Longitudinal dispersion affected by willow patches of low areal coverage

Abstract

Vegetation notably influences transport and mixing processes and can thus be used for controlling the fate of substances in the hydro-environment. Whilst most work covers fully vegetated conditions, the novelty of this paper is to focus on flows with real-scale flexible willow patches. We aimed to investigate how longitudinal dispersion varies according to the spatial distribution, density and coverage of the patches and to evaluate the explanatory power of predictors that consider the hydraulics, vegetation and channel geometry. Salt tracer experiments were performed in a trapezoidal channel where we established 3-4 m long and 1-1.6 m wide patches of artificial foliated willows that reproduced the shapes and plant densities observed on woody-vegetated floodplains. We examined sparsely distributed patches with low areal/volumetric coverage of 6-11%, and non-vegetated conditions for reference.Flow depths and surface widths were 0.7-0.9 and 6-7 m, respectively, and the mean flow velocities ranged at 0.3-0.6 m/s. The emergent patches generated from a negligible to over a four-fold increase in the longitudinal dispersion when compared with non-vegetated conditions. The patches with a preferential location in low-velocity areas, such as near banks, or with a high plant density and a blockage of the crosssectional flow area ⪆0.4, led to the largest dispersion and residence times. Patches under such configurations enhanced the normalized differential velocity defined as the difference between the highest (90th percentile) and lowest (10th percentile) cross-sectional flow velocities divided by the mean velocity, thus increasing shear dispersion. As existing analytical predictors failed to estimate the effect of different patch configurations, we proposed the change in the normalized differential velocity between vegetated and corresponding non-vegetated conditions as a basic predictor of the reach-scale longitudinal dispersion coefficient under patchy vegetation. In contrast, we observed no clear relationship between flow resistance and dispersion. Thus, our findings indicated that bankside vegetation may allow for reduced peak concentrations and lengthened residence times, supporting pollutant management, while ensuring good flow conveyance. Such rare field-scale analyses improve the estimation of solute transport in real vegetated flows.

Longitudinal dispersive coefficient in channels with aquatic vegetation: A review

“…However, the study of Vastila [41] showed that p I had noticeably poorer predictive power than that observed in the sparse vegetated patches condition of Park and Hwang [43] , with 1-3 orders magnitude lower value. The results indicated that both patch density and location control dispersion.…”

“…When vegetated patches grow in channels, they create stem-scale vortexes and patch-scale vortexes that impact dispersion [92] . The highest dispersive fractions of vegetated patches are typically larger than those of randomly distribution stems [41,93] . Two problems have drawn the attention of researchers.…”

“…The predictor of Perucca et al [73] indicated that low-coverage patches can lead to a maximum of three times larger impact on reachscale dispersion than homogeneous vegetation on floodplains, which can increase x k of compound channels by 1.5-fold. Dang et al [91] , Vastila et al [41] both illustrated a positive relationship between x k and the coverage of vegetation patches based on numerical and experimental results, respectively. To some extent, the increase in the vegetation coverage indicates a higher canopy density, and the relationship between dispersion and vegetation density can be predicted.…”

“…When considering the scenario with vegetation patches in the compound channels, the location of patches compared with the fully covered vegetation is also a crucial factor. Patches located in low-velocity areas, such as near the river bank, usually generate larger dispersion and longer residence time [41] . This phenomenon may be due to the fact that vegetated regions and the areas between longitudinally subsequent wake flows contribute to the dispersion process.…”

“…For example, the velocities in the vegetated regions are constricted by the drag force of canopies, leading to the regions with low or no flow, termed dead zones, which remarkably increases the resistance time of fine particles and contaminants in the river, floodplains, and wetlands [39] . Considering that the flexible plants in natural rivers are more common and grow in community patterns [40] , many studies have focused on the dispersion coefficient with plant clumps or flexible canopies [41][42] . However, these studies all need to deal with the bottleneck that is how to describe and quantify the vegetated arrangement and flexibility.…”