Abstract:
In this study, a watershed model SWAT-MT was developed on the basis of SWAT to simulate the non-point source pollution of heavy metals. In the development of SWAT-MT, we refer to the export coefficient model and use a dissolution rate to describe the leaching effect of heavy metals in surface runoff. The new model retains the advantages of both the export coefficient model and the process-based mechanism model, and can be applied in basins with poor data conditions. The SWAT-MT was used to evaluate the non-point source lead load in a subbasin of Xiangjiang River in Zhuzhou City, Hunan Province, and the sensitivity analysis and uncertainty analysis were also carried out. The simulated erosion load of lead is 1.86 t/a and the erosion intensity is 4.70 kg/km
2/a. The erosion intensity has significant spatial heterogeneity, and the erosion intensity in urban areas is as high as 2.23 times of the average level. Therefore, priority should be given to the control of urban non-point source pollution. The simulated lead load at the watershed outlet is 0.70 t/a, which is approximately 38% of the surface erosion load. Monte Carlo simulation indicates that there is significant uncertainty in the model output. Under 95% confidence interval, the erosion of lead is 1.39~2.31 t/a, and lead load at the watershed outlet is 0.43~1.00 t/a. Sensitivity analysis indicates that in order to reduce the uncertainty and improve the reliability of the model, modelers should pay attention to the assignment of three parameters, namely, the lead concentration in urban surface sediments, the dissolution rate of lead and the settling rate of lead in streamflow. The observation data in this study are very limited. The calibration of water and sediment simulation and the determination of heavy metal parameters were carried out with reference to the observation data and literature research of similar watersheds. The modeling paradigm in this study can provide a reference for the non-point source simulation of heavy metals in basins with poor data conditions.