In river systems, transported suspended sediment interacts with in-channel and riparian vegetation. The involved physical processes are complex and still poorly reproduced by numerical models. This study compares the performance of background horizontal eddy viscosity models in reproducing flow fields and suspended sediment transport processes inside partially vegetated flows. Particularly, we selected an experiment with a partly-vegetated flume which we numerically replicated with the Delft 3D-2DH model. We applied three existing horizontal eddy viscosity approaches: constant value, Elder model and hybrid model. Besides, we represented the vegetation by using Baptist formulation. The results show that the Elder viscosity model reproduces the explicit development of vortices along the flume, while all three viscosity models show an underestimate of the sediment deposition in the vegetated area and an overestimation of sediment deposition in the non-vegetated area. Further investigation is needed to reproduce the experimental introduction of the sediment and to link transversal suspended sediment dynamics with the resolved dynamics in numerical models.

The presence of instream and riparian vegetation significantly affects the flow field of rivers, which in turn impacts sediment transport (Vargas-Luna et al., 2016, Calvani et al., 2023). Few studies, however, have investigated the interaction between horizontal flow structure and suspended sediment transport. Hamidifar (2019) conducted a flume experiment investigating the flow field structure in a halfvegetated channel configuration and found strong horizontal vortices. Other experimental works observed the strong transverse sediment transport process along the interface area between vegetation and open channel (Box et al., 2018; Xu et al., 2022). As for numerical simulations, few studies focus on the reproduction of the physical processes between the partially channel vegetated flow and suspended solids. The present simulation studies mainly focus on the small-scale large eddy simulation model, which has not been applied in practice (Wang et al. 2021). This study focuses on the effect of horizontal viscosity models on the simulation of suspended solid transport in a partly vegetated channel in the most applied Reynolds-averaged Navier-Stokes model. Three viscosity models: the constant model, the Elder model, and the Hybrid viscosity model are applied to reproduce an experimental work selected from the literature.

Baptist’s method, Drag Force and Single-Stem approaches are the commonly used tools implemented in Delft3D to model water and sediment transport processes in vegetated channels. Despite their wide application, the model reliability has seldom been tested against data of controlled flume experiments with solid suspension. Here, we investigate the ability to reproduce suspended sediment transport through emergent vegetation by comparing the results of 2D simulations to existing experimental data. The results show that in low vegetation density, the Baptist and Drag Force approaches are not sensitive enough to density variations. The Single-Stem approach reproduces detailed flow structure and sediment deposition around stems, but its high computational time is a limitation for long-term simulations or dense vegetation. Furthermore, we observed that the simplification of 2D depth-averaged models and the non-equilibrium of sediment transport in both experiments and numerical simulations may also affect the overall performance of the vegetated modelling approaches.

Working as natural filter, well-designed vegetation schemes have been widely applied to improve the quality of water (Aiona, 2013; Stefanakis, 2015). Proper design, however, requires appropriate physics-based modelling of their filtering capacity. Several theoretical models predicting sediment transport in vegetated flow have been proposed: Baptist (2005); Yang and Nepf (2018); Wu et. al. (2021); Tseng and Tinoco (2021); Yagci and Strom (2022); Wang et. al. (2023). Some of them have been implemented in numerical tools (e.g. Caponi et al., 2022; Li et al., 2022) and in particular in Delft 3D (Deltares, 2014). However, they have been mostly designed and verified based on bedload processes, and their performance for suspended load should be further investigated.
This work compares different approaches on their ability to reproduce the effects of vegetation on suspended solids concentration in two-dimensional models built in Delft3D. The work focuses on emerging vegetation, represented as rigid cylinders, and sediment deposition. Comparisons are based on the ability to reproduce flume experiments available in the literature by analysing both flow field and sediment deposition results.