INTRODUCTION
In response to changes in land use activity during the next few decades administrators of Clarke County and the Lord Fairfax Planning District Commission have designed a groundwater protection plan in an attempt to insure adequate future groundwater resources for Clarke County. Current groundwater sources in the county include both springs and wells. This study was initiated during the formulation of Clarke County’s groundwater protection plan in order to provide planners with the geologic data and interpretations essential to their study.
Approximately 74 percent of the county is located on folded and fractured carbonate rocks that exhibit karst development. The term “karst” refers to terrain characterized by the differential solution of bedrock, development of underground drainage, and distinctive features such as pinnacled bedrock, sinkholes, and caves (Hubbard, 1983). Soil cover in Clarke County is relatively thin, but significant variations in thickness occur over horizontal distances of a few feet. Dolomitic beds generally have little or no soil cover while adjacent limestone beds are generally overlain by many feet of soil. Surficial basin-like, funnel-shaped, or vertical-sided depressions in soils or sediments are termed sinkholes. Sinkholes usually form by the subsidence or collapse of unconsolidated materials into voids resulting from the dissolution of carbonate bedrock but the location and size at which sinkholes form can be affected by man’s activities. These subsidence features may be significant input points for surface waters entering the subsurface drainage system. Pollutants entering sinkholes can access the groundwater system without undergoing filtration, through the soil profile, and cause contamination of the groundwater.
The subsurface “plumbing” system in the carbonate bedrock is due to solutional enlargement along bedding and joint and cleavage fractures. Solution openings provide flow paths in an aquifer dominated by a diffuse sponge-like network containing a number of large solution conduits, indicated by large springs and a small number of caves.
The extreme variability and unpredictability of groundwater flow paths, combined with the potential for rapid influx of pollutants into the subsurface drainage network, should cause concern for the susceptibility of groundwater to pollution hazards in karst terrains. The generally thin soils and enhanced permeability, because of carbonate rock dissolution, in karst areas provide little filtering of bacteria or viruses from agricultural or domestic sewage effluents. Once in the subsurface, bacteria and viruses are protected from the ultraviolet radiation of the sun, which effectively reduces their numbers in shallow surface waters (Aley and Thomson, 1981). The poor filtering characteristics of thin soils and sinkholes allow sewage nutrients and toxic and carcinogenic chemical pollutants to readily degrade groundwater quality.
Pollutants in the groundwater system may move miles in a few hours through solution conduits (Crawford, 1984). Seemingly confined concentrations of pollutants floating on the surface of groundwater may migrate during groundwater level fluctuations. Floating pollutants trapped in solution pockets may be forced into upper-level subsurface overflow conduits or back upstream in conduits during periods of high-water levels. Accumulations of hydrocarbon or other toxic fumes may be forced out of solution pockets and rise into homes and wells by high groundwater conditions (Crawford, 1984 and 1986). Potentially explosive or toxic pollutants may remain absorbed in extensive fine-grained sediment accumulations in conduits after high groundwater levels have receded (Crawford, 1984) and may result in fumes in houses or fumes and liquid impurities in wells. Droughts reverse the process and may lower groundwater levels sufficiently to allow floating pollutants to escape ceiling pockets in solution openings.
A number of hydrologic and hydrogeologic studies have been conducted in and around Clarke County. An early hydrologic study of the groundwater resources of the Shenandoah Valley (Cady, 1936) did not include Clarke County specifically, but made note of the lack of perennial surface drainage in the county and implied a dependence of residents on groundwater resources. A report on the groundwater hydrology of adjacent Jefferson County, West Virginia (Hobba, 1981) noted that much of that county is underlain by fractured and cavernous limestone aquifers that are susceptible to contamination. High nitrate concentrations were “almost always associated with limestone aquifers.” Attempts to trace the movement of groundwater by the injection of dye into two sinkholes were not successful even after eight weeks of monitoring discharge points, indicating the unpredictability of groundwater velocity and direction in this karst terrain. A hydrogeologic study was conducted of the National Fisheries Center in Leetown, Jefferson County, West Virginia, approximately 9 miles north of Clarke County (Jones and Deike, 1981). The geology of this area of Jefferson County is very similar to that of western Clarke County. Findings of the study indicated that all springs and wells in the vicinity of the Fisheries Center originate from a common aquifer which has a diffuse flow pattern through a complex system of solution-enlarged joints and fractures. Recharge for the area is itself diffuse and dye tracing indicates multiple resurgences and relatively long groundwater residence times (60 to 80 days), atypical of karst regions (Jones and Deike, 1981). Dye tracing in the Leetown area did not reveal any specific integrated solution networks but the report did acknowledge the presence of solution pockets and cavernous zones.
A study by Schnabel Engineering Associates (1983) addressed the hydrogeology of the Prospect Hill Spring within Clarke County. Parameters such as rock type, fracture trends, solutional modification of the carbonate rocks, soil characteristics, and the southwestward regional groundwater flow trend were examined. Only the soil characteristics parameter and surface drainage divides were indicated on the land-use map illustrating the report’s recommended precautions. Recommendations on how to protect the spring included the designation of land-use restriction zones and the installation of a network of observation wells. The delineation of the restricted zones and the location of the observation well sites apparently was based upon isotropic (and probably unrealistic) groundwater flow and “approximate drainage divides” (Schnabel Engineering Associates, 1983).
Two hydrologic studies were conducted in Clarke County in 1986. Sandberg (1986) discussed septic drainage field contamination, leaking underground storage tanks, and the gypsum method aerial spraying program in the Blue Ridge Mountains of the county. Smith (1986) generated a water-table contour map and discussed some pollution problems in the valley region of Clarke County.
An overview of the groundwater resources of Clarke County (Jones, 1987) included information on thehydrogeologic setting, geologic influences, karst landforms, development of secondary fractures in carbonate rocks, groundwater theory, karst drainage basins, pollution risk assessment, and tracer tests. Jones utilized findings of an earlier study (Jones and Deike, 1981) in Jefferson County, West Virginia, with research conducted within Clarke County to classify the carbonate aquifer as having anisotropic permeability and diffuse infiltration developed in a slightly enlarged fracture network. Dye tests in Jefferson and Clarke counties indicated a half-radial groundwater flow pattern parallel to the local strike and down the hydraulic gradient. The Clarke County dye traces indicated groundwater flow to the east and southeast with about a one-and two-thirds longer travel time than for similar distances for the earlier Jefferson County traces. The longer travel times are explained in terms of the “exceptionally tight rainfall” experienced during the test period. Jones (1987) reported that elevated bacteria and nitrogen levels and a few localized oil spills are associated with Clarke County groundwater. Measures to protect the carbonate rock aquifer include the exclusion of landfills or other dumping from carbonate rock areas and the prohibition of waste disposal in sinkholes or caves. Additional concerns about storage tanks, septic drain fields, runoff from areas containing agricultural wastes, and the use of fertilizers and pesticides are described (Jones, 1987).
The U.S. Geological Survey Water Resources Division in Richmond, Virginia has completed in situ hydrological study of the county, which includes a description of the groundwater flow system and the distribution of the parameters describing the water quality in wells and springs (Winfield Wright, 1990, personal communication).
The direct test results of Jones (1987) and Jones and Deike (1981) indicate that the diffuse groundwater movement in the Clarke County aquifer is slower than the flow expected from a well developed conduit karst system. The small number of known caves in Clarke County can be interpreted as an indication that large solution conduits are not as prevalent as in other karst areas. The fact that some caves and large carbonate-rich springs exist in the county is evidence of solution conduit development and the potential for rapid transport of pollutants along unknown routes. In addition, the diffuse movement of groundwater that is characteristic of groundwater flow within the county indicates that there is a potential for widespread dispersal of pollutants over a period of months before a problem would be detected.
Clarke County Pollution Problem
A number of groundwater pollution problems have been reported in Clarke County. Problems include the presence of fecal bacteria and high nutrient (nitrate) concentrations, toxic and carcinogenic chemicals, and hydrocarbon fumes and residues.
A review of the Environmental Protection Agency’s STORET System of well data for Clarke County by the Lord Fairfax Planning District Commission found 47 wells with fecal-coliform bacteria (Most Probable Number Method) and 9 wells with nitrate levels greater than 10 mg/L (Lord Fairfax Planning District Commission, 1985, written communication). Both the fecal-bacteria and nitrate levels could be the result of either sewage tank absorption fields or agricultural sources.
In October of 1980, E. O. Gooch and Associates conducted a study of the Berryville sewage lagoon, that was reportedly leaking at an approximate rate of 55,000 gallons per day. The study reported, “In those portions of the lagoon bottom that were not covered with thick layer of sludge we found areas in which the liner was either missing (rock exposed) or the liner was not the required two feet thick.” Additionally, a number of “disturbed Shelby tube samples” were “too soft to test” or had unacceptably high permeabilities (E. O. Gooch and Associates, 1980). The Berryville sewage lagoons were constructed in 1968 on carbonate rocks of the Cambrian-age Elbrook Formation (Edmundson and Nunan, 1973).
Complaints of strange odors and taste prompted testing of the Town of Berryville’s water wells (The Clarke Courier, October 9, 1980). Testing revealed that the wells were contaminated with phenols, xylene, nitrates (State Water Control Board, December 30, 1980, written communication), creosol, and trimethylbenzene (The Clarke Courier, March 26, 1981). At that time, the wells were “within several hundred feet of” “a facilities company, two-cold storage facilities, a box factory, a basket factory, and an oil distributor.” Additionally, a storage tank was located approximately 500 feet away from the northern-most well and was found to contain the fertilizer “liquid N” during the initial on-site investigation (State Water Control Board, October 9, 1980 and December 30, 1980, written communications). An “open drainage way flows to the west of the well sites. During our visit of December 9, 1980 there was a significant flow through the channel. However, at a point approximately 100 feet west of the North Well, flow terminated and apparently was infiltrating through the bottom of the channel… It is highly possible that runoff waters entering the drainage channel are the source of the contaminants” (State Water Control Board, December 30, 1980, written communication). The Berryville town wells were constructed in 1953 to depths of 365 feet (north well) and 230 feet (south well), and grouted to depths of 100 feet and 53 feet, respectively. Both wells have static water levels 93 feet below the surface (State Water Control Board, 1986, written communication). During the years that these wells were in operation, about six sinkholes formed to the east of the well sites (Jim O’Brien and Jim Sipes, 1985, personal communication). Approximately 500 feet to the southwest, Dog Run (known locally as Town Run) has been observed to sink in its channel during periods of low flow (Jim O’Brien, 1985, personal communication).
Complaints of hydrocarbon contamination have been investigated at three sites in Clarke County. Occasional complaints of odors and the taste of hydrocarbons from well water have been investigated at Waterloo and White Post. Water samples taken at these locations have yet to confirm problems. Both sites were near operating or abandoned service stations (Clarke County Health Department, November 25, 1985, written communications; State Water Control Board, February 5, 1986 and March 31, 1986, written communications). The third site of hydrocarbon pollution is the Pine Grove area of the county (Ken Hinkle, 1986, personal communication) which overlies clastic rocks of the Cambrian-age Harpers Formation.
Description of Map Components
The karst component map contains the locations of sinkholes, caves, springs, and LANDSAT lineament zones. The sinkholes indicated on this map were transferred from maps by Edmonds and Stiegler (1982) andHubbard (1983). A total of 806 sinkholes are indicated on this map. The locations of the five caves plotted on this map are from Douglas (1964), Hollinger (1975), and Winfield Wright of the U.S. Geological Survey (1986, personal communication). The spring locations were provided by the U.S. Geological Survey (1986, written communication) and are depicted on this map as having elevations of less than 10, 10 to 100, greater than 100 gpm, or no data available. LANDSAT lineament zones represent linear areas observed to contain two or more parallel to subparallel lineaments. These zones were identified by T. M. Gahrigh, II (1986, personal communication) using orthoscopic and oblique viewing of band 7, unenhanced, black and white, LANDSAT imagery. The geologic significance of LANDSAT lineaments is not well understood, however, the lineament zones depicted may contain areas of dense fracturing of the bedrock and may serve as recharge areas into the groundwater system. The systematic recognition of lineaments from conventional aerial photography was inhibited by the extensive agricultural use of the county.
Discussion of Map Components
Approximately 50 percent of the sinkholes displayed are within the drainage basin of Opequon Creek, a tributary of the Potomac River. The drainage basin of Opequon Creek within Clarke County represents only 20 percent of the county area. Most of these sinkholes are associated with the main belt of soil types classified as having less than severe restrictions for septic tank absorption fields (Edmonds and Stiegler, 1982). These soils overlie the Stonehenge and Rockdale Run Formations.
Other concentrations of sinkholes are found north and south of U.S. Highway 50 adjacent to the Shenandoah River, and are developed on the Tomstown and Waynesboro Formations. The southern of these two groups may be related to steep drainage gradients. The northern group is developed on terrace deposits in a meander bend of the river. Sinkhole development may be attributed to flow along solutionally enlarged fractures through rocks of the meander neck.
A small cluster of sinkholes located about a mile northeast of Double Tollgate is separated from a similar cluster to the north by the Opequon Creek divide. This southern cluster may have been part of Opequon Creek basin until its drainage was pirated by Borden Run, a tributary of the Shenandoah River. Geomorphic evidence of this drainage shift is the continuation of a low ridge that comprises the Opequon Creek divide to the north, located to the east of this southern cluster.
A concentration of sinkholes north and southwest of Cool Spring in the northeastern corner of the county appears to be lithologically controlled. These sinkholes are located in a narrow belt around the noses of the anticline at Cool Spring and the adjacent syncline to the southwest. Structural control of sinkhole development includes increased fracture permeability related to axial-plane cleavage and to the longitudinal, transverse, and diagonal fractures developed during folding. Axial aligning and fold nose clustering of sinkholes reflects this structural influence. Examples include sinkholes on the syncline east of Calmes Neck located between the Shenandoah River and Willow Lake and on the southern end of the anticline through Wadesville.
As previously noted, sinkholes should be regarded as potential sites for the influx of pollutants into the groundwater supply. The 806 sinkholes indicated on this map are only a fraction of the total number of sinkholes in the county. Areas that contain high densities, or linear trends, of sinkholes probably will contain more sinkholes than were detected in this study. Potential sources of pollution in these areas should be monitored closely.
Another problem associated with sinkholes is subsidence or collapse. The development of new sinkholes is more likely to occur in the vicinity of other sinkholes. The occurrence or reactivation of sinkhole subsidence or collapse may result from a number of conditions, including changes in the water-table elevation. Rapid and extreme groundwater fluctuations associated with the pumping of wells, either during well development or subsequent heavy use, has been established as a cause of sinkhole subsidence or collapse. Sinkhole development associated with the previously mentioned Berryville wells is one example of this hazard. Other documentation includes: Foose (1968), Newton and Hyde (1971), and Newton and others (1973). Other causes of sinkhole subsidence or collapse include increases in loading, vibration, and increases in storm water runoff. The use of sinkholes as drainage outfalls, for disposal of channelled effluents, is discouraged. Increased hydrologic input into sinkholes can result in induced subsidence or collapse (Kennedy, 1980) as well as in the propagation of new sinkholes. The larger the volume of the hydrologic input the greater the risk.
The elevation of springs and the static water level in wells serve to define the potentiometric groundwater surface (water-pressure surface) which may have extreme variability in relief. Water-bearing zones perched on impermeable or low permeability rocks may further complicate the mapping of this surface. Maps of the potentiometric surface can be used to predict the direction of groundwater flow and indicate general areas of recharge. Conduit flow paths may vary significantly from the general flow direction indicated by these maps. Large springs and high-yield wells are indicators of the presence of conduit flow.
Wells depicted as having had fecal coliform contamination (Most Probable Number Method) are widely distributed over the county. These and nitrate-contaminated wells do not show a close relationship to sinkhole occurrences. The distribution of wells with high bacteria and nitrate levels indicates places where these pollutants were detected, but not necessarily where they entered the groundwater system. Where polluted wells have high yields and conduit flow of groundwater is suspected, pollution sources may be located miles away. Where polluted well yields are low, diffuse groundwater flow may be more prevalent and pollution sources are more likely to be local. Fractured carbonate bedrocks overlain by shallow soils provide little if any filtering of domestic or agricultural sewage effluents and thus may allow pollution direct access to the groundwater.
The LANDSAT lineament zones may represent corridors of higher permeability as a result of fractures or solutional enlargement of fractures in the bedrock. At present, however, there is little or no data that would demonstrate a relationship between groundwater production and LANDSAT lineament zones in Clarke County. Fewer sinkholes occur within lineament zones than in surrounding areas as demonstrated by the two sinkhole concentrations adjacent to the Shenandoah River. These concentrations are separated by a lineament zone, that projects to a traverse fault. A number of springs occur along these lineament zones, however, andwater-table elevations and yields of wells located in these zones should be compared with surrounding well data to determine the importance of these zones to groundwater flow patterns and recharge.
The integration of information about the distribution of sinkholes, caves, springs, and lineament zones with a potentiometric groundwater surface map and dye tracing studies will further the understanding of groundwater resources and problems.
HYDROGEOLOGIC REFERENCES
Aley, Tom, and Thomson, K. C., 1981, Hydrogeologic mapping of unincorporated Greene County, Missouri, to identify areas where sinkhole flooding and serious groundwater contamination could result from land development: Project Summary prepared for Greene County Sewer District, by Ozark Underground Laboratory, Protem, Missouri, under contract with Missouri Department of Natural Resources, 11 p.
Cady, R. C., 1936, Ground-water resources of the Shenandoah Valley, Virginia: Virginia Geological Survey Bulletin 45, 137 p.
Crawford, N. C., 1984, Toxic and explosive fumes rising from carbonate aquifers: A hazard for residents of sinkhole plains, in Beck, B. F., ed., Sinkholes: their geology, engineering and environmental impact: Proceedings of First Multidisciplinary Conference on Sinkholes, Florida Sinkhole Research Institute, University of Central Florida, Orlando, p. 297–304.
Crawford, N. C., 1986, Groundwater pollution from leaking underground storage tanks in karst terrain [abs.]: American Association for the Advancement of Science, Abstract of Papers, 152nd Annual Meeting, p. 28.
Douglas, H. H., 1964, Caves of Virginia: Falls Church, Virginia, Cave Survey, 761 p.
Edmonds, W. J., and Stiegler, James, 1982, Soil survey of Clarke County, Virginia: U.S. Department of Agriculture, Soil Conservation Service, 211 p.
Edmundson, R. S., and Nunan, W. E., 1973, Geology of the Berryville, Stephenson, and Boyce quadrangles, Virginia: Virginia Division of Mineral Resources Report of Investigations 34, 112 p.
Foose, R. M., 1968, Surface subsidence and collapse caused by groundwater withdrawal in carbonate rock areas: Proceedings, 23rd International Geological Congress, Prague, v. 12, p. 155–166.
Gooch, E. O., and Associates, 1980, Flow deficit study for Berryville Lagoon: unpublished report to John McNair and Associates, December, 1980, Charlottesville, Virginia, 7 p.
Hobba, W. A., Jr., 1981, Ground-water hydrology of Jefferson County, West Virginia: West Virginia Geological and Economic Survey, Environmental Geology Bulletin EGB-16, 16 p.
Holsinger, J. R., 1975, Descriptions of Virginia caves: Virginia Division of Mineral Resources Bulletin 85, 450 p.
Hubbard, D. A., Jr., 1983, Selected karst features of the northern Valley and Ridge province, Virginia: Virginia Division of Mineral Resources Publication 44, map.
Jones, W. K., 1987, Overview of the ground water resources of Clarke County, Virginia with emphasis on the carbonate aquifers west of the Shenandoah River: Report to Clarke County Administration, 22 p.
Jones, W. K., and Deike, G. H., III, 1981, A hydrogeologic study of the watershed of the National Fisheries Center at Leetown, West Virginia: Report prepared for U.S. Department of Interior, U.S. Fish and Wildlife Service, Frankford, West Virginia, 79 p.
Kemmerly, P. R., 1980, Sinkhole collapse in Montgomery County, Tennessee: Tennessee Division of Geology, Environmental Geology Series No. 6, 42 p.
Newton, J. G., and Hyde, L. W., 1971, Sinkhole problems in and near Roberts industrial subdivision, Birmingham, Alabama: Geological Survey of Alabama, Circular 68, 42 p.
Newton, J. G., Copeland, C. W., and Scarborough, L. W., 1973, Sinkhole problem along proposed route of Interstate Highway 459, near Greenwood, Alabama: Geological Survey of Alabama, Circular 83, 63 p.
Sandberg, Scott, 1986, An assessment of the potential for groundwater contamination in the Blue Ridge Mountains of Clarke County, Virginia: Report to Clarke County Administration, 28 p.
Schnabel Engineering Associates, 1983, Hydraulic and engineering study, Prospect Hill Spring, Clarke County, Virginia: Report to Prospect Hill Spring Committee, Town of Berryville, Virginia, Richmond, Virginia, 25 p.
Smith, Craig, 1986, A hydrogeologic study in the Valley Region of Clarke County during the summer of 1986: Report to Clarke County Administration, 12 p.