Indiana Geological Survey project - Jackson County Project

Image showing IGS seal and the words Indiana Geological Survey, a research institute of Indiana University. Click to go to the home page.

Click for information about these images.

Analysis of Nitrate in Groundwater in Jackson County, Indiana

by Edwin J. Hartke and Denver Harper
December 2003

Overview

An unpublished study conducted by the Jackson County Department of Health (JCDH) in the early 1970s found nitrate concentrations to be near or above U.S. Environmental Protection Agency (USEPA) Maximum Contaminant Level (MCL) in all domestic wells in a major surficial aquifer north of the East Fork White River in Jackson County, Indiana (JCDH, oral communication, 1999). The results of the JCDH study created considerable concern among farmers and other homeowners for whom this aquifer was their only source of drinking water. A follow-up study by the Indiana Department of Environmental Management (IDEM) by Harris (1997) confirmed the results of the work done by the JCDH.

The purpose of this project was to identify sources of nitrogen that are responsible for contamination of the aquifer. Three techniques were employed to identify possible sources of nitrate contamination: (1) an assessment of the hydrogeologic setting, (2) identification of nitrogen inputs related to land-use activities, and (3) analyses of water chemistry and of nitrogen isotopes
(¹⁵N/¹⁴N).

The highly conductive aquifer, which is covered by a permeable layer of alluvium, is rapidly recharged by precipitation -- nearly half of total annual precipitation is recharged to the aquifer. The aquifer discharges to the river except under flood conditions, when flood waters recharge the aquifer.

The sources of nitrate in the shallow ground water are indicated by spatial variations in nitrate concentrations, as well as isotopic results, and their relationships to farm fields receiving fertilizers and to animal-waste sites. Samples collected from wells installed in farm fields indicate inorganic fertilizers or mixed inorganic and animal-waste sources, whereas wells adjacent to and down-gradient from concentrated animal feeding operations (CAFOs) indicate animal-waste sources.

Denitrification occurs in the aquifer as the vertical component of the ground-water flow system carries the nitrate downward through the redoxcline (namely, the zone where the water becomes anoxic). This zone lies between 2.5 and 4.7 m (8 and 13 ft) beneath the water table. Water beneath the redoxcline is essentially free of nitrate and most other toxic contaminants, although most of the deeper wells produce some hydrogen sulfide and precipitates of iron sulfide.

With the implementation of best management practices (BMPs), naturally occurring denitrification and recharge from precipitation make this aquifer a candidate for remediation.

Statement of Problem

Nitrate concentrations at or above USEPA MCL were discovered during a JCDH study in the early 1970s and were confirmed by an IDEM study in 1997 (Harris, 1997). During these studies, water samples were collected from domestic wells in the shallow outwash aquifer that lies north of the East Fork of the White River and northwest of Seymour, Jackson County, Indiana. This is a shallow unconfined aquifer that is very sensitive to contamination. It is also very productive and the sole ground-water source in the area.

Most of the residents of this agricultural area relied on shallow wells for drinking water prior to the discovery of nitrate contamination but now use ground water only for farm needs. Most study-area residents now purchase bottled water for consumption, although a few have installed reverse-osmosis systems. The local school has been connected to a public water-supply system.

This study was commissioned (1) to verify that nitrate contamination is a continuing problem, (2) to determine the sources of nitrogen contamination, and (3) to determine if other inorganic contaminants of concern, or bacteria, are present at levels that are hazardous to human health.

The contaminated wells were located in a 32 km² area that lies within the 14,882 km² East Fork White River watershed in southeastern Indiana. The study area is about 6 km northwest of Seymour in Jackson County in the modern floodplain of the river.

Research Methods

Ground water in the study area is subject to nitrate contamination from several sources, including inorganic nitrogenous fertilizers applied to crops, septic systems (the area is unsewered), farm livestock, and two concentrated animal feeding operations (CAFOs). The latter include a small facility that houses pullets and a large facility that houses hens for egg production.

More than 40 monitoring wells at 23 locations were established in the 40 km² (12-square-mile) study area to identify and map aquifer characteristics, to provide water samples, and to allow continuous monitoring of water levels at 1-hour intervals. A continuously recording precipitation gage was installed in the center of the study area.

At all but five sites, the wells were paired with one well screened at the water table and the other well screened at the base of the aquifer. Screens were 1.6 m (5 ft) in length for all paired wells. Paired wells were installed to determine if the aquifer is continuous, to examine the age difference of the shallow and deep water, and to determine differences in water chemistry with depth.

At Sites 33 and 36, which are situated adjacent to and down-gradient from the egg-wash lagoons and spray field, respectively, of the large CAFO, sets of triple wells with 0.3-meter (1-foot) screens were installed. At Sites 5, 42, and 43, which are in the vicinity of the small CAFO, single wells were installed with continuous screens that extended 9.5 m (30 ft) into the aquifer. The single and triple wells were installed to better define the depth to and thickness of the zone of denitrification (redoxcline).

The 5.1-centimeter (2-inch) PVC monitoring wells were installed through a 15.2-centimeter (6-inch) hollow-stem auger. During withdrawal of the auger flights, the coarse-grained aquifer material collapsed into the annulus as the auger was withdrawn, so that an additional sand pack was not required. The wells were sealed at the surface with bentonite to prevent vertical flow down the annulus. Samples collected from each 1.5-meter (5-foot) auger flight were examined and described to help define aquifer characteristics. The well was then developed using a high-volume suction pump.

Each well received a dedicated submersible pump; a transducer and data logger were installed in at least one well at each site. Each well was protected with an 20.3-centimeter (8-inch) PVC casing founded in a concrete pad and supplied with a lockable cap.

Water samples were collected at least once per quarter throughout the study, which lasted from September 1, 1999 to June 30, 2002. Sample timing was based on fertilizer applications and likely recharge events. The initial sample set was collected in late autumn of 1999, prior to application of fertilizers for the 2000 crop year and during a period when the water table was abnormally low as a result of scant precipitation from summer through autumn. These samples were analyzed for the complete suite of anions and cations, plus heavy metals and Escherichia coli. Later analyses were limited to analytes that are indicators of sources of nitrogen or of the denitrification process.

On-site analyses were accomplished in the field using a YSI multiprobe and flow-through device to measure pH, temperature, specific conductance (SpC), dissolved oxygen (DO), and standard redox potential. The samples were then delivered to the chemisty laboratory of the Indiana State Department of Health (ISDH) for the analyses of inorganic constituents and E. coli bacteria. Selected samples were also sent to the University of Nebraska Water Science Laboratory for isotopic analyses.

Aquifer characteristics were determined through a 24-hour pumping test using an already operational irrigation well as the pumping well and a nearby monitoring well (Site 22) and specially installed piezometer to record drawdown. Water-level data collected from the data loggers, in addition to data from the pumping test, provided flow direction and an approximate flow rate for the ground water in the aquifer.

Nitrogen isotope ratios (¹⁵N/¹⁴N), nitrate concentrations, the hydraulic characteristics of the aquifer (ground-water flow), and denitrification within the aquifer were used to determine the source and fate of the nitrogen.

Samples were collected for tritium analyses from both shallow and deep wells by a staff member of the concentrated animal feeding operation. These samples were analyzed by Tritium Analytics in West Lafayette, Indiana. The results, recorded in Tritium Units (TUs), were used to evaluate the age of the ground water.

Findings

Hydrogeologic Setting

The study area is part of a glacial valley-train aquifer that consists of fine to coarse sand with some gravel. Drilling logs show that the aquifer ranges in thickness from 0 m at the northern boundary to 22 m (70 ft) in the study area, and is overlain by 2 to 3.7 m (6 to 12 ft) of river-deposited overbank material (fine to medium sand with variable silt, clay, and some gravel). The aquifer material was deposited on discontinuous lake clay that is up to 11 m (36 ft) thick. The wide but shallow bedrock valley (approximately 31 m [100 ft] deep) in which this aquifer lies was incised into a thick section of organic-rich shale that is predominantly black to dark gray in the study area. The aquifer thins toward the north and pinches out at the edge of the bordering glacial-till- and loess-covered upland. It is thickest near the center of the study area where the underlying clay layer thins or is absent in places.

That the aquifer is continuous both vertically and laterally is shown by the essentially identical water levels recorded in the paired shallow and deep wells. The depth to the water table in this unconfined aquifer varies spatially and temporally with surface elevation and in response to precipitation events and seasonal changes. The water level ranges from a low of 2 to 3.8 m (6 to 12 ft) beneath the surface during the growing season to above the surface during flooding of low-lying areas (Sites 11 and 12). When the study began in the fall of 1999, the water level in the aquifer was unusually low following several months of drought; the level was about 45 cm (18 inches) below the seasonal lows of 2000 and 2001. The water level followed normal seasonal patterns in 2000 and 2001, falling during the growing season and rising to annual highs during the late fall and winter.

The water table slopes toward the river and slightly down-valley in the direction of surface-water flow with an average gradient (S) of about 0.0011. Results of pumping tests indicate that the hydraulic conductivity (K) of the aquifer is about 1x10^-1 cm/sec (8x10^-2 ft/sec), whereas the hydraulic conductivity of the overlying river-deposited material is about an order of magnitude less (1x10^-2 cm/sec [8x10^-3 ft/sec]). The ground-water flow rate (K x S) within the aquifer, therefore, averages about 34 m/yr (110 ft/yr) to the south-southwest toward the river. Vertical velocity was not determined in this aquifer, but Engesgaard and Jensen (1990) recorded vertical velocity in an aquifer in a very similar hydrogeologic setting (Rabis aquifer, Rutland, Denmark) as about 0.75 m/yr (2.4 ft/yr).

During flood stage, the water levels in the two wells nearest the river at Sites 11 and 13 (at distances of approximately 560 m [1,800 ft] and 840 m [2,700 ft], respectively) rose in response to the rise in river level. Records from these two wells also indicate that the water level rose above the ground surface during a flood in December 2001. In effect, the earth materials were completely saturated to the land surface, so that the ground water was essentially continuous with the flood water.

Precipitation during the period of the study averaged about 100 cm/yr (39.4 in/yr), and recharge to the aquifer averaged about 46 cm/yr (18 in/yr). The recharge rate was determined by comparing precipitation to rises in water level in the aquifer. Because the cover material is quite permeable and the land surface practically flat, recharge to the aquifer is direct and rapid if sufficient antecedent moisture is present in the vadose zone. The rapidity with which recharge occurs indicates that the aquifer is very sensitive to contamination. For example, a 7-cm (2.7-in) rainfall that occurred during a 2-day period in the fall of 2000 when vegetation was dormant caused a 28-cm (11-in) rise in the water table within 2 days.

Tritium is a naturally occurring radioisotope that was also produced by testing of thermonuclear bombs in the atmosphere after 1952. Tritium was used as a tool to determine time constraints on recharge to the aquifer. Tritium values determined from water samples taken from both shallow (6 to 8 TUs) and deep (9 to 15 TUs) wells indicate that the shallow water is less than 5 to 10 years old, while the deep water is on the order of 15 to 30 years old. The age difference does not explain the absence of nitrate at depth because nitrogen fertilizers have been in use for more than 30 years. Therefore, denitrification within the aquifer must be considered.

Nitrogen Application

Application rates of inorganic nitrogenous fertilizers ranged from lows of 1 to 28 kg/ha/yr (1 to 26 lbs/acre/yr) on soybeans to highs of 380 kg/ha/yr (340 lbs/acre/yr) on an experimental field of high-density corn. Other cornfields received between 110 and 200 kg of nitrogen/ha/yr (101 and 180 lbs of nitrogen/acre/yr) for year 2000.

All chicken manure from the CAFOs is stored under cover on a concrete floor and is periodically removed from the study area. Water that has been used in the egg-washing process and that contains nitrogen is stored in lagoons and periodically applied to the land through a center-point irrigation system located in a 22-hectare (55-acre) field immediately to the west of the large CAFO. Since 2001, the lagoons, which were formerly lined with clay, have been lined with a synthetic liner. The rate of application of the nitrogen-containing egg-wash water to the land is unknown, except for the period from September 1997 through June 1998. During that period, an average of about 190 kg (420 lbs) of nitrogen/acre/month were applied, based on records on file at IDEM.

Water Chemistry

Analyses of samples collected from the system of monitoring wells indicate high nitrate concentrations in the shallow wells across much of the study area. These high nitrate concentrations occur as a result of nitrification, which takes place under aerobic conditions in the unsaturated zone and in the upper part of the aquifer. Nitrification is the process through which ammonia is oxidized to nitrite and nitrate. Nitrosomonas bacteria convert ammonia in soil to the nitrite ion (NO₂^-), which generally is considered to be an intermediary nitrogen product. Nitrobacter bacteria quickly convert NO₂^- to nitrate (NO₃^-) (Cornell Cooperative Extension, 1997).

Samples from the deep wells have very low levels of nitrate. This is a result of denitrification, which occurs in the anaerobic zone when DO concentrations drop below 2 mg/L or Eh potentials fall below 0.7 v (Clark and Fritz, 1997). Denitrification is generally carried out by the bacterium Thiobacillus denitrificans in the presence of organic carbon, although others can denitrify in the absence of organic carbon, using electron sources such as Mn²⁺, Fe²⁺, sulfide, and CH₄ (Scholten, 1991). Nitrogen gas (N₂^-) is returned to the atmosphere as a result of this process.

Isotopic end-member values were determined for various potential nitrate sources at the outset of the study. Samples of each type of fertilizer, a sample of cow manure collected from the farmers, and a sample of chicken manure from the CAFO were analyzed at the University of Nebraska Water Sciences Laboratory. The average ¹⁵N/¹⁴N for chicken manure was 7.9, for cow manure was 2.6, for pelletized inorganic fertilizer was -1.8, and for liquid organic fertilizer was 0.8.

Nitrogen isotope ratios (¹⁵N/¹⁴N) determined from shallow-well samples show that two or more sources are contributing nitrogen to the aquifer. Most of the wells installed in, or adjacent to, fertilized fields produced ¹⁵N/¹⁴N values that indicate an inorganic (fertilizer) source, although some show mixed inorganic and organic (human or animal waste) sources. Wells installed adjacent to and down-gradient on the water table from the confined feed lots indicate an organic nitrogen source. Two wells installed near homes indicate mixed sources, but these wells are in close proximity to septic systems and one is also near a small cattle-holding pen. Certain other wells installed adjacent to farm fields provided mixed-source ¹⁵N/¹⁴N ratios. The source of organic nitrogen found in these wells is unknown.

Very high levels of nitrate were discovered in shallow wells installed adjacent to and downdip from the small CAFO and the egg-wash waste-water spray field. Samples collected from Site 36, which is adjacent to the spray field, returned nitrate values of as high as 55 mg/L NO3-N at a depth of 2.6 m (6 ft) beneath the water table. Samples from Site 43, which is adjacent to the small CAFO, indicated a nitrate concentration at the water table of 114 mg/L. Water samples from Site 42, which is 400 m (1,300 ft) downdip from the small CAFO had nitrate concentrations of 15 mg/L at the water table and 38 mg/L at both 2.2 m (7 ft) and 2.8 m (9 ft) beneath the water table. Isotopic values, which were light at the water table and heavy at the 2.2- and 2.8-m (7- and 9-ft) levels indicate a fertilizer source at the water table and an animal-waste source at depth.

The seasonal application of inorganic fertilizers does not produce any clear short-term changes in nitrate concentrations. The maintenance of nitrate concentrations at approximately the same level throughout the study suggests that an approximate equilibrium has been reached between nitrogen loading to the system and the ability of the system to dilute or denitrify that load. This state of approximate equilibrium may be a product of increased evapotranspiration and reduced recharge during the late spring and early summer seasons when most fertilizer is applied, compared to greater recharge with some continued leaching of nitrate during the fall and winter. These conditions, in combination with the relatively slow lateral flow of water in the aquifer (34 m/yr [112 ft/yr]), may act to maintain the nitrate concentrations at a reasonably constant level with only minor fluctuations.

Summary

Nitrate contamination is present in the shallow part of the aquifer across much of the study area even though nitrogen application rates have recently been reduced somewhat (Jackson County farmers, oral communication, 1999). The absence of any trend in nitrate concentrations during the study indicates that some medium-term balance had been achieved for nutrient loading in the aquifer. Nitrate was either below detection levels or present in minor concentrations in water samples from the deep monitoring wells.

Both inorganic nitrogen (chemical fertilizer) and organic nitrogen (animal wastes) were contributing to the problem. Nitrate in shallow wells installed in cropped fields was derived from fertilizer, whereas nitrate in shallow wells immediately down-gradient from the CAFOs was derived from animal wastes. Nitrate in wells influenced by septic systems or farm animals and fertilizer showed mixed sources.

Nitrate is being removed through denitrification in the anoxic zone of the aquifer as the water migrates downward. This reaction takes place between about 2.5 and 5 m (8 and 15 ft) beneath the water table.

Wells that are down-gradient from the two CAFOs (Sites 36, 42, and 43) indicate that contaminant plumes containing elevated concentrations of nitrate originated from beneath these facilities and were migrating toward the river. These plumes had components of both horizontal and vertical flow and were driven through the aquifer toward the river by recharge from precipitation and gravity flow. The tops of the plumes dipped further beneath the water table as they flowed toward the river and were diluted by recharge from precipitation. The bases of the plumes were determined by the position of the redoxcline, except beneath the small CAFO where the redoxcline appeared to be unable to fully denitrify the very highly concentrated nitrate. A more detailed study could define the full extent of these plumes. Based on the horizontal flow rate calculated during this study and projected vertical flow rate, inferred from Engesgaard and Jensen (1990), and continued denitrification in the redoxcline, these plumes could dissipate in 8 to 10 years if no further nitrogen is introduced to the system.

Nitrate concentrations in water samples from shallow wells fluctuated from one sampling period to the next and did not exhibit any clear trends. The lack of any trend may indicate a balance between the addition of nitrogen through fertilizers and animal wastes and the removal of nitrogen through uptake by plants and downward percolation of infiltrating precipitation.

The balance of the nitrate concentration in ground water can be affected by changes in the types and amounts of fertilizers that are applied and the timing of their application. More careful attention to such details is referred to as "best management practices" (BMPs). Based on infiltration rates and aquifer hydraulics, implementation of BMPs has the potential to reduce nitrate concentrations below EPA MCL (10 mg/L NO3-N) in the contaminated zone of the aquifer within 8 to 10 years.

Project Participants and Presentation of Findings

Noel Krothe, Department of Geological Sciences (DOGS), Indiana University, was a co-investigator on the project. Brian Motzel, formerly a graduate student in DOGS, produced an M.S. thesis based on the project. Steve Loheide and Catherine Talbot, also former graduate students in DOGS, collected and analyzed data. John Haddan, Indiana Geological Survey, supervised the field program.

The Cortland Neighborhood Association (CNA) and Rose Acre Farms (RAF) participated in the field operations of the study. CNA farmers were particularly supportive providing access to their fields for installation of monitoring wells and for providing detailed information about the crops that were planted and the agricultural chemicals that were applied during the term of the study. RAF environmental personnel collected duplicate water samples and supplied the tritium data that was used in the study.

General chemical and heavy metal analyses were performed by the Chemistry and Environmental Laboratory of the Indiana State Department of Health (ISDH), and the isotope samples were analyzed by the Water Sciences Laboratory at the University of Nebraska. E. coli testing was done at the Environmental Microbiology Laboratory at the ISDH.

Funding was provided by the Indiana Department of Environmental Management (IDEM) as FFY 2002 Section 319 Grant No. C9995008-99-0 (ARN 99-206). A final report was submitted to IDEM in July, 2002.

References

Clark, I.D., and Fritz, P., 1997, Environmental isotopes in hydrology: Boca Raton, Florida, CRC Press, Lewis Publishers, 328 p.

Cornell Cooperative Extension Service, 1997, Something to grow on (the nitrogen cycle): Cornell University, Department of Floriculture and Ornamental Horticulture.

Engesgaard, P., and Jensen, K.H., 1990, Flow and transport modeling, Rabis Creek, Forsk. fra Miljostyrelsen B13: Copenhagen, Environmental Protection Agency.

Harris, J., 1997, High nitrates, Jackson County: Indiana Department of Environmental Management, Ground Water Section case summary report.

Motzel, B.C., 2001, A hydrochemical and stable isotopic study of ground water impacted by nitrogen and sulfur sources in an unconfined outwash aquifer in Jackson County, Indiana: Bloomington, Indiana University, M.S. thesis, 218 p.

Motzel, B.C., Krothe, N.C., Hartke, E.J., Harper, D., Spalding, R.F., Talbot, C.B., and Loheide II, S.P., 2001, The use of nitrogen and sulfur isotopes in a shallow outwash aquifer: XXXI International Association of Hydrogeologists Congress, Vol. 1, p. 595-599.

Scholten, S.O., 1991, The distribution of nitrogen isotopes in sediments: Netherlands, Rijksuniversiteit, Ph.D.dissertation, 99p.

For additional information, you may contact either
Denver Harper (e-mail: dharper@indiana.edu) or
Edwin Hartke (e-mail: ehartke@indiana.edu)
at the Indiana Geological Survey, 611 N. Walnut Grove, Bloomington, IN 47405.