Contaminant attenuation processes in the vadose zone of a crude oil spill site near Bemidji, MN have been simulated with a reactive transport model that includes multicomponent gas transport, solute transport, and the most relevant biogeochemical reactions. Dissolution and volatilization of oil components, their aerobic and anaerobic degradation coupled with sequential electron acceptor consumption, ingress of atmospheric O₂, and the release of CH₄ and CO₂ from the smear zone generated by the floating oil were considered. The focus of the simulations was to assess the dynamics between biodegradation and gas transport processes in the vadose zone, to evaluate the rates and contributions of different electron accepting processes towards vadose zone natural attenuation, and to provide an estimate of the historical mass loss. Concentration distributions of reactive (O₂, CH₄, and CO₂) and non-reactive (Ar and N₂) gases served as key constraints for the model calibration. Simulation results confirm that as of 2007, the main degradation pathway can be attributed to methanogenic degradation of organic compounds in the smear zone and the vadose zone resulting in a contaminant plume dominated by high CH₄ concentrations. In accordance with field observations, zones of volatilization and CH₄ generation are correlated to slightly elevated total gas pressures and low partial pressures of N₂ and Ar, while zones of aerobic CH₄ oxidation are characterized by slightly reduced gas pressures and elevated concentrations of N₂ and Ar. Diffusion is the most significant transport mechanism for gases in the vadose zone; however, the simulations also indicate that, despite very small pressure gradients, advection contributes up to 15% towards the net flux of CH₄, and to a more limited extent to O₂ ingress. Model calibration strongly suggests that transfer of biogenically generated gases from the smear zone provides a major control on vadose zone gas distributions and vadose zone carbon balance. Overall, the model was successful in capturing the complex interactions between biogeochemical reactions and multicomponent gas transport processes. However, despite employing a process-based modeling approach, honoring observed parameter ranges, and generally obtaining good agreement between field observations and model simulations, accurate quantification of natural attenuation rates remains difficult. The modeling results are affected by uncertainties regarding gas phase saturations, tortuosities, and the magnitude of CH₄ and CO₂ flux from the smear zone. These findings highlight the need to better delineate gas fluxes at the model boundaries, which will help constrain contaminant degradation rates, and ultimately source zone longevity.