The predictive power of bioaccumulation models may be limited when they do not accountfor strong sorption of organic contaminants to carbonaceous materials (CM) such as black carbon, and when they do not include metabolic transformation. We tested a food web accumulation model, including sorption to CM, on data from a model ecosystem experiment with historically contaminated sediment. In combination with measured CM contents of the sediment, the model gave good fits for the biota that are known not to metabolize PAHs (macrophytes, periphyton, floating algal biomass). The same model was applied to invertebrates and fish but now with optimization of their metabolic transformation rates (k(m)). For fish, these rates correlated empirically with log K(OW): Log k(m) = -0.8 log K(OW) + 4.5 (r2 adj = 0.73). For invertebrates, log k(m) did not correlate with logK(OW). Sensitivity analysis revealed that the model output is highly sensitive to sediment CM content and sorption parameters, moderately sensitive to metabolic transformation rates, and slightly sensitive to lipid fraction of the organism and diet-related parameters. It is concluded that CM-inclusive models yield a better assessment of accumulation than models without sorption to CM. Furthermore, inclusion of CM in a model enables metabolic transformation rates to be calculated from the remaining overestimation in the model results when compared to measured data.