Identification of historical PCDD/F sources in Newark Bay Estuary subsurface sediments using polytopic vector analysis and radioisotope dating techniques

• Published in: Chemosphere (Oxford) 1998-03, Vol.36 (6), p.1167-1185
• Main Authors: Huntley, S.L., Carlson-Lynch, H., Johnson, G.W., Paustenbach, D.J., Finley, B.L.
• Format: Article
• Language: English
• Subjects: Applied sciences; Brackish; Earth sciences; Earth, ocean, space; Engineering and environment geology. Geothermics; Exact sciences and technology; Marine; Pollution; Pollution sources. Measurement results; Pollution, environment geology; Soil and sediments pollution
• ISSN: 0045-6535; 1879-1298
• DOI: 10.1016/S0045-6535(97)10008-X

Fifty sediment cores, comprising a total of 194 individual samples, were collected throughout the Newark Bay Estuary including the Arthur Kill, Rahway River, Elizabeth River, Hackensack River, Kill van Kull, and Passaic River between November 1991 and March 1993. All samples were analyzed for polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs). An additional six to ten samples were collected from each core and analyzed for the radioisotopes 210Pb and 137Cs. The 210Pb and 137Cs results were used to estimate sediment accumulation rates for each core and approximate dates of sediment deposition for individual samples analyzed for PCDD/Fs. Sediment cores for which radioisotope results were inadequate for radiodating were excluded from the analysis, leaving a total of 26 cores, comprising 50 individual samples, for further statistical analysis. These remaining samples were grouped into three categories (models) based on their estimated dates of deposition. The time intervals represented by each model were 1950–1965 (model 5065), 1970–1979 (model 7079), and 1980–1989 (model 8089). Polytopic vector analysis (PVA) was performed separately on each model to determine the number of dominant fingerprint patterns (end members) present within each model. PVA results indicate that the 5065 model comprises five distinct end members, while the 7079 and 8089 models comprise six distinct end members. Comparison of end member patterns with source-specific fingerprint patterns found three PCDD/F congener patterns common to all models: combustion sources, sewage sludge sources, and sources associated with polychlorinated biphenyls (PCBs). In addition, each of the three models contained more than one sewage sludge or combustion pattern as well as at least one pattern that could not be identified. In many cases, the location of specific land-based sources (i.e, combined sewer overflow (CSO) outfalls and PCB-contaminated sites) was consistent with the identified end members. For example, the highest proportion of the PCB source end member was present in a sample collected adjacent to a PCB-contaminated site on the lower Passaic River, and the highest proportions of the sewage sludge source end members were present in samples collected adjacent to CSO and sewage treatment plant outfalls. The combustion and sewage sludge source end members were present in relatively high proportions in nearly all samples. These results are consistent with a recent study of surface sediment