Characterization of Atrazine in Source Water: A Case Study From Kansas

Table of Contents

Abstract
Introduction
Atrazine, A Triazine Herbicide
Methods of Study
Results of Sampling
Conclusions and Implications
Acknowledgments
References

Figures

Figure 1. Map showing location of the Little Arkansas River Basin and pertinent sites and boundaries.
Figure 2. Graph showing relation between triazine concentrations determined by enzyme-linked immunosorbent assay and atrazine concentrations determined by gas chromatography/mass spectrometry, January 1995 through December 1997.
Figure 3. Graphs showing daily mean stream discharges and atrazine concentrations, 1995-97, for two surface-water sampling sites in the Little Arkansas River Basin.
Figure 4. Graphs showing daily atrazine loads from the Little Arkansas River at Highway 50 near Halstead and the Little Arkansas River at Sedgwick, 1995-97.

Table

Table 1. Table showing stream discharges and atrazine concentrations, loads, and yields.

Abstract

Introduction

Currently (1998), water supply for the city of Wichita in south-central Kansas comes from two primarysources-the Wichita well field and Cheney Reservoir (fig. 1). The well field withdraws water from the Equus Beds aquifer, which is the easternmost part of the High Plains aquifer in Kansas (fig. 1).

Because of expected population growth in the region, the available water supply is not expected to meet future demands. The ongoing Equus Beds Ground-Water Recharge Demonstration Project is evaluating two ground-water recharge and recovery techniques, surface-spreading basins and direct-injection recharge wells, to help meet these demands. The project is a collaborative effort among the following agencies and engineering consulting firms: city of Wichita, Groundwater Management District No. 2 (Halstead, Kansas), Bureau of Reclamation and U.S. Geological Survey (both U.S. Department of the Interior agencies), U.S. Environmental Protection Agency, various Kansas State agencies, Burns and McDonnell Engineering Consultants (Kansas City, Missouri), and Mid-Kansas Engineering Consultants (Wichita, Kansas).

The characterization of atrazine in the Little Arkansas River, presented herein, included an examination of daily stream discharge, atrazine concentrations, seasonal atrazine load distribution, and atrazine yields as related to hydrologic conditions and herbicide-application practices from January 1995 through December 1997. These data will be useful in determining the need for surface-water treatment to minimize the effects of atrazine before recharging the aquifer.

Land use in the Little Arkansas River Basin is extensively agricultural, with about 78 percent cropland and 19 percent grassland. The percentage of cropland in each subbasin is about the same, as are the percentages of each crop type-about 63 percent wheat, 27 percent sorghum, 5 percent corn, and 5 percent soybeans. Crop data were estimated from county data compiled by the Kansas Department of Agriculture and U.S. Department of Agriculture.

Atrazine, A Triazine Herbicide

Methods of Study

The following approach for the characterization of atrazine was developed for use during the Equus Beds Ground-Water Recharge Demonstration Project; however, the processes may be applied to other basins as well. As part of the project, data were collected from January 1995 through December 1997 to document the quality and quantity of water in the Little Arkansas River. The two surface-water sampling sites established for the study of atrazine are U.S. Geological Survey (USGS) gaging stations on the Little Arkansas River at Highway 50 near Halstead (site 07143672) and the Little Arkansas River near Sedgwick (site 07144100) (fig. 1). The Halstead gage has a contributing drainage area of about 686 mi ² (square miles). The gage near Sedgwick has a contributing drainage area of about 1,163 mi ². Both sampling sites are affected by ground-water withdrawals, surface-water diversions, and return flow from irrigated areas.

Stream water-surface elevation (stage) was determined at the two gaging stations with nonsubmersible, pressure transducers and was measured to the nearest 0.01 foot. Stage data were electronically recorded and transmitted by satellite to a downlink site and then to the computer at the USGS office in Lawrence, Kansas.

Methods used to determine streamflow are described in Buchanan and Somers (see references). Streamflow measurements were made at least monthly by determining the cross-sectional area of the stream and measuring the flow velocity for at least 10 vertical transects in the cross section. A stage-discharge relationship was developed on the basis of discharge measurements and the stage of the stream at the time of measurement.

Most of the surface-water samples were collected by the USGS using automated samplers. The automated samples allowed for the collection of several samples per day during the growing season. Sample bottles were collected from the automated sampler periodically, labeled, chilled to 4°'C, and transported for processing and analysis. Results of samples collected by automated samplers were compared to results from samples collected manually using depth- and width-integrating techniques. Both types of samples were analyzed by the USGS laboratory in Lawrence, Kansas, and the USGS National Water-Quality Laboratory in Arvada, Colorado.

During the growing season, March through September, automated samplers collected two or more samples per day at the two sampling sites. During periods of low flow, October through February, fewer samples were collected, usually one per month. The automated sampler at the Halstead sampling site was discontinued in October 1996, although manual samples were still collected periodically.

About 1,800 samples were analyzed for triazine herbicides using enzyme-linked immunosorbent assay (ELISA), with a turnaround time of about 2 days per sample. However, the ELISA procedure is not specific to atrazine. Other triazine herbicides such as ametryn, propazine, prometryn, and prometon may be cross reactive with atrazine and are indistinguishable from atrazine using the ELISA procedure. To determine specific atrazine concentrations, 191 samples were also analyzed by gas chromatography/mass spectrometry (GC/MS).

The relation between triazine-herbicide concentrations determined by ELISA and atrazine concentrations determined by GC/MS for the 191 pairs is shown in figure 2. The slope of the regression line, 0.81, indicates that the relationship between ELISA-determined triazine values and the GC/MS-determined atrazine values are nearly one to one. Not only are the results of the two procedures similar, but the ELISA analysis allowed for many samples to be analyzed at a low cost. Therefore, the triazine-herbicide concentrations determined by ELISA are referred to as atrazine in this proceeding.

Daily mean atrazine concentrations are the average concentrations of samples collected that day. During the growing season, linear interpolation was used to estimate concentrations between days with samples collected. On days of low flow during October through February when no sample was collected, the daily mean atrazine concentration was arbitrarily assigned a value equal to the ELISA detection limit of 0.10 µg/L. Annual mean atrazine concentrations are the averages of the daily atrazine concentrations for each year.

Results of Sampling

Annual mean atrazine concentrations did not exceed the MCL for atrazine for any year from 1995 through 1997 at either sampling site (table 1). However, daily mean atrazine concentrations in water samples from the sampling sites were as high as 33.5 µg/L during spring and summer runoff (fig.3). Concentrations larger than 3.0 µg/L did not occur at other times of the year.

The Sedgwick site had larger annual mean stream discharges than the Halstead site (table 1). The largest annual mean discharge for both sites occurred in 1995; however, that year had the smallest annual mean atrazine concentration. Therefore, the largest atrazine concentrations did not necessarily correspond with the largest discharges. This indicates that one or more combinations of physical factors were affecting the concentration of atrazine detected in the surface-water samples. Possible physical factors include the timing of the runoff and atrazine application, crop or tillage type, method of application of atrazine, amount of atrazine applied, and other land-management techniques.

Atrazine loads were calculated to determine when large amounts of atrazine are transported in the Little Arkansas River (fig. 4). The computation of atrazine loads allow for an easier comparison between years because they account for hydrologic variability and allow for a mass-to-mass comparison. Daily atrazine loads (in pounds) were calculated by multiplying daily mean atrazine concentrations (in micrograms per liter) by daily mean discharge (in cubic feet per second) and by a unit conversion factor of 0.00538.

Daily atrazine loads increase with an increase in either discharge or atrazine concentration. Atrazine is typically applied in early spring when crops are planted; however, this is also when rainfall generally is the most intense and likely to produce runoff. As a result, daily atrazine loads were largest during May through July when increases occurred in both discharge and atrazine concentration. In general, the Sedgwick site had larger daily atrazine loads in 1995 and 1996 than the Halstead site, resulting more from an increase in discharge at the downstream Sedgwick site than from an increase in atrazine concentration.

Annual mean atrazine loads were calculated by summing the daily loads. The largest annual mean atrazine loads of 2,100 and 4,100 lb (pounds) for the Halstead and Sedgwick sites, respectively, occurred in 1995. The smallest loads of 1,500 and 2,200 lb occurred in 1996 (table 1). For the years in which data were available, the largest annual mean atrazine loads occurred in years of largest discharge, whereas the smallest loads occurred in years of smallest discharge.

Annual atrazine yield was determined by dividing the annual load by the drainage area. Annual yield was calculated to account for the difference between the two drainage areas. If the annual yields were significantly different between the two sampling sites, this could reflect a difference in land use or land-management practices. In 1995 the larger annual yield occurred at the Sedgwick site, whereas in 1996 the larger annual yield occurred at the Halstead site. Because the downstream Sedgwick site had larger discharges both years, the larger annual yield at Halstead in 1996 is probably not due only to hydrologic conditions but rather a combination of other physical factors. These physical factors would be the same as those which affect the atrazine concentrations, such as method of application and crop or tillage type.

Conclusions and Implications

Acknowledgments

References