High-resolution simulation of free-surface flow and tracer retention over streambeds with ripples

Abstract

This study presents a novel high-resolution simulation of free-surface flow and tracer retention over a streambed with ripples based on varying ripple morphologies, surface hydraulics and the transport of a tracer pulse from surface water to surface dead zone. For the simulations, the computational fluid dynamics (CFD) model OpenFOAM was used to solve the threedimensional Navier-Stokes equations in combination with an implemented transport equation. Pressure gradients at the streambed were used to account for hyporheic exchange, assuming water flow from high pressure zones to low pressure zones. Flow velocities, ripple sizes and spacing showed to significantly affect these pressure gradients, but also the transport of a passive tracer at the streambed, which was not investigated so far. Due to the velocity field, large parts of the tracer mass were transported alongside the main stream above the ripples. Tracer mass reaching the space between the ripples was temporarily retained due to low velocities and recirculations. It was shown that the retention is depending on the ripple size and space between the ripples as well as on the flow velocity. Decreasing ripple sizes and higher flow velocities lead to a smaller tracer retention. Furthermore we showed that the ripple length to height ratio controls the generation of recirculation zones which affect the residence time of the tracer significantly. Ripple spacing leads to temporarily higher tracer concentration at the streambed, but smaller tracer retention. We conclude that the impact of the streambed morphology on the hydraulics in combination with tracer retention should be addressed for a comprehensive understanding of compound movement, exchange and transformation within the hyporheic zone.

Effect of Surface Hydraulics and Salmon Redd Size on Redd‐Induced Hyporheic Exchange

“…Simulation performance was quantified with Nash‐Sutcliffe coefficients (NSC) whose values indicate the quality of the model performance: very good (NSC > 0.75), good (0.65 < NSC ≤ 0.75), satisfactory (0.5 < NSC ≤ 0.65), and unsatisfactory (NSC ≤ 0.5) (Moriasi et al., 2007). The accuracy of the comparison by visual inspection between predicted and measured pressure distributions (Figure 3a) is comparable to that reported in the literature (Broecker et al., 2018; Cardenas & Wilson, 2007a; Reeder et al., 2018). The NSC of 0.7 and 0.6 for flows with velocities of 0.29 and 0.44 m/s, respectively, also support the visual inspection.…”

“…These results are consistent with observations from other flume experiments (Huettel & Gust, 1992), which have reported an adverse pressure gradient near the pressure dip caused by eddy detachment. Cardenas and Wilson (2007a) have also used Fehlman's experiments for validation and have recognized the difficulty in capturing the crest as a singularity with an adverse pressure gradient both numerically and experimentally (Broecker et al., 2018; Cardenas & Wilson, 2007a; Reeder et al., 2018). The overall performance of our simulation is comparable to the published works.…”

The role of bed-penetrating Kelvin–Helmholtz vortices on local and instantaneous bedload sediment transport

“…2012; Broecker et al. 2018; Zgheib & Balachandar 2019). We should note here that some EE simulations (Cheng, Hsu & Chauchat 2018) consider the sediment–fluid interface to be the location where the sediment bed attains its maximum packing fraction.…”

“…A non-zero fluid velocity at the bed surface has important implications on EE simulations, since most EE simulations use no-slip and no-penetration conditions (Janssen et al 2012;Broecker et al 2018;Zgheib & Balachandar 2019). We should note here that some EE simulations (Cheng, Hsu & Chauchat 2018) consider the sediment-fluid interface to be the location where the sediment bed attains its maximum packing fraction.…”

Integral Flow Modelling Approach for Surface Water-Groundwater Interactions along a Rippled Streambed

“…This solver was developed by Oxtoby et al [34] and is based The mesh has been discretized using the three-dimensional finite element mesh generator gmsh. The ripple parameters are based on the approach of Broecker et al [35]. Table 1 summarizes the most important ripple parameter values.…”