Soil quality standards are based on partitioning and toxicity data for laboratory-spiked reference soils, instead of real world, historically contaminated soils, which would be more representative. Here 21 diverse historically contaminated soils from Sweden, Belgium, and France were obtained, and the soil-porewater partitioning along with the bioaccumulation in exposed worms (Enchytraeus crypticus) of native polycyclic aromatic compounds (PACs) were quantified. The native PACs investigated were polycyclic aromatic hydrocarbons (PAHs) and, for the first time to be included in such a study, oxygenated-PAHs (oxy-PAHs) and nitrogen containing heterocyclic PACs (N-PACs). The passive sampler polyoxymethylene (POM) was used to measure the equilibrium freely dissolved porewater concentration, C-pw, of all PACs. The obtained organic carbon normalized partitioning coefficients, K-TOC, show that sorption of these native PACs is much stronger than observed in laboratory-spiked soils (typically by factors 10 to 100), which has been reported previously for PAHs but here for the first time for oxy-PAHs and N-PACs. A recently developed K-TOC model for historically contaminated sediments predicted the 597 unique, native K-TOC values in this study within a factor 30 for 100% of the data and a factor 3 for 58% of the data, without calibration. This model assumes that TOC in pyrogenic-impacted areas sorbs similarly to coal tar, rather than octanol as typically assumed. Black carbon (BC) inclusive partitioning models exhibited substantially poorer performance. Regarding bioaccumulation, C-pw combined with liposome-water partition coefficients corresponded better with measured worm lipid concentrations, C-lipid (within a factor 10 for 85% of all PACs and soils), than C-pw combined with octanol-water partition coefficients (within a factor 10 for 76% of all PACs and soils). E. crypticus mortality and reproducibility were also quantified. No enhanced mortality was observed in the 21 historically contaminated soils despite expectations from PAH spiked reference soils. Worm reproducibility weakly correlated to C-lipid of PACs, though the contributing influence of metal concentrations and soil texture could not be taken into account. The good agreement of POM-derived C-pw with independent soil and lipid partitioning models further supports that soil risk assessments would improve by accounting for bioavailability. Strategies for including bioavailability in soil risk assessment are presented.
2014. Vol. 48, no 19, 11187-11195 p.