Hongwu Tang

Hydrodynamics and contaminant transport on a degraded bed at a 90-degree channel confluence

Channel confluences at which two channels merge have an important effect on momentum exchange and contaminant diffusion in both natural rivers and artificial canals. In this study, a three-dimensional numerical model, which is based on the Reynolds Averaged Navier–Stokes equations and Reynolds Stress Turbulence model, is applied to simulate and compare flow patterns and contaminant transport processes for different bed morphologies. The results clearly show that the distribution of contaminant concentrations is mainly controlled by the shear layer and two counter-rotating helical cells, which in turn are affected by the discharge ratio and the bed morphology. As the discharge ratio increases, the shear flow moves to the outer bank and the counter-clockwise tributary helical cell caused by flow deflection is enlarged, leading the mixing happens near the outer bank and the mixing layer distorted. The bed morphology can induce shrinkage of the separation zone and increase of the clockwise main channel helical cell, which is initiated by the interaction between the tributary helical cell and the main channel flow and strengthened by the deep scour hole. The bed morphology can also affect the distortion direction of the mixing layer. Both a large discharge ratio and the bed morphology could lead to an increase in mixing intensity.

Response to: “Discussion of: hydrodynamics and contaminant transport on the degraded bed at a 90-degree channel confluence” by Seastien Pouchoulin, Pedro Xavier Ramos, Emmanuel Mignot, et al.

The authors would like to thank the discussers for their interests and useful discussions regarding the paper. The main point raised in the discussion concerns the rotated direction of the helical cell in scenario S3/S7. The discussers find that the rotated direction of the helical cell in the literatures are opposite to that in this study. And in the literature, the rotated direction of the helical cell is clockwise, with the water from the tributary channel flowing towards the outer bank in the near-surface layer, then dropping along the opposing wall, then returning towards the inner bank along the near-bed layer, and finally closing the loop. It is important to note that there is a pair of counter-rotating helical cells in this study. As mentioned in the second part of Sect. 3.1 in the paper, one of them is originated from the deflection of the tributary flow, it locates in the upper part of the water column and rotates anti-clockwise. The other is initiated by the interaction between the tributary helical cell and the main channel flow, it locates in the near-bed layer with clockwise rotated direction. This counter-rotating helical cells also occur in the work of Schindfessel et al. [1] (as shown in their Fig. 4) and Riviere et al. [2] (as shown in their Fig. 7), the width to depth ratio in their works are nearly 2.5 (same as this study). Whereas, there is only one helical cell with clockwise rotated direction (denoted as classical rotation direction in the discussion) occurs in the work of Weber et al. [3], Huang et al. [4], Shakibainia et al. [5] and Yang et al. [6], the width to depth ratios are nearly 3. Then the width to depth ratio should be responsible for the different flow patterns in these studies. The secondary currents is more complex as the width to depth ratio is smaller. Due to the different formation reasons of the counter-rotating helical cells, the size of the helical cells are affected by the flow condition, bed morphology and channel geometry at channel confluence. The flow condition and channel geometry are different with each other among the

Simulation of Natural River Flow by a Three-Dimensional Hydrodynamic Model

A fully calibrated three-dimensional hydrodynamic model, namely Curvilinear Hydrodynamics in 3-Dimensions (CH3D), has been developed to determine changes of flow velocity in a natural river. The objective of this paper is to present the application of the CH3D modeling. A grid scheme was constructed by using SMS software along an example river-the Detroit River from Lake St. Clair to Lake Erie with two natural curved banks as the transversal boundaries. The water depth (vertical ?-) was divided into several layers from water surface to the river bottom. A modification to the program was made by the authors to enable flexible Manning’s roughness by applying the Strickler’s formula. The National Oceanic and Atmospheric Administration (NOAA, USA) bathymetry data were post-processed for water level corrections by using MapInfo software. The data were also extended to areas where there were no NOAA measurements in the Detroit River basin. Modeling results show that computed velocities and velocity measurements at various river cross-sections are in a good agreement typically below 10% error. Comparison between the computational water surface elevations versus that of records in gauge stations along the Detroit River is also in a good agreement.