Identifying polychlorinated biphenyl dechlorination pathways, processes and implications for risk management

For polychlorinated biphenyl (PCB) remedial actions and fate models to be truly risk based, they must consider the full complexity of the congeners and the uncertainty surrounding their measurement, and the transformation reactions connecting them to one another. To date, fate models have been limited in the number of congeners considered, the number of pathways simulated, and their consideration of congener-specific uncertainty. The present work stresses the inclusion of all 209 PCB congeners and 840 dechlorination pathways. The most common method for quantitating congener concentrations is on a gas chromatograph with electron capture device. This method quantitates fewer than 209 congeners due to co-elution of PCBs with similar structures. As a result, potentially significant uncertainty is introduced to co-eluting congener concentrations due to their interpretation in Aroclor ratios. The use of interpreted results should explicitly consider concealed congeners, and congeners which have very different response factors and weight percentages in the average Aroclor distribution. Additionally, surrogate variables routinely used to describe dechlorination can be biased through the inclusion of these uncertain congener measurements. The present work first assesses error introduced to surrogate variables as a result of such interpretations.;Dechlorination patterns, termed processes, are a useful tool for researchers in the absence of identified PCB-dechlorinating microorganisms and the biogeochemical conditions that influence them. Application of the classification tree method to eight incomplete dechlorination pattern observations generated a suite of potential pathways that the governing microorganism(s) can catalyze. The presented Dechlorination Process Estimator (DPE) identifies the occurrence of one or more known dechlorination patterns, or a pattern that has yet to be identified. It is a significant improvement on current qualitative and expert-informed methods. The DPE uses the Bayes Monte Carlo method, which is a replicable and quantitative approach. It considers congener-specific uncertainty and expert opinion, which is subsequently updated with laboratory data. Application of these methods to 124 sediment core samples from the Grasse River demonstrated that surrogate variables calculated from interpreted co-eluting congeners largely underestimate the occurrence of dechlorination and that one or more dechlorination processes sharing pathways with known Processes N, P and Q, may have occurred in a laboratory study of Grasse River sediment.