In this paper we develop and test a rigorous modeling framework, based on Duhamel’s Theorem, for the unsteady one-dimensional transport and mixing of a solute across a flat sediment-water interface (SWI) and through the benthic biolayer of a turbulent stream. The modeling framework is novel in that it allows for depth-varying diffusivity profiles, accounts for the change in porosity across the SWI and captures the two-way coupling between evolving solute concentrations in both the overlying water column and interstitial fluids of the sediment bed. We apply this new modeling framework to an extensive set of previously published laboratory measurements of turbulent mixing across a flat sediment bed, with the goal of evaluating four diffusivity profiles (constant, exponentially declining, and two hybrid models that account for molecular diffusion and enhanced turbulent mixing in the surficial portion of the bed). The exponentially declining profile is superior (based on RMSE, coefficient of determination, AICc, and model parsimony) and its reference diffusivity scales with a dimensionless measure of stream turbulence and streambed permeability called the Permeability Reynolds Number, . The diffusivity’s dependence on changes abruptly at , reflecting different modes of mixing below (dispersion) and above (turbulent diffusion) this threshold value. The depth-scale over which the diffusivity exponentially decays is about equal to the thickness of the benthic biolayer (2 to 5 cm), implying that turbulent mixing, and specifically turbulent pumping, may play an outsized role in the biogeochemical processing of nutrients and other contaminants in stream and coastal sediments.