The application of the Bp neutral network in the estimation of non-point source pollution
暂无分享,去创建一个
Abstract In view of the relatively large basin area in China and the application limitations of small-scale, non-point source (NPS) pollution models, we developed an estimation platform for NPS based on a back propagation (Bp) network with the combination of Visual Basic 6.0 (Microsoft Corporation) and Supermap Objects 2.1 (Supermap Corporation). In addition, the total nitrogen (TN) load and its spatial characteristics in the Songliao Basin were estimated and analyzed with our estimation platform. Analysis of the estimation platform demonstrated that it is well suited to nonlinear mapping and spatial analysis. The estimation of NPS in the Songliao Basin revealed that the TN load generally decreased in the basin from 1980 to 2000. In addition, the TN load in the Songliao Basin was characterized by being significantly influenced by loads in neighboring zones. Concerning administrative areas, the TN load in Jilin Province was the highest in 1985 and 1995, while it was highest in the Inner Mongolia autonomous region in 2000. Concerning water systems, the Second Songhuajiang Basin had the highest nitrogen load in 1985 and 1995, while the Daliao Basin had the highest nitrogen load per unit area. Similarly, the Eerguna River Basin had the highest nitrogen load and nitrogen load per unit area in 2000. Finally, the applicability of the NPS platform in small- and medium-sized basins of semi-arid regions in the northwestern, such as the Weihe River Basin, was studied
[1] Zhenli He,et al. Water quality prediction of marine recreational beaches receiving watershed baseflow and stormwater runoff in southern California, USA. , 2008, Water research.
[2] L. Ermini,et al. Artificial Neural Networks applied to landslide susceptibility assessment , 2005 .
[3] M. Hosomi,et al. Novel application of a back-propagation artificial neural network model formulated to predict algal bloom , 1997 .