Gas diffusivity in soils compared to ideal isotropic porous media |
| |
| Authors: | Anke Kühne Helmer Schack‐Kirchner Ernst E Hildebrand |
| |
| Institution: | Institute of Soil Science and Forest Nutrition, Albert‐Ludwigs‐Universit?t Freiburg, Bertoldstra?e 17, 79085 Freiburg i. Br., Germany |
| |
| Abstract: | Evidence of anisotropy is reported for advective air and water permeabilities in soils. Thus, anisotropy is likely to exist also for diffusive gas fluxes. Information about direction‐dependent soil gas diffusivity is scarce and most modeling approaches assume isotropy. At hundreds of closely lying positions in a compacted and adjacent undisturbed forest soil, gas diffusivity (Ds/D0) was measured either in vertical or horizontal direction. The volume‐independent diffusion efficiency (i.e., diffusivity divided by air‐filled porosity) was fitted by a generalized additive model (GAM). Significant regressors were air‐filled porosity (?), soil depth, and the discrete diffusion direction. The model yields in all cases higher vertical diffusion efficiencies. The compaction factor did not yield a significant regressor of its own, i.e., the reduction of diffusivity in the compacted soil was the same as in low‐porosity samples of the undisturbed profile. To elucidate the role of sharing vertically and horizontally orientated pore space and a potential competition between diffusivity in different spatial directions, simple geometric models consisting of 3‐dimensionally crossed pores have been parameterized. These models provided a good explanation of the typical nonlinear Ds/D0(?) relationship. By simple one‐parameter correction (linear or power function), this mechanistic model could be fitted to the data. The one‐parameter correction of the geometric model could be a straightforward approach to consider direction dependence of measured diffusivities. However, by applying this approach to the observations the anisotropy effect was not clearly evident, which could be attributed to a changing Ds/D0(?) relationship with depth. As a reason for the preference of the vertical gas diffusion the dominance of vertical stresses and the activity of anecic earthworms are discussed. Direction dependency of gas diffusivity seems to be a basic feature of natural pore systems and has to be considered for modeling gas fluxes in soils. Generally, a preferential vertical diffusion direction reduces horizontal balancing and increases the heterogeneity of gas concentrations in the soil air. |
| |
| Keywords: | diffusion efficiency anisotropy generalized additive model (GAM) Menger sponge |
|
|
