Effectiveness of remediation at contaminated sediment sites based on measurements of bioavailability: testing a conceptual approach

Remediation of contaminated sediment sites is complex and expensive, critical to the restoration of damaged ecosystems, and essential to the reduction of threats to human health. Along with the challenges of remediating a contaminated site is determining if the remediation has been effective. In this investigation, the use of bioavailability-based metrics was evaluated as a tool for assessing remedial effectiveness. Metrics included surface water concentrations and bioaccumulation in proxy species. Following the identification, collection, and curation of relevant retrospective bioavailability metric datasets, statistical analyses were performed comparing pre-remediation and post-remediation time periods. Statistical analyses used (i) hypothesis testing to identify significant reductions in bioavailability and (ii) regression to assess significant negative slopes. Datasets from three Superfund sites were evaluated in this investigation: Lower Grasse River, New Bedford Harbor, and Stauffer Chemical Company. Both the Lower Grasse River and New Bedford Harbor demonstrated decreases in the bioavailability of total PCBs as the remediations proceeded. Remedial effectiveness was more apparent for the Lower Grasse River compared to New Bedford Harbor most likely because remediation at the former had been completed for several years while it was still on-going at the latter. In contrast, effectiveness of remediation at Stauffer Chemical Company was less obvious based on statistical analyses of the bioavailability metrics. We speculate this reflects the more complicated environmental behavior of mercury compared to PCBs. This preliminary use of retrospective datasets of bioavailability metrics from Superfund sites to assess remediation effectiveness has shown promising results and merits further evaluation at other sites. In addition, contaminants like PCBs appear to be good candidates with this approach while mercury may not be. Also, datasets with good temporal separation of the pre-remediation and the post-remediation time periods are better candidates for this approach.