Purpose and Scope

The purpose of this report is to summarize the findings of this study, including descriptions of the topographic, lithologic, physiographic, and structural controls on fracture occurrence and distribution. Understanding these processes controlling fracture distribution across West Virginia can help to better assess potential depths of active fracturing of the bedrock, the depth at which fractures diminish or become so tight as to preclude storage or yield of water to a well, for assessing the potential fate and transport of contaminants within fractured-rock aquifers in the event of an unanticipated contamination event, and to guide management and protection of fractured-rock aquifer resources within the State.

Description of Study Area

West Virginia encompasses an area of 24,230 square miles (fig. 1; Cardwell and others, 1968). The region is largely present in two physiographic provinces, the Appalachian Plateaus Physiographic Province to the west and the Valley and Ridge Physiographic Province to the east (Fenneman and Johnson, 1946). Along the West Virginia and Virginia border in Jefferson County, West Virginia, there is also a thin outcropping of rocks within the Blue Ridge Physiographic Province. Elevation is highly variable across West Virginia. In the western part of the State and within the Appalachian Plateaus Province, there are flat-lying but extensively eroded rocks (Fenneman and Johnson, 1946). The eastern part of the Appalachian Plateaus Province in the central part of the state includes higher elevations above 3,000 feet (ft) above mean sea level. The Valley and Ridge Province in the eastern part of the State is characterized by long parallel ridges and valleys (Fenneman and Johnson, 1946) resulting from past orogenic (mountain building) events in the area, having moderate to high elevations along the ridges above 3,000 ft above mean sea level. The Blue Ridge Province in West Virginia only occupies a small strip of land separating West Virginia from Virginia (Fenneman and Johnson, 1946). The highest point in West Virginia is 4,863 ft above mean sea level on Spruce Knob in Pendleton County, and the lowest point is 240 ft above mean sea level along the Potomac River near Harpers Ferry in Jefferson County (fig. 1; U.S. Geological Survey, 2023).

Physiographic Provinces

West Virginia has three physiographic provinces (fig. 2)—the Appalachian Plateaus, Valley and Ridge, and Blue Ridge (figs. 1, 2). The Appalachian Plateaus Province occupies about two-thirds of West Virginia, and the Valley and Ridge Province comprises the additional one-third. The Blue Ridge Province is limited in exposure to a thin strip along the West Virginia and Virginia border in Jefferson County, West Virginia.

Topography

Topography is a major factor affecting the yield and depth of water-bearing fractures within West Virginia, especially in areas within the Appalachian Plateaus Province. The effects of topography on well yields vary greatly among hilltop, hillside, and valley settings within the Appalachian Plateaus Province in southern West Virginia (Kozar and others, 2020). Kozar and others (2020) determined that for 777 wells in southern West Virginia, median well yields were 2.9 gallons per minute (gal/min) for hilltop wells, 12 gal/min for hillside wells, and 30 gal/min for valley wells. This correlation between location and yield was first discussed by Ferguson (1967) and by the research of Wyrick and Borchers (1981). Ferguson (1967) and Wyrick and Borchers (1981) postulated that stress relief caused by erosional unloading of weight of eroded rock above valleys resulted in upward arching of bedrock in valley bottoms accompanied by vertical stress-relief fracturing on hillsides was responsible for the variance in well yield (fig. 3). The groundwater-flow system is extensively modified where active or abandoned mines are present (fig. 4). Groundwater in interconnected abandoned mine workings typically flows down-dip below overlying ridges, which leads to inter-basin transfer of groundwater from one surface watershed to another (fig. 4).

However, the effect of stress-relief fracturing processes on well yields common in the southern Appalachian Plateaus Province was unknown elsewhere in the State. The topography in northern West Virginia is not as steep as in the southern part. A primary focus of this study was to investigate the effects of topography on the occurrence and distributions of fractures in the State’s bedrock aquifers, especially for wells in the Valley and Ridge Province that previously lacked data for assessing such trends.

Geology

West Virginia geology is complex with sedimentary rocks of Mississippian to Permian age outcropping in the Appalachian Plateaus Physiographic Province, sedimentary rocks of Cambrian to Mississippian age outcropping within the Valley and Ridge Physiographic Province, and moderately metamorphosed sedimentary rocks and a few rocks of igneous origin of Neoproterozoic to Cambrian age cropping out in the Blue Ridge Province (figs. 2 and 5; Cardwell and others, 1968). This geologic complexity is related to past orogenic processes that resulted in the State’s major physiographic provinces. Generally, rocks within West Virginia are predominantly sedimentary in origin and consist of alternating sequences of sandstone, shale, siltstone, mudstone, limestone, dolomite, and coal. Various gradations exist among these primary rock types within the Appalachian Plateaus Province, but higher percentages of shale, siltstone, and limestone are in the Valley and Ridge Province. The Blue Ridge Province consists of various slightly metamorphosed sedimentary rocks such as schist, phyllite, and gneiss but contains a few igneous plutonic rocks, such as greenstone (Cardwell and others, 1968). Limestone and, to a lesser degree, dolomite bedrock form various carbonate units that are often karstic and crop out in various areas within the State. The more important limestone units are the Ordovician age units in the Shenandoah Valley and thinner units such as the Tonoloway and Helderberg limestones that crop out in thin bands of Silurian and Devonian age rocks within the Valley and Ridge Province. The various formations within the Greenbrier Group crop out mostly within the Greenbrier valley in Monroe, Greenbrier, and Pocahontas Counties, and in Canaan Valley in Tucker County, West Virginia.

Hydrogeologic terrain, a concept first applied in West Virginia by Schwietering (1981) and adopted as standard nomenclature by Puente (1985), is one of the most important factors for assessing whether an aquifer is intrinsically susceptible to contamination from near-surface processes. Kozar and Paybins (2016) defined hydrogeologic terrains as bedrock or alluvial aquifers within a physiographic province that share similar hydraulic properties, groundwater-flow processes, recharge rates, groundwater ages and residence times, lithology, and water quality. Hydrogeologic terrains are commonly used to differentiate aquifers within West Virginia. Kozar and Paybins (2016) also state that in addition to the three distinct physiographic provinces within the State, the Appalachian Plateaus, Valley and Ridge, and Blue Ridge (fig. 1), there are also three primary rock types and alluvial sand-and-gravel deposits in the State and compose the State’s principal aquifers. Kozar and Paybins (2016) defined the five major hydrogeologic terrains in West Virginia (fig. 6) as

1. (1) alluvial aquifers bordering the Ohio and Kanawha Rivers and other major tributary streams;

2. (2) karst aquifers in the Shenandoah River Valley of West Virginia’s eastern panhandle and in the Greenbrier River valley, and thin limestone outcrops in the Valley and Ridge Province;

3. (3) fractured crystalline metamorphic and igneous rock aquifers in a thin band within the Blue Ridge Province in Jefferson County, in the easternmost part of West Virginia;

4. (4) abandoned underground coal-mine aquifers in southern West Virginia (fig. 7); and

5. (5) fractured sandstone, shale, coal, thin limestone, and associated sedimentary-rock aquifers in most of the State, including parts of the Appalachian Plateaus and Valley and Ridge Provinces (fig. 6).

Previous Investigations

A sparse number of previous borehole investigations have been completed in West Virginia. A borehole geophysical assessment of lithologic and structural controls on groundwater flow processes was studied for the USGS facility at Leetown, West Virginia (Kozar and others, 2007). This study defined a complex flow system consisting of a highly permeable epikarstic zone underlain by fractured but less extensively solutioned limestone and dolomite. A sparser network drains this flow system of interconnected strike-parallel and cross-strike faults and fractures commonly enlarged by the dissolution of the limestone that acts as drains for the broader fracture-dominated parts of the aquifer. Borehole geophysics was used to document the overall orientation of dominant fractures within the study area and determined that rocks in the area had been subjected to folding and faulting, resulting in bedrock fractures with moderate to high-angle dip commonly exceeding 60 degrees (°). Dominant fracture orientation is parallel to bedrock strike at approximately N. 15° E. and perpendicular to strike at S. 70° E.

A second borehole geophysical assessment was undertaken on abandoned underground coal mine aquifers in the Elkhorn area of McDowell County, West Virginia (Kozar and others, 2012). This Elkhorn area borehole geophysical assessment documented large-scale dewatering of overburden strata and subsequent concentration of groundwater flow at depth within the Pocahontas Number (No.) 3 and, to a lesser extent, the Pocahontas No. 4 coal seams. The Pocahontas No. 3 coal seam has been largely mined out in the area and left to collapse as overburden strata and remaining coal pillars collapsed (gob) because of the weight of overburden rock above the mined coal seam. Overburden collapse has created a large network of interconnected underground abandoned mine workings that function similarly to deeply buried coarse alluvial aquifers. The dominant orientation of fractures mapped was N. 0–30° E. and N. 20–45° W. Most of the dip angles were relatively flat at 0–40°, but a few were higher angle fractures, 50–90°. The trends are generally aligned with the strike of bedding. Most of the fractures are strike-parallel joints and cleavage that dips generally 15–60°. Bedding-plane fractures noted in caliper logs had shallower dips and were less common than the dominant trends shown by joint and cleavage fractures. Hydrologically active high-angle joints and cleavage likely provide pathways for groundwater flow linking them to more horizontally oriented bedding-plane partings. Logs collected for the Elkhorn study suggest sandstones in the study area are poorly fractured and limit vertical flow between more permeable yet finer grained units and coal seams. Coal seams or abandoned underground coal mines are the major aquifers in the region.

Borehole geophysical logs are routinely collected by private entities to support the needs for subsurface oil and gas exploration and for assessing coal reserves for potential mining operations. These borehole geophysical surveys are largely proprietary and unavailable for public access. Kozar and others (2007) and Kozar and others (2012) are the only known publicly available published studies containing significant geophysical data collection and interpretation within West Virginia.

Studies of Silurian and Devonian fractured-carbonate units in West Virginia have shown that these systems may produce abundant water for public and domestic use (Rega, 2011; Donovan, 2012). Rega (2011) described lineament density in Silurian and Devonian carbonate aquifers in Mineral County, West Virginia, and concluded that lineaments were associated with groundwater discharges and groundwater discharges from carbonate aquifers were higher than from adjacent groundwater discharges from siliciclastic rocks such as shale, siltstone, and sandstone. Donovan (2012) found that drilling wells through overlying siliciclastic formations along the east side of an anticline into the underlying Silurian and Devonian fractured carbonate aquifers produced the greatest yield for public supply wells along Knobley Road in Mineral County, West Virginia.

Geophysical Methods

Multiple borehole geophysical logs were collected during this study. These geophysical logs are described in this section and summarized in table 1. Geophysical tools included a mechanical 3-arm caliper, an electromagnetic (EM) induction tool, a multi-parameter E-logging tool, acoustic (ATV) and optical (OTV) televiewer tools, an EM-flowmeter tool, and a full-wave sonic waveform tool.

Field Methods

Specific sequential procedures were followed for every well logged. Prior to initiating borehole geophysical logging, each well was visually inspected with a borehole camera equipped with a pan and tilt revolving camera head. The camera was lowered down the wellbore to document the depth and condition of cased- and open-hole intervals in the well, identify fractures and fracture zones penetrated by the wellbore, and assess the wellbore’s suitability for logging (wells with pumps, wiring, drop pipe, or other obstructions were not acceptable candidate wells for logging). Numerous wells were not logged after camera inspection for various reasons, mostly because of obstructions in the well or liners in the well that would prevent the use of the ATV or OTV tools for assessing fractures.

Upon completion of the camera inspection, the EM-induction tool and the 3-arm caliper were calibrated prior to collecting borehole log data. The remaining tools did not require on-site calibration and were maintained according to manufacturer’s specifications. The magnetometers on the ATV or OTV tools were checked periodically with a compass and protractor to ensure they were operating according to the manufacturer’s specifications, and the EM-flowmeter tool was calibrated annually at the beginning of the logging season to ensure the tool was collecting accurate flow readings. Appropriately sized rubber diverters were installed on the EM-flowmeter tool to seal tightly against the wellbore, thus forcing water to flow through the core of the flowmeter tool.

Site and date-specific information was also recorded for each well logged and included the following information:

Borehole geophysical investigations were completed for 126 wells (fig. 2; Kozar and others, 2025) using multiple borehole geophysical tools. Fractures were assessed in 115 of 126 wells (Kozar and others, 2025). Water-bearing fractures were identified in 109 of the 115 assessed wells; 6 wells contained no identified water-bearing fractures ( Kozar and others, 2025). Data collected for the borehole geophysical surveys are summarized in a USGS data release (Kozar and others, 2025) and include lithologic interpretations, the logged interval for each tool deployed, the USGS 15-digit station number, the USGS local well number, the latitude and longitude location of the well logged, the date the well was logged, the water level in the well at time of logging, the elevation of land surface at the well, the magnetic declination for the location on the date the wells were logged, and the geologic unit and type of rock in which the well was completed.

All logs were archived in the USGS Virginia-West Virginia Water Science Center borehole log computer archives. The original logs are also available on the USGS GeoLog Locator website (U.S. Geological Survey, 2025b). All well construction data and other miscellaneous site data are stored in the USGS NWIS database (U.S. Geological Survey, 2025c). A list of the wells logged is presented in the data release (Kozar and others, 2025).

Methods of Geophysical Log Analysis

Various methods were used to analyze borehole geophysical data collected for this study, including processing the borehole data using WellCAD (Advanced Logic Technology, 2023), a borehole geophysical analysis software package that allows for processing and analyzing of borehole geophysical data, and by graphical and statistical analysis of the data resulting from the processed data derived from the WellCAD analysis. For this study, WellCAD was used for several purposes, including fracture mapping analysis, lithologic identification, and computation of porosity within the wells logged.

Data from prior USGS studies were augmented by logging 100 additional wells between 2015 and 2019. Of the 126 wells available for analysis, 115 had sufficient logs to aid in the analysis of fracture distribution and the assessment of geologic, lithologic, topographic, and physiographic processes controlling fracture occurrence and distribution for fractured-rock aquifers.

Topography

Topography has long been known to be a primary factor governing the yield of wells in hilltop, hillside, and valley settings in the Appalachian Plateaus Province of West Virginia (Wyrick and Borchers, 1981; Kozar and others, 2020). Wyrick and Borchers (1981) found that isostatic rebound, resulting from erosion of overburden strata, causes an upward arching of beds in valley bottoms and corresponding high-angle stress-relief fractures on hillsides, resulting in higher well yields in valley settings, intermediate yields on hillsides, and lowest well yields in hilltop settings. Kozar and others (2020) examined 777 wells in southern West Virginia and found this general pattern of well yields to be true with respect to topographic setting. The highest median well yield for the 777 wells of 30 gal/min was in valley settings, the lowest median well yield of 2.9 gal/min for wells in hilltop settings, and an intermediate median well yield of 12 gal/min for wells on hillside settings (fig. 8).

The yield of wells in fractured-rock aquifers is dependent on the transmissivity of water-bearing fractures within the aquifer. Data collected for this study have provided considerable additional evidence to support earlier work by Ferguson (1967), Wyrick and Borchers (1981), and Kozar and others (2020). A plot of transmissivity derived for 84 wells that contained sufficient data from analysis of discrete-zone borehole fracture and flowmeter data for mapped water-bearing fractures using the USGS FLASH software (fig. 9) shows this same classic pattern of highest well transmissivity in valley settings, intermediate transmissivity in hillside settings, and lowest transmissivity in hilltop settings (fig. 9).

Transmissivity data from wells in the Appalachian Plateaus Province assessed with the USGS FLASH software (fig. 10) indicate similar topographic influence on transmissivity for wells with highest median well transmissivity (355 ft²/d) in valley settings, intermediate median transmissivity (45 ft²/d) in hillside settings, and lowest median transmissivity (6.7 ft²/d) in hilltop settings (fig. 10). Of water-bearing fractures for valley wells in the Appalachian Plateaus Province, 51 of 75 (68 percent) have transmissivity that equals or exceeds 100 ft²/d, but most hilltop wells (15 of 20 or 75 percent) and hillside wells (15 of 22 or 68 percent) have transmissivity less than 100 ft²/d.

Transmissivity data from wells in the Valley and Ridge Province show similar topographic controls as was observed in wells of the Appalachian Plateaus Province (fig. 11) with one major exception. For Valley and Ridge wells, 11 of 30 valley wells (37 percent) had water-bearing fractures with transmissivity in excess of 100 ft²/d, 14 of 19 hillside wells (74 percent) had fractures with transmissivity in excess of 100 ft²/d, and for hilltop wells only 1 of 15 water-bearing fractures (7 percent) had a transmissivity in excess of 100 ft²/d. In the Valley and Ridge Province (fig. 11), wells on hillsides had the highest median transmissivity (528 ft²/d), wells on hilltop settings had the lowest median transmissivity (5.7 ft²/d), and wells in valley settings had intermediate median transmissivity (16.9 ft²/d).

For this study, the data for wells on hillside settings were skewed toward high-yield wells, specifically within thin limestone units such as the Tonoloway and Helderberg Limestone Formations at depth, and may not be indicative of similar trends in transmissivity of fractures for other hillside wells completed within siliciclastic rocks within the Valley and Ridge Province.

Physiography

Physiography strongly affects fracture distribution. For this study, data for 2,253 discrete fractures identified in 77 wells with mapped fractures logged in the Appalachian Plateaus Province were graphically compared with 1,150 discrete fractures identified in 38 wells with mapped fractures logged in the Valley and Ridge Province. Tadpole plots were used to show the distribution of fractures for the wells logged for the study and show the dip of the fracture from 0° to 90° and the direction of the dip of the fracture indicated by the tail on the tadpole from 0° to 360°. Tadpole plots for the 77 wells assessed in the Appalachian Plateaus Province show several trends (fig. 12). A large concentration of fractures has low dip angles of less than 30°. This pattern of fracturing is not unexpected because bedrock within the Appalachian Plateaus is typically flat to slightly dipping, and a preponderance of low-angle fractures consistent with the dominance of bedding planes on control of fractures is expected. Mapped fracture dips ranged from 0° to 360°. Because the dataset is statewide and variability in regional stress fields caused fracturing, faulting, and folding of the rocks within the State, such a wide distribution in apparent direction of fractures is to be expected. The distribution of water-bearing fractures mapped shows a decreasing trend with depth. Fewer water-bearing fractures are below depths of 215 ft than at shallower depths.

The same fracture data plotted in the tadpole plots can be assessed by plotting the fracture data on a stereographic projection of poles to planes, commonly referred to as polar plots. In a polar plot, the fracture data are plotted as poles at a 90° angle to the fracture or bedding plane. To visually display a planar surface such as a bedding plane in bedrock with a northerly strike dipping due west at 45° (Lisle and Leyshon, 2004), a plot of poles to planes would place the pole at a 90° angle to the bedding plane, pointing downwards, and would plot in the southern hemisphere of the polar stereographic projection due east at 45°. A more rigorous explanation of stereographic projections of fracture data is provided in Lisle and Leyshon (2004).

A polar plot of fracture data for the Appalachian Plateaus Province (fig.13) shows wide variability in dip direction and angles. This was expected because of the major tectonic stress resulting in bedrock folding, faulting, and fracturing changes from the southern to the northern part of the State within the Appalachian Plateaus Province. However, the gently dipping bedding characteristic of the Appalachian Plateaus Province is evident in the plot by the large cluster of points near the center of the plot dipping from 0° to 40°. Overall, no other trends appeared in fracture orientation visible on the plot. Fracture patterns are more distinct within the abandoned underground coal mine aquifers and karst aquifers in the Greenbrier valley because fracture distribution is specific to the areas analyzed, whereas overall data distribution for the entire Appalachian Plateaus Province is highly variable as is consistent with the data shown on figures 12 and 13. Additional details about Greenbrier valley fractures will be discussed in the “Selected Areas of Specific Investigation” section of this report.

Tadpole plots for the 38 wells assessed in the Valley and Ridge Province also show several trends (fig. 14). There is a higher distribution of high-angle fractures in the Valley and Ridge Province than for fractures assessed in the Appalachian Plateaus Province. This is expected because the Valley and Ridge Province experienced significant folding and faulting during orogenic (mountain building) events. This folding and faulting resulted in long linear ridges and valleys in the Valley and Ridge Province rather than the slightly dipping strata characteristic of the Appalachian Plateaus. Additionally, the dip direction of strata in the Valley and Ridge Province is more consistent than the Appalachian Plateaus, and trend dominantly northwest to southeast, perpendicular to the general northeast-to-southwest strike of bedrock. Finally, the distribution of water-bearing fractures decreases with depth. Few water-bearing fractures are below a depth of 215 ft, and even fewer below a depth of 400 ft.

A polar plot of fracture data for the Valley and Ridge Province (fig.15) shows less variability in dip direction and angles and a strong clustering of points in the southwest and northwest quadrants of the plot. This is characteristic of the dip of bedding being primarily at right angles to the strike or major orientation of bedrock ridges within the Valley and Ridge Province, which trend southwest to northeast. This was expected because the major tectonic stress resulting in bedrock folding and faulting was from the southeast. Another trend in the fracture data visible on the plot is the larger distribution of fractures with steeper dip. Much of the bedrock in the Valley and Ridge Province has been subject to extreme stress caused by folding and faulting of the rock during orogenic processes. Overall, several apparent trends in fracture orientation are visible on the plot resulting from extensive folding and faulting due to past orogenic activity within the region. The large clustering of points in the northwest quadrant of the plot is indicative of the southeast dip of bedding prevalent within the study area.

Rose diagrams were used to assess dip angle (fig. 16) and dip direction (fig. 17) for fractures assessed within the Appalachian Plateaus Province. A rose diagram shows the predominant dip angle for fractures assessed within the Appalachian Plateaus Province are low-angle dips (fig. 16) with a large percentage of dip angles less than 15°. A rose diagram for dip azimuth (fig. 17) shows the wide distribution of dip azimuth for fractures assessed within the Appalachian Plateaus Province, owing to the low-angle dip of fractures within the province (mean dip angle of 2.2°) and variability in tectonic stress fields that varies across the State.

A rose diagram of the dip angle for fractures assessed within the Valley and Ridge Province revealed a large percentage of high-angle fractures, showing more than 50 percent of the fractures have a dip angle between 55° and 85° (fig. 18). Additionally, the rose diagram of dip directions for fractures in the Valley and Ridge Province shows a strong trend between 120° and 150°, perpendicular to the general strike of bedrock, which trends northeast to southwest (fig. 19). A smaller proportion of fractures trend 180° opposite to the general trend between 315° and 325°.

The dip angle and the azimuth of water-bearing fractures are heavily skewed with respect to the physiographic province, showing much steeper fracture dips in the Valley and Ridge Province and mean fracture dips of 42.0 and median fracture dips of 41.9 degrees, than in the Appalachian Plateaus Province, with mean fracture dips of 23.9 and median fracture dips of 17.5 degrees. Fracture depth is slightly higher in the Valley and Ridge Province (mean fracture depth of 180.5 ft and median fracture depth of 133.3 ft) than within the Appalachian Plateaus Province (mean fracture depth of 140.9 ft and median fracture depth of 99.1 ft). These trends are also evident in the data plotted in figures 13, 14, 15, and 16.

Transmissivity also varies with respect to topographic and physiographic controls (fig. 20). Boxplots of water-bearing fractures with quantifiable transmissivity show distinct differences in transmissivity between hilltop, hillside, and valley wells (fig. 20A). Grouping fractures by physiographic province reveals an obvious difference in the distribution of water-bearing fractures (fig. 20B). The plot for transmissive water-bearing fractures for wells logged in the Appalachian Plateaus Province (fig. 20B) looks similar to the plot for well yields for 777 wells in 13 counties within the Appalachian Plateaus Province in southern West Virginia (fig. 8; Kozar and others, 2020). Median transmissivity was 355 ft²/d for valley, 45.0 ft²/d for hillside, and 6.7 ft²/d for hilltop wells for wells logged in the Appalachian Plateaus Province. Median transmissivity was 16.9 ft²/d for valley, 528 ft²/d for hillside, and 5.7 ft²/d for hilltop wells for wells logged in the Valley and Ridge Province.

The distribution of transmissivity as a function of topography is different for water-bearing fractures in the Valley and Ridge Province than in the Appalachian Plateaus Province (fig. 20C). Within the Valley and Ridge Province, transmissivity for water-bearing fractures is highest in hillside wells, having a higher median transmissivity than either hilltop or valley wells. The reason for this distribution is that the dataset of wells logged for the Valley and Ridge Province contained several wells that were drilled on hillsides, specifically targeting thin limestone units such as the Tonoloway and Helderberg Limestone Formations at depth. Therefore, some bias may be in the dataset of water-bearing fractures for hillside wells within the Valley and Ridge Province. Hilltop wells in the Valley and Ridge Province show similar lowest median transmissivity for hilltop wells, similar to that for the Appalachian Plateaus Province wells logged (fig. 20B). Additional discussion of these thin Silurian and Devonian carbonate aquifers will be discussed in more detail in the “Selected Areas of Specific Investigation” section of this report.

Depth Below Land Surface

Statistical results showing the total number of low transmissive and water-bearing fractures are summarized in table 2. A total of 3,403 fractures were identified based on this study’s analyses. Only 252 (7.4 percent) were identified as water-bearing. For the Appalachian Plateaus Province, 2,253 total fractures were identified, 2,072 (92.0 percent) of those fractures are low-transmissive fractures, and only 181 (8.0 percent) are water-bearing fractures. For the Valley and Ridge Province, 1,150 total fractures were identified, 1,079 (93.8 percent) of those fractures being low-transmissive fractures and only 71 (6.2 percent) being water-bearing fractures.

A histogram of the distribution of mapped fractures for the Appalachian Plateaus (fig. 21A) and the Valley and Ridge Provinces (fig. 21B) clearly shows that most of the fractures are shallow features, typically less than 250 ft in depth. A similar histogram for only water-bearing fractures for the Appalachian Plateaus (fig. 22A) and the Valley and Ridge (fig. 22B) shows a similar pattern with most water-bearing fractures for Appalachian Plateaus aquifers and Valley and Ridge aquifers being less than 250 ft.

Results of comparisons between characteristics of water-bearing fractures at depth in logged wells in the Appalachian Plateaus and the Valley and Ridge Provinces were assessed by Wilcoxon signed-rank tests for statistical significance. The tests indicate a statistically significant difference (p-value=0.001) in the number of water-bearing fractures in the depth range from 0 to 215 ft in contrast to the number of fractures at depths greater than 215 ft. Histograms of the fracture data alone do not allow a full assessment of the depth at which water-bearing fracture distribution begins to decrease. Cumulative distributions of water-bearing fractures with depth of well were prepared to better assess at what depth fractures, especially water-bearing fractures, start to diminish significantly. The cumulative distribution plots indicate a distinct break in slope beginning at a depth of 215 ft for the Appalachian Plateaus (fig. 23A) and Valley and Ridge (fig. 23B) Provinces, below which the number of water-bearing fractures diminishes quickly. Most water-bearing fractures are found at depths shallower than 215 ft within the Appalachian Plateaus (81 percent; fig. 23A) and Valley and Ridge (72 percent; fig. 23B) Provinces. These statistically significant differences observed in each cumulative frequency distribution plot were confirmed by Wilcoxon signed-rank tests of statistical significance (JMP Statistical Discovery LLC, 2022). In the Appalachian Plateaus Province, differences in frequencies of water-bearing fractures, above and below the 215-ft threshold, were found to be statistically significant, having probability values (p-values) of less than 0.0001 (p<0.0001). Similarly, differences in frequencies of water-bearing fractures in the Valley and Ridge Province were also found to be statistically different above and below a depth of 215 ft, having probability values (p-values) of less than 0.0001 (p<0.0001). In both instances, the small p-values indicate significant differences in frequencies of water-bearing fractures above and below a depth of 215 ft.

Lithostratigraphy

The distribution of fractures mapped for the wells logged within the Appalachian Plateaus Province for this study shows the overall distribution, water-bearing, and low-transmissive fractures with their corresponding lithology (fig. 24), and detailed summaries of the data with respect to lithology are shown in table 2. Lithology was determined by interpretation of the borehole geophysics logs collected for the study as previously discussed in the “Methods” section of the report. Four lithologies dominate the overall occurrence of fractures within the Appalachian Plateaus Province, argillaceous or shaly sandstone, massive sandstone, shale, and siltstone or silty shale. The distribution of the number of fractures for each lithologic unit is summarized in table 2. Statistical summaries of the data by major rock type and the ratio of the number of water-bearing to low-transmissive fractures are summarized in table 3. The highest ratio of water-bearing to low-transmissive fractures are found within coal seams within the Appalachian Plateaus Province (table 3), which emphasizes the importance of these relatively thin but prolific aquifers—typically only present in areas of active or legacy underground coal mining in the southern part of West Virginia.

The distribution of fractures mapped for the wells logged within the Valley and Ridge Province (fig. 25) shows the overall distribution of low-transmissive and water-bearing fractures and the corresponding lithology for the fractures. Limestone is the dominant lithology for low-transmissive and water-bearing fractures within the Valley and Ridge Province, but argillaceous or shaly limestone and shale also play a prominent role with respect to producing low-transmissive and water-bearing fractures within the region. The Valley and Ridge Province contains prolific karst limestone aquifers of Ordovician age within the eastern panhandle (fig. 5) of West Virginia. Shale bedrock fractures easily and is commonly thought of as an aquitard, but the preponderance of fractures in shale units often results in them being prolific aquifers.

Patterns exist among fracture distribution and five West Virginia rock types—sandstone, shale, carbonates (limestone and dolomite), and metamorphic rocks (table 3). Regardless of rock type, the overall distribution of water-bearing fractures to the total number of all fractures assessed by rock type is low, ranging from 4.0 to 11.0 percent. Thus, only 1 in 10 to 1 in 20 fractures, depending on rock type and physiographic province, are water-bearing fractures, except coal seams.

Even though the total number of fractures mapped for wells logged within the Appalachian Plateaus Province for coal seams is low (only 1.0 percent of all fractures mapped), the ratio of water-bearing to the total number of fractures for coal seams is high (60.9 percent), which demonstrates the importance of these relatively thin but prolific aquifers.

Sandstone is the dominant rock type within the Appalachian Plateaus Province, comprising 49.5 percent of all fractures and 54.1 percent of water-bearing fractures. Shale or siltstone rocks hold 43.1 percent of all mapped fractures and 30.9 percent of water-bearing fractures. Finally, carbonate rocks comprise 6.4 percent of all mapped fractures and 7.2 percent of identified water-bearing fractures.

For logged wells in the Valley and Ridge Province, carbonate rocks are the dominant rock type, comprising 58.6 percent of all identified fractures and 53.5 percent of water-bearing fractures. Sandstone comprises 15.0 percent of all identified fractures but 26.8 percent of water-bearing fractures. Shale and siltstone comprise 19.7 percent of all identified fractures and 12.7 percent of water-bearing fractures. Finally, metamorphic rocks comprise 6.7 percent of all identified fractures and 7.0 percent of water-bearing fractures.

Prior to this study, sparse information was available to definitively link water-bearing fractures, or the distribution of highly transmissive fractures, to specific lithology across West Virginia. Work completed by the USGS in the Leetown (Kozar and others, 2007) area in Jefferson County found that certain carbonate-dominant geologic formations, especially those containing limestone with a high calcium carbonate content, could have wells with fractures exhibiting extremely high transmissivity. However, other limestone formations with lower carbonate content produced wells with much lower or little-to-no effective transmissivity. The same study showed that the Martinsburg Shale Formation, which is highly fractured in Jefferson County, produced relatively high median transmissivity when compared to the other geologic formations in the area but lacked the extremely high transmissivity that could be found in the more karstic limestone formations. Data collected for this study expands on the trends previously determined in earlier studies.

Stratigraphy can control the distribution of water-bearing fractures. Groundwater commonly flows from stress-relief fractures or bedding-plane partings (Wyrick and Borchers, 1981) where a permeable unit, such as a fractured sandstone or karstic limestone unit, sits atop a low-permeable bed, such as an unfractured shale or chert unit (fig. 26). Before this study, the prevalence or occurrence and distribution of such stratigraphic controlled water-bearing fractures within West Virginia was only based on anecdotal references. For this study, each water-bearing fracture was viewed with respect to its location in stratigraphic boundaries to determine the role of stratigraphy in controlling the occurrence and distribution of water-bearing fractures (Kozar and others, 2025). Of 252 identified water-bearing fractures with corresponding lithology, only 33 (13.1 percent) were located at an identifiable contact at or near two lithologies. Most fractures (205 of 252 or 81.3 percent) were within an individual lithologic unit. Coal mines were a special category having 10 fractures (4 percent) associated with abandoned coal mine workings and another 14 fractures (5.6 percent) attributed to mine subsidence fractures above mined coal seams. These results support the hypothesis that stratigraphy influences the overall distribution and occurrence of water-bearing fractures, but stratigraphic contacts may not be as important as previously thought.

The Appalachian Plateaus and Valley and Ridge Province have wide lithologic variability. Fourteen lithologic types were identified for the Appalachian Plateaus Province wells assessed (fig. 24), and 17 lithologic types were identified for the Valley and Ridge wells assessed (fig. 25). Other than schist, gneiss, and slate (which are all metamorphic rocks and are found solely in the Blue Ridge Province) and coal (which is of sedimentary origin in the Appalachian Plateaus Province) the type of rocks for which water-bearing fractures were identified were similar in the Appalachian Plateaus and Valley and Ridge Provinces (table 3). However, the distribution of carbonate rocks (argillaceous or shaly dolomite, argillaceous or shaly limestone, dolomite, and limestone) was dominant in the Valley and Ridge Province, having 674 of 1,150 (58.6 percent) fractures being in carbonate rocks whereas in the Appalachian Plateaus Province, sandstone (comprising 49.5 percent of all identified fractures and 54.1 percent of water-bearing fractures) and shale or siltstones (comprising 43.1 percent of all identified fractures and 30.9 percent of water-bearing fractures) were dominant. Basic statistical summaries of the water-bearing fracture data are presented in table 4 and can provide insight into the occurrence and distribution of water-bearing fractures in the complex fractured bedrock aquifers within the State of West Virginia.

Computed transmissivity ranged widely for the 189 identified water-bearing fractures assessed using the USGS FLASH software, from a minimum of 0.02 ft²/d to a maximum of 109,600 ft²/d (table 4). The mean and median computed transmissivities also varied greatly because of the non-normal distribution of the dataset with the mean transmissivity being 2,508 ft²/d and the median transmissivity being 63.0 ft²/d. The standard deviation of the dataset was large at 10,268 ft²/d (table 4). The median transmissivity of water-bearing fractures was highest for wells logged in the Appalachian Plateaus Province (102 ft²/d) and lowest in the Valley and Ridge Province (30 ft²/d). Mean well yields were reversed. Valley and Ridge wells had a higher mean transmissivity (4,307 ft²/d) than wells in the Appalachian Plateaus (1,565 ft²/d). Wells in the Valley and Ridge Province have a higher percentage of wells drilled in karst carbonate aquifers, which can often exhibit extremely high transmissivity (maximum of 109,600 ft²/d) compared to wells within the Appalachian Plateaus Province (maximum of 31,800 ft²/d) and is responsible for the vast difference between median and mean transmissivities for the two principal physiographic provinces within the study area. Two fracture trace and lineament analysis studies were completed in Jefferson and Berkeley Counties in West Virginia (McCoy and others 2005a; 2005b) and clearly show the substantial variability of transmissivity for wells in karst aquifers of the Shenandoah Valley.

Porosity of Fractured-Rock Aquifers

There are two types of porosity. The first is primary porosity, which is associated with the intergranular spaces between individual grains of clay, silt, sand, or gravel, and the second is secondary porosity, which is associated with fractures such as bedding planes, joints, cleavage planes, and faults that store and yield water to wells (Freeze and Cherry, 1979). In West Virginia, the only rock types that typically contain appreciable primary porosity are shales, siltstones, sandstones, and varying degrees of rock grading between these end members. While the primary porosity between the individual grains of clay or silt can be high in siltstones and shales, the rocks themselves typically do not yield water readily from these pore spaces, so the effective porosity is low. Sandstone and conglomerate rocks can have high porosity up to 30 percent, but many are so well cemented with silica cement as to have very low primary porosity (Freeze and Cherry, 1979). Most carbonate rocks such as limestone and dolomite are so well cemented with calcite or lithified that they also have little to no primary porosity. Shale and siltstone have a high primary porosity, but their pore spaces are so small that their effective porosity is low (Freeze and Cherry, 1979). This study concluded that fractured-rock aquifers in West Virginia have secondary porosity features such as bedding planes, joints, cleavage planes, faults, and other fractures in the rock, which provide most of the appreciable porosity for the siliciclastic and carbonate rocks comprising them.

Borehole geophysical analyses completed for this study provided the first comprehensive assessment of aquifer porosity for the fractured-rock aquifers in West Virginia. The primary borehole geophysics tool employed for these investigations was the full-wave sonic porosity tool. The tool emits a pulse of sound into the borehole, and return echoes are recorded at three separate receivers on the tool. The return times or echoes received by these three receivers can provide estimates of primary and secondary porosity.

Sonic porosity was evaluated in 74 boreholes in fractured-rock aquifers across West Virginia. Median and average values of porosity were assessed for formations greater than 5 ft in thickness and grouped into seven rock types (table 5). Sandstone rock type represents aquifers primarily composed of sandstone and sand-sized particles with less than 10 percent shale volume. The median and average total shale-corrected sonic porosity of sandstone aquifers was estimated to be 14 percent. The argillaceous or calcareous sandstone aquifers are composed of sandstone with fine-grained or calcareous proportions greater than 10 percent. Shale-corrected sonic porosity calculated for fine-grained lithologies had lower values ranging from 5 percent in argillaceous sandstones to 0 percent porosity calculated for siltstones and shales. This shows that sandstone aquifers in West Virginia likely have higher effective porosity than fine-grained or mineralized aquifers such as those with substantial argillaceous or calcareous proportions.

Total and shale-corrected values of porosity for carbonate aquifers, including limestone and dolomite, had median and average values of less than 4 percent and generally can be considered mineralized with low effective porosity. Porosity slightly increased to 4 to 5 percent for argillaceous or sandy limestone in contrast to clean limestone (2 to 4 percent), but these values may be affected by a high shale content. Shales and siltstones had a relatively high total sonic porosity when compared to other lithologies, but the compared lithologies often have bound water that is not effectively part of the groundwater flow system. Considering the shale-corrected values for shale and siltstone (0 to 2 percent), fine-grained lithologies should be considered minimally effective or not effective at yielding water beyond that produced as secondary porosity in bedding planes and associated fractures such as joints, faults, and cleavage planes.

Hydraulic Head

Hydraulic head was calculated based on results of EM-flowmeter logs, monitored pump discharge rates, and groundwater-level measurements made prior to and during the pumping of wells logged for EM-flowmeter analysis. Head estimates derived from FLASH analysis of EM-flowmeter logs, pumping rates, and water-level measurements made assessing whether water was moving upward or downward within a borehole possible. Flow, either upward or downward within the borehole, was assessed by topographic setting and physiographic province (table 6).

Directions of flow were assessed for 118 borehole intervals; downward flow was exhibited for 61 intervals, and upward flow was exhibited for 57 intervals overall. In simple hydrogeologic systems, areas receiving active recharge often have a downward flow, and groundwater discharge areas often have an upward flow. However, complex lithostratigraphic influences and geologic structure make this simple assessment less obvious. For example, valley settings are commonly thought to be groundwater discharge areas, so upward flow would be expected in valley settings. Conversely, hilltops are commonly thought to be groundwater recharge areas, so downward flow would be expected in hilltop areas. Hillside settings could exhibit either upward or downward flow, depending upon local geologic and lithostratigraphic controls.

Generally, the data in table 6 show a much larger number of intervals (52) exhibiting upward flow in valley or hillside settings than in hilltop settings (only 5 intervals with upward flow). When the data are parsed with respect to physiographic province, the relations between upward and downward flow and topographic setting become more complex.

For the Appalachian Plateaus Province, in valley settings, 22 of 61 borehole intervals (36 percent) exhibit upward flow, and 39 of 61 borehole intervals (64 percent) exhibit downward flow. In hillside settings, 11 of 12 borehole intervals (92 percent) exhibit upward flow, and 1 of 12 borehole intervals (8 percent) in hillside settings exhibit downward flow. In hilltop settings, 3 of 9 (33 percent) borehole intervals exhibit upward flow, and 6 of 9 (67 percent) borehole intervals exhibit downward flow. Thus, downward flow is dominant in hilltop and valley settings within the Appalachian Plateaus Province. The dominant number of borehole intervals with upward flow in hillside settings within the Appalachian Plateaus Province was unexpected but may simply indicate the dominance that perched aquifers have on the overall hydrogeologic setting within the Appalachian Plateaus Province, given that perched intervals are common in bedrock aquifers within the Appalachian Plateaus. Within the Appalachian Plateaus Province in West Virginia, the alternating sequences of sandstone, shale, siltstone, coal, and limestone often result in perched aquifers. Downward flow in valley settings may represent recharge to deeper subregional or regional aquifers.

For the Valley and Ridge Province, in valley settings, 7 of 15 borehole intervals (47 percent) exhibit upward flow, and 8 of 15 borehole intervals (53 percent) exhibit downward flow. In hillside settings, 12 of 14 borehole intervals (86 percent) exhibit upward flow, and 2 of 14 borehole intervals (14 percent) exhibit downward flow. In hilltop settings, 2 of 7 (29 percent) borehole intervals exhibit upward flow and 5 of 7 (71.4 percent) exhibit downward flow. Thus, downward flow is dominant in hilltop settings within the Valley and Ridge Province. Groundwater flow processes within the Valley and Ridge are governed by flow down dip along bedding that then discharges upward near valley floors or in hillsides immediately adjacent to valley bottoms, and the water often undergoes geothermal heating as it flows to depth within the bedding of permeable layers such as thin limestones of the Helderberg or Tonoloway Limestone Formations or within permeable sandstones such as the Oriskany sandstone (Hobba and others, 1979; Lessing and others, 1991). Thus, the high percentage of borehole intervals exhibiting upward flow in hillside settings within the Valley and Ridge Province may be characteristic of the deep circulation of groundwater along bedding.

Abandoned Underground Coal Mine Aquifers

Abandoned underground coal mine aquifers have been a prolific source of groundwater for public supply in southern West Virginia and elsewhere (fig. 7; Kozar and others, 2012). The deep interconnected passageways in the underground mines provide substantial storage for groundwater and are commonly referred to as “mine pool aquifers” or simply “mine pools.” However, this study has documented that mine pool aquifers often are not pools of water at all, but rather collapsed overburden strata and pillars that form a very high-porosity rubble zone commonly referred to as gob. Therefore, the term “mine pool” or “mine pool aquifers” is not accurate. “Abandoned underground coal mine aquifer” is a more representative term because it covers not only the recently abandoned mines where pools of water may actually exist, but also most older mines where overburden strata and pillars have collapsed, leaving behind a high-porosity, high-permeability zone of rubble (gob). Abandoned room and pillar mines, especially when the pillars are mined on the retreat phase of mining, only remain open for short and (or) undocumented periods. Longwall coal mining is designed to recover as much coal as possible, while the resulting overburden strata are designed to collapse as the shear (a rotating metal drum with cutting teeth) advances along the longwall panel, creating extensive zones of overburden collapse (Shultz, 1988).

Well logs from abandoned underground coal mine aquifers were assessed to better investigate their porosity and composition. A total of 56 wells were logged in geologic formations known to include coal seams (Kozar and others, 2025). However, not all the wells logged included coal seams, and of the 56 wells potentially containing coal seams, only 8 were documented as either containing a coal seam or were completed in abandoned mine workings. Of those 8 wells, 4 were completed in abandoned underground coal mines—wells Bar-0152, Mcd-0206, Mcd-0225, Mcd-0226, and Wyo-0290 (Kozar and others, 2025).

Close examination of the optical televiewer log for well Wyo-0290 (fig. 27) shows a distinct collapse zone of overburden strata in the interval between 274.9 and 276.4 ft below land surface that was once occupied by a coal seam that was completely mined out in this interval, leaving the rubble zone visible in the image. Collapse of the overburden strata resulted in subsidence fractures above the roof of the mine at 266.3, 267.4, and 267.8 ft below land surface.

A mined-out area similar to the area in well Wyo-0290 was observed in well Bar-0152 in Audra State Park (fig. 28) at a depth between 79.7 and 85.7 ft below land surface; however, in well Bar-0152, there was less aperture for the overlying subsidence fractures above the collapsed mine workings at depths of 67.0, 70.6, 75.8 ft below land surface, potentially due to several factors including: the age of the abandoned mine workings; type of mining activity (room and pillar versus longwall); or composition of the overburden strata. For this well, the overburden strata are comprised of a hard, competent, somewhat argillaceous or shaly sandstone, and the composition of the overburden strata may be responsible for the lack of subsidence fractures with visible aperture above the abandoned underground coal mine workings.

Well Mcd-0225 was completed to the abandoned mine workings of a very old (greater than 50 years since mining ceased) abandoned coal mine in the Elkhorn area in McDowell County in southern West Virginia. There is little evidence that the 6 ft thick Pocahontas No. 3 coal seam even existed in this interval because the coal pillars were evidently pulled and the entire overburden strata above the mine pancaked down into the abandoned underground coal mine workings (fig. 29). This type of collapse feature would be typical of wells completed in abandoned longwall coal mines where the overburden strata are designed to collapse upon retreat of the shear as it passes the face of the mine. For the Elkhorn area, the abandoned underground coal mine workings largely dewatered overlying intervals of bedrock, and the Pocahontas No. 3 coal seam became the interval controlling groundwater levels within the aquifer (Kozar and others, 2012).

An optical televiewer log for well Mcd-0226 (fig. 30) in the Elkhorn area shows a section of open mine entry, or void, in a section of the mine in which the overburden strata and coal pillars have not completely collapsed. The overburden strata for this well is composed of a massive sequence of slightly argillaceous sandstone, which was competent and resisted collapse. In this well, the thickness of the Pocahontas No. 3 coal seam for the mines in the Elkhorn area can be observed because in this well, the thickness of the coal seam was greater than 5 ft.

The above examples from borehole logs completed in abandoned underground coal mines should help better characterize the nature of abandoned underground coal mine aquifers. Extensive pools of water may reside in abandoned coal mine entries, as evidenced by the optical televiewer image for well Mcd-0226 (fig. 30), and mine entries can remain open for periods of time immediately upon cessation of mining. However, in many cases such open pools of water do not exist. Often, open pools of water are replaced by extensive areas of abandoned collapsed overburden strata and collapsed coal pillars forming a very porous rubble zone akin to a glacial outwash aquifer, as evidenced by the optical televiewer logs for wells Wyo-0290, Bar-0152, and Mcd-0225 (figs. 27, 28, and 29). Regardless, these extensive abandoned underground coal mine aquifers are often formed by large interconnected abandoned coal mine workings where several adjacent abandoned mines are hydraulically connected. These areally extensive interconnected abandoned mine workings result in prolific aquifers for water supply in areas where the coal seams contain a low pyrite or sulfur content, such as is common in the Pocahontas coal seams of southern West Virginia (Kozar and others, 2012).

Areally Extensive Carbonate Aquifers

In West Virginia, two principal areas comprise large, areally extensive carbonate aquifers. These include the Cambro-Ordovician age karst aquifers in the eastern panhandle in Jefferson and Berkeley Counties and the Mississippian age karst aquifers in the southern part of the State in Monroe, Greenbrier, and Pocahontas Counties (Jones, 1997). Both karst aquifers are prolific sources of water for public water systems and agricultural, commercial, and individual residential homeowner use (U.S. Geological Survey, 2025c).

Shenandoah Valley Karst Aquifer

The Cambro-Ordovician age karst aquifer system within the Shenandoah Valley in West Virginia crops out in the eastern panhandle counties of Jefferson and Berkeley. This karst aquifer system has outcrops of Ordovician age limestone and dolomite that crop out in thin bands elsewhere, such as in Monroe County along the thrust fault separating the Appalachian Plateaus and Valley and Ridge Provinces, and can form prolific aquifers (Kozar and others, 2023). The USGS has completed multiple investigations of karst aquifers in Jefferson and Berkeley Counties, but only one of those studies in the Leetown area of Jefferson County included collection of borehole geophysical data (Kozar and others, 2007) for assessing controls on fracture occurrence and distribution. Additional wells were logged during the current study to provide a more thorough understanding of groundwater-flow processes in this complex, fractured, karst-dominated aquifer system.

Twenty-four fractures were identified in sixteen wells, 13 in Jefferson County and 3 in Berkeley County (Kozar and others, 2025; table 7), and were available for assessment of the various factors controlling the occurrence and distribution of fractures in the Shenandoah Valley karst aquifer. Transmissivity was computed for 17 fractures from 11 of the 16 wells logged and ranged from a minimum of 0.80 to a maximum of 109,600 ft²/d (table 7). Mean transmissivity for the 17 discrete fractures for which transmissivity was computed was 13,853 ft²/d, but the median value was only 58 ft²/d, which is indicative of the extreme variability in fracture transmissivity for karst aquifers. Higher transmissivity values are commonly associated with large water-filled solution conduits or caves, and lower transmissivity values are characteristic of the more diffuse fracture-dominated parts of the aquifer (table 7).

Table 7.    

Table of water-bearing fractures, depth, azimuth, dip, lithology, and transmissivity identified as part of the statewide analysis of fracture data from 16 wells in the Shenandoah Valley karst aquifer in the Valley and Ridge Physiographic Province in the eastern panhandle of West Virginia.

[Aquifer data from Kozar and Paybins (2016). Physiographic province data from Fenneman and Johnson (1946). ft²/d, square feet per day; max., maximum; —, no transmissivity estimate for these fractures; NA, not applicable; min., minimum]

Table 7.    Table of water-bearing fractures, depth, azimuth, dip, lithology, and transmissivity identified as part of the statewide analysis of fracture data from 16 wells in the Shenandoah Valley karst aquifer in the Valley and Ridge Physiographic Province in the eastern panhandle of West Virginia.

Local well number	Fracture depth, in feet	Fracture azimuth, in degrees	Fracture dip, in degrees	Lithology and location of water-bearing fracture	Transmissivity of water-bearing fracture, in ft2/d
Ber-0445	128.6	115.8	83.1	in limestone solution conduit	—
Ber-0851	91.0	133.8	60.6	in argillaceous limestone	7.07
Ber-0851	372.8	328.1	30.1	in argillaceous limestone	6.79
Ber-0852	294.0	112.8	51.7	in argillaceous limestone	267
Ber-0852	312.2	324.9	78.7	in calcareous sandstone	384
Jef-0585	42.8	240.3	68.2	in argillaceous limestone	5.00
Jef-0586	149.4	128.8	81.9	in limestone solution conduit	2,000
Jef-0587	93.9	358.4	47.4	in limestone	—
Jef-0587	140.8	110.4	71.9	in limestone	—
Jef-0587	394.4	119.4	57.0	in limestone	57.5
Jef-0589	63.8	342.2	20.6	in limestone	30
Jef-0590	40.1	292.8	52.1	in limestone solution conduit	109,600
Jef-0590	70.0	359.4	14.1	in limestone solution conduit	30,900
Jef-0592	138.0	33.2	71.7	in limestone	2,000
Jef-0601	111.4	26.1	58.3	in limestone	30.0
Jef-0602	37.2	263.6	36.1	in limestone	—
Jef-0603	340.6	144.1	58.8	in argillaceous limestone	—
Jef-0752	145.1	75.4	21.4	at limestone argillaceous limestone contact	26,200
Jef-0753	88.7	241.4	7.6	in limestone solution conduit	64,000
Jef-0797	59.7	219.6	55.1	in limestone	—
Jef-0797	68.4	240.2	55.3	in limestone	—
Jef-0800	121.6	99.0	1.9	in argillaceous dolomite	0.80
Jef-0800	231.1	151.3	65.6	in dolomite	3.01
Jef-0800	642.7	121.4	16.0	in argillaceous dolomite	6.12
Max	642.7	359.4	83.1	NA	109,600
Min	37.2	26.1	2.0	NA	0.80
Mean	174.1	190.9	48.5	NA	13,853
Median	125.1	147.7	55.2	NA	58
Count	24	24	24	NA	17

Bedrock strike is the direction or trend that a structural surface, for example, a bedding or fault plane takes as it intersects the horizontal, and bedrock dip is usually 90° perpendicular to bedrock strike (Gary and others, 1972). The distribution of water-bearing and low-transmissive fractures for the Shenandoah Valley karst aquifer are shown in tadpole plots (fig. 31). The azimuth and dip angle, as depicted in the tadpoles, show the high variability in dip angle and azimuth of the fractures and a higher distribution of high dip angles for deeper fractures. The same data are shown in rose plots of dip angle (fig. 32) and dip azimuth (fig. 33). A high distribution of water-bearing fractures are shallow in the aquifer (fig. 31), which is consistent with data for the Valley and Ridge Province (fig. 23B), and the rose diagram of dip angle (fig. 32) shows a high density of high-angle dipping fractures for the aquifer and a strong correlation to overall dip of bedding perpendicular to the strike of bedrock within the region, as shown in a rose plot of dip azimuth (fig. 33).

Kozar and others (2007) provide a thorough description of the karst aquifer within the Shenandoah Valley of West Virginia based on borehole and surface geophysical research and installation of wells and piezometers collected for hydrogeologic investigations in or near Leetown, West Virginia. Kozar and others (2007) state numerous fractures intersect a borehole, but usually only one or two fractures are large water-bearing fractures. According to Kozar and others (2007), many smaller fractures provide storage potential for water but are not major avenues of groundwater flow. The less dominant fractures provide groundwater storage capacity and yield water to larger, more dominant water-bearing fractures, which in turn are drained by solution conduits within the carbonate rock commonly found along thrust and cross-strike faults or along bedding planes and drain water over a wide area. Kozar and others (2007) also conclude that the smaller diffuse-flow-dominated network of fractures that provides most of the storage occupies a larger proportion of the total aquifer volume, but solution conduits that occupy a much smaller proportion of the overall volume of aquifer are important because they constitute the primary drains for the groundwater flow system. Kozar and others (2007) also state that an active epikarstic layer within the karst ranges in depth from a minimum of 0 to a maximum of 60 ft below land surface and is highly permeable due to fluctuations at the water table over time. These intervals are typically cased off during drilling because of the large amount of residual mud and clay that is present within the epikarst.

From borehole geophysics studies completed for the Leetown study (Kozar and others, 2007), the dominant orientation of fractures aligns with bedrock strike with fracture dips ranging from 15° to 70° to the northwest and southeast. These trends are consistent with surficial mapping of strike parallel fractures in the Shenandoah Valley of West Virginia and with the data collected for this study and previously discussed, but numerous high-angle cross-strike joints and faults observed in outcrops are not well represented by ATV or OTV log analysis for either the Leetown study or for this investigation. The discordance is likely a result of sample bias because vertical boreholes are not effective for imaging high-angle fractures.

Greenbrier Valley Karst Aquifer

The second principal karst aquifer system in West Virginia is the Mississippian age karst and siliciclastic aquifers comprised of formations within the Greenbrier Group, Bluefield Formation, and Hinton Formation that crop out extensively in Greenbrier, Monroe, and Pocahontas Counties with lesser exposures in other counties within the State including Preston, Randolph, Tucker and Summers Counties. This report does not distinguish between the carbonate and siliciclastic units comprising the aquifer because even the Greenbrier Group itself contains non-carbonate siliciclastic units. These aquifers are also heavily relied on as a source of groundwater for public and residential supplies. An intensive county-wide hydrogeologic study in Monroe County focused on groundwater budgets, groundwater flow, borehole geophysics, and potentiometric-surface (water-table) mapping (Kozar and others, 2023). For the Monroe County study, wells were logged in the Mississippian age Greenbrier Group, Bluefield Formation, Hinton Formation, and Ordovician age karst rocks. Karst feature mapping and mapping of the Greenbrier Group at the formation level was an important part of the study. Additional wells were logged in Greenbrier County specifically for this study and the combined dataset was available for analysis.

Based on the logs for 15 wells in the Greenbrier valley aquifer logged as part of this and the recently completed Monroe County hydrogeologic assessment, 30 water-bearing fractures were identified (table 8). The wells logged were not all in carbonate rocks but also included wells logged in siliciclastic non-carbonate strata. As such, the logs are not specifically representative of the karst carbonate geologic formations within the Greenbrier Group but also contain a mixture of carbonate and non-carbonate siliciclastic rocks and included wells completed in the Bluefield and Hinton Formations. Transmissivity was calculated for 27 of the 30 fractures (table 8), and, as would be expected for a mixture of siliciclastic non-carbonate and karst carbonate strata, there was a high degree of variability in the computed transmissivities, ranging from a minimum of 0.02 ft²/d to a maximum of 18,500 ft²/d, and having a mean transmissivity of 1,831 ft²/d and median transmissivity of 159 ft²/d. These computed transmissivities are not as large as the ones presented for the Shenandoah Valley karst aquifer (table 7) and are probably not indicative of the potential range of expected transmissivities, which would be expected to be similar or even larger for the Greenbrier aquifer than for the Shenandoah Valley aquifer. Because the Greenbrier aquifer is significantly more conduit dominated than the Shenandoah Valley karst aquifer, maximum, median, and average transmissivities would be expected to be larger for the Greenbrier aquifer.

However, the dataset for the Greenbrier valley aquifer is skewed simply because of the smaller number of water-bearing fractures actually logged in karst carbonate strata (9 wells) compared to non-carbonate strata (18 wells; table 8). For the 5 fractures assessed only in limestone units, mean transmissivity was 8,950 ft²/d and median transmissivity was 9,520 ft²/d, significantly higher than the mean and median transmissivity for all Greenbrier valley aquifer wells of 1,831 and 159 ft²/d, as would be expected. This sparsity of data is a function of the limited amount of work conducted by the USGS in the Greenbrier valley aquifer compared to the much larger number of studies completed for the Shenandoah Valley karst aquifer. Other karstic limestone areas comprising the Greenbrier valley aquifer, including in Pocahontas and especially in Greenbrier County, have not been studied sufficiently to fully characterize the transmissivity of karstic areas of the Greenbrier valley aquifer. Additional well logging and hydrogeologic investigations would be needed to better understand transmissivity distribution in these areas.

Table 8.    

Table of water-bearing fractures, depth, azimuth, dip, lithology, and transmissivity identified as part of the statewide analysis of fracture data from 15 wells in the Greenbrier valley karst aquifer in southeastern West Virginia.

[Physiographic province data from Fenneman and Johnson (1946). Fracture depth, azimuth, and dip data are reported to one significant figure. ft²/d, foot squared per day; —, no transmissivity estimate for these fractures; NA, not applicable]

Table 8.    Table of water-bearing fractures, depth, azimuth, dip, lithology, and transmissivity identified as part of the statewide analysis of fracture data from 15 wells in the Greenbrier valley karst aquifer in southeastern West Virginia.

Local well number	Fracture depth, in feet	Fracture azimuth, in degrees	Fracture dip, in degrees	Lithology and location of water-bearing fracture	Physiographic province	Transmissivity of water-bearing fracture, in ft2/d
Grb-0317	40.8	261.6	19.6	in shale	Appalachian Plateaus	670
Grb-0317	75.3	220.0	62.8	in shale	Appalachian Plateaus	500
Grb-0317	108.0	350.9	21.0	in shale	Appalachian Plateaus	630
Grb-0318	38.1	165.9	22.5	in argillaceous sandstone	Appalachian Plateaus	670
Grb-0318	74.1	194.9	58.9	in argillaceous sandstone	Appalachian Plateaus	500
Grb-0318	109.8	322.1	36.7	in sandstone	Appalachian Plateaus	630
Mnr-0154	19.9	182.4	9.9	in argillaceous limestone	Appalachian Plateaus	9,520
Mnr-0154	104.0	187.8	37.2	in limestone	Appalachian Plateaus	18,500
Mnr-0163	218.3	138.7	6.7	in siltstone	Appalachian Plateaus	0.19
Mnr-0163	649.4	245.6	41.0	in siltstone	Appalachian Plateaus	0.02
Mnr-0164	77.9	253.9	20.2	at sandstone limestone contact	Appalachian Plateaus	159
Mnr-0165	26.8	44.8	12.0	in limestone	Appalachian Plateaus	1,630
Mnr-0166	706.8	246.0	12.8	in shale	Appalachian Plateaus	—
Mnr-0167	26.7	310.7	25.1	in sandy or silty shale	Appalachian Plateaus	575
Mnr-0184	126.1	346.5	50.3	in argillaceous limestone	Appalachian Plateaus	11,470
Mnr-0184	143.3	113.3	36.3	in limestone	Appalachian Plateaus	3,620
Mnr-0283	138.8	91.5	11.2	in shale	Appalachian Plateaus	5.73
Mnr-0283	340.2	212.2	4.5	in siltstone	Appalachian Plateaus	0.12
Mnr-0368	51.0	158.5	21.6	in argillaceous sandstone	Appalachian Plateaus	0.07
Mnr-0368	114.9	225.2	38.5	at argillaceous sandstone shale contact	Appalachian Plateaus	0.07
Mnr-0369	22.1	66.6	5.9	in sandy or silty shale	Appalachian Plateaus	9.0
Mnr-0369	34.6	210.1	11.4	in sandy or silty shale	Appalachian Plateaus	33.3
Mnr-0369	47.9	27.3	22.0	in sandy or silty shale	Appalachian Plateaus	258
Mnr-0432	150.6	222.4	6.6	in argillaceous sandstone	Appalachian Plateaus	4.0
Mnr-0432	189.3	179.1	13.9	in argillaceous sandstone	Appalachian Plateaus	6.0
Mnr-0432	369.5	322.1	89.5	in argillaceous sandstone	Appalachian Plateaus	2.9
Mnr-0434	114.5	119.0	7.6	in argillaceous sandstone	Appalachian Plateaus	49
Mnr-0434	536.3	86.8	14.7	in sandstone	Appalachian Plateaus	1
Mnr-0433	61.2	64.9	22.1	in sandy limestone	Valley and Ridge	—
Mnr-0433	62.3	111.1	38.4	in sandy limestone	Valley and Ridge	—
Max	706.8	350.9	89.5	NA	NA	18,500
Min	19.9	27.3	4.5	NA	NA	0.02
Mean	159.3	189.4	26.0	NA	NA	1,831
Median	106.0	191.3	21.3	NA	NA	159
Count	30	30	30	NA	NA	27

Tadpole plots of water-bearing and low-transmissive fractures mapped within the Greenbrier valley aquifer (fig. 34) show similar patterns to that for the Appalachian Plateaus Province (fig. 12) in which the Greenbrier valley is situated. Fracture depth decreases with depth similar to the Appalachian Plateaus Province and a higher proportion of fractures have shallow dip angles, as would be expected for bedding-plane dominated aquifers such as those in the Greenbrier valley aquifer. The water-bearing fractures (blue square tadpoles in fig. 34) are largely a shallow phenomenon and decrease in frequency rapidly below a depth of about 215 ft, similar to what was documented for the Appalachian Plateaus Province (fig. 23). Deep water-bearing fractures may be tapping the geologic contact between the Hillsdale Limestone and underlying Maccrady Shale Formations, which is a common practice for agricultural wells in the area.

Dip angles for fractures in the Greenbrier valley aquifer are mostly very gentle. A rose diagram (fig. 35) illustrates the percentage of dip for each 5° interval between 0° and 90°. Most dip angles are between 0° and 20°. Such slight dip angles represent largely bedding planes, but higher angle fractures represent joints, faults, or other fractures. The azimuth or direction of dip of fractures in the Greenbrier valley aquifer varies substantially, but the northwest and southeast azimuth angles are dominant as would be expected for bedding dipping perpendicular to the general northeast-to-southwest strike of bedrock in the Greenbrier valley aquifer (fig. 36). However, numerous fractures with dip azimuths are orthogonal to bedrock strike and those azimuths are representative of the higher angle fractures shown in figure 35.

Previous studies in Monroe and Greenbrier Counties discuss the distribution of fractures in detail, and the hydrogeologic report summarizing the groundwater hydrology and geology of Monroe County (Kozar and others, 2023) provides a thorough literature review of pertinent groundwater investigations in Monroe and Greenbrier Counties. One of the studies summarized in the Monroe County hydrogeology report (Hirko, 2012) stated that folds in the southern Appalachian Mountains trend approximately N. 60° E., whereas folds in the central Appalachians trend approximately N. 10°–30° E. The transition between the central and southern Appalachian Mountains is located where the Roanoke Recess, bounded by the Covington lineament to the north, converges with the Saint Clair thrust fault and Glen Lyn syncline. Dean and others (1988) state that this same feature in Monroe County marks the proximal extent of the Appalachian Mountain structural front. Hirko (2012) also states that linear features in central Monroe County correlate with sinkhole collapses, trending N. 10°–20° E., N. 30°–40° E., and N. 50°–80° E. and that these linear features parallel major fractures in limestone bedrock of the area as documented by Lessing (1979; 1981). Ogden (1976) states that bedding with a low plunge angle generally dips 15° or less but increases slightly along fold limbs, fold axes, and thrust fault margins. Ogden (1976) also states that plunged directions of major folds are to the southeast, and folds parallel to the Saint Clair thrust fault trend N. 55° E. Ogden (1976) also states that many of the caves in Monroe County form along the strike of bedrock, and the most form in the Denmar Formation and Union Limestone, which are the most calcite dominant carbonate units in the Greenbrier Group. These formations are often adjacent to shale confining units. Ogden (1976) also notes that although caves tend to form along strike, that trend is not statistically significant. Joints are also a controlling factor, especially those oriented N. 58° W. and N. 27° W. Stratigraphic dip averages about 15° but may be near vertical or overturned in areas. The fracture data collected from borehole geophysical surveys of wells logged within the Greenbrier valley aquifer for this study are in general agreement with the trends documented by previous studies in the region.