Estimation and interpretation of spatially varying bottom friction coefficients in Bohai Bay using A-4DEnVar

IntroductionThe bottom friction coefficient (BFC) is critical for numerical simulations of tides and coastal ocean engineering. In situ observations indicate that BFC is spatially variable; however, it is challenging to represent it accurately in tidal models.MethodsIn this study, two principal tidal constituents (M2 and K1) were synchronously simulated in Bohai Bay using the Princeton Ocean Model with a generalized coordinate system. The tidal harmonic constants obtained from the NAO.99jb tidal harmonic constants dataset and tidal gauge stations are assimilated into the tidal model. An analytical four-dimensional ensemble variational data assimilation scheme is used to jointly estimate the open boundary conditions, BFCs, and bathymetry of the tidal model. Additionally, a random forest approach is applied to quantify the relative contributions of form drag, suspended sediment concentration, tidal current speed, and water depth to the spatial distribution of BFC in Bohai Bay.ResultsThe tidal model performance improved after data assimilation, and the multi-parameter joint optimization scheme outperformed optimizing BFC alone. The spatial distributions of BFCs estimated through different observations were nearly consistent and robust, as verified by sensitivity experiments. The model explains approximately 76% of the BFC spatial variance, with suspended sediment concentration and form drag identified as the dominant controlling factors. The relationships between BFC and each influencing factor, as derived from partial dependence analysis, are consistent with the conclusions reported in existing studies.DiscussionThe results are beneficial for determining accurate spatially varying BFCs in the tidal model of Bohai Bay.