Investigation of ethylene dibromide (EDB) in ground water in Seminole County, Georgia

U. S. - G E 0 ..1 0 G I C A L SURVEY CIRCULAR

Prepared in cooperation with Exposure Assessment Branch Hazard Evaluation Division Office of Pesticide Programs Office of Pesticides and Toxic Substances U.S. Environmental Protection Agency

SURVEY U. S.

Prepared in cooperation with Exposure Assessment Branch Hazard Evaluation Division OHice of Pesticide Programs OHice of Pesticides and Toxic Substances U.S. Environmental Protection Agency

~t-32

r'I

J

systems, multicropping, and increased applications of EDB and other pesticides.

In Seminole County, the principal crops are vegetables, peanuts, soybeans, and pecans during the summer months and grains during the winter months. EDB (trade names Soilbrom and Dowfume) is used extensively on vegetable and peanut crops and to a lesser extent on soybean crops to control nematodes. EDB is applied several inches below the soil surface as a liquid at the rate of 1 to 2 gal (15 to 30 lbs of active ingredients) per acre.

Prior to this investigation, a researcher from Florida State University, Tallahassee, had detected EDB in water samples collected from three irrigation wells and one domestic well south of Donalsonville near Buck Hole in Seminole County (fig. 8). Buck Hole is one of many sinkholes in a swampy depression that is connected to Fishpond drain, a major drainage that flows from north to south through the center of the county. The researcher initially found EDB in water samples from these wells during a method development study for synthetic organic compounds. In subsequent sampling, the researcher confirmed the presence of EDB in these wells. The findings are listed in the following table:

Two analytical approaches were used by the Florida State University researcher to determine EDB concentrations in the water samples: extraction of the sample followed by gas chromatographic analysis of the extract, and a purge-and-trap technique using a Tenax trap. Detection limits were about 0.2 11g/L.

The University informed the EDB manufacturers of the findings in May 1982. The manufacturers resampled the wells, confirmed the presence of EDB in ground water from BH1 and BH2, and notified the EPA Office of Pesticide Programs of the contamination. The manufacturers submitted data to the Office of Pesticide Programs in September 1982 that showed EDB had been used annually for several years as a soil fumigant for control of nematodes in the fields where the BH1 and BH2 irrigation wells are located. The history of EDB use in the field where BG 1 is located is uncertain. The domestic well is located in the vicinity of fields where EDB has been applied.

After the contamination was reported, the EPA informally requested that the manufacturers design and conduct a ground-water-quality monitoring study in southwest Georgia, but no study has been initiated to date. Recent reports of EDB contamination of ground water in Florida, Georgia, California, and Hawaii prompted the EPA to conduct its own study in Seminole County, Ga, as quickly as possible to better define the nature of the contamination problem. In August 1983 the U.S. Geological Survey, in cooperation with the Exposure Assessment Branch of EPA, began an investigation of EDB in ground water in Seminole County.

The source of most domestic, municipal, industrial, and irrigation water in southwest Georgia is the principal artesian aquifer. This carbonate aquifer system extends from South Carolina through Florida. In the Dougherty Plain (fig. 1), the upper surface of the aquifer dips generally southeastward and ranges from about 300 ft above sea level in the northern part of the area to about sea level in the southern part (fig. 2). However, the surface of the aquifer is highly irregular because of differential weathering and solution-cavity collapse. The aquifer ranges in thickness from 25 ft on the northwestern edge of the Dougherty Plain to more than 350 ft on the southern edge (fig. 3). In many areas, the limestone is largely fractured, and solutioning has created a labyrinth of subterranean channels which result in high transmissivities for wells that penetrate the channels. Regionally, estimates of transmissivity values range from 3,000 average velocities of 3 ft/d near the Flint River upstream from Bainbridge and 0.2 ft/d in the northem part of the Dougherty Plain, away from streams. Average velocities of ground-water flow vary greatly in the principal artesian aquifer because the limestone acts as both a free-flow (channel flow) and a diffuse-flow system. Actual ground-water velocity may be more or less than the average values, depending on the flow path followed and local geohydrologic conditions (Hayes and others, 1983). The maximum potential rate of lateral water movement in the vicinity A sandy-clay residuum overlies the principal artesian aquifer, and in Seminole County it ranges in thickness from about 50 to 100 ft (fig. 7). Water levels in the residuum respond to rainfall and are highest in March and April and lowest in November and December. According to Hayes and others (1983), where the residuum is relatively thick and impermeable, the water table is believed to be a subdued replica of the topography; where the residuum is relatively thin and wells that tap the principal artesian aquifer. One sample of ponded surface water was collected from Fishpond Drain near Buck Hole. The locations of the sampling sites are shown in figure 8. In addition, nine core samples for EDB and particle-size analysis were collected from a 42-foot test hole at 4-foot intervals beginning a foot below land surface.

Selection of sampling sites for EDB was based on available information for EDB in ground water, hydrogeologic data and soil surveys, EDB application information obtained from the county agent and landowners, land-use data, and landowners cooperation. Land use was determined from 1982 high-altitude photographs by the Environmental Photographic Interpretation Center of the EPA Office of Research and Development. Wells W4(BH2) and W5(BH1), in the vicinity of Buck Hole where the highest concentrations of EDB in ground water had been reported by Florida State University, were selected for resampling in this study. Other sites were selected upgradient and downgradient from Buck Hole in fields with different soil types, some of which had not received EDB applications. All the sites, except one well in southernmost Miller County, were in Seminole County. The site selected for the collection of core samples was about 150ft south of well W4 (fig. 8).

Ground-water samples from irrigation wells were collected at water spigots on the delivery lines between the pumps and the sprinkler systems. Irrigation systems that were not operating were started and allowed to pump at a rate of several hundred gallons per minute for at least 15 min. prior to sample collection. Samples from domestic wells were also collected at a water spigot after about 15 min. of continuous pumping. Domestic wells sampled tapped the same aquifer as the irrigation wells, and the water was untreated.

Water samples were collected in 40-mL glass vials and sealed with Teflon-lined silicone septa and plastic screwcaps. Water was delivered to the vials through a piece of small-diameter silicon tubing attached to the spigots. The tubing was inserted into the vials and the vials were flushed with several volumes of sample water. The tubing was then slowly removed while the water was flowing to allow the vials to completely fill before capping. Care was taken to ensure that the water samples were free of air bubbles. To reduce the possibility of cross contamination, new tubing was used each time a sample was collected. The surface-water sample was collected with a brass "grab type" sampler and transferred to a 40-mL vial through silicon tubing to minimize aeration of the sample. The samples were stored in ice until analyzed.

Core samples were collected with a split-spoon coring device 18 in. long by 2 in. in diameter. A truck-mounted flight auger was used to bore to the desired sampling depth. After the loose cuttings were removed from the hole with the auger, the split-spoon was attached to a solid steel rod, lowered to the bottom of the hole, and driven into the formation. The coring device was then lifted to land surface and split apart, and a central section of the core (4 to 6 in. long) was removed for EDB and particle-size analysis. Core samples for EDB analysis were sealed in clean glass jars and stored in ice until analyzed. Core samples for grain-size analysis were stored in plastic containers. Before each core sample was collected, the core barrel was rinsed with acetone and then with chemically pure water.

During the boring, the cuttings that were removed from the hole were used to describe the lithology of the residuum. After the core samples were collected, a continuous natural gamma radiation log was made of the hole with a borehole geophysical logger according to procedures described by Keys and MacCary (1971). Natural gamma radiation was detected by a sodium iodide crystal that was lowered into the bore hole on a wire line. The wire line also served as an electrical connector between the crystal and a stripchart recorder at the land surface. The natural gamma radiation log supplemented the core data by providing additional data to identify sand, clay, and limestone in the borehole.

Guidelines used for the collection of EDB samples in ground water were provided by U.S. Geological Survey, Water Quality Branch Memorandum 83.12, "Guidelines for the Collection of Ground-Water Samples for the Analysis of Organic Compounds.'' Recommended sampling e

wa

lt J

I0

mode. Fluorobenzene was added to all samples as a surrogate.

Analytical conditions were as follows:

For water samples, the gas chromatograph was calibrated daily using a standard solution containing 43.4 pg/p.L of EDB in hexane. Calibration was checked with the same standard solution at least twice daily.

Spiked reagent water samples, prepared in the laboratory by pipetting microliter quantities of a solution of EDB in methanol into 37 mL of water contained in sample vials, were analyzed and confirmed by the same methods used to analyze the samples. Recovery data for extraction and purge-and-trap methodologies are reported in the following table:

Laboratory and field blanks were analyzed with the samples, and EDB was not detected in either set of blanks. Fluorobenzene was added to all samples as a surrogate prior to GC/MS confirmation. Recoveries fell within the acceptable limits

(U.S. Environmental Protection Agency, 1982).

Duplicate water samples were collected from five wells in the field and analyzed as blind samples by the laboratory. Results of blind duplicate analyses were as follows:

Standard solutions for soil samples were prepared and used as described for water samples. Spikes were not used for soil samples.

Sampling locations, land-surface altitudes, available well construction data, and pumping information are listed in table 1. EDB application information and soil charateristics are shown in tables 2 and 3, respectively. The results of the water-sample analyses for EDB are listed in table 4.

A synopsis of data gathered during this study is presented in figure 9. Included in figure 9 are map locations of sample sites, EDB concentrations in water samples, and information on land use, soil permeability, crop type, and EDB applications. Potential nonagricultural sources of ground-water contamination such as gasoline storage tanks and the Seminole County landfill are also shown.

Irrigation and domestic wells in the study area commonly are 150 to 200 ft deep (see table 1). Wells generally are cased to the top of the limestone and are open-hole from the top of the limestone to the bottom of the well. Yields of 1,200 gallmin from the irrigation wells are common.

TABLE !.-Location and information for wells sampled [Sample site: numbers used to identify wells in report and to locate wells on map in fig. 8. Use: I, irrigation; D, domestic]

Soils in Seminole County and throughout a large part of the Dougherty Plain consist of sand, sandy-clay, and clay and are low in organic matter. The characteristics of major soil types in Seminole County are shown in table 3. These data were developed by the U.S. Department of Agriculture Soil Conservation Service (Middleton and Smith, 1976) and consider only the upper 60 to 80 in. of soil.

EDB was found in 6 of 19 wells (table 4). The highest concentrations of EDB in ground water in Seminole County were found in the Buck Hole

area of Fishpond Drain. Concentrations at wells W4 (BH2) and W5 (BH1), east and west of Buck Hole, were 7.1 and 11.8 p.g/L, respectively, much higher than at the other wells sampled. However, concentrations at these two sites have decreased substantially since May 1982. (See table in "Background" section.)

At well W4, three samples were collected in succession after 5, 15, and 25 min. of continuous pumping. During this period, concentrations of EDB decreased from 7.1 to 4.5 p.g/L. A decrease in concentrations with increased pumping time would be expected where high levels of contaminant are localized.

North of Buck Hole (upgradient in the aquifer), EDB concentrations in ground water were much less than at wells W4 and W5 and in most samples were below the level of detection (less than 0.01 p.g!L). The concentration at well W14, a domestic well about one-half mile north of well W5, was roughly 400 times lower than at well W5. The concentration in well W13, a domestic well about 3 mi north-northwest of well W5, was below the level of detection. At well W6, about 4 mi northwest of well W5, the concentration was about 200 times less than at well W5. EDB was not detected in the remaining six wells to the north. A sample of ponded surface water collected from Fishpond Drain near Buck Hole did not contain detectable levels of EDB. South of Buck

~Limestone

(figs. 5 and 6) suggest that the aquifer discharges into Fishpond Drain. Sinkholes, which are prevalent throughout this area, act as vertical conduits connecting the principal artesian aquifer, the water-table aquifer, and the surface-water drainage system. During the summer and fall, pumping temporarily lowers the water level locally in the aquifer, and ground-water discharge to Fishpond Drain decreases or ceases. A lowering of the water level may accelerate the downward movement of water and contaminants from the overlying residuum into the aquifer.

Lithologic data obtained at the coring site near well W4 indicate the presence of clay layers in the residuum. A layer of dense clay exists just above the top of the aquifer between depths of about 29 and 35 ft below land surface. The clay layers would probabiy retard the downward movement of contaminated water into the aquifer. Perhaps EDB that was applied to fields reached the aquifer by moving laterally downgradient through permeable zones in the residuum toward Fishpond Drain and Buck Hole or other sinkholes where the absence of clay permitted it to flow directly into the aquifer.

Possibly, EDB in the residuum moved down the annular spaces between the well casings and the surrounding soil and residuum at wells where the annular spaces were not sealed with grout during construction. The contamination of ground water by EDB at wells W6, W7, and W9 may have resulted from leakage of locally applied EDB downward through the residuum, or from movement of EDB downward through the annular spaces between the well casings and the residuum into the aquifer. Contamination at wells W7 and W9 also may have been from the movement of a plume of EDB in the ground water from the Buck Hole area to those wells. Ground-waterlevel data are not available to define hydraulic gradients during periods of heavy pumping. These data are needed to assess the direction of movement of a potential contaminant plume. Movement of a contaminant plume southward from the Buck Hole area is plausible because that is the natural direction of ground-water flow. The presence of a plume to the south of Buck Hole is suggested by the low concentration of EDB detected at well W9, which is in a new field where EDB has never been applied. However, no EDB was found in the ground water at well W8 threefourths of a mile southeast. Farther south, EDB was detected at well W7, where applications of EDB have been made one or two times annually since 1979. Locally applied EDB, rather than a plume, may have been the source of contamination at well W7.

Data are needed that will describe the pathways and rates of movement of contaminants into the aquifer and their fate in the ground-water system. These data will provide information that can be used to assess potential contamination problems from pesticide use in Seminole County and other areas having similar hydrogeology.

Data that will provide information on the pathways, rates of movement, and fate of contaminants in the study area could be obtained by: (1) sampling a network of residuum and aquifer wells and surface-water sites for EDB, (2) measuring ground-water levels in the residuum and the aquifer to define the local seasonal hydraulic gradients during pumping, and (3) collecting residuum core data to describe local lithology and to determine EDB concentrations in the residuum.

Contamination of the aquifer with EDB seems to be limited to about a 4-square-mile area in central Seminole County in the vicinity of Buck Hole. Five of the six wells contaminated with EDB are within high EDB-use areas where the soil permeability ranges from moderate to rapid. The sixth well, W9, is in an area having moderate to rapid soil permeability and is downgradient from high-use areas. Considering the contamination pattern, the soil-core data, local hydrogeology, the location of the county landfill with regard to the area of contamination, and the lack of evidence of gasoline spills, the ground-water contamination in Seminole County probably is due to soil fumigation with EDB. However, because the high concentrations of EDB in the aquifer seems to be localized in the Buck Hole area, the possibility of contamination from an EDB fumigant spill cannot be disregarded at this time.

The levels of EDB found in this study are representative of a single point in time. It is likely that repetitive sampling would show significant temporal variability, particularly on a seasonal basis, owing to the high transmissivities, seasonal fluctuations in water levels and recharge rates, and seasonal variations in irrigation withdrawals.

Where this page came from

This page was imported from U.S. Geological Survey. Published by the U.S. Geological Survey and, as a work of the United States government, in the public domain.

Nobody has written it yet — it is the source material at a new address, which is why search engines are asked to skip it and why no one earns from it. It is up for grabs: take it on, and it is yours to rewrite and to earn from.