Abstract
Modeling fluid mud in estuaries and coastal areas is a complicated task due to its non-Newtonian characteristics. This study adopts a continuous modeling approach to investigate the laminar flow of fluid mud by introducing a three-dimensional (3D) form of the Herschel–Bulkley rheological model into the Finite Volume Coastal Ocean Model (FVCOM). The model was validated against flume experiments, demonstrating its ability to accurately reflect velocity and density distributions from the near-bottom to the upper water column. Comparisons between the 3D rheological model and a 1DV model reveal significant differences, highlighting the critical role of horizontal apparent viscosity in simulating mud flow through complex geometries.
Introduction
Fluid mud is a high-concentration suspension of cohesive sediment commonly found in estuaries and river deltas. Unlike standard seawater, it is non-Newtonian, meaning its viscosity changes based on the shear rate. Historically, coastal models have treated fluid mud as a distinct layer or used 1DV approaches that only consider vertical shear. However, real-world estuaries involve complex lateral spreading where horizontal viscosity becomes significant. This paper addresses the lack of 3D rheological accounting in current hydrodynamic models, which often leads to inaccuracies in predicting how mud flows into wider basins or around complex coastal features.
Methodology
The authors built their model on FVCOM, which utilizes unstructured triangular grids to fit complex coastlines. The core innovation is the implementation of a 3D Herschel–Bulkley rheological model that incorporates the full 3D strain rate tensor. This allows the model to calculate apparent viscosity based on velocity gradients in all directions ($x, y,$ and $z$).
To handle depth variations, the model employs a $\sigma$ vertical coordinate system. This spatial organization supports a "continuous modeling approach," where the transition from clear water to dense mud is represented as a smooth gradient rather than a hard interface.
For horizontal calculations, a second-order spatial reconstruction method is used for partial derivatives on the triangular grids. This ensures that horizontal shear forces—often ignored in simpler models—are accurately captured as the fluid mud spreads laterally.
Validation and Results
The model was validated against two laboratory datasets: gravity currents on a slope (van Kessel and Kranenburg, 1996) and viscous propagation on a flat bottom (Chowdhury and Testik, 2012). In both cases, the model's predictions for the mud front position and vertical profiles of density and velocity matched experimental data with high precision.
A key insight was gained from a synthetic "Flume-Basin" test case. The 1DV model, which ignores horizontal shear, predicted a faster-moving mud front. In contrast, the 3D model showed a slower, more realistic propagation because it accounted for the energy lost to horizontal viscous forces as the mud entered the wider basin. The 3D model also produced a more realistic rounded "head" for the mud flow.
Conclusion
Integrating a 3D rheological model into a hydrodynamic framework like FVCOM significantly improves the accuracy of fluid mud simulations in complex coastal environments. By accounting for horizontal velocity shear and treating the water-mud system as a continuous fluid, the model avoids the instabilities of layered approaches. Future work will focus on incorporating turbulent flow models and sediment thixotropy to further refine predictions in dynamic estuarine settings.
Summary updated