USGS Fact Sheet 134-98
December 1998

Herbicides in Midwestern Reservoir Outflows, 1992-93

Table of Contents

Introduction
Results
Implications
Additional References

Introduction

Reservoirs are an important resource in the 11 Midwestern States that comprise the upper Mississippi River Basin (see index map) often referred to as the Corn Belt (see map). Reservoirs impound and store water for such beneficial uses as flood control, hydropower, irrigation, public supply, recreation, and wildlife habitat. However, reservoirs are also repositories for contaminants that are introduced into Midwestern streams. These contaminants, primarily generated from agricultural activities, include nutrients (nitrogen and phosphorus), pesticides (herbicides and insecticides), sediment, and sediment-associated contaminants (U.S. Environmental Protection Agency, 1984; Humenik and others, 1987).

Although most of the sediment entering reservoirs is permanently trapped and deposited along reservoir bottoms, chemicals such as soluble herbicides generally remain in the water column and are stored temporarily until they are flushed from the reservoir. This can substantially affect the water quality downstream from reservoirs. Unregulated streams in the Midwest generally have high concentrations of herbicides for short periods of time immediately following planting as spring storms wash newly applied herbicides into streams (Wauchope, 1978; Thurman and others, 1991). In contrast, herbicides in streams regulated by reservoirs have smaller peak concentrations that can persist for a much longer period of time (Leung and others, 1982; Stamer and others, 1995).

The length of time that reservoirs discharge water with elevated concentrations of herbicides depends on a number of factors. These factors include the time that water remains in a reservoir (residence time), the timing of inflow to the reservoir, the land use and herbicide use in the contributing drainage area, the timing and intensity of rainfall, and the physical and chemical characteristics of the herbicides themselves.

The Corn Belt is the largest and most intensive crop-producing region of the United States, accounting for about 65 percent of the total harvested cropland and about 60 percent of the herbicide use in the Nation (Battaglin and Goolsby, 1995). The three major row crops in the Midwest with the highest rates of herbicide use are corn, soybeans, and sorghum. In 1992, use of the herbicides atrazine (21.3 million kilograms), alachlor (20.9 million kilograms), metolachlor (18.6 million kilograms), and cyanazine (11.3 million kilograms) on these three crops in the Corn Belt alone accounted for 25 percent of the 285 million kilograms of herbicides used in the United States (U.S. Department of Agriculture, 1993).

Although herbicides have many benefits, they may also produce a wide range of toxic side effects that may pose a potential hazard to human health and the environment. For instance, the U.S. Environmental Protection Agency (1996) classifies the herbicide alachlor as a human carcinogen, and several other herbicides including atrazine, cyanazine, and metolachlor are classified as possible human carcinogens. In addition, the effects of long-term, low concentrations of herbicides or combinations of herbicides on aquatic ecosystems are largely unknown.

To assess the effects of reservoirs on downstream concentrations of herbicides, 76 reservoirs, out of the more than 500 in the Midwest with a capacity of more than 5,000 acre-feet, were selected for study (see map). The purpose of the study was to evaluate the occurrence and temporal distribution of herbicides in Midwestern streams and to relate concentrations to human and natural factors. The drainage-basin areas of the reservoirs ranged from 44 to 64,560 square kilometers with a mean area of 6,140 square kilometers. Basins were highly variable with respect to land use, soil type, and climatic conditions. Cropland represented more than 50 percent of the land area in about 25 percent of the basins. Less than 5 percent of the basins were represented by less than 5 percent cropland. Reservoir outflows were sampled eight times (approximately bimonthly) from April 1992 through September 1993 for a total of 608 samples. The timing and frequency of sampling made it possible to determine approximately when maximum and minimum concentrations of herbicides occurred in the reservoir outflows. All samples were collected using noncontaminating containers and analyzed for 11 herbicides and 6 herbicide metabolites (table 1). A metabolite is a compound that is formed from the breakdown of the parent herbicide.

                      ¹Concentration at or above reporting limits.
                      ²Based on 608 reservoir outflow samples.
                      ³Based on 511 reservoir outflow samples.

Results

Most of the reservoirs sampled during this study had detectable concentrations of at least one or more herbicides and (or) metabolites (table 1). Atrazine, its two metabolites, deethylatrazine and deisopropylatrazine, and the alachlor metabolite alachlor ESA were the most frequently detected compounds, all being detected in more than 60 percent of the samples. Cyanazine and metolachlor were also detected in about 50 percent of the samples. The highest herbicide concentration detected was for atrazine [12.4 micrograms per liter (mg/L)], and the highest metabolite concentration was for alachlor ESA (19.7 mg/L). Of the parent herbicides, atrazine, the most extensively applied herbicide in the Corn Belt, occurred at the highest frequency (82 percent) and in the largest concentration (mean of 1.36 mg/L for samples with detectable concentrations). Alachlor, the second most extensively applied herbicide after atrazine in 1992, was detected less frequently (29.8 percent of samples) and at smaller concentrations (mean of detectable concentrations of 0.27 mg/L) than either metolachlor or cyanazine. However, alachlor ESA was detected in a majority of the samples (70.2 percent) and had the highest mean concentration (1.74 mg/L) of all of the parent herbicides or metabolites analyzed. Alachlor ESA is more mobile and thought to be less toxic than alachlor, and its higher frequency of detection and larger concentrations in streams as compared to alachlor have been documented in other studies (Thurman and others, 1996; Kalkhoff and others, 1998).

During the 18-month study, the mean sum of the individual concentrations of alachlor and alachlor ESA was 1.9 mg/L when at least one of the two compounds was detected. The mean sum of atrazine and its metabolites was also 1.9 mg/L, implying that alachlor and atrazine were used in nearly equal amounts in the study area, a finding consistent with use data (U.S. Department of Agriculture, 1993). Similarly, the mean sum of the individual concentrations of cyanazine and its metabolites was 1.0 mg/L, about one-half the sum for alachlor and atrazine plus their metabolites. This result is also consistent with the fact that cyanazine usage in the study area was about one-half that of atrazine or alachlor.

The temporal variability in the concentrations of two herbicides and two metabolites in reservoir outflows are shown in the graphs. Concentrations of the parent herbicides, in this case atrazine and alachlor, are generally smallest in the winter and early spring prior to planting. Concentrations of atrazine peaked in the summer gradually decreasing through the fall and winter. Concentrations of alachlor also peaked in the summer and were small or undetectable for the rest of the year. Concentrations of the metabolites deethylatrazine and alachlor ESA also varied seasonally but tended to peak later in the summer and appeared to be less variable as compared to the associated parent herbicides.

The variability in the concentrations of herbicides and metabolites in reservoir outflow is bly affected by the reservoir residence time. Lake Vermillion in Illinois is a small reservoir with a volume of about 11.4 million cubic meters (9,200 acre-feet) and a relatively short residence time (about 9 days). As a result, water is flushed through the reservoir rapidly, and concentrations of herbicides and associated metabolites exhibit a temporal distribution similar to that of unregulated streams (see graphs). During both 1992 and 1993, concentrations of both herbicides and metabolites in Lake Vermillion outflow were low in late April, peaked rapidly in early July, and then decreased rapidly by early October. In contrast to Lake Vermillion, Hillsdale Lake in Kansas (see photograph) has a much larger volume of about 89 to 160 million cubic meters (72,000-130,000 acre-feet) and a residence time of nearly 8 months. As a result, herbicides and metabolites delivered to Hillsdale Lake in the spring and early summer are stored and subsequently delivered downstream in smaller concentrations and for longer periods of time.

Statistical models were used to estimate mean herbicide concentrations in each reservoir outflow as a function of the physical characteristics of the reservoir, the land and chemical use in the drainage basin, soil characteristics, topographic conditions, and climatic variables. Results from the models indicate that the amount of herbicide used in a drainage basin is the primary process affecting the concentration of the herbicide in reservoirs. The models also indicate that when drainage basins have steep slopes and poorly drained clay-rich soils, the receiving reservoirs tend to have higher herbicide concentrations. These findings suggest that best-management practices targeted at reducing the use of herbicides and reducing the loss of herbicides to surface- and ground-water systems will be the most successful in lowering herbicide concentrations in reservoirs.

Implications

The U.S. Environmental Protection Agency (1996) has established either a Maximum Contaminant Level (MCL) or a Health Advisory Level (HAL) for 9 of the 11 parent herbicides analyzed in this study (table 1). MCLs are legally enforceable drinking-water regulations, whereas HALs are drinking-water criteria and nonenforceable. Because MCLs and HALs are based on annual average concentrations, one or more exceedances of the specified value does not necessarily indicate noncompliance. MCLs or HALs have not been established for any of the six herbicide metabolites analyzed during this study or for the combination of the parent herbicides and associated metabolites. However, if the concentrations of the associated metabolites are added to the concentrations of the parent herbicides, the number of reservoirs with mean annual outflow concentrations that exceeded the parent compound MCL or HAL increases. For example, in 1992, 17 reservoirs would have exceeded the MCL for atrazine on the basis of atrazine and its associated metabolites compared to 6 reservoirs on the basis of atrazine only. Similarly, 14 reservoirs would have exceeded the HAL for cyanazine in 1992 on the basis of cyanazine and its associated metabolites compared to 6 reservoirs on the basis of cyanazine only. If alachlor ESA were added to the concentration of alachlor, 11 reservoirs would have exceeded the MCL in 1992, and 16 reservoirs would have exceeded the MCL in 1993. However, unlike the atrazine and cyanazine metabolites, alachlor ESA is a chlorine-free compound upon degradation; thus, it is considered to be less toxic than the parent compound (Baker and others, 1993). Therefore, a comparison of the sum concentration of alachlor plus alachlor ESA to established MCLs for alachlor probably is not valid.

--John K. Stamer, William A. Battaglin, and Donald A. Goolsby

This Fact Sheet is based largely on information contained in the following publications and references cited therein:

Battaglin, W.A., and Goolsby, D.A., 1996, Using GIS and regression to estimate annual herbicide concentrations in outflow from reservoirs in the Midwestern USA, 1992-93, in Hallam, C.A., and others, eds., GIS and water resources, Proceedings: American Water Resources Association Annual Symposium, Fort Lauderdale, Florida, September 22-26, 1996, p. 89-98.

Goolsby, D.A., Battaglin, W.A., Fallon, J.D., Aga, D.S., Kolpin, D.W., and Thurman, E.M., 1993, Persistence of herbicides in selected reservoirs in the Midwestern United States--some preliminary result, in Goolsby, D.A., Boyer, L.L., and Mallard, G.E., compilers, Selected papers on agricultural chemicals in water resources of the midcontinental United States: U.S. Geological Survey Open-File Report 93-418, p. 51-74.

Scribner, E.A., Goolsby, D.A., Thurman, E.M., Meyer, M.T., and Battaglin, W.A., 1996, Concentrations of selected herbicides, herbicide degradation products, and nutrients in outflow from selected Midwestern reservoirs, April 1992 through September 1993: U.S. Geological Survey Open-File Report 96-393, 128 p.

Additional References

District Chief
U.S. Geological Survey
4821 Quail Crest Place
Lawrence, Kansas 66049-3839
(785) 842-9909
email: waucott@usgs.gov