论文部分内容阅读
基于单位面积负荷法和洛杉矶巴洛纳河上游流域土壤特性的Browne经验关系式,利用GIS平台应用了一个暴雨径流预测模型。通过把流域划分成多个子流域并对每个子流域应用集总参数,使流域的非均匀性得以量化。通过零阶正则化和采用边界约束有限内存BFGS(L-BFGS-B)优化算法,来表征特定土地利用类型产生的污染物负荷占总负荷的比重。以总Zn为例,模型预测的污染物负荷与在流域出口排放点实测的污染物负荷一致。使用开发的水量模型预测的径流量与实测数据吻合良好,相关系数R2达0.86。水质模型的均方根误差仅为9kg,低于每场暴雨77kg的平均排放量。
Based on the Browne empirical relationship between unit area loading and soil properties in the upper reaches of the Balona Basin in Los Angeles, a stormwater runoff forecasting model was applied using the GIS platform. The watershed heterogeneity is quantified by dividing the basin into multiple sub-basins and applying lumped parameters to each sub-basin. The zero-order regularization and Boundary Constrained Finite-Memory BFGS (L-BFGS-B) optimization algorithm are used to characterize the proportion of pollutant load to total load generated by a particular type of land use. Taking total Zn as an example, the pollutant load predicted by the model is consistent with the measured pollutant load at the discharge point at the catchment. The predicted runoff using the developed water model is in good agreement with the measured data, with a correlation coefficient R2 of 0.86. The root mean square error of the water quality model was only 9 kg, less than the average of 77 kg per rainstorm.