Nitrate at Wastewater-Application Sites

Terrace Farm Site

Water levels measured from wells at the Terrace Farm wastewater-application site indicated that ground water generally flows towards the northwest (fig. 2A). Depths to water at the Terrace Farm site ranged from about 15 m at well MW-40 to 26 m at well MW-53 and then shallowed to around 8 m below land surface downgradient of the cattle feedlot (fig. 2B).

Apparent recharge ages ranged from 1.6 to 49 years in water samples from wells in the wastewater-application site and averaged about 8 years for water samples from wells downgradient of the feedlot (fig. 2; table 1). Apparent recharge ages were uncorrected for terrigenic helium because the samples may have been fractionated. The tritium concentration values were consistent with piston flow at several of the wells (Terrace-1, Terrace-2, and MW-53) because the tritium concentration values were about equal to those expected in precipitation for their given recharge year.

Conversely, at wells MW-39 and MW-40, the tritium concentration values indicated a mixture of young and old waters. The δ¹⁸O and δ²H ratios of water sampled from wells were along the local meteoric water line (constructed from samples collected from nearby rivers and streams), indicating that the aquifer water was not notably affected by evaporation (fig. 3).

Constituent-constituent plots that were constructed for relatively conservative ions at the Terrace Farm site indicated that three identifiable, potential end members were present (fig. 4). The end members included a group with low concentrations of sulfate, chloride, and calcium (less than 100 mg/L)and low δ¹⁸O of water; a group with higher concentrations of sulfate, chloride, calcium and slightly higher δ¹⁸O of water; and Terrace Farm water, which had a high concentration of sulfate, but a low concentration of chloride and a much higher δ¹⁸O of water. The median concentrations of constituents in water from wells MW-39 and MW-40 were selected to represent the group with lower concentrations of sulfate, chloride, and calcium. This end member likely is representative of the upgradient ground water because these wells generally are on the upgradient side of the study area and away from the irrigation canal. The median concentrations of constituents in water from wells MW-22 and MW-52 were selected to represent the water with higher concentrations of sulfate, chloride, and calcium and slightly higher δ¹⁸O ratio of water. The source of this end member is unknown, but may be ground water flowing from west of the site, and (or) may be affected by irrigation canal leakage or application of road salt because these wells are screened near the water table and are near the irrigation canal and road. On a molar scale, the largest increase in concentration was for the chloride ion (4.5 millimolar) between wells MW-22 and MW-52 relative to wells MW-39 and MW-40 (table 2). Additionally, the increases in concentrations of calcium, magnesium, alkalinity, and sulfate were not consistent with the dissolution of a combination of carbonate minerals and gypsum. It is unlikely that geochemical reactions with aquifer minerals were responsible for the increases in the concentrations of most major ions in water from wells MW-22 and MW-52.

Well MW-53 was selected to represent the Terrace Farm water and because it was geochemically isolated from other wells sampled. The water in well MW-53 likely represents the ground water from the wastewater application on the Terrace Farm, because of its young apparent age, lack of mixing with older waters, and heavier δ¹⁸O and δ²H values, which may indicate recharge during summer when the wastewater is applied. The increase in calcium and sulfate concentrations in water from well MW-53 relative to wells MW-39 and ‑40 is consistent with the dissolution of about 230 mg/L of gypsum. Terrace-1 was excluded from the EMMA because it seemed to be affected by heterogeneity or a possible fourth source of water that may originate from the west side of the study area; Terrace-1 frequently plotted outside the mixing lines for the other three end members.

The PCA selected for use in the EMMA incorporated four constituents (calcium, chloride, sulfate, and δ¹⁸O of water) for the Terrace Farm site because the first two principal components explained about 99 percent of the variability, and the remaining conservative constituents were strongly correlated with at least one of the four incorporated constituents. For each constituent included in the EMMA, the linear-regression relations between the predicted and measured concentrations produced slopes between 0.93 and 1.0 and coefficient of determination (R²) values between 0.98 and 0.99 (table 3). The R² values and slopes close to 1 indicated that the EMMA model is a strong predictor of ground-water concentrations for the conservative solutes. To determine how robust the analysis was, the results from the EMMA also were used to predict concentrations for constituents that were not included in the EMMA. The conservative constituents (magnesium, bromide, and δ²H) had high R² values and slopes near 1, similar to the constituents that were included in the EMMA. Bicarbonate, nitrate, sodium, and potassium, however, had much lower slopes and R² values, indicating that these constituents probably were nonconservative.

EMMA indicated that most samples included a mixture of water from upgradient and western ground water that was affected by irrigation-canal leakage. A large proportion of water from Terrace Farm (table 1) was inferred by the EMMA only in samples collected from two wells (MW-51 and MW-53). The sample from well MW-51 contained a mix of about 60 percent upgradient ground water and 40 percent Terrace Farm water, which is consistent with its location on the upgradient side of the study site next to the wastewater-storage lagoon. As much as 5 percent of water from the Terrace Farm was detected in three wells (MW-38, Terrace-2, and Terrace-3). All these wells were on the downgradient side of the wastewater-application area.

The δ¹⁵N of nitrate generally ranged between +2 and +9 ‰ and the δ¹⁸O of nitrate generally ranged between -2 and -6 ‰ (table 1). Well MW-53, which represented the wastewater, had a δ¹⁵N of +2.36 ‰ and a δ¹⁸O of -2.06 ‰. The measured isotope values at well MW-53 overlap with previously reported values for soil nitrogen (N), animal and septic waste, and ammonium fertilizer (fig. 5). Values in this range generally are the result of nitrification. Isotope values resulting from nitrification of organic nitrogen in the food-processing wastewater determined in this study also are in this range.

Because of the overlap in the potential nitrate sources, the proportion of each sample attributed to the Terrace Farm ground water was compared to the isotope values to determine any statistically significant relation (fig. 6). No significant relation was determined between either δ¹⁵N or δ¹⁸O and the proportion of the Terrace Farm water (p-values = 0.68 and 0.07, respectively). For water samples from wells in the wastewater-application area with no Terrace Farm water, the nitrate-isotope values likely were primarily the result of ammonium fertilizer or nitrification of soil N.

North Farm Site

Interpretation of water-quality data from the North Farm site was more difficult than interpretation of water-quality data from the Terrace Farm site because the types and amounts of water applied to the North Farm site changed over time. From 1972 to 1992, about 3.4 million m³ of process water were applied each year to the North Farm site with an additional 33,300 m³ of supplemental surface-water application. The alluvial wells at the wastewater-processing plant are the original sources of pre-processed water. Beginning in 1992, the amount of processed water applied annually to the North Farm wastewater-application site decreased to 284,000–370,000 m³ with an additional 1.4–1.9 million m³ of supplemental surface-water application (Emmett Walker, ConAgra Foods, oral commun., July 2006). Because the source of water has changed over time, application of the EMMA would not be appropriate at this site. Piper diagrams and constituent-constituent plots were examined to determine which wells, if any, were affected by the wastewater applications.

Ground-water levels collected at the North Farm wastewater-application site indicated that ground water flowed radially away from the center of the ground-water mound beneath the site with a southwesterly flow direction beneath the general study area (fig. 7A). Depths to water generally ranged from about 26 m on the edge of the wastewater application area to more than 45 m southwest of the Umatilla Ordnance Depot landfill (fig. 7B). Unlike the water samples from wells collected at the Terrace Farm site, several samples collected downgradient of the North Farm site were below the local meteoric water line and may be affected by evaporation (fig. 3). Whether the evaporative signal of these samples was the result of changes to the ground water prior to pumpage and use in food processing or changes to the water after it was applied to the site is not known.

Wastewater applied at the North Farm site is particularly rich in potassium and sodium (fig. 8) compared with most of the ground-water samples. Wells MW-3, MW-4, MW-33, and 11-5 all appear to be affected by the processed water because they plot between the wastewater samples and remaining well samples as shown in figure 8.

Sulfate levels generally were low in the wastewater and supplemental surface-water irrigation (fig. 9). With the exception of samples from well MW-10, the samples clustered into two groups: (1) wells with low sulfate concentrations, which seemed to be affected by the wastewater, and (2) wells with high sulfate concentrations, which did not seem to be affected by the wastewater. Samples from wells 11-3, 11-4, 11-5, MW-3, MW-4, and MW-33 all had calcium and sulfate concentrations that were between those detected in the wastewater and those in the supplemental-irrigation water.

Well MW-10 is on the edge of the wastewater-application area and is screened within a silt layer. Compared to wells affected by wastewater (low sulfate and high potassium), the water chemistry of this young water (5 years) from the shallow well MW-10 is perplexing. The highest concentrations of calcium, magnesium, alkalinity, chloride, and nitrate were detected in water from well MW-10 for the North Farm site, but potassium concentrations were within the range determined for wells affected by the wastewater. When compared to the wastewater affected wells, water from well MW-10 contained significantly more calcium, magnesium, alkalinity, and nitrate with lesser amounts of chloride and sulfate (table 4). The confinement of this water in the silt layer may have allowed geochemical reactions such as dissolution of gypsum and calcium-magnesium carbonates and the remineralization of total Kjeldahl nitrogen to nitrate to proceed faster relative to the flow of water. The addition of significant amounts of calcium, magnesium, and alkalinity shifted MW-10 water to a mixed-cation/bicarbonate type (fig. 8). Alternatively, water from MW-10 may reflect flow from surface-water irrigation applied north of the North Farm site.

The apparent-recharge age of the water just downgradient of the wastewater-application area in water samples from well MW-10 was 5 years. Further downgradient from the wastewater-application area, the apparent recharge dates of the three Umatilla Ordnance Depot wells, wells 11-3, 11-7, and MW-33, were greater than 50 years (table 1). The old apparent recharge age in these wells likely is because the source of most of the water applied to the North Farm site originally was ground water, which in this case appears to be tritium dead. The young age of the water in well MW-10 likely was due to surface-water irrigation either on the North Farm or the area north of the North Farm site. The lack of tritium in water samples from wells MW-33, 11-3, and 11-7 combined with δ¹⁸O and δ²H ratios of water below the local meteoric water line indicated that the water in these wells did not have a substantial amount of supplemental surface-water irrigation. The water in these wells likely was from the wastewater applied prior to 1992 when the supplemental surface water application amounts were less than 1 percent of the total water applied.

The δ¹⁵N of nitrate in sampled wells generally ranged between +2 and +6 ‰ and the δ¹⁸O of nitrate generally ranged between -3 and -7 ‰ (fig. 10). The heaviest isotope values, with a δ¹⁵N of +6.27 ‰ and a δ¹⁸O of -3.06 ‰, and the highest nitrate concentration were detected in the water sample collected from well MW-10. The samples from wells MW-3, MW-4, MW-33, 11-3, and 11-4, which are believed to be affected by the process water, also had slightly heavier values of δ¹⁸O than the remaining downgradient samples, but the values of δ¹⁵N were similar. The samples affected by wastewater generally did not have distinctly different isotope values of nitrate when compared with samples that were not affected by wastewater. As at the Terrace Farm wastewater-application site, the measured isotope values at the North Farm wastewater-application site overlap with previously reported nitrate isotope values for soil N, animal and septic waste, and ammonium fertilizer (Kendall, 1998) (fig. 5).