New framework for optimizing best management practices at multiple scales

Abstract The optimal design of best management practices (BMPs) is of vital importance for the control of nonpoint source (NPS) pollutants, but proper methods for multiple-scale BMP design are still lacking, especially for large-scale watersheds. In this study, a new framework (MS-BMPs) was proposed for multiscale BMP design by combining multiscale design ideas, hydrological models, cost estimators and shuffled frog leaping algorithms. In addition, the computation efficiency of BMP design was improved by modifying the memetic evolution process. This new framework was then tested in the Daning River Watershed, Three Gorges Reservoir Region, China. Based on the results, the new framework could provide more powerful convergence and optimization ability, while the rationality of BMP configuration was increased by four times. From the Pareto front of the MS-BMPs, the maximum removal of TN and TP improved by 19.42% and 14.36%, indicating a more cost-effective BMP design. This study also highlighted the priority of fertilization management in the regulation of NPS pollution. This new framework can be easily extended to any watershed to assist managers with the optimal design of BMPs.