Biodegradation is an important process contributing to the natural attenuation (NA) of organic contaminants in groundwater. A numerical model was created to describe anaerobic phenol biodegradation data from an aquifer-derived laboratory scale microcosm. The dynamic behaviour of the system was simulated by considering a two-step syntrophic biodegradation model with fermentation and respiration steps, both simulated kinetically, and with hydrogen and acetate as intermediate species, and additionally, other geochemical reactions including aqueous speciation, surface complexation, mineral dissolution and precipitation. The model suggested microbial competition between respiration processes using different electron acceptors was important. In contrast, a partial equilibrium approach, considering only thermodynamics, and not kinetics, for respiration, did not explain the data.
The laboratory scale biodegradation model was transferred to a field scale reactive transport model of the phenol plume at Four Ashes, UK. The effects of acclimatisation, toxicity, and bioavailability on microbial kinetics were considered. The simulations suggest that plume core processes are much more important than previously thought, possibly with a greater impact than plume fringe processes.
The field scale model was computationally demanding due to the biogeochemical complexity. Two strategies for dealing with high computational demands are (i) parallel processing, where the workload is shared between multiple processors, and (ii) locally adaptive remeshing, where a refined area of the grid tracks moving plume fringes through the domain. A new code was developed using the partial differential equation software toolbox, UG, and tested against other biodegradation simulators. The relative efficiency of parallel, adaptive methods for multispecies biodegradation simulations was measured.
It appears, in general, that relatively complex models are required for the realistic, quantitative assessment of NA at field scale, and that parallel, adaptive numerical methods provide appropriate efficiency benefits for such simulations.
