A contaminated alluvial aquifer, in a former gasworks factory and discharging to an adjacent river, was the object of field and laboratory investigations to assess pollutant attenuation and dispersion. Various organic and inorganic compounds were found in the aquifer in concentrations exceeding regulatory values, among them, benzene, which was presenting the major worry for off-site dispersion, mainly due to its mobility and high concentration, i.e. up to 750 mg L-1 in the source zone. However, benzene could never be detected near the river which is about 160 m downgradient the main source. Due to redox conditions of the aquifer, heavy metals were almost immobile, thus not posing a major risk of dispersion off-site the brownfield. Benzene concentrations together with redox conditions in the aquifer, suggested that benzene degradation was mainly occurring within 100 m distance from the contaminant source under anoxic conditions, and most probably with sulphate as main oxidant. A numerical groundwater flow and transport model, calibrated under transient conditions, was used to simulate benzene attenuation in the alluvial aquifer towards the Meuse River. The mean benzene degradation rate used in the model was quantified in situ along the groundwater flow path using compound-specific carbon isotope analysis (CSIA). The results of the solute transport simulations confirmed that benzene concentrations decreased almost five orders of magnitude 70 m downgradient the source. Simulated benzene concentrations were found to be below the detection limit in the zone adjacent to the river and consistent with the absence of benzene in downgradient piezometers located close to the river. In a transient model scenario including groundwater-surface water dynamics, benzene concentrations were observed to be inversely correlated to the river water levels, leading to the hypothesis that benzene dispersion is mainly controlled by natural attenuation and river fluctuations.