A gas plume from the Foxhall landfill in Suffolk, UK contains elevated concentrations of methane and carbon dioxide in unsaturated unconsolidated sands which comprise Pleistocene “Red Crag” deposits. The plume emanates from a relatively restricted zone along one side of the landfill, but extends over 100 m from the site boundary. The reduction in the methane to carbon dioxide ratio with distance and the systematic changes in their carbon and hydrogen stable isotope ratios are evidence of microbially mediated methane oxidation. A one-dimensional advection–diffusion model was used to describe the combined concentration of methane and carbon dioxide in the plume. Diffusion alone underestimated the concentration profile, but a good fit to the data was achieved with an advective flux of 4.5 m yr–1, indicating that advection due to a pressure gradient from the landfill as well as diffusion should be considered in gas migration modelling. The kinetics of methane oxidation was studied by parameter-fitting a reaction rate into the advection–diffusion equation with first-order decay. A decay constant of –0.063 yr–1 (half-life 11 yr) produced a poor fit to the methane profile, suggesting that oxidation may not be constant throughout the plume. However, the stable isotope data allowed two rates of oxidation to be inferred. A slow rate of oxidation with a half-life of the order of 4.3 to 7.6 yr was inferred in the centre of the plume where oxygen was absent. A much faster rate with a half-life no longer than 0.76 to 1.21 yr occurred beyond 60 m of the landfill and around the top fringe of the plume where oxygen was present. These rates are considered to reflect the difference between aerobic and anaerobic oxidation, the latter utilizing iron(III) in the sediment as an electron acceptor. The shape of the plume is asymmetrical, indicating a geological control on gas migration. In a two-dimensional model a poor fit to the observed data was obtained when the sand was assumed to be homogeneous and where the gas entered from a restricted part of the landfill boundary. However, a better model was produced by varying the diffusion coefficient in the sands over the range 5 × 10–7 to 2 × 10–6 m2 s–1 without the need to restrict the zone of gas release along the landfill boundary. Such a range in transport properties could be accounted for by normal variability in the porosity, tortuosity and water content of the sand. The long-term dissipation of the plume assuming only diffusion was predicted to take up to 30 yr for the gas concentration to reduce to 10% of its initial value. However, the plume disappeared within a year after pumping from gas wells in the landfill, indicating that advection under an imposed pressure gradient was a major control on remediation. This study shows that models can be used to explain landfill gas migration and to infer oxidation rates which can be used to predict gas migration at other sites. However, the need to obtain field data on gas permeabilities and diffusivities will always be a major limitation in predicting gas migration in permeable formations.