Nathan E. Derby, Raymond E. Knighton, and Dean D. Steele
Research Specialist, Soil Science Department;
Associate Professor, Soil Science Department; and
Assistant Professor, Agricultural Engineering Department;
North Dakota State University, Fargo, ND 58105
Non-point source pollution of surface and ground water has become a topic of increased controversy and importance in recent years, with agriculture being fingered as a primary contributor. The threat of ground water contamination increases under irrigation on sandy soils because mobile contaminants such as nitrate are more easily leached through sandy soils, compounded with possible excess water applications. In an effort to reduce non-point source pollution from agricultural lands, producers are encouraged to adopt Best Management Practices, i.e., farming practices capable of reducing nutrient contamination of surface and ground water. These practices are based on research results and field experience and may be as simple as following fertilizer recommendations and irrigation scheduling.
A field scale study of Best Management Practices (BMP's) for irrigation and nitrogen fertilizer use on corn was initiated in 1990 near Oakes, ND (Steele et al, 1992). A primary objective of the study is to develop BMP's and to determine the impacts of those BMP's on ground water quality in the shallow aquifer underlying sandy soils typical of the Oakes Test Area.
The study area comprises approximately 53 irrigated hectares instrumented to monitor soil, soil leachate, ground water, and tile drain effluent quality. Instrumentation in the irrigated portion of the site consists of 16 undisturbed profile lysimeters, 4 disturbed profile lysimeters, 4 multi-level osmotic well samplers, 54 ground water observation wells, and two sampling points on each of two subsurface tile drains. The study is funded by the U.S. Bureau of Reclamation, operated by NDSU Departments of Soil Science and Agricultural Engineering, and farmed by Hokana Farms (Herman Meyer, landowner).
Quality of the ground water, soil leachate, and soil with respect to nitrate have been intensively monitored over the course of the project. Although the purpose of this paper is to give a preliminary summary of observations made in regard to nitrate over the first four years of this project, a brief description of the instrumentation is included for those not familiar with the project.
Field scale soil and ground water monitoring.
Prior to initiation of research on the BMP field, background information regarding soil and ground water nitrate concentrations was needed. In the fall of 1988, a 100 meter grid was established and temporary ground water observation wells installed (Figure 1). Soil samples from the bore holes were collected and analyzed for inorganic nitrogen. Water samples from the ground water were taken for nitrate analysis. The grid was established again in the fall of 1992 to determine the extent of changes in soil and ground water nitrate over the first three years of the study.
Tile drains
Two sub-surface headwater tile drains underlay the BMP study field running from south to north. Each drain has two access manholes where samples for quality analysis are taken and flows are read.
Nested observation wells.
Ground water quality on two irrigated transects has been intensively monitored since installation of nested observation wells in the fall of 1988. Two east-west transects, one 200 meters south of the north edge of the field (C transect) and the other 200 meters north of the south edge of the field (G transect), were instrumented with permanent observation well nests at 100 meter intervals. Each nest of wells consists of three wells; one screened at the water table (shallow), one screened at approximately 0.5 meters into the aquifer (medium), and one screened just above the clay till barrier (deep). These wells have been sampled for nitrate approximately monthly since the project started. Typical well nests are depicted in Figure 2, denoted as C6 and C7.
Multi-level samplers.
Four wells containing multi-level osmotic samplers (Olson and Knighton, 1992) were installed between two well nests on the C transect (6A, 6B, 6C, and 6D in Figure 2). Each sampler consists of 8 sample collection vials at 0.15 m increments starting at the water table to a depth of approximately 1.2 m into the aquifer. The purpose of the multi-level sampler wells is to more closely monitor the stratification of nitrate in the upper layers of the ground water.
Lysimeters.
Undisturbed profile lysimeters (Derby and Knighton, 1992) were installed in the fall of 1989 along the same east-west transects as the nested observation wells (eight per transect) to monitor leachate quality and quantity. These lysimeters consist of a soil core contained in a 2 meter length of 60 cm diameter PVC pipe and instrumented to allow drainage samples to be collected from both a gravity drainage and vacuum system. They are located approximately 30 cm below grade to allow normal cropping practices to continue.
Four disturbed profile lysimeters (two per transect) are also installed on the C and G transects. They are reconstructed soil profiles contained in a 1.2 m X 1.5 m steel box instrumented to collect drainage water via gravity flow for quality analysis. They are also buried below grade.
All lysimeters have been sampled on a weekly basis during the growing season since 1990 to monitor the quantity and quality, primarily nitrate, of the soil water draining past the root zone.
Irrigation scheduling.
The BMP field is divided into quasi-quadrants, each receiving a different water management treatment resulting from four irrigation scheduling methods. Two quadrants were managed as water balance regimes, one to replace 80 percent of estimated ET and one which replaced 100 percent of estimated ET up to the blister kernel stage and 60 percent ET replacement after blister. The third quadrant irrigation was scheduled according to CERES-Maize model predictions. The last scheduling method was based on real-time sensor feedback, utilizing infrared canopy temperature and tensiometer measurements. Because the past growing seasons have been wetter and cooler than normal, the four scheduling methods resulted in small differences of applied water. Hence, the effect of irrigation scheduling method on nitrate distribution will not be discussed further in this report.
Field scale soil and ground water nitrate.
The initial survey of soil nitrate on the 100 m grid in November 1988 showed higher concentrations in the 0-15 and 30-60 cm depth increments (Figure 3). The highest nitrate concentration of 60 mg/kg was found in the 30-60 cm depth increment. In October 1992, when the 100 m grid was reestablished, soil nitrate was found to have decreased in the 0-15 and 30-60 cm depth increments but increased slightly at the lower 180-210 cm depth, with a maximum concentration of 24 mg/kg in the 180-210 cm depth increment. This trend would suggest that with intensive soil sampling and aggressive fertilizer nitrogen management, we have been able to crop down the soil nitrate.
Ground water nitrate concentrations were relatively low (<40 mg/L) in April 1989 when the initial samples were taken from the temporary observation wells on the 100 m grid. The next time the grid wells were installed and sampled in April 1993, the ground water nitrate concentrations had increased to a maximum of 120 mg/L, although highly variable over the entire field (Figure 4). It should be noted that only the upper 60 cm of the ground water was sampled with the temporary wells.
Tile drains.
Tile drainage effluent nitrate concentrations have shown a very slight increase since 1989. This would follow from the increase in ground water nitrate measured at the temporary observation wells. The average nitrate concentration at the four tile drain access points was approximately 3 mg/L in the fall of 1989. Over the past four years, that average has increased to approximately 5 mg/L, still well below the EPA standard for drinking water of 10 mg/L. Although the nitrate concentrations have been fairly stable over the past 4 years, one dramatic spike was observed at all four access manholes in the spring of 1991 when spring fertilizer application was followed by an intense rainfall event (Figure 5).
Two hypothesis are being considered to explain the spike in nitrate concentration in the tile drain effluent. One explanation is the existence of preferential flow pathways in which the soil water bypasses the majority of the soil mass and proceeds directly to the ground water in times of very intense percolation. This in result transporting nitrate from the upper layers of the soil profile to the water table.
The other hypothesis is based on the assumption that a bio-film is present in the tile drain and is immobilizing or reducing (denitrification) the nitrate present in the effluent. The increase in flow in effect overloaded the organisms capacity to remove nitrate. This hypothesis would also explain the discrepancy in the tile drainage nitrate versus the nitrate in the ground water observation wells, as they are essentially sampling the same portion of the aquifer. Comparison of nitrate in the wells versus tile drainage to chloride in the wells versus tile drainage supports the hypothesis that the bio-film is present and responsible for the low tile drain effluent nitrate concentrations (Figure 6).
Total nitrate transported from the quarter section through the tile drains in 1990, 1991, 1992, and through September of 1993 was 122, 258, 233, and 170 kilograms, respectively. It is impossible to report nitrate loss per hectare as the exact area of influence of each tile is unknown.
Nested observation wells.
Water quality samples are taken from the nested wells approximately monthly for nitrate analysis. The ground water nitrate concentrations were relatively low in 1990 and have fluctuated substantially in subsequent years, although the average trend has been an increase in nitrate. Even though the nitrate concentration in the ground water has been increasing, essentially all of the nitrate has remained stratified in the upper 30-60 cm of the aquifer. Average, maximum, and minimum nitrate concentrations in the ground water measured in samples from nested wells is shown in Table 1. The first section of Table 1 includes all wells (deep, medium, and shallow) while the second section is limited to the shallow wells. Note that the maximum nitrate concentration is found in the shallow wells with the exception of 1990 and 1991 when the maximum nitrate was detected in a medium well.
Table 1. Nitrate concentrations in ground water observation well water samples for 1990- 1993 seasons.
Deep, medium, and shallow well nitrate Year Average Maximum Minimum mg/L mg/L mg/L 1990 13.59 148.60 0.00 1991 21.43 140.60 0.00 1992 23.74 168.40 0.00 1993 20.24 124.68 0.00 Shallow well nitrate Year Average Maximum Minimum mg/L mg/L mg/L 1990 22.76 107.20 0.02 1991 38.53 118.56 0.00 1992 37.77 168.40 0.00 1993 27.26 124.68 0.04
Multi-level samplers.
The multi-level samplers have shown a great fluctuation in nitrate concentrations since they were installed in 1991. The trend of nitrate stratification observed in the nested wells is also evident in the multi-level sampler wells. The well designated 6D is located next to a tile drain and indicates the mixing effect the tile drain has on ground water nitrate concentration, i.e. the nitrate is no longer stratified but is distributed deeper into the ground water near the tile drain (Figure 7).
Lysimeter leachate.
The sixteen undisturbed profile lysimeters installed along two lines transecting the field from east to west, have been sampled weekly between April 1 and October 31 each growing season. Leachate nitrate concentration from the undisturbed profile lysimeters was greatest in 1990 with a maximum of approximately 300 mg/L and has decreased each successive year to an average of <20 mg/L in the fall of 1993 (Figure 8). Total nitrate lost in the leachate in 1990 averaged 188 kg/ha and has declined to 31 kg/ha lost in 1993.
The peak of nitrate concentration in 1990 represents the combined effects of increased mineralization on newly irrigated soils and high residual soil nitrate in the area where the existing corn crop was destroyed prior to lysimeter installation.
The disturbed or re-constructed profile lysimeters located along the same transects have also been monitored for leachate nitrate on a weekly basis. These lysimeters show the same trend as the undisturbed lysimeters in that the nitrate concentration came to a peak and is subsequently decreasing. The nitrate peak, however, came in 1991 for the disturbed lysimeters (Figure 9). In 1991, an average of 84 lb N/acre was lost through the disturbed lysimeters compared to only 24 lb/acre in 1993.
Best Management Practices have shown to be effective in reducing nitrate leaching through sandy soils under irrigation as illustrated by lysimeter leachate data. Now that the slug of nitrate due to increased N mineralization on newly irrigated soil has been taken up or leached, it is expected that the ground water nitrate will begin to decline and fall in line with the lysimeter leachate concentrations. It is important that work of this nature be continued to determine whether the ground water nitrate concentrations will in fact begin to decline to acceptable levels. Studies are also in place to study the tile drain bio-film and the preferential flow hypothesis.
It is obvious that research of this nature is necessary in developing BMP's and in determining the effects of irrigation and fertilization practices on ground water quality, especially under sandy soils.
Derby, N.E, and R.E. Knighton. 1992. Monitoring leachate losses with large undisturbed profile lysimeters. In Proceedings of North Dakota Water Quality Symposium, Mar. 25-26, 1992. Bismarck, North Dakota.
Steele, D.D., E.C. Stegman, L.D. Prunty, and R.E. Knighton. 1992. Overview of a field study of best management practices for improved irrigation and fertilizer use efficiencies. In Proceedings of North Dakota Water Quality Symposium, Mar. 25-26, 1992. Bismarck, North Dakota.
Olson, J.M. and R.E. Knighton. 1992. Monitoring and stratification of nitrogen in a surficial unconfined aquifer. In Proceedings of North Dakota Water Quality Symposium, Mar. 25-26, 1992. Bismarck, North Dakota.