The Integration of Data Review, Remote Sensing and Ground Survey for a Regional-Level Karst Assessment ROBERT KENNETH DENTON JR.1 GeoConcepts-Engineering, Inc., 19955 Highland Vista Drive, Suite 170, Ashburn, VA 20147
ASHLEY HOGAN Geocapital-Engineering LLC, 4545 42nd St. N.W., Washington DC 20016
RONALD DREW THOMAS ECS Mid-Atlantic LLC, 14026 Thunderbolt Place, Chantilly, VA 20151
Key Terms: Karst, Sinkhole, Remote Sensing, Pipeline, Isopod, Habitat, Conservation Plan ABSTRACT Detailed karst terrain assessments require the identification and survey of surface features such as closed depressions, sinkholes, and cave entrances. Typically, surveys are carried out at sites encompassing several hundred acres or less; however, the traditional methods have proven impractical from a time and expense viewpoint for extensive, regional-level surveys in well-developed karst terrain. This subject survey covered a 76-mi (122.3-km) length of a natural gas transmission pipeline. The goal of the study was to assist the United States Fish and Wildlife Service in assessing habitat vulnerability for the Madison Cave Isopod, a federally protected threatened species found only in phreatic ground water of the Great Valley of Virginia and West Virginia. The survey used an integrated approach involving the evaluation of topographic maps, digital elevation models, shaded relief maps, satellite imagery, and historic aerial photographs to identify “concentrations” of karst features. Ground surveys were then undertaken by walking sections of the pipeline right-of-ways that occurred within karst concentrations, and documenting the features using GPS instrumentation. The karst survey encompassed 48,640 acres (19.2 hectares), and was the largest utilityassociated karst assessment ever conducted in the Commonwealth of Virginia. Approximately 216 closed depressions and 28 cave entrances were located, identified, and described based on their geology, physical appearance, and drainage characteristics. The use of an integrated approach significantly reduced the time and cost of the study. The study’s findings and 1
Corresponding author email:
[email protected].
recommendations have been used to develop conservationbased avoidance and minimization measures intended to limit the impact to the species’ habitat. INTRODUCTION AND BACKGROUND The cirolanid isopod Antrolana lira Bowman 1964, commonly known as the Madison Cave Isopod (MCI), is a free-swimming crustacean native to the phreatic karst aquifer of the Great Valley of Virginia. First collected in 1958 by Thomas Barr of the University of Kentucky, the holotype and six paratype specimens were subsequently described by Thomas Bowman of the United States National Museum in 1964. The type locality was limited to two pools in Madison Cave (Figure 1) and a small pool in an adjacent cave named Steger’s Fissure, both located in the Grottoes area of Augusta County, Virginia (Bowman, 1964). Studies suggested that A. lira was threatened by human visitation to its only known habitat, and by mercury pollution of the nearby South River (Bolgiano, 1980; Collins, 1982). As a result, the taxon was proposed as a threatened species by the U.S. Fish and Wildlife Service (USFWS) on January 12, 1977 in the Federal Register (42FR 2507-2515). The threatened status of A. lira was finalized in a rule issued by the USFWS on October 4, 1982 (47FR 43699-43701). Subsequent investigations (Orndorff and Hobson, 2007) have extended the range of A. lira beyond its original type locality to a much larger area of the Great Valley (Figure 2). The MCI appears to have a specialized habitat preference for the high conductance, ionically saturated phreatic waters reposing in a specific group of carbonate rocks ranging in age from the Lower Cambrian through the Middle Ordovician, and solely within the Great Valley section of the Valley and Ridge Physiographic Province.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
81
Denton, Hogan, and Thomas
Figure 1. The carbonate saturated water of the East Lake in Madison Cave, the type locality of A. lira. Note the calcite “rafts” floating on the surface of the water in the right center part of the photograph. (Plastic tubes are housings for pressure and conductivity sensors.)
It is notable that A. lira has not been found in the western portion of the Great Valley north of the North Fork of the Shenandoah River. Nevertheless, the carbonate rocks in which the preferred habitat is located are contiguous along the strike through this section. Thus, it is difficult to explain the absence of the MCI beyond the North Fork, especially considering that other consequent streams and rivers (i.e. streams that cross the bedding of the local rock), such as the Maury River in Rockbridge County, have not acted as a boundary to the MCI’s range. Nevertheless, the area north of the North Fork is still considered part of the taxon’s “suspected” range. The phreatic aquifer in which A. lira lives is connected to the surface through the epikarst zone. There are few perennial surface streams in karst areas, and much of the surface runoff is diverted underground through closed drainages. Recharge to the
82
Figure 2. Map showing the original type locality of A. lira near Grottoes, Augusta County, Virginia, and the inferred range of the taxon based on subsequent studies, with the three sections of the karst survey AOI, a one mile wide corridor paralleling the existing pipeline ROW.
subsurface can be rapid, specifically where water plunges into sinkholes with open throats, percolates through the bed of losing streams, or flows into cave entrances. In contrast, groundwater recharge can also be slow in places where water percolates down through the surface soils and infiltrates a diffuse network of cracks and fissures located at the soil/ bedrock interface. In many areas where the bedrock is mantled with clay-rich, cohesive soils, the infiltration would be expected to be extremely slow; however, periods of extended drought can open up deep fissures in the soil that can rapidly channel water directly to the underlying bedrock. Similarly, bedrock pinnacles that extend above the soil surface can channel significant quantities of water to the underlying epikarst, particularly if the surrounding soils dry out, shrink, and contract away from the surface of the pinnacles. All of these factors can affect the ionic concentration of the underlying phreatic aquifer and the habitability of the aquifer for the MCI.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
Figure 3. Idealized geological column of the three segments of the survey area.
Because of the myriad of variables associated with the habitat of A. lira, effective survey methods to determine the taxon’s abundance and suitable habitat have yet to be developed. Thus, the USFWS assumes the presence of the species if a project is located within a potential habitat area. Currently, potential habitat areas include the phreatic aquifers of any of the Cambrian and Ordovician carbonate rocks within the Great Valley. Accordingly, the USFWS recommends implementation of measures to minimize effects on A. lira and its habitat, among which are karst assessments intended to identify all surface karst features within an area of potential impact. GeoConcepts Engineering was contracted in June 2009 by The Conservation Fund on behalf of the USFWS, to provide a karst survey for the NiSource/ Columbia Gas Transmission Pipeline sections that extend through potential habitat areas of A. lira in the Great Valley. The following report summarizes the procedures and findings of that survey with comments on the utility of integrating the data review, remote sensing, and subsequent ground survey phases of the study.
GEOLOGICAL SETTING According to the Geological Map of Virginia (Virginia Division of Mineral Resources, 1993), the project is located entirely within Great Valley Section of the Valley and Ridge Physiographic Province. The Great Valley is a generally downwarped trough (synclinorium) of Paleozoic limestones, shales, and sandstones; it lies between the Blue Ridge Massif on the east and the Allegheny Mountains to the west. The Valley extends between the two mountain uplands from northeast to southwest, parallel with the strike of the bedrock. Specifically, the survey area has been mapped as underlain by a series of carbonate and clastic rocks ranging in age from Lower Cambrian to Middle Ordovician. A geologic column for the study area is presented as Figure 3. Karst Geology As in any region where soluble bedrock is present, a karst landform regime has developed in the Great Valley. All of the bedrock units present throughout
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
83
Denton, Hogan, and Thomas
the project area are capable of undergoing dissolution with the exception of the clastic rocks of the Middle and Upper Martinsburg Formation. Folding and faulting of the local carbonate rocks has opened up numerous fractures both parallel with the axis of the geologic structures, as well as perpendicular to them. Surface fractures and joints weather differentially, producing a pinnacled or “sawtooth” profile at the bedrock/soil interface (referred to as the epikarst zone). In contrast, rockenclosed fractures can be secondarily enlarged by the action of carbon dioxide charged groundwater, in some cases forming water-filled or air-filled conduits. As the regional terrain is “mature” karst, nearly all the fractures have undergone successive cycles of sediment filling and flushing. In areas such as the survey area, where there is little topographic relief and a relatively minimal groundwater gradient, the great majority of solution fissures are sediment-filled (Ford and Williams, 1989; Moore and Wade, 1989). METHODOLOGY For ease of explanation, the Area of Interest (AOI) for the assessment comprised an area of “covered lands” encompassing K mi (0.804 km) on either side of the NiSource Pipeline in sections that crossed potential or suspected MCI habitat. Within the covered lands was the pipeline Right of Way (ROW), which encompassed 80 ft (24.3 m) on either side of the pipeline. The entire length of the pipeline mapped as passing through known or suspected MCI habitat was divided into three discreet survey sections (Figure 2): 1. Northern Section (20 mi; 32.2 km) passing through Clarke, Warren, and Shenandoah counties; 2. Central Section (17 mi; 17.3 km) passing through Rockingham and Page counties; 3. Southern Section (39 mi; 62.8 km.) passing through Augusta and Rockbridge counties.
Data Review and Remote Sensing Phase (DR-RSP) Because the AOI encompassed approximately 76 sq mi (196 km2), it was beneficial to identify potential karst features remotely and/or by database review, and then confirm their actual presence in the field. This process significantly reduced the amount of time spent on location and on survey tasks. An inventory of known karst features located within the covered lands was reviewed from the following sources:
84
1. Caves of Virginia (Douglas, 1964); The Cave Database of the Virginia Speleological Survey (VSS; proprietary); 2. Maps of selected karst features (sinkholes, caves, and springs) available from the Virginia Division of Mines and Mineral Resources and the United States Geological Survey (USGS); 3. 2-ft and 4-ft contour interval maps for the counties containing AOIs (to determine the presence of surface karst features not included in the previously listed databases based on the presence of closed, descending contours or other suspect karst “fingerprint” features). Contour maps were generated from Triangular Irregular Network (TIN) data for each county available from the Virginia Geographic Information Network (VGIN); 4. Aerial photographs (both recent and historical) available from VGIN; 5. USGS 7.5-minute topographic quadrangles. In addition, readily available geological literature for bedrock and structural characteristics was reviewed (Virginia Division of Mineral Resources, 1993; Orndorff and Goggin, 1994; Gathright et al., 1996; Orndorff et al., 1999; Rader and Gathright, 2001; Campbell et al., 2006; 2007; and Southworth et al., 2009). The survey relied upon the highest resolution mapping available for the particular AOI being examined. The result of the DR-RSP was to identify areas of surface karst “concentrations” or clusters within the AOI that were then subject to closer scrutiny through field reconnaissance. Initial Assessment The initial assessment of karst concentration locations was determined using two studies undertaken by the Virginia Division of Mines and Mineral Resources (Hubbard, 1983; 1988). At the time of our investigation (2009) the Hubbard studies had not yet been rendered as digital files, and the published, paper maps lacked the resolution necessary to identify specific features smaller than several hundred feet in diameter. Nevertheless, the Hubbard maps would provide important clues to areas of significant karst feature concentrations (Hubbard, 2003). Cave location data was also used with the knowledge that many of the cave entrance coordinates were determined by manual approximation using 7.5-minute topographic quadrangles. Thus, reported cave entrance locations were expected to have significant errors of up to several hundred feet or more.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
Figure 4. (a) Excerpt of the USGS New Market, Virginia 7.5minute topographic quadrangle. Red polygons are sinkholes referenced from the Virginia Department of Mines, Minerals and Energy (DMME) datasets (Hubbard, 1983). (b) DEM dataset showing suspect closed depression (yellow dashed circle). This feature did not appear on the topographic quadrangle shown in Figure 4a and was not included in the reference DMME karst dataset.
Cartographic Data Review Four principal forms of cartographic data were used in the DR-RSP of the survey: 1. 2-ft (0.61 m), and 4-ft (1.22 m) contour maps of the various counties under study; 2. USGS 1/3 Arc-Second National Elevation Dataset (NED) digital elevation models (DEMs), Earth Resources Observation Systems (EROS) Web Map Service, and NED Shaded Relief Map series; 3. USGS topographic maps; 4. Aerial photographs, both recent and historical. The route of the pipeline ROW and covered lands was mapped using Esri’s ArcMapH geographic
Figure 5. (a) One-meter contour map derived from the DEM shown in Figure 4b of a previously unmapped feature in the New Market, Virginia Quadrangle (all contour elevations in meters). (b) Two-foot (0.61-m) contour map of the suspect closed depression shown in Figure 5a. Contours derived from TIN county data for the Commonwealth of Virginia.
information system (GIS) software, and then each section was scrutinized by toggling among each of the four data sources. Initial review consisted of examining topographic maps and comparing them to the areas of significant karst concentrations indicated in the Hubbard studies (Figure 4a). Then, the DEMs were examined to determine if smaller features not visible in the 7.5minute topographic quadrangles could be resolved (Figure 4b). Ultimately, final reliance was placed upon close-interval contour maps (Figure 5a and b), which were carefully examined for areas of closed descending contours. The contour maps were also rendered as hillside shaded relief maps for rapid scanning and evaluation. In this way, nearly the entire length of the covered lands could be scrutinized within a reasonable amount of time.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
85
Denton, Hogan, and Thomas
Figure 6. (a) Aerial photograph of the location of suspect feature shown in Figures 4a to 5b. (b) Google Earth imagery of the suspect closed depression shown in Figures 4a to 6a derived from DEMs. This feature was subsequently confirmed by field observation as a sinkhole approximately 100 ft in diameter lying within a larger close depression approximately 300 ft in diameter and encompassing approximately 1.15 acres (0.404 hectares).
Once an area was identified as a closed depression, it was then compared to the other data sources, in particular high-resolution aerial photographs (Figure 6a and b). Many areas which appeared as putative sinkholes could be eliminated as closed depressions related to farm ponds or man-made excavations in this way, the usual signature being a berm on the downgradient side of the depression. Finally, suspect closed sinkholes were examined using the historical capabilities of the Google EarthTM application. This allowed us to compare closed depressions during relative “wet” and “dry” years, based on comparative climate data. Typical wet years used for this analysis were 2010, 2007, 2003 and 1997; dry years were 2005, 2002 and 2000. In many cases, structures that appeared as relatively nondescript vegetated closed depressions during dry years
86
exhibited the presence of standing water, or active resurgence flow, during wet years (Figure 7a and b). In cases where the depression exhibited an active outflow of water, the structure could be classified as an “estavelle”, i.e., a sinkhole that acts as a recharge point during dry seasons and a resurgence point during wet seasons. In some instances there was no outflow during wet seasons, but the depression clearly contained standing water, qualifying it as a “turloch” (i.e., dry lake). When a feature was identified as a sinkhole with reasonable certainty it was then outlined using the GIS software. Attributes were then assigned including county ID (a unique identifier created for this study), USGS quadrangle, bedrock geology, area (in square meters) and the northing and easting of the feature’s centroid. The entire set of closed depressions was then used to determine the specific areas of the pipeline ROW that would be examined closely during the field survey phase of the project. It should be noted that for the purposes of this study we considered any enclosed topographic basin with no sign of external drainage, regardless of its size, as a closed depression. However only features that could be verified as dolines (conical or funnelshaped features formed in the carbonate bedrock) or cover collapse sinkholes (in soil) were designated as “sinkholes”. We referred to any opening within a depression leading into the subsurface that could accept drainage as a “throat” rather than “ponor”. Definitions used herein were based the United States Environmental Protection Agency lexicon of cave and karst terminology (Field, 2002). Field Reconnaissance Phase (FRP) Upon completion of the DR-RSP for each section, we commenced the field reconnaissance phase. Field reconnaissance was based on two principal criteria: 1. Any suspect closed-depression or cave entrance that intercepted or appeared to receive drainage from the ROW was to be examined, regardless of whether there was a significant karst concentration in the area; 2. The ROW was to be examined by direct observation in any area where there was a significant concentration of karst features identified during the DR-RSP, regardless of whether they intercepted the ROW. In addition, suspect closed depressions were examined by direct observation if they occurred near, or could be reasonably accessed from, an intersecting roadway or residence. In all cases, the field staff
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
entrance was located within a sinkhole, then the bounds of the sinkhole were mapped as a polygon. It should be noted that many of the cave entrance locations in our database were originally mapped using manual methods; thus, the coordinates for the entrances had a considerable degree of error. In a number of cases, the cave entrances had been obliterated by development, filled by the owner, or could not be located; however, the location of the now filled entrance was retained in the database. In all cases where a cave entrance was located, the entrance was mapped using GPS, and the cave data attribute table was edited using the corrected entrance locations. INTEGRATION OF THE DATA REVIEWREMOTE SENSING PHASE AND THE FIELD RECONNAISSANCE PHASE
Figures 7. (a) Area of suspected flooding sinkholes (estavelles) south of Forestville, Shenandoah County, Virginia; parapets of the sinkholes are within the dashed lines. Photograph (a) was taken during a relatively dry period in late December 2002; photograph (b) was taken during late April 2007, during a particularly wet spring season (Google Earth, 2011).
contacted the landowners to gain access to the property to inspect the suspect karst structures, particularly those that fell outside of the ROW but within the covered lands. Much to their credit, all the contacted landowners readily gave permission to access their property, allowing the field staff to examine nearly all of the closed depressions and caves located within the covered lands that were identified during the DR-RSP. The actual field survey was performed by walking the ROW, observing and noting any previously identified surface karst features such as sinkholes, subsidence areas, or cave entrances. Any feature noted within the examined area was recorded using a handheld sub-meter accuracy GPS unit. Cave entrances were recorded as points, and sinkholes were recorded as vertex-bounded polygons. If a cave
The DR-RSP of the survey commenced in midAugust 2009, with the supporting field work commencing during September 2009. As each section of the survey area was examined during the DR-RSP, field crews would be dispatched to examine any suspect karst feature concentrations, and these data were used to correct and amend the findings of the DR-RSP observations. The entire field verification of the DR-RSP was completed in less than 2 months, with field work concluding by late October 2009. The final report summarizing the survey results was delivered in December 2009. The integration of the two methods allowed for a rapid turnaround in the delivery of a final report to our client. In the process, several important deficiencies of each method (if they had been performed individually) were revealed: 1. Remote sensing and data review, especially the use of 2-ft contour maps and DEMs, often misidentified man-made features (e.g., farm ponds, silage pits) as natural closed depressions. The integration of field verification with remote sensing allowed for the refinement of the DR-RSP data. 2. Field reconnaissance teams, without the benefit of the DR-RSP generated maps, would have been compelled to walk the entire length of the survey’s extent, including the 1/2-mi (0.80-km) wide covered land zone. We estimate this would have required at least 6 to 8 months of field work. 3. Some of the karst features that were surveyed did not appear on USGS topographic maps; however, they were observed and noted during the DR-RSP. This allowed the field crews to focus their efforts on suspect structures with greater accuracy.
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
87
Denton, Hogan, and Thomas Table 1. Summary of areal extent of geological formations and karst features. Percentage of Closed Depressions in Areal Extent Geology in Covered Number of Closed Geology per square Number of (mi2/km2) Lands Depressions (N) mile Caves (N)
Geology Martinsburg & Oranda Formations Middle Ordovician Limestones Beekmantown Group Conococheague Formation Elbrook Formation Waynesboro Formation & Tomstown Dolomite Shady Dolomite
13.8/35.74
6.1
4
0.3
0
15.16/39.26 27/69.93 27.02/69.98 24.63/63.79
6.7 11.9 11.9 10.9
88 71 18 14
5.8 2.6 0.7 0.6
13 4 5 6
14.15/36.65 4.7/12.17
6.2 2.1
10 11
0.7 2.3
0 0
4. There were instances where the karst features were actually too large or so obscured by vegetation that, without the DR-RSP-generated maps, they would have been bypassed by the field teams.
SURVEY RESULTS It is difficult to concisely present the findings of a survey of this size; however, patterns regarding the distribution of the various karst features that could be potentially affected by the pipeline can be summarized as follows. The areal distribution of geological formations throughout the survey area is shown in Table 1. As can be seen from these data, the extent of the potentially karst-forming units within the covered lands is distributed unevenly, with the majority (35%) of areal extent being underlain by the Elbrook, Conococheague, and Beekmantown units. Interestingly, the distribution of the closed depressions/ sinkholes did not follow the same trend (Table 1). Of the 227 features identified in the DR-RSP of the project, 216 were natural karst features, while the remaining 11 were man-made depressions, primarily farm ponds. Of the 33 caves documented in the various databases, 27 entrances were observed and located during the study, with the remaining entrances assumed to have been filled or otherwise destroyed. The mapped closed depressions/sinkholes demonstrated a clustering of surface-expressed karst features in the Middle Ordovician Limestones, with nearly twice the number per square mile than in the Beekmantown Group (Figure 8a, b, and c). These two units showed the most surface karst development within the study area. In general, the majority of the sinkholes overall were bowl-shaped, soil-bottomed and vegetated, with gradual slopes leading down from their well-defined edge or “parapet”, but lacked any
88
Caves in Geology per Square Mile 0 0.9 0.1 0.2 0.2 0 0
obvious throat or opening into the subsurface. Sinkholes of this type were especially common in the northern and central sections of the survey (i.e., the “Lower” Great Valley), typically considered a mature karst terrain. In contrast, sinkholes with open throats were more prevalent in the southern section of the survey. A total of eight sinkholes with open and/or debris-clogged (but readily apparent) throats were observed and documented exclusively in the Rockbridge County section of the survey. There was an interesting pattern regarding the number of closed depressions/sinkholes detected in the DR-RSP phase of the project versus the features that had been mapped for the same areas in prior studies. Of the 113 features observed and documented in the northern section there were at least 73 features that had not been documented in any prior study. In contrast, of the 22 features in the central section, and 91 in the southern section, there were 17 and 8 features, respectively, that had not been documented previously. The reason so many features had been previously documented in the southern section may be because the closed depressions/sinkholes in this area (Rockbridge County) tended to be better developed; i.e., deeper and more conical in shape (see Figure 9). It should be noted that one of the reasons for the low number of surface features and caves we documented in the Lower Cambrian Waynesboro Formation and Tomstown/Shady Dolomite was that these units were only present in the current survey’s study area near the eastern edge of the Great Valley along the western base of the Blue Ridge. This area is characterized by two very different types of surface karst features: sinkholes and caves located in widely scattered hills that are the erosional remnants of the original limestone upland; and broad shallow dishshaped alluvium-filled sinkholes in the gently rolling terrain between the hills. The alluvium-filled sinkholes are particularly common in the easternmost portion of the Great Valley, where outwash from the eroding
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
Figure 9. Well-developed conical sinkholes southwest of Lexington in Rockbridge County, Virginia, representative of the typical karst landform in this section of the Shenandoah Valley. The majority of open-throat sinkholes were found in this section of the current survey.
Blue Ridge Mountains provided the gravelly sand overburden that filled previously existing karst depressions. The fact that these sinkholes formed prior to the deposition of the alluvium is supported by the discovery of fossilized pollen dating from the Early Tertiary to Late Cretaceous in sediment
Figure 10. Broad, shallow sinkhole developed in the alluvial sediments in eastern Page County, Virginia near the western pediment of the Blue Ridge Mountains.
Figure 8. (a) Karst feature locations (green centroids) documented during the current study in the northern section of the route overlaid on the regional karst-forming carbonate bedrock. (b) Central section, (c) Southern section. (Geological Abbreviation Key: s 5 Shady Dolomite; s 5 Rome Formation; wb 5
r Waynesboro Formation; wbt 5 Waynesboro & Tomstown Formations; e 5 Elbrook Formation; O co 5 Conococheague Formation; Ob 5 Beekmantown Group; Oeln 5 New Market, Lincolnshire and Edinburg Formations; DSu 5 Silurian-Devonian Carbonates, undivided.)
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
89
Denton, Hogan, and Thomas
Figure 11. Entrance to Buzz Hole-GeoConcepts Cave located in Rockbridge County, Virginia. Previously unexplored, this cave was explored and partially mapped as part of the current study. It is being monitored during high water periods for the intrusion of the phreatic aquifer as possible habitat of the MCI.
preserved in the sinkholes below the overburden (Pierce, 1965; Tschudy, 1965). Many of the features we mapped in the eastern part of the survey within Page and Augusta counties were far too shallow to be readily apparent using 10- or 20-ft (3.04- or 6.09-m) contour interval topographic maps, aerial photographs, or the available 1/3 arc-second DEMs. However they were detected by examination of 2-ft (0.61-m) and 4-ft (1.22-m) contour interval maps
generated from TIN data. Nevertheless, we assume that only the largest buried subsurface karst features were expressed as surface depressions in the alluvium (Figure 10), and that smaller features in the underlying carbonates were hidden. Thus, the burial of the underlying karst probably accounted for the lower percentage of surface features in the eastern valley. Similar to the data for closed depression density, the highest occurrence of caves (i.e., openings into the subsurface large enough to allow a human to enter) per square mile was noted in the Middle Ordovician Limestones; however, caves in the Beekmantown Group were clearly outnumbered by the caves within the Conococheague and Elbrook Formations. This was an intriguing pattern, especially considering that the Conococheague and Elbrook Formations showed relatively few surface karst features, compared to the overlying Ordovician units. One of the ancillary benefits of our study was to provide the VSS with accurate cave entrance coordinate data for the mapped caves. Many of these cave entrances had been located long before the advent of handheld GPS survey equipment, and the entrance locations were grossly inaccurate. We provided the VSS with accurate locations for more than 20 caves within the survey area, as well as survey information for a “new” cave and the relocation of a “lost” cave. While surveying in the Lexington area of Rockbridge County our survey team noted that the pipeline ran along the upper parapet of a steeply
Figure 12. Survey map of Buzz Sink (GeoConcepts) Cave.
90
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
Regional Karst Assessment
of a triangular, 6-ft (1.82 m) deep sinkhole, partially filled with leaves (Figure 13). The actual entrance passage to Madden’s Cave was found to be clogged with dirt and plant debris that had washed in from above. It did not appear that anyone had entered the cave for many years. Madden’s Cave is approximately 100 ft (30.4 m) in depth and 275 ft (83.8 m) in length, and has been completely mapped. It is considered “very significant” by the VSS and the Virginia Cave Board because it is the type locality for two unique species: a cave beetle and a pseudoscorpion. There is also a record of the Indiana Bat (Myotis sodalis), a federally listed Rare, Threatened and Endangered Species (RTES) for this cave. Madden’s Cave is not expected to be affected by the pipeline because of its upgradient location and distance from the ROW. Figure 13. Entrance County, Virginia.
to
Madden’s
Cave
in
Shenandoah
ACKNOWLEDGMENTS sloping sinkhole near a newly constructed home. At the southeastern terminus of the depression they discovered a steep-walled sinkhole, with an open throat at its base that appeared large enough to permit human entry (Figure 11). This structure was reported to Bob Thren, chairman of the Rockbridge County section of the VSS, who was aware of its existence, but admitted that it had never been investigated or mapped. The cave was registered with the VSS and dubbed “Buzz Sink (GeoConcepts) Cave” in honor of the landowner’s nickname and the exploratory crew from GeoConcepts. A team from GeoConcepts, the Virginia Department of Conservation and Recreation–Virginia Natural Heritage Program, and the VSS returned to the newly discovered cave on November 3 and again on December 15, 2009, to determine its extent (Figure 12). The cave was primarily vertical and several hundred feet of passage were mapped to a depth of approximately 80 vertical ft (24.3 m). This was particularly intriguing, because caves of this type (i.e., vertical or semi-vertical caves that descend rapidly to the phreatic water table) are where populations of the MCI have most often been found. There is evidence that the lower levels of the cave are completely flooded during seasonal water table high stands. The cave is still under exploration and is planned to be monitored periodically for the presence of the MCI. To the north in Shenandoah County, in the heavily forested bluffs above the North Fork of the Shenandoah River, the survey relocated the entrance to Madden’s Cave. Madden’s had been considered a lost cave because the entrance could not be located when the VSS resurveyed Shenandoah County in the 1980s and 1990s. The entrance was located at the base
The authors would like to thank NiSource Gas Transmission and Storage for allowing the survey data compiled for the Multi-Species Habitat Conservation Plan to be presented in this report. We would also like to thank Wil Orndorff of the Virginia Department of Conservation and Recreation–Natural Heritage Program, for providing invaluable data, guidance and assistance. Mapping of Buzz Sink (GeoConcepts) Cave could not have been completed without the help of Bob Thren, Virginia Speleological Survey–Rockbridge County Chairman and Nathan Farrar, Virginia Speleological Survey–Page County Chairman. We would also would like to thank Cody Sheaffer and Sam Consolvo (GeoConcepts) for assistance in the preparation of various tables and figures. Finally, we want to thank the management of GeoConcepts Engineering for their endless patience and unflagging support of this work. REFERENCES BOLGIANO, R. W., 1980, Mercury Contamination of the South, South Fork Shenandoah, and Shenandoah Rivers: Virginia State Water Control Board Basic Data Bulletin 47. BOWMAN, T. E., 1964, Antrolana lira, a new genus and species of troglobitic cirolanid isopod from Madison Cave, Virginia: International Journal Speleology Vol. 1, pp. 229–236. CAMPBELL, E. V.; EVANS, N. H.; SPENCER, E. W.; AND WILKES, G. P., 2007, Geologic Map of Rockbridge County, Virginia: Department of Mines, Minerals and Energy, Division of Mineral Resources Publication 170, Plate 1. CAMPBELL, E. V.; WILLIAMS, S. T.; DUNCAN, I. J.; HIBBITS, H. A.; FLOYD, J. M.; REIS, J. S.; AND WILKES, G. P., 2006, Interstate 81 Corridor Digital Geologic Compilation: Virginia Department Mines, Minerals, and Energy—Division of Mineral Resources Open File Report 06-01. COLLINS, T. L., 1982, An Ecological Study of the Troglobitic Cirolanid Isopod, Antrolana Lira Bowman, from Madisons Saltpetre Cave And Stegers Fissure, Augusta Co., Virginia:
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
91
Denton, Hogan, and Thomas Unpublished MS Thesis, Old Dominion University, Norfolk, VA, 41 p. DOUGLAS, H. H., 1964, Caves of Virginia: Virginia Cave Survey, Falls Church, VA, 761 p. FIELD, M., 2002, A Lexicon of Cave and Karst Terminology with Special Reference to Environmental Karst Hydrology: United States Environmental Protection Agency, EPA/600/R-02/003. FORD, D. C. AND WILLIAMS, P. W., 1989, Karst Geomorphology and Hydrology: University Press, Cambridge, U.K. GAITHRIGHT, T. M. II; MCDOWELL, R. C.; ORNDORFF, R. C.; AND RADER, E. K., 1996, Geologic Map of Clarke, Frederick, Page, Shenandoah, and Warren Counties, Virginia: Lord Fairfax Planning District, U.S. Geological Survey Publication 143. GOOGLE EARTH, 2011, “Timberville Sinkholes” 38.679224u N and 78.750552u W. Imagery Dates: December 30, 2002 and April 29, 2007. Accessed: September 26, 2011. HUBBARD, D., Jr., 1983, Selected Karst Features of the Northern Valley and Ridge Province, Virginia: Virginia Division of Mineral Resources Publication 44. HUBBARD, D., Jr., 1988, Selected Karst Features of the Central Valley and Ridge Province, Virginia: Virginia Division of Mineral Resources Publication 83. HUBBARD, D. A., Jr., 2003, Use of regional sinkhole mapping for sinkhole susceptibility maps, In Beck, B. F. (ed.), ASCE Geotechnical Special Publication No. 122, pp. 61–71. MOORE, C. H. AND WADE, W. J., 2013, Carbonate Reservoirs: Porosity and Diagenesis in a Sequence Stratigraphic Framework, 2nd Ed., Elsevier. ORNDORFF, R. C.; EPSTEIN, J. B.; AND MCDOWELL, R. C., 1999, Geologic Map of the Middletown Quadrangle, Frederick, Shenandoah, and Warren Counties, Virginia: United States Geological Survey, Geologic Quadrangle Map GQ-1803. ORNDORFF, R. C. AND GOGGIN, K. E., 1994, Sinkholes and KarstRelated Features of the Shenandoah Valley in the Winchester 309 x 609 Quadrangle, Virginia and West Virginia: U. S.
92
Geological Survey Miscellaneous Field Studies Map, No. 2262. ORNDORFF, W. D. AND HOBSON, C. S., 2007, Status Survey for the Madison Cave Isopod (Antrolana lira) in Virginia, 2005–2007: Natural Heritage Technical Report 07-11. Virginia Department Conservation Recreation, Division Natural Resources, Richmond, VA, 17 pp. PIERCE, K. L., 1965, Geomorphic significance of a Cretaceous deposit in the Great Valley of southern Pennsylvania: U. S. Geological Survey Professional Paper 525-C, pp. C152 –C156. RADER, E. K. AND GATHRIGHT, T. M., 2001, Geologic Map of the Augusta, Page, and Rockingham Counties Portion of the Charlottesville 30 x 60 Minute Quadrangle: Virginia Department Mines, Minerals, and Energy—Division of Mineral Resources Publication 159 SOUTHWORTH, S.; ALEINIKOFF, J. N.; BAILEY, C. M.; BURTON, W. C.; CRIDER, E. A.; HACKLEY, P. C.; MOOT, J. P.; AND TOLLO, R. P., 2009, Geologic Map of the Shenandoah National Park Region, Virginia: U. S. Geological Survey Open File Report 2009–1153. “THE CAVE DATABASE OF THE VIRGINIA SPELEOLOGICAL SURVEY,” Virginia Speleological Survey Proprietary Database, accessed October, 2009. THE NATIONAL MAP (Digital Elevation Models, 7.5-Minute Topographic Quadrangles) Courtesy of the United States Geological Survey. Retrieved from http://viewer.national map.gov/basic/ TSCHUDY, R. H., 1965, An Upper Cretaceous Deposit in the Appalachian Mountains, in Geological Survey Research 1965: U.S. Geological Survey Professional Paper 525-B, pp. B-64–B-68. VIRGINIA DIVISION OF MINERAL RESOURCES, 1993, Geologic Map of Virginia: scale 1:500,000. VIRGINIA GEOGRAPHIC INFORMATION NETWORK, 2009, Triangular Irregular Network Dataset for Virgina Counties. Retrieved from http://vgin.maps.arcgis.com/home/
Environmental & Engineering Geoscience, Vol. XXII, No. 1, February 2016, pp. 81–92
