An evaluation of the ability of chemical measurements to predict polycyclic aromatic hydrocarbon-contaminated sediment toxicity toHyalella azteca

The present study examined the ability of three chemical estimation methods to predict toxicity and nontoxicity of polycyclic aromatic hydrocarbon (PAH) -contaminated sediment to the freshwater benthic amphipod Hyalella azteca for 192 sediment samples from 12 field sites. The first method used bulk sediment concentrations of 34 PAH compounds (PAH34), and fraction of total organic carbon, coupled with equilibrium partitioning theory to predict pore-water concentrations (KOC method). The second method used bulk sediment PAH34 concentrations and the fraction of anthropogenic (black carbon) and natural organic carbon coupled with literature-based black carbon-water and organic carbon-water partition coefficients to estimate pore-water concentrations (KOCKBC method). The final method directly measured pore-water concentrations (pore-water method). The U.S. Environmental Protection Agency's hydrocarbon narcosis model was used to predict sediment toxicity for all three methods using the modeled or measured pore-water concentration as input. The KOC method was unable to predict nontoxicity (83% of nontoxic samples were predicted to be toxic). The KOCKBC method was not able to predict toxicity (57% of toxic samples were predicted to be nontoxic) and, therefore, was not protective of the environment. The pore-water method was able to predict toxicity (correctly predicted 100% of the toxic samples were toxic) and nontoxicity (correctly predicted 71% of the nontoxic samples were nontoxic). This analysis clearly shows that direct pore-water measurement is the most accurate chemical method currently available to estimate PAH-contaminated sediment toxicity to H. azteca.