loading page

On the Control of Soil Heterogeneity, Peclet number and Spatially Variable Diffusion over Unsaturated Transport
  • Christopher Vincent Henri,
  • Efstathios Diamantopoulos
Christopher Vincent Henri
Geological Survey of Denmark and Greenland

Corresponding Author:[email protected]

Author Profile
Efstathios Diamantopoulos
University of Copenhagen
Author Profile

Abstract

Physical properties of soils are ubiquitously heterogeneous. This spatial variability has a profound, yet still partially understood, impact on conservative transport. Moreover, molecular diffusion is often a disregarded process that can have an important counter-intuitive effect on transport: diffusion can prevent non-Fickian tailing by mobilizing mass otherwise trapped in low velocity zones.
Here, we focus on macroscopically homogeneous soils presenting small scale heterogeneity, as described by the Miller-Miller method. We then analyze the dynamic control of soil heterogeneity, advection and diffusion on conservative transport. We focus especially on the importance of diffusion and of its tortuosity-dependent spatial variability on the overall transport.
Our results indicate that high Peclet number systems are highly sensitive to the degree of heterogeneity, which promotes non-Fickian transport. Also, diffusion appears to have a profound impact on transport, depending on both the degree of heterogeneity and the Peclet number. For a high Peclet number and a very heterogeneous system, diffusion leads to the counter-intuitive decrease of non-Fickian macrodispersion described previously. This is not observed for a low Peclet number due to the non-trivial impact of the spatial variability in the diffusion coefficient, which appears to be a significant controlling factor of transport by promoting or preventing the accumulation of mass in low velocity zones.
Globally, this work (1) highlights the complex, synergistic effect of soil heterogeneity, advective fluxes and diffusion on transport and (2), alerts on potential upscaling challenges when the spatial variability of such key processes cannot be properly described.