The researchers used a model called SWAT and GIS technology to study pollution in the Qingjiang River Basin. They found that input parameters play a big role in accurately predicting runoff, sediment, and nutrient levels in the watershed. By analyzing sensitive parameters, they were able to calibrate the model and validate it with real data. The study revealed that non-point source pollution is highest during the flood season, with critical areas of soil erosion identified. The simulation showed that average annual runoff and nutrient levels varied based on factors like land use and rainfall.