Level I and II fugacity approaches were used to model the environmental distribution of benzene, anthracene, phenanthrene, 1-methylphenanthrene and benzo[a]pyrene in a four phase biopile system, accounting for air, water, mineral soil and non-aqueous phase liquid (oil) phase. The non-aqueous phase liquid (NAPL) and soil phases were the dominant partition media for the contaminants in each biopile and the contaminants differed markedly in their individual fugacities. Comparison of three soils with different percentage of organic carbon (% org C) showed that the % org C influenced contaminant partitioning behaviour. While benzene showed an aqueous concentration worthy of note for leachate control during biopiling, other organic chemicals showed that insignificant amount of chemicals leached into the water, greatly reducing the potential extent of groundwater contamination. Level II fugacity model showed that degradation was the dominant removal process except for benzene. In all three biopile systems, the rate of degradation of benzo(a)pyrene was low, requiring more than 12 years for soil concentrations from a spill of about 25 kg (100 mol) to be reduced to a concentration of 0.001 mu g g(-1). The removal time of 1-methylphenanthrene and either anthracene or phenanthrene was about 1 and 3 years, respectively. In contrast, benzene showed the highest degradation rate and was removed after 136 days in all biopile systems. Overall, this study confirms the association of risk critical contaminants with the residual saturation in treated soils and reinforces the importance of accounting for the partitioning behaviour of both NAPL and soil phases during the risk assessment of oil-contaminated sites. (C) 2008 Elsevier Ltd. All rights reserved.}