The analysis of nitrogen load and simulation uncertainty using SWAT in a catchment with paddy field in China.

Excessive load of nitrogen from anthropogenic sources is a threat to sustaining a healthy aquatic ecosystem. The difficulty in identifying the critical source areas (CSAs) of nitrogen load and apportioning the in-stream nitrogen to individual sources spatially and seasonally has made the Soil and Water Assessment Tool (SWAT) useful for analyzing nitrogen load at the catchment scale. However, the uncertainty of the nitrogen load simulated by SWAT has rarely been analyzed. The two simulations with the highest or the lowest PBIAS of total nitrogen (TN) load were proposed in this study to represent the range of the prediction uncertainty and therefore were used to generate the uncertainty of CSAs and nitrogen source apportionment. The model was set up for the Yuan River Catchment, which is under threat of extensive nitrogen load. Results indicated the highest nitrogen load was from downstream paddy fields with a denser population and 85% of the load was from fertilizer and feedlots. The relatively high prediction uncertainty was observed on both CSAs and source apportionment, which emphasizes the limitation of calibration only based on certain processes and the necessity to consider parameter uncertainty in the application of nitrogen load simulation.