A Source-Receptor Method for Determining Non-Point Sources of PAHs to the Milwaukee Harbor Estuary

• Published in: Water science and technology 1993, Vol.28 (8-9), p.91-102
• Main Authors: Singh, A K, Gin, M F, Ni, F, Christensen, E R
• Format: Article
• Language: English
• Subjects: Aromatic hydrocarbons; Atmospheric particulates; Chi-square test; Coal; Coal tar; Computer software; Correlation coefficient; Correlation coefficients; Detection; Dust storms; Estuaries; Estuarine dynamics; Exhaust emissions; Exhaust gases; Fuel oils; Gasoline; Gasoline engines; Mathematics; Point source pollution; Pollutants; Pollution sources; Polycyclic aromatic hydrocarbons; Receptors; Regression analysis; River banks; Riverbanks; Rivers; Sediment; Sediment samplers; Sediment samples; Sediments; Storage tanks; Tar; Underground storage; Underground storage tanks
• ISSN: 0273-1223; 1996-9732
• DOI: 10.2166/wst.1993.0607

The purpose of this work is to determine non-point sources of PAHs to the Milwaukee Harbor Estuary. Sediment samples were collected from the Milwaukee River, the Menomonee River, the Kinnickinnic River, and the Inner and Outer Harbor of the Milwaukee Harbor Estuary. Source samples were collected from a major highway, a leaking underground storage tank, and two river banks. All samples were analyzed for sixteen PAHs, which are identified as priority pollutants by the U.S. EPA. In addition to this, literature values of concentrations of PAHs in known non-point sources viz. coal tar, coal tar air emissions, No. 2 fuel oil, and gasoline engine exhaust tar were considered. A computer program, using the mathematics of least-squares, was adopted to determine the best combination of sources to the mixed PAH signature of the sediment, and to calculate the contribution of each source to the concentration of total PAHs. Chi square and the multiple correlation coefficient were used to evaluate the fits. For apportionment of sediment PAHs into sources, only those PAHs are considered whose concentrations are above the instrument's (GC-MS) detection limit. Of the 211 regressions considered, about half present good fits to the proposed sources. Highway dust is found to be a significant source in 88%, gasoline engine exhaust tar in 33%, coal tar air emissions in 23%, and coal tar in 17% of the successful regressions.