Abstract
Suspended sediments, as one of the most important factors affecting the water environment of inland lakes, are closely related to the various pollutants migration and interaction. Thus, the simulation and prediction of suspended sediment concentration is important. Existing studies show that the prediction accuracy of suspended sediment concentration can be effective predicted based on assimilation methods coupled with hydrodynamic models. However, in the process of assimilation of hydrological simulation, the existing perturbation generation methods consider that the perturbation error is a random Gaussian distribution, which does not consider the spatial variation characteristics of errors. In this paper, a new method to generate the perturbation field for assimilation simulation was proposed. This method uses hybrid error to generate perturbation field for assimilation simulation instead of using random error. The proposed approach was validated through its application to assimilation simulation of suspended sediment concentration of Taihu Lake in China, and five assimilation experiments was conducted. The proposed method was compared with existing methods for perturbation field generation. After three days and 72 time steps of assimilation simulation based on hybrid perturbation generation, we found that the proposed assimilation method provided results that were more consistent with buoy-measured data. The accuracy of the two assimilation methods based on hybrid perturbation is improved. Compared with the assimilation method based on random perturbation, the mean values of RMSE(root mean square error) decreased from 9.56 to 8.70 and from 9.55 to 8.68, respectively. The results show that the proposed hybrid perturbation generation method has higher simulation accuracy than other methods. This study shows that the method is effective and provides a new idea for the assimilation simulation of suspended sediment concentration in inland lakes.
An Assimilation Simulation Approach for the Suspended Sediment Concentration in Inland Lakes Using a Hybrid Perturbation Generation Method
