Vertical Distribution Rule Analysis and Simulation of Suspended Sediment Concentration in Hangzhou Bay

王繁,周斌,徐建明,宋立松,王新
DOI: https://doi.org/10.3785/j.issn.1008-973x.2008.11.033
2008-01-01
Abstract:Aimed at monitoring the suspended sediment concentration(SSC) in water column by remote sensing,a back propagation(BP) model of artificial neural networks was applied to simulate the vertical concentration distribution of suspended sediment in Ganpu of Hangzhou bay.When the gravity function of sediment particles was ignored,the complex movement and diffusion mechanism of high-energy current were approximated into the dynamics of different tide situations which could be represented by some macroscopical variables.Based on the correlation analysis of vertical SSC,the SSC of surface water(0.2H) and the tide dynamic were selected as main input factors to simulate the vertical average SSC.Results showed that the dynamical characteristic of tide force influences SSC remarkably,which fluctuates according to the change of tide flood and ebb.The vertical distributions of SSC have observable correlation and the coefficients between the SSC at 0.2H and the vertical average SSC all attain 0.8 in total tide situations.The BP model has precise prediction capability,whose performance index(PI) is 0.04,better than those of two linear regression models,which are 0.06 and 0.07,respectively.BP model can provide a simple and convenient method for monitoring the undersurface SSC by remote sensing.
What problem does this paper attempt to address?