Population growth, urbanization, and intensified agriculture have resulted in mobilization of nitrogen and phosphorus, which is the main cause of river water quality deterioration. Environmental regulation has expedited the necessity for agricultural producers to design and implement more environmentally suitable practices. Therefore, there is a need to identify critical nutrients and their loss/transport potential. Watershed model can be used to better understand the relationship between land use activities/management and hydrologic processes/water quality changes that occur within a watershed. The objective of the study is to test the performance of the SWAT model and the feasibility of using this model as a simulator of water flow and nitrogen and phosphorus yields over the Dongjiang River basin in South China.

Spatial data layers of land slope, soil type, and land use were combined with geographic information system (GIS) to aid in creating model inputs. The observed streamflow and sediment at Boluo station in the Dongjiang River basin were used to calibrate and validate the model. Time series plots and statistical measures were used to verify model predictions. Predicted values generally matched well with the observed values during calibration and validation (R²≥0.6 and Nash-Suttcliffe Efficiency ≥0.5) except for underestimation of sediment peaks and overestimation of sediment valleys at Boluo. This study shows that SWAT is able to predict streamflow, sediment generation, and nutrients transport with satisfactory results.