Numerical analysis of soil microstructure reveals conditions for natural attenuation in aged tar oil contaminated soil

Understanding of ongoing biogeochemical processes (natural attenuation) within contaminated soils is crucial for the development of plausible remediation strategies. We studied a tar oil contaminated soil with weak grass vegetation at a former manufactured gas plant site in Germany. Despite of the apparent toxicity (the soil contained up to 120 g kg^-1 petroleum hydrocarbons, 26 g kg^-1 toxic metals, and 100 mg kg^-1 polycyclic aromatic hydrocarbons), the contaminated layers have 3-5 times as much cell counts as an uncontaminated control soil nearby. To test, if the geometry of the pore space provides favourable living space for microorganisms, we applied scanning electron microscopy to the thin sections and calculated on sets of 15 images per layer three specific Minkowski functionals, connected to soil total porosity, interface, and hydraulic parameters.

Our investigation showed that the uncontaminated control soil has a relatively low porosity of 15-20 %, of which 50-70 % is comprised of small (< 15 µm) pores. These pores are poorly connected and show high distances between them (mean distance to the next pore 10 µm). The dominating habitats in the control soil are therefore created by small pores. They provide good protection from predators and desiccation, but input of dissolved organic C and removal of metabolic products are diffusion limited. Coarser pores (>15 µm) provide less space (< 50 % of total porosity) and solid surface area (< 20 %), are prone to desiccation and offer less protection from predators. However, they serve as preferential flow paths for the soil solution (input of nutrients) and are well aerated, therefore we expect the microbial activity in them to appear in “hot moments”, i.e. after rain events.

All layers of the contaminated profile have higher porosities (20-70 %) than the control. Coarse pores comprise 83-90 % of total pore area and create 34-52 % of total interface. Pores are also more connected and tortuous than in the control soil, which implies a better aeration and circulation of soil solution. The loops of pore channels may retain soil solution and be therefore preferably populated with microorganisms. The small (< 15 µm) pores comprise less than 17 % of total porosity but represent a substantial proportion of the interface (48-66 % vs 82-91 % in control). In the uppermost layer of the contaminated profile, such pores occur in plant residues, are close to the largest pores (mean distance to the next pore 4 µm) and therefore, along with good protection, are supplied with air, water, and non-tar C. In the middle of the profile, the small pores, presumably constantly filled with water, are located within dense tar pieces remote from the neighbouring pores (mean distance to the next pore 22 µm), and therefore, with hindered aeration and no supply of non-tar C, may create anaerobic domains of tar attenuation.

Our results show that the contaminated soil offers more favourable conditions for microorganisms than the control soil, probably because the hydrocarbons provide suitable energy and nutrition sources and a beneficial pore space geometry.