PCB sources and degradation in river sediments determined by receptor modeling (Ohio, Wisconsin)

Sources and environment degradation of polychlorinated biphenyls (PCBs) were investigated in the sediments of two highly polluted Great Lakes rivers, namely, Ashtabula River, Ohio, and Fox River, Wisconsin, USA. A factor analysis (FA) model with non-negative constraints was modified to improve the resolution with which PCB sources and major congener patterns are identified in the sediments. This was done through incorporation of scaling on the data matrix and back-scaling on the resulting factor loading matrix, as well as the use of diagnostic tools. Additionally, an anaerobic dechlorination model was developed to work with the FA model to identify possible patterns of anaerobic dechlorination of PCBs in the sediments, and to quantify the relevant dechlorination pathways. This model uses a least squares method to alter original Aroclor profiles according to six major dechlorination activities (M, Q, H, H ′, P, N) to simulate congener patterns found in the environment. Validation of both models was successfully performed using artificially created data sets, as well as laboratory and in situ dechlorination data from the literature.; The PCB data sets used in this study were composed of 31 congeners, analyzed in samples from deep, dated sediment cores; collected from four sites in Ashtabula River, eight from Fox River and two reference sites from Green Bay. Results from the FA model of Ashtabula River samples indicate Aroclor 1248 to be the major PCB contaminant. Another congener pattern, significantly different from that of Aroclor 1248 was also identified by the FA model. Simulation of anaerobic dechlorination on the original Aroclor 1248 profile according to dechlorination activities H/H′ yielded a congener profile very similar to that of this pattern. This indicated the likelihood of anaerobic dechlorination of PCBs in Ashtabula River sediments.; Two different PCB data sets were evaluated from lower Fox River sediments; one obtained from the Wisconsin Department of Natural Resources, and the other was a result of our recent sampling and chemical analysis. The FA model results for both data sets, collected about 10 years apart, show Aroclor 1242 to be the major PCB source, but indications of the presence of Aroclors 1254 and 1260 also exist. The congener profiles indicate the effect of both physicochemical weathering mechanisms and anaerobic dechlorination activity. Loss of di- and tri-chlorobiphenyls (e.g., #8, #18) from the sediments suggest desorption from the sediments. On the other hand, observation of elevated amounts of certain congeners such as #25 and #26, indicated the presence of anaerobic dechlorination activity. The anaerobic dechlorination model demonstrated significant similarities between Aroclor 1242 profile altered according to dechlorination activity Q, and the dominant congener patterns obtained from the FA model using the recent PCB data set.; The PCB congener profiles of Green Bay sediments show significant effects of physicochemical weathering and long range transport of low chlorinated congeners. Any anaerobic dechlorination activity, if present, is difficult to determine due to the effect of weathering mechanisms.; This study is the first application of receptor modeling for identifying and quantifying possible anaerobic dechlorination of PCBs in contaminated sediments. The improved factor analysis model, coupled with the proposed anaerobic dechlorination model, presents a viable tool for such analysis of environmental data.