Spatial optimization of best management practices (BMPs) is an effective way to select and allocate BMPs for watershed management such as soil and water conservation and nonpoint source pollution reduction. The commonly used spatial units for BMP configuration (or BMP configuration units) include subbasins, hydrologic response units (HRUs), farms, and fields. Normally, these spatial units are not homogeneous functional units from the perspective of physical geography at the hillslope scale (in terms of geomorphic and hydrologic conditions of the hillslope, for example), and thus cannot effectively represent the spatial relationships between BMPs and spatial locations with respect to hillslope processes from upstream to downstream. This makes it difficult to efficiently and rationally construct spatial optimizations for watershed BMPs. This paper proposes a spatial BMP optimization approach based on slope position units, which are homogeneous spatial units with physical geographic features. In the proposed approach, slope position units are used as BMP configuration units by which the relationships between BMPs and slope positions along a hillslope can be explicitly considered during BMP scenario initialization and optimization via genetic algorithm (i.e., NSGA-II). A distributed and physically based watershed model was used to evaluate the environmental effectiveness (i.e., the reduction rate of soil erosion), and a simple estimation method was developed to calculate the net cost of BMP scenarios. A case study was conducted in a small hilly watershed in the typical red-soil region of the Fujian Province in southeastern China, which suffers severely from soil erosion. A simple system of three types of slope positions (i.e., ridge, backslope, and valley) was used to delineate BMP configuration units. Four BMPs that are used in actual Chinese red-soil regions (closing measures, arbor-bush-herb mixed plantation, low-quality forest improvement, and orchard improvement) were considered in the proposed approach to achieve the multiple optimization objectives, which included maximizing the reduction ratio of soil erosion and minimizing the net cost of the BMP scenario. The proposed approach was compared with the standard random optimization approach, which selects and allocates BMPs randomly on BMP configuration units. The results show that the proposed approach is more effective and efficient for proposing practical and effective BMP scenarios than the random approach.
Read full abstract