Longitudinal dispersion affected by willow patches of low areal coverage

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 cross-sectional 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.