气候变化领域集成服务门户

The Soil and Water Assessment Tool (SWAT) has been widely used to estimate pollutant losses from various agricultural management practices. Although many studies have shown good performance in simulating total nitrogen (TN) and dissolved nitrogen (N), the model performed poorly in many other applications, particularly on dissolved N. Poor performance on dissolved N could be attributed to landscape (in-field) processes and/or in-stream N processes in the model. Therefore, the overall goal of this study was to evaluate SWAT N simulations with in-stream processes and without in-stream processes. Sensitivity analysis results showed that when in-stream processes were not simulated, denitrification threshold water content (SDNCO), N in rainfall (RCN) and N percolation coefficient (NPERCO) were the most sensitive parameters to dissolved N losses. However, when in-stream processes were simulated, the most sensitive parameters changed to initial organic N concentration in soil layers (SOLORGN) and organic N enrichment ratio (ERORGN); and the impact of SDNCO, RCN, and NPERCO was greatly decreased. Furthermore, fertilizer timing and amount had little impact on N simulations. SWAT under-estimated dissolved N, but over-estimated organic N and TN. Further calibration could improve the simulation of dissolved N, but would degrade the simulations of organic N and TN.