Best management practices for agricultural non-point source pollution in a small watershed based on the AnnAGNPS model

Best management practices (BMPs) are implemented to reduce non-point source pollution (NPSP) worldwide. In this study, the AnnAGNPS model was adapted and used to evaluate the effectiveness of BMPs for NPSP control in the Three Gorges Reservoir Area in China. Through field investigation and experimental observation, the AnnAGNPS model database was constructed. The model best simulated run-off and its simulation of total nitrogen (TN) was better than for total phosphorus (TP). Simulation of the spatial distribution of different pollutant loads showed that the annual outputs of TN and TP per unit area were the highest, with an average of 63 and 6 kg hm^?2 per year, respectively. An optimal scenario was developed based on adjustments to the existing land structure which were most effective at reducing N and P loads. The reduction rates of TN and TP loads were 54% and 63%, respectively, while the reduction rates of dissolved and particulate N and P loads were between 47% and 68%. After 27% of arid slope fields were converted into terraces in the study area, the modelled particulate N and P outputs were reduced by 44% and 46%, respectively, showing that terracing sloping land could be effective at reducing particulate N and P losses. These findings indicated that slope-shifting technology is the most important method for NPSP control in the TGRA. In addition, reduced fertilizer use and conservation tillage measures should also be adopted to effectively reduce TN and TP losses.