CASE STUDIES IN SELECTED KARST REGIONS
Twenty-two case studies were selected from the Tennessee Division of Superfund files to evaluate chlorinated-solvent contamination in the karst regions of Middle and East Tennessee (fig. 13, and table 5). The Division of Superfund files are available for review at the Division of Superfund central office in Nashville, Tennessee, and in the Division's regional field offices. For a thorough review, both sources should be consulted. Additional information for a few sites was gathered from other State agencies. State Superfund sites are organized by county and individual site number.
These sites are at different stages of the investigation-remediation process, and the amount of data varies from site to site. Available site data was evaluated to select the case studies. The information included in this report is based on available data at the time the case study evaluation was completed (March 1997).
Criteria used to select the case studies for this report consisted of the following: (1) the site is located in a karst region of Tennessee, (2) the volume and type of release of chlorinated solvents is known, (3) the vertical and horizontal extent of the dissolved chlorinated solvents is defined, (4) dye-tracer tests were completed for the site, and (5) chlorinated DNAPL is documented at the site. During the selection process, even the most extensively documented case studies did not meet all of these criteria. A discussion of the selection criteria is presented below. Detailed discussions of five case studies follow the discussion of the criteria.
Case studies were selected in five karst regions: the inner Central Basin (3 sites), the outer Central Basin (2 sites), the Highland Rim (7 sites), the Valley and Ridge (9 sites), and the western toe of the Blue Ridge (1 site). Most of the case studies reported springs, sinkholes, or seeps present at or near a site; however, four case studies did not have any documented evidence of karst features. No case studies were selected in the Pennyroyal Plateau part of the Highland Rim or in the coves and escarpments of the Cumberland Plateau due to the lack of sites with documented contamination from chlorinated solvents.
The volume of chlorinated solvents and types of release were documented at some sites. At other sites this information was estimated from the number of drums of solvent used or removed. Fifteen of the selected case studies did not have sufficient information to estimate the amount of released solvent.
The reported vertical and horizontal extent of the dissolved chlorinated solvents was determined from the chemical analyses of ground-water samples collected from monitoring wells and springs on or near the sites. When available, additional information of contaminant extent was provided by analyses of soil, surface-water, and domestic-well water samples. In some case studies, dissolved contaminants were present in creeks or streams. Documentation of the vertical extent of dissolved chlorinated solvents for each case study was limited by the depths and screened intervals of monitoring wells completed at a site. In karst, the effectiveness of monitoring wells depends on intercepting the localized ground-water flow paths along which contaminants may migrate. Dye-trace study results were available for 4 of the 22 selected case studies at the time of this study.
Documented evidence of chlorinated DNAPL was absent in most of the case studies chosen for this report. Only 4 of the 22 case studies reported visual evidence of DNAPL, usually described as a separate-phase liquid, denser than water, having a dark brownish/blackish color. Chlorinated solvents such as TCE and PCE are colorless or clear when pure or unused (Lucius and others, 1992).
The lack of evidence of DNAPL does not necessarily indicate its absence at a site but may reflect the complexity of the site and of the DNAPL distribution. The U.S. Environmental Protection Agency (U.S. EPA) published a guidance document (U.S. EPA, 1992b) that suggests a method to estimate the potential for DNAPL occurrence at a site.
The method, as described by the U.S. EPA (1992b), emphasizes the historical use of DNAPL and site characterization data. Relevant historical data include the industry type, processes, and waste-disposal practices for DNAPL and DNAPL-related chemicals. Important site characterization data include the observation of DNAPL and chemical analyses of ground water or soil samples collected at a site. Two important and often used chemical analytical conditions that suggest the presence of DNAPL include: (1) concentrations of DNAPL-related chemicals that are greater than 1 percent of the solubility limit (see table 2) in the ground water, and (2) concentrations of DNAPL-related chemicals that are greater than 10,000 milligrams per kilogram in the soil. Using this information, the potential for the occurrence of DNAPL is divided into three main categories: I--confirmed or high potential, II--moderate potential, and III--low potential. Based on the available data, combinations of these categories are possible.
Because four of the case studies already had documented evidence of DNAPL occurrence, the U.S. EPA method was applied to the remaining 18 sites to estimate potential DNAPL occurrence. These sites rated "high" or "moderate" potential for the occurrence of DNAPL. A limiting factor in the U.S. EPA method is that the approach to site characterization outlined is designed for granular porous media and may not be applicable in karst or fractured rock settings. Field (1988a), Quinlan and others (1992), and Quinlan (1994) address the problems associated with applying characterization techniques designed for porous media to karst settings.
The 22 case studies selected for review in this report are grouped by karst regions of Tennessee (table 5 and Appendix). Each case study is designated with a Tennessee county number from 01 (Anderson County) to 95 (Wilson County) and a unique site number. Site background, hydrogeology, and contaminant characteristics allow a comparison of each case study (table 5). Five case studies were selected for discussion in detail to illustrate the relevance of their specific site characteristics to the previously described conceptual models. Case study summary sheets for each of the 22 sites are presented in the Appendix.
Case study 26-505 provides an example of DNAPL accumulation on a low permeability layer in the regolith. This site was used as a leach pit for the disposal of waste solvents and acids from 1950 to 1972. An unknown quantity of chlorinated solvents were disposed at the site. PCE and 1,1,1-TCA are the main contaminants at the site. PCE has been measured in the ground water at 140 milligrams per liter (mg/L), approximately 70 percent of its water-solubility limit. 1,1,1-TCA has been measured in the ground water at 290 mg/L, approximately 22 percent of its solubility limit. A 3-m column of DNAPL has been documented in a shallow recovery well onsite.
This site is located in the Highland Rim. The site stratigraphy consists of regolith of up to 26 m in thickness over fractured, cherty limestone bedrock of the Fort Payne Formation. The regolith is composed of clay with silt, sand, and varying amounts of chert grading in size from clay-size particles to large gravels. The size and quantity of chert tend to increase with depth. Isolated sandy clay and sand layers also are reported to be present.
Both the DNAPL and dissolved-phase contamination at this site appear to be limited to shallow depths and have not been documented at depths below about 11 m. DNAPL reportedly appears to have pooled on top of a clay layer (capillary barrier) underlying a sand layer. This DNAPL pool will most likely provide a continuing source for dissolved-phase contaminant for a long time. DNAPL recovery operations are currently underway at this site. Of note, at two other nearby sites in the same geologic setting (26-501 and 26-502), contamination has been documented to move downward through the regolith into water-bearing zones at the top of rock and in fractures in the rock. Therefore, the significant DNAPL pooling observed in the regolith at this site has not been documented at the other similarly situated sites nearby. These data highlight the extreme variability of karst settings and the need for careful and thorough investigation of each individual contaminated site.
Case study 59-502 illustrates DNAPL accumulation at the top of rock and its persistence within this zone for an extended period. This site is a manufacturing plant that was established in 1937. In 1980, an underground pipe connecting a storage tank to an onsite degreaser burst inside a utility trench. An estimated 13,000 liters of TCE was lost and drained into the underground sewer pipes. Clean-up and containment activities at the site have consisted of: two shallow recovery trenches, two deep pump-and-treat recovery well systems, a weir in a nearby surface stream, and treatment of contaminated soils not covered by the site building.
This site is located in the inner Central Basin and is underlain by the Lebanon, Ridley, and Pierce Limestones (Crawford, 1992; Crawford and Ulmer, 1994). The Ridley Limestone is a karst aquifer at least 30 m thick at the site and is confined above and below by the relatively impure Lebanon and Pierce Limestones, respectively. The confining units vary from 5 to 20 m thick. Fractures and dissolutionally enlarged openings that occur along the bedding planes of the Ridley Limestone facilitate horizontal and vertical ground-water flow. Crawford (1992) suggested that caves and cave streams can develop at the contacts between the aquifer and confining units.
TCE and its breakdown products (DCE and vinyl chloride) were detected in soils, ground water, and a surface stream. Water samples from 53 monitoring wells indicate that aquifer contamination has spread at least 150 m across the site. Vertically, TCE and its breakdown products were detected in ground-water monitoring wells with total depths as great as 68 m below land surface. The highest concentration of TCE (950,000 mg/L) was detected in 1987, when a 61-cm column of DNAPL was measured in a shallow well completed above the bedrock. Nine years later in 1996, TCE (150 mg/L) was still detectable in a top-of-rock well. Chlorinated solvents detected in deep wells completed below the top of rock indicate that the Lebanon Limestone confining unit did not prevent downward vertical migration of the contaminants.
Case study 82-516 provides an example of DNAPL distributed throughout a network of fractures. This site reportedly was used from 1967 to 1974 for the disposal of drummed and bulk liquids. An estimated 79,000 liters of liquid wastes were discharged to an open pit where they were burned. Clean-up activities in 1984 and 1989 removed nearly 600 buried or partially buried drums from the site. The primary contaminant at the site is TCE. Sampling has detected maximum concentrations of TCE in the ground water at 960 mg/L and in the soil at 5.34 milligrams per kilogram. The dissolved concentration measured is nearly 90 percent of the water-solubility limit for TCE, indicating the probable presence of DNAPL. DNAPL has been detected and recovered from six onsite wells.
This site is located in the Valley and Ridge karst region. Site stratigraphy consists of less than 1.5 m of shaley silt loam overlying the shale of the Sevier Formation. The top 21 m of bedrock shows low fracture density and poorly developed ground-water flow. The next 60 m contains highly developed fracture networks with well-developed ground-water flow. At depths greater than 90 m, the shale is dense and hard allowing for little ground-water flow.
The six wells where DNAPL has been detected are all located near the center of the site. They delineate an area of approximately 230 by 90 m where DNAPL has been found in fractures. DNAPL has been detected in every well completed in this area. The maximum depth where DNAPL has been detected is 72 m, near the bottom of the highly developed fracture zone. Surrounding this DNAPL area is a larger area of dissolved contamination measuring approximately 550 by 335 m and extending to a depth of 108 m. The DNAPL distributed in the fractures most likely will provide a persistent source for dissolved-phase contamination.
At this site, the upper 21 m is considered poorly fractured and not conducive to ground-water flow, but this zone did not inhibit penetration of DNAPL to the deeper, more fractured ground-water flow zone. The DNAPL migrated downward with little lateral spreading, contaminating the underlying, more productive ground-water zone.
Case study 89-504 illustrates DNAPL accumulation in a perennially saturated zone at top of bedrock. This site is a manufacturing plant where TCE and PCE have been stored and used since 1960. Releases of solvents to the environment are known to have occurred. The actual quantities released are unknown.
The site is located in the Eastern Highland Rim and is underlain by approximately 8 to 9 m of St. Louis Limestone residuum, consisting of clay, silt, sand, and chert gravel. This regolith overlies less than 2 m of a calcareous siltstone confining unit that separates the regolith from the middle Warsaw Limestone. The middle Warsaw Limestone is an aquifer characterized by dissolutionally enlarged fractures and joints. According to the case study file reports, perched water in the regolith can breach the confining unit where it is thin and enter cave streams formed in the middle Warsaw Limestone. Wells completed in the bedrock aquifer have been dry on different occasions, indicating seasonal saturated and unsaturated conditions.
Twenty-five wells were installed to determine the extent of contamination at the site. Concentrations of TCE and some of its breakdown products were detected in the ground water to 22 m. The highest dissolved TCE concentrations (0.7 to 5 mg/L) were detected in a spring discharging from the base of the regolith approximately 200 m from the source area. The highest dissolved chlorinated solvent, 1,2-DCE (8.82 mg/L) was detected in a shallow well completed above the bedrock. In 1996, a 20-cm-thick black/dark brown column of DNAPL was detected in a bailer removed from this well. No DNAPL has been detected in wells completed in the limestone aquifer.
Case study 95-501 demonstrates the occurrence of elevated chlorinated-solvent concentrations in the saturated zone of the regolith, without confirmed presence of DNAPL. This site is a manufacturing facility that was in operation from 1961 to 1986. In 1989, a site investigation determined that an unknown amount of TCE had leaked from an onsite vapor degreaser. Additionally, usage and storage practices contributed to extensive spillage at the facility. Remedial activities at this site have included soil removal and a recovery well system to prevent further contaminant migration.
The site is located in the inner Central Basin. Stratigraphy consists of approximately 3 m of silty clayey regolith overlying thick beds of fossiliferous Ordovician limestone interspersed with shale. According to the file reports, horizontal bedding planes in the top 20 m of the bedrock exhibit dissolutionally enlarged openings and interconnected fractures and joints. Ground-water elevations in the regolith and upper 20 m of the bedrock are similar, indicating hydraulic connection between the two units. Pumping-test data indicate a hydraulic connection with a creek approximately 30 m away. Below 20 m, the bedrock is characterized by low fracture density and thin shale partings. Several deep borings completed within this zone were dry.
Six open-hole bedrock wells and 35 screened wells were installed to characterize the contamination at the site. Dissolved chlorinated solvents, including TCE and its breakdown products (vinyl chloride and cis-1,2-DCE), were detected in the regolith (horizontal extent about 60 by 430 m), and in the upper 10 to 20 m of the bedrock (horizontal extent about 200 by 360 m). The highest dissolved TCE concentration (290 mg/L) was detected in a shallow (less than 1 m deep) well at 26 percent of the solubility limit. However, the presence of DNAPL is not mentioned in the file reports. TCE concentrations in water samples collected from the upper 10 to 20 m of the bedrock were less than 2 mg/L. Neither DNAPL nor chlorinated-solvent concentrations above 0.005 mg/L were detected in water samples collected from bedrock wells below 20 m. The interconnected fractures in the upper bedrock have likely facilitated the lateral movement of the dissolved contaminants at the site. TCE was detected in a water sample collected from the nearby creek at a concentration of 1.2 mg/L.
| Next | Title Page | Table of Contents | List of Illustrations | List of Tables | Conversion Factors and Vertical Datum | Glossary |