A methodology for estimating the parameters in three-dimensional cohesive sediment transport models by assimilating in situ observations with the adjoint method

The model parameters in the suspended cohesive sediment transport model are quite important for the accurate simulation of suspended sediment concentrations (SSCs). Based on a three-dimensional cohesive sediment transport model and its adjoint model, the in situ observed SSCs at four stations are assimilated to simulate the SSCs and to estimate the parameters in Hangzhou Bay in China. Numerical experimental results show that the adjoint method can efficiently improve the simulation results, which can benefit the prediction of SSCs. The time series of the modeled SSCs present a clear semidiurnal variation, in which the maximal SSCs occur during the flood tide and near the high water level due to the large current speeds. Sensitivity experiments prove that the estimated results of the settling velocity and resuspension rate, especially the temporal variations, are robust to the model settings. The temporal variations of the estimated settling velocity are negatively correlated with the tidal elevation. The main reason is that the mean size of the suspended sediments can be reduced during the flood tide, which consequently decreases the settling velocity according to Stokes's law, and it is opposite in the ebb tide. The temporal variations of the estimated resuspension rate and the current speeds have a significantly positive correlation, which accords with the dynamics of the resuspension rate. The temporal variations of the settling velocity and resuspension rate are reasonable from the viewpoint of physics, indicating the adjoint method can be an effective tool for estimating the parameters in the sediment transport models.