Methods

Ground-water levels and samples were collected from the shallow ground water at the Terrace Farm site in August 2004 and at the North Farm site in January and April 2004. Samples were collected from the ground-water system, rather than directly from the process water, because the process water is nitrate poor and the isotopic values of the nitrate would change as the organic nitrogen undergoes nitrification to become nitrate.

Water samples from monitoring wells were collected by using submersible pumps. Each monitoring well had a screen-interval length of 3.0 m, and samples were collected at each well only after at least three casing volumes of water were purged and basic field parameters had stabilized. Water samples were collected from temporary push-probe wells at three Terrace Farm sites (Terrace-1, Terrace-2, and Terrace-3; fig. 2).

Hydrogen-isotope ratios of water were analyzed by isotope-ratio mass spectrom­etry following hydrogen-water equilibration (Coplen and others, 1991) and are expressed relative to Vienna Stan­dard Mean Ocean Water (VSMOW), such that delta-2 hydrogen (δ²H) of Standard Light Antarctic Precipitation (SLAP) reference water is -428 parts per thousand (‰) (Coplen, 1996). Oxygen-isotope ratios of water were determined by the carbon dioxide-water equilibration (Epstein and Mayeda, 1953) and are expressed relative to VSMOW, such that δ¹⁸O of SLAP reference water is ‑55.5 ‰ (Coplen, 1996).

Samples for hydrogen-3/helium-3 (³H/³He) age determination (³H, δ³He, ⁴He, and Ne analyses) were collected in 80-cm-long copper tubes with pinch-off clamps and 500 mL plastic bottles, and analyzed by using mass-spectrometric procedures (Ekwurzel and others, 1994; Ludin and others, 1998) at the Noble Gas Laboratory, Lamont-Doherty Earth Observatory, of Columbia University, Palisades, New York. Gases were extracted from the water samples in the copper sample tubes by using a vacuum extraction line and were transferred into flame-sealed glass ampoules. After separation of water vapor and dissolved gases into several cryogenic traps, He isotopes were measured on a VG 5400 mass spectrometer. After release from the cryogenic trap, neon was measured in a quadrupole mass spectrometer. The gas-free water sample was flame-sealed in a glass bulb, and after about 6 months, the tritiogenic ³He was extracted from the glass bulb and measured using a mass spectrometer. The ratio of ³H/tritiogenic ³He then was used to calculate the apparent recharge age of the water (Schlosser and others., 1988, 1989).

Unfiltered samples were analyzed for concentrations of nutrients and major ions by the U.S. Environmental Protection Agency (USEPA) Manchester Environmental Laboratory in Port Orchard, Washington. Samples for the analysis of nitrite-plus-nitrate, ammonia, and phosphate were preserved with sulfuric acid and analyzed by using USEPA Methods 353.2, 350.1, and 365.1, respectively (U.S. Environmental Protection Agency, 1993). Samples for the analyses of magnesium, calcium, potassium, and sodium were preserved with nitric acid (pH less than 2) and analyzed by using USEPA Method 200.7 (U.S. Environmental Protection Agency, 1991). Samples for the analysis of chloride and sulfate were untreated and were analyzed using USEPA method 300.0 (U.S. Environmental Protection Agency, 1993). Filtered water (0.45 mm) was analyzed for isotope ratios of nitrogen and oxygen in dissolved nitrate by bacterial conversion of nitrate to nitrous oxide and subsequent measurement on a continuous-flow isotope-ratio mass spectrometer at the U.S. Geological Survey (USGS) Stable Isotope Laboratory (Sigman and others, 2001; Casciotti and others, 2002; Revesz and Casciotti, 2003). Nitrogen-isotope ratios were reported in parts per thousand relative to nitrogen gas (N₂) in air (Mariotti, 1983). Oxygen-isotopes of nitrate were expressed in the same manner as those of water and were corrected for drift, blank, and exchange with water by using USGS 34 and USGS 35 (Böhlke and others, 2004). Data from field and laboratory measurements for sites listed in table 1 are available at http://waterdata.usgs.gov/wa/nwis/qw. Major ions and nutrient results from samples collected from the Umatilla Ordnance Depot landfill wells on the North Farm site were obtained from SCS Engineers (Mark Varljen, SCS Engineers, written commun., November 2004).

End-member mixing analysis (EMMA) (Christophersen and Hooper, 1992) was used to estimate the water contribution to each well sample from the various end members determined at each site. The EMMA model was developed by: (1) constructing constituent-constituent plots to estimate the number of identifiable end members present at each site and the representative solute concentrations of each end member; (2) completing a principal component analysis (PCA) (SAS Institute, 1990) for various constituent combinations; (3) selecting constituents with the largest proportion of variance explained by the first two components for the end-member mixing analysis; (4) determining the proportions of each end member in each well sample by using a Microsoft EXCEL spreadsheet (available at www.cof.orst.edu/cof/fe/watershd/Documents/Shortcourses/shortcourse/EMMA.xls; and (5) comparing the concentrations predicted by EMMA with the measured concentrations by using linear regression.