Search

Mixing between water column and sediment is a crucial physical process for the fate of nutrients and for the biogeochemical reactions in the stream ecosystem. There are two factors that characterize this mixing process: the mixing magnitude at the sediment water interface, which controls the amount of solutes exchange between water column and its underlying sediment layer; and the mixing extent that defines how quickly the mixing rate decays with depth of the sediment laye. The vertical mixing process has been shown to be well represented by a 1-D equivalent diffusivity model with exponentially decaying profile in the vertical direction. This exponential profile is controlled by two parameters: the effective diffusion at the sediment water interface, that is equivalent to the mixing magnitude, and the decay coefficient of the exponential, that represents the reciprocal of the mixing extent.The 1-D diffusivity model was applied to a large dataset of experiments from literature that were conducted with conservative solutes on three different morphology types: flat beds, dunes and ripples, and alternate bars. The mixing magnitude and extent appears to be the largest in alternate bars, the smallest in flatbed and intermediate in dunes and ripples. Afterwards, the dependence of the mixing magnitude and mixing extent on stream and sediment characteristics was studied to derive a regression model to infer the values of the mixing controlling parameters from stream and sediment characteristics. This regression model was developed in dimensionless form using the Multiple Linear Regression (MLR) technique. The regression formulae demonstrate no dependence of the mixing magnitude on morphology type, while it is significantly correlated to the permeability Reynolds number (Re_k) that depends on sediment permeability, shear velocity and water viscosity. On the other hand, the mixing extent of solutes within the benthic biolayer can be categorized based on the morphology type. Specifically, for flat beds the mixing extent exhibits different trends of dependance on the permeability Reynolds number due to the dominance of different physical process for each interval of Re_k. Instead, for dunes and ripples the mixing extent is monotonically correlated to the bedform (ripple or dune) wave number, as expected from literature. Finally, due to the limited number of experiments on alternate bars we were not able to derive a robust and reliable regression formula for this morphology type. Therefore, more experiments under different conditions should be performed with alternate bars. These results help to unravel the influence on mixing processes of different characteristics of streams and their sediments.