Management of polychlorinated biphenyl (PCB)-contaminated sediment has proven to be a widespread, complex, and costly issue. PCBs are a primary contaminant driving risk at many Department of Defense facilities. There is a need for sound science and effective tools to characterize and manage these sites and their associated risks. Monitored natural attenuation, sediment capping, and cap-and-treat approaches are potentially viable and attractive treatment alternatives to dredging contaminated sediments. The effectiveness of these less costly alternatives and the decision to use them will ultimately rely on the ability to adequately characterize the natural or enhanced biodegradation potential and rate at a specific site. Spatial and temporal variability, inadequate understanding of the factors influencing the rate and extent of PCB dechlorination, and the lack of time and resources for extensive characterization at all sites introduces significant uncertainty in assessing biodegradation potential.

The objectives of this project were to provide a better understanding of the role of site-specific differences in geochemistry and microbial populations in the transformation of PCBs in sediments and to develop advanced modeling tools to enhance the ability to predict dechlorination and evaluate the likelihood of natural attenuation at specific sites.

Technical Approach

Laboratory analysis of sediment samples and microcosm-based PCB dechlorination experimentation were used to enhance the understanding of microbial populations that are capable of dechlorinating PCBs. Surficial sediments from the Grasse River and the Hudson River were collected and analyzed for congener-specific PCB concentrations, critical PCB transformation markers (e.g., molar dechlorination product ratio (MDPR), chlorines per biphenyl (CPB), homolog concentrations), and bacterial species of relevance to reductive dechlorination. An intact core from the Grasse River was also analyzed for these criteria.

The development of advanced modeling tools was used to enhance prediction of dechlorination and evaluate the likelihood of natural attenuation at specific sites. Classification trees were developed to predict PCB pathways that are likely to be observed when specific processes are active in contaminated sediments. A Bayesian Monte Carlo method was used to identify the occurrence of one or more of the eight known dechlorination processes. The model, termed the Dechlorination Process Estimator (DPE), uses a discrete chemical reaction equation for each of the 209 congeners such that mass from parent congeners is transferred to child congeners across 840 dechlorination pathways.


Microbial diversity indicators were found to be useful tools to evaluate differences in microbial populations that result from differences in sediment geochemistry, whether in native sediment samples or in amended microcosms. Results indicated that PCBs present in the Grasse River sediments have undergone reductive dechlorination for decades and that the Grasse River contains abundant PCB degrading populations throughout the sediment. Core analysis also was determined to be a useful technique in evaluating MNA potential. From the microcosm studies, PCB dechlorination was significantly affected by sediment biogeochemistry (including sulfate and iron concentrations) as well as by initial sediment microbial community (the Grasse River was found to have more dechlorinators than the Hudson River). Initial PCB congeners, sediment biogeochemistry, and microbial community all affect the rate, extent, and nature of PCB dechlorination in ways that can be quantified through careful experimentation.

Classification tree dechlorination process generalizations reflect the capability of microbes that have not been identified or are not well understood and can provide insight into congener endpoints in laboratory experiments and in the field, particularly in sediments undergoing monitored natural attenuation. The DPE model is capable of simulating large, complex, and partly understood chemical transformations in a statistically rigorous manner. Application of the DPE to a laboratory experiment suggests that it is suitable for identifying process occurrence.


This project improved understanding of the features controlling the transformation of PCBs in sediment systems. Extensive sediment characterization can aid in determining the suitability of monitored natural attention as a remedy at a specific field site. Sediment amendments may be useful to seed microbial populations or change pathways and end points for reductive dechlorination. Utilization of the modeling tools developed in this project will enable prediction of PCB end points and the effect of amendments on the rate, extent, and end point of reductive dechlorination. Carefully designed microbial experiments, paired with statistically valid models accounting for uncertainty, can provide unique insights into potential remediation strategies.