Trophic transfer of polychlorinated biphenyls (PCBs) in the food web of the Anacostia River (Washington, D.C.).

摘要

The Anacostia River, which flows through Washington, D.C., USA, is contaminated with polychlorinated biphenyls (PCBs). Median sediment concentrations are approximately 225 ng/g dwt in the main portion of the river, with concentrations of around 1,700 ng/g dwt in the hot spot located near the Navy Yard. Zooplankton, benthic organisms, and fish were analyzed for PCB concentrations, and these values were used to develop and evaluate a bioaccumulation model that predicts PCB concentrations in fish based on concentrations in the environment. On a wet weight basis, concentrations in zooplankton and benthic organisms were similar to the concentrations in the sediment and particulate matter, typically around 85% of those concentrations. However, when lipid and organic carbon normalized accumulation factors were calculated, they were in the range of 1 to 3 for these organisms. Four species of fish were modeled: the pumpkinseed (Lepomis gibbosus), the brown bullhead (Ameiurus nebulosus), the spottail shiner (Notropis hudsonius), and the white perch (Morone americana). PCB concentrations in the fish ranged from 115 ng/g wwt to 1420 ng/g wwt, with some differences between species, age classes and sites. Seasonal differences in PCB concentrations were also noted. A model for trophic transfer of PCBs from the prey organisms to the fish was developed, based in large part on bioenergetics equations for largemouth bass (Micropterus salmoides). Growth of the fishes was modelled using species-specific von Bertalanffy equations derived from the data collected during the project. Using median measured prey values as input, model predictions proved to be close to measured concentrations, with 25% error. Based on model calculations, bioaccumulation in this system is driven primarily by the balance between dietary uptake and growth dilution, with only a small component (1%) derived from bioconcentration. Linking the empirically derived bioaccumulation factors and the fish model allowed prediction of fish whole body concentrations from sediment and particulate values. Using measured concentrations in those compartments, predictions for the fish were again close to measured values, with an error of 35%. Using the model to evaluate current conditions in the river and several reduction scenarios as compared to human health and ecological benchmarks showed the greatest reductions in the fish would occur if both sediment and source reduction measures were implemented. (Abstract shortened by UMI.)