Abstract:The water body's suspended concentration reflects many coastal environmental indicators, which is important for predicting ecological hazards. The modeling of any concentration in water requires solving the settling-diffusion equation (SDE), and the values of several key input parameters therein (settling velocity ws , eddy diffusivity Ds , and erosion rates p(t) ) directly determine the prediction performance. The time-consuming large-scale simulations would benefit if the parameter values could be estimated through available observations in the target sea area. The present work proposes a new optimization method for synchronously estimating the three parameters from limited concentration observations. First, an analytical solution to the one-dimensional vertical (1DV) SDE for suspended concentrations in an unsteady scenario is derived. Second, the near bottom suspended sediment concentration (SSC) profiles are measured with high-resolution observation. Third, the key parameters are optimized through the best fit of the measured SSC profiles and those modeled with the unsteady solution. Nonlinear least square fitting (NLSF) is introduced to judge the best fits automatically. The high-resolution concentration measurements in a specially-designed cylindrical tank experiment using the Yellow River Delta sediments test the proposed method. The method performs well in the initial period of turbulence generation when sediment resuspension is significant. It optimizes p(t) , ws , and Ds with reasonable values and uniqueness of their combination. The proposed theory is a practical tool for quickly estimating key substance transport parameters from limited observations; it also has the potential to construct local parametric models to benefit the 3D modeling of coastal substance transport. Although the present work takes SSC as an example, it can be extended to any suspended particulate concentration in the water.

Optimization of suspended particulate transport parameters from measured concentration profiles with a new analytical model