USGS2001
Free-phase hydrocarbons are present in the alluvial aquifer at Mandan, North Dakota. A large contaminant body of the hydrocarbons [light nonaqueous phase liquid (LNAPL)] floats on the water table about 20 feet below land surface. The main LNAPL body is about 6 feet thick, and the areal extent is about 657,000 square feet. A study was conducted to describe the hydrologic setting and characterize the geochemical composition of the free-phase hydrocarbons in the alluvial aquifer. Most of the study area is underlain by alluvium of the Heart River Valley that ranges in thickness from about 25 to 109 feet. The alluvium can be divided into three stratigraphic units silty clay, silty sand, and sand and is underlain by shales and sandstones. Monitoring wells were installed prior to this study, to an average depth of about 29 feet. Regional ground-water flow in the Heart River aquifer generally may be from west-northwest to eastsoutheast and is influenced by hydraulic connections to the river. Hydraulic connections also are probable between the aquifer and the Missouri River. Ground-water flow across the north boundary of the aquifer is minimal because of adjacent shales and sandstones of relatively low permeability. Recharge occurs from infiltration of precipitation and is spatially variable depending on the thickness of overlying clays and silts. Although the general water-table gradient may be from west-northwest to east-southeast, the flow directions can vary depending on the river stage and recharge events. Any movement of the LNAPL is influenced by the gradients created by changes in water-level altitudes. LNAPL samples were collected from monitoring wells using dedicated bailers. The samples were transferred to glass containers, stored in the dark, and refrigerated before shipment for analysis by a variety of analytical techniques. For comparison purposes, reference-fuel samples provided by the refinery in Mandan also were analyzed. These reference-fuel samples included a current diesel fuel, a closely related but slightly broader refinery-cut fuel, a crude-oil composite, unleaded regular gasoline, and additives. Four principal analytical techniques were used for geochemical characterization: Purge-and-trap gas chromatography/mass spectrometry (volatile components); capillary gas chromatography/mass spectrometry (semivolatile components); isotope ratio mass spectrometry (carbon isotopes; whole oils); and liquid chromatography/mass spectrometry with electrospray ionization (additives and other organic components). Volatile analytes included solvents, disinfection byproducts, halogenated hydrocarbons, and alkylbenzenes, including benzene, toluene, ethylbenzene, and meta-, para-, and orf/zo-xylenes. Semivolatile analytes included rt-alkanes, isoprenoid alkanes, cycloalkanes, and polycyclic aromatic hydrocarbons and related compounds (naphthalenes, phenanthrenes, and dibenzothiophenes and their alkylated derivatives). Of the additives, only the diesel-fuel additive with the red dye marker was amenable to electrospray ionization. Results indicate the LNAPL consists of closely correlatable diesel fuel at various stages of degradation. All LNAPL samples contained the red dye marker for diesel fuel. None of the samples contained chlorinated solvents associated with industries such as drycleaning or automotive maintenance. Solvents such as acetone, dimethyl ether, and methylene chloride and the gasoline additives methyl-t-butyl ether (MTBE), ethyl-t-butyl ether (ETBE), and t-amyl-methyl ether (TAME) were not found. With one possible exception, no evidence of a different diesel or other hydrocarbon fuel contribution was identified. At one site near the north edge of the main LNAPL body, evidence exists for traces of possible gasoline components in addition to the diesel fuel. The geochemical analysis of the LNAPL and correlations with other fuel products and additives strongly suggest episodic releases of a single, local-source, diesel fuel into the aquifer over an extended period of time.