∙yr−1 to 1266 kg∙km−2∙yr−1 in 2000–2009. The TN load from atmospheric deposition accounted for 65 %-71 % of the total load, which implies that river water quality protection must include local basin eco-environmental protection and air pollution control over broader areas. This study showed that the distributed modelling of hydrological and biogeochemical processes is crucial to understand the spatiotemporal variability in nonpoint source nutrient loads over the watershed and is helpful to identifying key factors affecting river water quality.