Source Apportionment of Polychlorinated Biphenyls in the Tidal Delaware River

Abstract

The Delaware River, similar to many surface water bodies throughout the United States, is impaired due to polychlorinated biphenyl (PCB) concentrations that exceed the federal water quality standard. A total maximum daily load (TMDL) for PCBs in the Delaware was promulgated in 2003 after construction of a detailed water quality model that relied upon estimated PCB loadings from sources such as wastewater treatment plant effluents, contaminated sites, and tributaries. The purpose of this project was to corroborate this loading estimate by analyzing ambient water column data on PCB concentrations in the Delaware River via positive matrix factorization (PMF) in order to identify covarying congener patterns that are, in theory, associated with sources. The PMF program identified six factors (covarying congener patterns) that appear to be associated with sources, such as sediment resuspension, contaminated sites, and wastewater effluents and combined sewer overflows (CSOs). PCB 11 was found to be a useful tracer for CSO/ wastewater inputs despite the fact that no known dye manufacturers exist in the watershed. This analysis generally corroborates the PCB loading estimate used in the water quality model.

PCBs in Great Lakes sediments, determined by positive matrix factorization

“…In chemical mass balance (CMB) modeling the source profiles must be known while in factor analysis both source profiles and contributions can be determined, as long as minimal information about sources is available Henry, 2003;Magar et al, 2005;Bzdusek et al, 2006a,b;Du et al, 2008). This information can be that certain compound values in sources are zero or near zero or that there are several measurements that contain zero contribution for at least one source.…”

Sources, Fate and Effects of Contaminant Emissions in Urban Areas

“…This was the conclusion of Rodenburg and co-workers based on a variety of measurements and statistical methods that they used to study PCBs in the Hudson River estuary and the Philadelphia-Camden region (Totten et al, 2004;Asher et al, 2007;Du and Rodenburg, 2007;Du et al, 2008Du et al, , 2009). …”

“…Rodenburg and co-workers used positive matrix factorization (PMF) as a more formal statistical method to distinguish sources of PCBs based on congener profiles (Du and Rodenburg, 2007;Du et al, 2008Du et al, , 2009. Coupled with potential source contribution analysis, they were able to distinguish geographically the source(s) of PCBs in the PMF factors.…”

Nationwide PCB congener pattern analysis in freshwater fish samples in France

“…Moreover the negative eigenvectors sometimes obtained in PCA are not appropriate when looking for source profiles. While there are numerous examples of pattern analysis using models adapted from multivariate statistics in environmental matrices such as air (Logue et al, 2009;Pekney et al, 2006), soil (Skrbic and Durisic-Mladenovic, 2007) or bottom sediment (Bzdusek et al, 2006a;Bzdusek et al, 2006b;Du et al, 2008), there have been fewer attempts to develop this kind of approach on biota. Examples include an exploration of spatial differences in organochlorine chemical loads in herring gull colonies of the Great Lakes (MacDonald et al, 1992), a study of PCB accumulation features, but neither spatial nor source pattern analysis in the Baltic Sea area (Falandysz et al, 2002), an attempt to differentiate local PCB sources in the Hudson River and its estuary Monosson et al, 2003), and a study of PCB pattern variations across species, trophic levels, and wild versus farmed salmon (Yunker et al, 2011).…”