非点源污染模拟的空间分割优化

Optimization of watershed subdivisions in nonpoint sources pollution simulations

以水土评价模型SWAT在芦溪小流域的应用为例,讨论不同支持数据下模型计算结果与空间分割响应关系。结果表明,空间分割的持续细化并不能够持续地改善计算结果。利用地统计学工具对不同支持数据下研究区域的空间变异进行定量评价,并与SWAT模型输出空间分割的响应曲线对照,可以得出:研究区域的空间自相关尺度是确定合理的空间分割程度的一个关键依据。合理的子流域面积特征尺度A应当在空间自相关尺度L的1/2～1/3之间。

A SWAT model of the Luxi watershed was used as an example with various levels of support data to evaluate the model results as a function of watershed subdivision size. The results show that the outputs are not always improved by finer subdivisions. Geostatistics tools were used to assess the spatial heterogeneity of the study areas with two sets of support data. Comparison of the model output to the watershed subdivisions shows that the correlation length, L, of the study area is the key factor determining how the watershed should be partitioned. A reasonable value of A, where A is the average areas of the sub-basins, should be larger than L/3 but less than L/2.