Purpose and Scope

The purpose of this report is to summarize findings of a reconnaissance study with the goal of locating and assessing potential sources of groundwater to augment the existing public-supply well for the City of Gary (L. Barber, Treasurer for the City of Gary, oral commun., 2023). Nine potential sources of groundwater were identified from abandoned underground mines (table 1) that may serve as a source of water to augment the existing public-supply well for the City of Gary, as the existing public-supply well has experienced increasing turbidity in recent years. This increased turbidity has resulted in additional water-treatment costs for the City of Gary (L. Barber, Treasurer for the City of Gary, oral commun., 2023) to ensure the public water meets EPA drinking water standards (U.S. Environmental Protection Agency, 2025). The alternate groundwater sources were evaluated for water-quantity yield and water-quality suitability with respect to the EPA drinking water standards. Water-quality and flow data collected for the study are available from the USGS NWIS database (U.S. Geological Survey, 2025a) and are also presented in appendix 2 of this report. An alternate source of water with better water quality could be blended with or replace the existing source to reduce treatment costs.

Description of Study Area

The City of Gary is in McDowell County in the southernmost part of the State of West Virginia. The Gary area is in the heart of the historically important Pocahontas coalfields, which provided low-sulfur coal for steel production to companies such as U.S. Steel Mining Corporation (U.S. Steel), which was the dominant employer in the City of Gary from the early 1900s until 1985, when U.S. Steel ceased all coal mining operations in its Gary district (Schust, 2005). Coal mined in U.S. Steel’s Gary mines provided coal for coking and subsequent steel production at their facilities in the Pittsburgh, Pennsylvania area and other states (Schust, 2005). Widespread underground mining of coal in the Pocahontas No. 3 and No. 4 coal seams has created a vast network of interconnected, abandoned mine shafts and tunnels, which has created a prolific groundwater aquifer that is the source of multiple public groundwater supplies in southern West Virginia (Kozar and others, 2020).

The groundwater flow processes within Pocahontas No. 3 and No. 4 coal mines were studied extensively in the Elkhorn area (Kozar and others, 2012), only 7.8 miles from the City of Gary, and are indicative of groundwater flow processes in the Gary area. Kozar and others (2012) found that groundwater flow within the abandoned underground coal mines happens within abandoned mine entries that are often filled with collapsed overburden strata and collapsed coal pillars, forming an aquifer comprised of extensive interconnected abandoned underground coal mine workings. Groundwater flows downdip in these abandoned mine workings and can flow beneath surface-water drainage divides, resulting in inter-basin transfer of groundwater from one surface-watershed to another.

Geology

The geology of the Gary area are sedimentary rocks composed of sandstone, shale, siltstone, mudstone, and prominent coal seams such as the Pocahontas No. 3 and No. 4. Bedrock in the Gary area dips slightly to the northwest (fig. 2) and is composed of interlayered sedimentary rocks of the late Mississippian Princeton Sandstone and Bluestone Formations and younger rocks of the early Pennsylvanian New River and Pocahontas Formations (fig. 3; West Virginia Geological and Economic Survey, 1968; 2025a).

The West Virginia Geological and Economic Survey hosts an online interactive geologic map that presents geologic information of the area (West Virginia Geological and Economic Survey, 2025a). Additionally, the West Virginia Geological and Economic Survey hosts a detailed interactive web service that presents the stratigraphy of the Princeton Sandstone, Bluestone, New River, and Pocahontas Formations and other coal-bearing strata within West Virginia in detail (West Virginia Geological and Economic Survey, 2025b).

The New River Formation is predominantly sandstone, having some shale, siltstone, and coal, but transitions to nearly all sandstone in the subsurface. The formation extends from the base of the Pocahontas No. 8 coal seam to the top of the upper Nuttall Sandstone Member and includes the Iaeger, Sewell, Welch, Raleigh, Beckley, Fire Creek, and Pocahontas Nos. 8 and 9 coal seams (West Virginia Geological and Economic Survey, 1968; 2025a).

The Pocahontas Formation consists of about 50-percent sandstone, and the rest is a mix of shale, siltstone, and coal. This formation extends from the base of the Pennsylvanian System (top of the Mississippian rocks of the Princeton Sandstone and Bluestone Formation) to the base of No. 8 Pocahontas coal seam or unconformity at the base of Pineville Sandstone Member of the New River Formation and includes from bottom upward the Pocahontas coal seams Nos. 1 through 7 (West Virginia Geological and Economic Survey, 1968; 2025a).

The Bluestone Formation comprises mostly red, green, and medium-gray shale and sandstone, and the Princeton Sandstone is underneath. There are no minable coal seams within the Bluestone Formation.

Coal Mining in the Gary Area

The first known written documented reference of the occurrence of coal in what is now southern West Virginia was made in 1750 by Dr. Thomas Walker, Chief Land Surveyor of the Loyal Company, as part of land surveys of 800,000 acres west of the Allegheny Mountains and north of the North Carolina border in Virginia (Schust, 2005). The first official geological survey of the region was conducted between 1836 and 1842 by Professor William B. Rogers, who was the State Geologist for Virginia at the time (Schust, 2005). The second geological survey covering what are present-day portions of McDowell, Mercer, Wyoming, and Raleigh Counties, W. Va., was overseen by Captain Isaiah Welch in 1873 and included the first geological survey of what would become the Pocahontas coalfields of southern West Virginia (Schust, 2005).

The Pocahontas coalfields of southern West Virginia and southwest Virginia started production when coal was first shipped from Pocahontas, Virginia (Schust, 2005). On December 31, 1901, United States Coal & Coke acquired a lease to 50,000 acres, and the company was incorporated on June 14, 1902 (Schust, 2005). The company would later merge with two other companies to become United States Steel Corporation (U.S. Steel) in 1950. The development of Gary as a coal mining hub began in 1902 when the first annual report for U.S. Steel Corporation, dated December 31, 1902, noted the construction of 3,000 beehive-shaped coking ovens (Schust, 2005) that converted the coal into high-carbon coke for use in steel making. U.S. Steel continued mining operations in the Gary area until 1985, when it closed all active mines or transferred the property to other companies. During the approximately 83-year span of operations, the U.S. Steel operations in the Gary area became the largest coal mining operation in the world, supplying high-grade, low-sulfur coal for U.S. Steel’s plants in the Pittsburgh area and elsewhere (Schust, 2005). The extensive and now abandoned underground coal mines in the Pocahontas No. 3 and No. 4 coal mines of U.S. Steel’s Gary works filled with water over time, creating a vast aquifer containing billions of gallons of high-quality groundwater for public supply in the area (Kozar and others, 2020).

Previous Investigations

Numerous hydrogeologic investigations have taken place in the Gary, W. Va., and the surrounding area. These previous hydrogeologic studies provide information on groundwater-flow processes and the quality of mine-derived groundwater sources in the region.

A study of a similar underground mine system was done for the nearby City of Welch, W. Va. (Ferrell, 1992). Ferrell (1992) found that rates of groundwater withdrawal from the City of Welch’s public supply wells, which are completed in the abandoned underground coal mines in the Pocahontas Nos. 3 and 4 coal seams (fig. 2) of the Exeter mine, ranged from about 1.6 to 1.8 million gallons per day (Mgal/d). The original pumping needed to keep the mines dry during mining was only 1.0 Mgal/d; thus, recharge to the mine had increased since the cessation of mining. Ferrell (1992) estimated that the abandoned mines could contain as much as 3.1 billion gallons of water (3.1×10⁹ gallons) stored in abandoned mines or associated fractures caused by the collapse of mine pillars or overburden strata. Ferrell (1992) noted that the quality of groundwater changed as groundwater levels rose in the abandoned coal mine complex. One notable change attributed to the rise in water level was the reduction in sulfate concentrations. Ferrell (1992) analyzed groundwater obtained from the City of Welch’s well number 3 on four separate occasions. Ferrell (1992) found that iron concentrations ranged from 330 to 1,660 micrograms per liter (µg/L) and had a median concentration of 365 µg/L; all four samples exceeded the 300 µg/L EPA secondary maximum contaminant levels (SMCL) for iron. Ferrell (1992) also found that manganese, another common contaminant in coal-mine waters, ranged from 10 to 110 µg/L and had a median concentration of 18 µg/L; only 1 of the 4 samples collected exceeded the 50 µg/L EPA SMCL. Ferrell (1992) analyzed sulfate in groundwater on three occasions and found concentrations of 122, 91, and 85 milligrams per liter (mg/L), far less than the EPA SMCL of 250 mg/L. Sodium concentrations measured by Ferrell (1992), however, were elevated, and all four sodium samples exceeded the 20 mg/L EPA health-based value (HBV) for individuals on a sodium-restricted diet. All four samples exceeded the EPA SMCL for total dissolved solids (TDS) of 500 mg/L, with maximum and minimum concentrations of 1,390 and 3,040 mg/L, with a median concentration of TDS of 2,480 mg/L.

The USGS completed numerous studies of surface and groundwater hydrology and water quality within coal-mining areas of the Appalachian region of the Eastern United States. The hydrology of Area 13, the eastern coal province of Kentucky, Virginia, and West Virginia, encompasses the drainage area of the Big Sandy River and its tributaries in eastern Kentucky, southwestern Virginia, and southern West Virginia. Kiesler and others (1983) investigated groundwater, surface-water, and coal resources of portions of the Tug Fork, Levisa Fork, Blaine Creek, and Big Sandy River watersheds in West Virginia, Virginia, and Kentucky; Kiesler and others (1983) is the primary publication for the USGS series of coal-hydrology reports for the Gary area and encompasses the Gary area and McDowell County, W. Va. Total dissolved solids concentrations ranged from less than (<)100 mg/L in streams draining undisturbed watersheds to 13,600 mg/L in mined watersheds (Kiesler and others, 1983). Most streams in the region had near-neutral pH, ranging from 7 to 8 standard units. They also found that streams draining areas of extensive mining had sulfate concentrations 18 times greater than streams draining unmined areas. In addition, Kiesler and others (1983) found that annual suspended-sediment yield in streams is related to the type of mining and to the percentage of the basin area disturbed by mining. For example, suspended-sediment yields, where surface-mining predominates within the basin, were found to be as much as 10 times greater than those where underground-mining methods were dominant and as much as 100 times greater than those of an unmined basin.

An assessment of hydrogeologic terrains, well-construction characteristics, groundwater hydraulics, and water-quality and microbial data for the determination of surface-water-influenced groundwater supplies in West Virginia was conducted to assess potential aquifer vulnerability and susceptibility to contamination from the surface (Kozar and Paybins, 2016). Kozar and Paybins (2016) identified abandoned underground coal-mine aquifers, such as the aquifer used by the City of Gary, to be only moderately susceptible and potentially less vulnerable to contamination than other West Virginia aquifers. However, Kozar and Paybins (2016) also found that abandoned underground coal mine aquifers that are derived from mine sources below the elevation of local tributary streams (below drainage) are more susceptible to contamination than abandoned underground coal mine aquifers with mine sources above the elevation of local tributary streams (above drainage) due to the potential for contaminated stream water to recharge underlying abandoned coal mines. The recharge area for the abandoned coal mine aquifer for the City of Gary is largely above local drainage elevation and therefore less susceptible to contamination from surface water than below drainage mine sources such as for the City of Welch, W. Va., which taps the same complex of abandoned underground coal mines but at a location where the coal seams dip below the elevation of tributary streams in the area.

Groundwater quality and geochemistry of West Virginia’s southern coal fields were recently investigated to assess the quality of groundwater upon which many residents of southern West Virginia rely as a source of water (Kozar and others, 2020). Kozar and others (2020) found that water-quality data from 60 different groundwater sites indicated that most waters sampled did not exceed thresholds for most EPA drinking-water standards and USGS health-based screening levels. However, they noted several exceptions, including samples where turbidity exceeded the 5-NTU EPA treatment technique (TT) drinking-water standard in 14 of 60 (23 percent) sites sampled and exceeded the 1-NTU TT standard in 51 of 60 (85 percent) sites sampled. Fourteen of the 60 (23 percent) sites sampled had detections of Escherichia coli (E. coli). Manganese and iron were prevalent contaminants in the groundwater samples collected for the study. Thirty of 60 (50 percent) sites analyzed for manganese and 25 of 60 (42 percent) sites analyzed for iron exceeded the proposed 50- and 300- μg/L SMCL drinking-water standards. Sodium concentrations exceeded the 20-mg/L EPA health-based value (HBV) for individuals on a sodium-restricted diet for blood pressure or other health reasons in 27 of 60 (45 percent) sites sampled.

A borehole geophysical assessment was performed on a well (USGS well Mcd-0226; USGS site 372259081334201; U.S. Geological Survey, 2025a), which is immediately adjacent to the Grapevine Well (Mcd-0109; fig. 2), completed in abandoned mines in the Pocahontas No. 3 coal seam in the Grapevine Hollow of Gary. The well is immediately adjacent (<10 ft) to the city’s current public supply well that provides the source water to the nearby water treatment plant. This well is now being used by a coal mining company, and the well and the City of Gary’s public supply tap the same supply source: a tunnel beneath the Tug Fork River connecting upgradient and downgradient coal mines in the Pocahontas No. 3 and No. 4 coal seams.

Field Inventories of Potential Supply Sources

Field inventories were done to locate potential water-supply sources to augment the City of Gary’s existing water supply, which is obtained from a well tapping abandoned underground coal mines in the Pocahontas No. 3 and No. 4 coal seams (fig. 2). Six potential water-supply sources were identified by field reconnaissance mapping of groundwater emanating from abandoned coal mine portals or emanating downstream from multiple such resurgences (sites 1–6, fig. 2). A seventh potential water-supply source, an abandoned air shaft tapping an abandoned underground coal mine about a quarter of a mile south of the City of Gary’s water plant, was also identified as a potential alternate source of water (site 7, fig. 2).

Measurement of Discharge

Discharge measurements were made for mine-outflow streams (sites 1–6, 8, and 9; fig. 2) by making wading measurements using a top-setting wading rod and a USGS-calibrated pygmy flow meter. Streamflow data were collected according to established USGS policies and procedures (Turnipseed and Sauer, 2010) to assess the amount of water that would be available during low-flow periods because that is a limiting factor in whether the site is viable as a potential water source to augment or replace the existing source for the City of Gary. Each of the eight stream channels was divided into segments, and the depth and velocity of each segment were measured; total discharge in each stream channel was a summation of the discharge in the various segments measured. Discharge for the eight sites where flow could be assessed (discharge could not be measured in the abandoned mine air shaft, site 7 on fig. 2) was measured on two separate occasions, from September 20–22, 2023, and again when sites were sampled for water quality, from October 16–October 18, 2023.

Water-Quality Sampling

Water samples were collected from the seven potential water-supply sources over the course of three days, from October 16–18, 2023. These samples were collected (U.S. Geological Survey, 2006) and processed (Wilde, 2002) according to USGS standard protocols. Bacteria were resampled at the abandoned mine air shaft (Mcd-0231) in the Sawmill Hollow area of Gary near Wilcoe, W. Va., on September 18, 2024, because the original bacteria sample collected on October 18, 2023, was over-incubated, and the results were deemed unusable. Water samples were packed in ice-chilled coolers and sent overnight to the USGS National Water Quality Laboratory for analysis of concentrations of common ions, trace metals, nutrients, volatile organic compounds (VOCs), semivolatile organic compounds (SVOCs), and then sent to a USGS contract laboratory for analysis of polychlorinated biphenyls (PCB) concentrations. A list of the constituents analyzed for this study and the constituent’s method reporting limit is shown in table 2 and appendix 1, table 1.1.

All samples were analyzed by the USGS National Water Quality Lab except for PCBs, which were analyzed by RTI Laboratories. Method detection limits and analytical methods for each constituent analyzed are available from the StarLIMS database and appendix 1, table 1.1. StarLIMS is the USGS laboratory information system database and can be queried to access specific information with respect to analytical methods and method detection limits. For query purposes, USGS schedules 4440, 2755, 2750, 2703, lab code 2707, and RTI Laboratories CIN schedule 50666 were analyzed for all samples collected for this study. This database is only accessible to USGS employees, so please contact the USGS Virginia and West Virginia Water Science Center should you require access to this resource.

Mine discharge sources were sampled by placing pre-cleaned sample bottles in the centroid of flow from the mine discharge source and by filling a previously cleansed and decontaminated churn in the centroid of flow for samples requiring filtration. Samples requiring filtration were processed inside a mobile field laboratory to prevent environmental contamination. Water temperature, dissolved oxygen, pH, and specific conductance were measured in the field by placing a calibrated multiparameter sonde in the centroid of flow and waiting for field parameters to stabilize prior to recording measurements. Turbidity was measured on site by collecting a sample of water from the centroid of flow and analyzing the sample on a Hach 2100P field portable turbidimeter.

Water in the abandoned coal mine air shaft (site 7, fig. 2) was sampled using a submersible pump with a Teflon line that was lowered into the air shaft to pump water from the air shaft. Field parameters were measured in a flow-through apparatus connected to the pump discharge line with an in-line multiparameter sonde to prevent contamination by aeration of the sample. Sample collection and processing for bacterial analysis were performed by lowering a pre-cleaned sterile poly-vinyl chloride (PVC) bailer into the air shaft and submerging the bailer several ft below the surface of the water within the air shaft. The bailer was then brought to the surface, and an aliquot of the sample was transferred to a pre-sterilized 100 mL sample bottle.

Bacteria samples were also collected on site, processed within 2 hours of collection, and incubated overnight, according to USGS standard protocols designed for assessment using the Colilert system. Collected water samples were also analyzed for bacteria following standard USGS methods (Myers and others, 2014) and processed using the Colilert system (IDEXX, 2019) according to established methods (American Public Health Association, American Water Works Association, and Water Environment Foundation, 2017). The Colilert system uses a defined substrate liquid-broth medium method for the determination of total coliform bacteria and E. coli. Water samples were collected by filling a pre-sterilized 100-milliliter (mL) sample bottle by immersion in the center of the flow from the potential water-supply source. Bacteria media was then added to the sample bottle and gently swirled to dissolve the media into the sample. Samples for bacterial analysis were then transferred into an IDEXX tray, sealed, and incubated at 35 °C for a period varying from 22 to 26 hours. Bacterial concentrations for E. coli and total coliform bacteria were then determined by examination of the IDEXX trays. Sample cells that were yellow indicated the presence of total coliform bacteria, and cells that glowed blue when examined with an ultraviolet light source were identified as positive for E. Coli bacteria.

Hydrogen sulfide gas odor was detected when site Mcd-0232 (table 1), discharge from two abandoned public supply wells at Havaco, W. Va., (fig. 2) was originally inventoried. Because site Mcd-0232 was the only site of the seven sites sampled for water-quality constituents that had a detectable odor of hydrogen sulfide gas, the site was sampled for hydrogen sulfide using a Hach hydrogen sulfide test kit in addition to being sampled for the full suite of chemical constituents listed in appendix 1. The sample to be analyzed for hydrogen sulfide was collected immediately as it discharged from one of the two abandoned free-flowing artesian public supply wells at that location. This sample was not collected at the same time as the original samples because of sampling errors, but was resampled on October 4, 2024, at flow conditions similar to those when previously sampled on October 16, 2023. Hydrogen sulfide was only measured at Mcd-0232 because the remaining six sites were open to the atmosphere and likely would not have been expected to contain hydrogen sulfide gas because of rapid degassing into the atmosphere. The only site with a noticeable odor of hydrogen sulfide gas was for site Mcd-0232, the discharge from the two abandoned public supply wells at Havaco, W. Va.

Quality Assurance and Quality Control

Two types of quality-assurance and quality-control (QA/QC) samples were collected during the study: a replicate and a field blank. A replicate sample was collected in conjunction with an environmental sample to assess the reproducibility of analytical methods and assess bias that may result from laboratory analysis. The field blank was run to assess any contamination that may result from field sampling methods, to evaluate decontamination procedures between sites, and to detect contamination on sampling equipment during transit to and from the sampling site.

Variability for a replicate sample pair was quantified by calculating the relative percent difference (RPD) of the samples. The RPD was calculated using the following formula (eq.1):

Borehole Geophysical Survey

A borehole geophysical survey was completed on a 6-inch diameter unused well (Mcd-0226; fig. 4) that taps the exact same source of water as the Grapevine Hollow well (Mcd-0109) that the City of Gary currently uses as the sole source of water for the city’s water plant. The well (Mcd-0226) is approximately 10 ft from the city’s production well (Mcd-0109) and was logged in July 2018 as part of a previous investigation. Both the city’s production well and the unused smaller diameter well tap a tunnel that connects upgradient abandoned mines with similar downgradient abandoned mines in the Pocahontas No. 3 coal seam on the southeast side of the Tug Fork (river; fig. 2). The tunnel connects the abandoned mines and crosses beneath the Tug Fork (river).

Various borehole-geophysical tools and methods were used during the borehole geophysical survey (table 3). Century Geophysical tools were used to collect most of the borehole logs for this study. Century Geophysical tools included a mechanical 3-arm caliper, an electromagnetic (EM) induction tool, a multi-parameter E-logging tool, an acoustic televiewer (ATV) tool, an EM flow-meter tool, and a full wave sonic waveform tool. The only exception to the use of Century Geophysical tools was an optical televiewer (OTV) obtained from Mount Sopris Instruments. All borehole geophysical tools were lowered into selected wells using a Century Geophysical 700M draw works winch and cable system. Data from the Century Geophysical logging tools were collected and processed into a computer via a Century Geophysical System VI borehole processing module and the associated Century Geophysical Log VI version 3.60Q software (Century Geophysical, LLC, 2025). The Mount Sopris optical televiewer utilized the same Century 700 M draw works, but data were collected using a Mount Sopris Matrix Geophysical Logging System with associated ALT Logger Suite software (Advanced Logics Technology, 2025).

Statistical and Graphical Analysis

Statistical analyses were computed using SAS-JMP (JMP Statistical Discovery LLC, 2022), a statistical software package. Statistical analyses include preparing data tables, summary tables, and summary statistics that include evaluations of mean, median, maximum, minimum, quantile, and standard deviations of available data.

The nature of data collected for this study cannot be effectively analyzed with bivariate statistical methods, such as linear regression, or multivariate statistical analysis, such as principal components analysis, due to the small sample size (n=7). Graphical analyses including boxplots and trilinear diagrams (Piper plots) were therefore the primary statistical analyses utilized to summarize the water-quality data collected for the seven sites sampled for the project and were created using SAS-JMP (JMP Statistical Discovery LLC, 2022).

Borehole Geophysics

Results of the borehole geophysical survey on well Mcd-0226 (immediately adjacent to the city’s Grapevine Well [Mcd-0109]), fig. 2) indicate the well is cased to a depth of approximately 111 ft below land surface and has an open interval of approximately 5 ft within the tunnel at the bottom of the well. Also, four low-angle fractures are present in the well at depths of approximately 111, 117, 123, and 124 ft. The lithology of the overburden strata is argillaceous sandstone. The open void in the tunnel is clearly visible in the optical televiewer image. Since the well was logged on July 9, 2018, a pump has been installed in the well, and as of January 7, 2024, two wells are withdrawing water from the tunnel: the City of Gary’s production well and the well currently being pumped as a source of water for a nearby mining operation. Whether the two wells pumping near one another are affecting turbidity levels in the City of Gary’s water supply is undetermined at the time of this writing.

Measured Discharge

Discharge for mine-discharge sites 1–6 (fig. 2) was measured on two separate occasions during the period of initial site reconnaissance between September 20 and 22 and then again during the period when the sites were sampled for water quality between October 16 and 18. Sites 8 and 9 were measured only once during initial site reconnaissance. All discharge data are summarized in table 1, and sites measured are shown in figure 2. Measured discharges ranged between 0.082 for site 5, the tributary below the mine discharge sites at Havaco, W. Va., to a maximum value of 3.685 ft³/s for site 6, Mcd-0232—the discharge from two abandoned public supply water wells near Havaco, W. Va. Of the six sites measured, only two, site 4, Harmon Branch at Thorpe, W. Va., and site 6, Mcd-0232 the discharge from two abandoned public supply water wells near Havaco, W. Va., had flow in excess of 1.0 ft³/s. Discharge in the abandoned underground mine air shaft (site 7, fig. 2) could not be assessed, but the air shaft drains an abandoned mine that likely contains water stored in approximately 1.7 square miles (mi²) of abandoned underground coal mines in the Pocahontas No. 3 coal seam, and possibly an additional 0.9 mi² of leakage from the overlying Pocahontas No. 4 coal seam.

Flow-duration statistics were calculated for the Tug Fork downstream of Elkhorn Creek at Welch, W. Va., streamgage (USGS site 03212750; U.S. Geological Survey, 2025a) using the percentiles function in Microsoft Excel and daily mean discharges for the period of record dating from January 29, 1985, through October 18, 2023, and were used to assess the flow duration respective of when the sites were measured. Discharges for the six study sites were measured between September 20, 2023, and October 18, 2023, and corresponded to the 12th to the 15th percentiles of flow duration for the Tug Fork downstream of Elkhorn Creek at Welch, W. Va., streamgage (USGS site 03212750) for these periods. Thus, the discharge measurements are respective of low to very-low streamflow and corresponding groundwater storage from which streamflow is based during baseflow conditions similar to during the study.

Quality Assurance Results

Analyte concentrations of replicate sample pairs typically differed by small amounts for the 35 inorganic contaminants analyzed (common ions, trace metals, nutrients, and pH). Only 4 of the 35 constituents analyzed had an RPD exceeding 2.8 percent. The few constituents that exceeded the 2.8-percent RPD threshold were for concentrations just above the method detection limit (refer to table 2). There was some variability between replicate pairs for manganese (RPD of 9.4 percent), nickel (RPD of 8.2 percent), fluoride (RPD of 8.8 percent), and pH (RPD of 6.4 percent). The RPD for pH was anticipated because pH of a sample will vary between the time the sample was collected and the time the sample was processed in the laboratory. The RPD for nickel and fluoride (8.2 and 8.8 percent, respectively) was likely due to the extremely low concentrations of nickel and fluoride in the environmental and replicate samples. The environmental and replicate concentrations for fluoride were 0.184 and 0.201 mg/L, respectively; the environmental and replicate concentrations for nickel were 0.881 and 0.812 µg/L, respectively. Therefore, QA/QC analyses did not indicate any bias in laboratory sampling methods for any constituent other than manganese. Even for manganese, the RPD of 9.4 percent was acceptable. The environmental and replicate concentrations for manganese were 271 and 297 mg/L, respectively. Relative percent differences were not calculated for any of the PCBs, VOCs, or SVOCs analyzed because constituent concentrations were all less than the method detection limit. Overall, the replicate and blank data did not indicate any significant bias with either field methods or laboratory analytical methods.

Field-blank values for one sample were less than the method detection limit for all 187 constituents analyzed, and there were no detections of any common ions, trace metals, PCBs, VOCs, or SVOCs (Data available on request from the USGS Virginia and West Virginia Water Science Center). These results indicate that field collection and processing procedures for samples and decontaminating equipment between sites were adequate to prevent cross-contamination of the environmental samples collected for the study.

Groundwater Quality in Relation to Drinking Water Standards

Although the mine discharge sources sampled for this study are unregulated, water-quality data collected for this study are related to EPA drinking-water standards and USGS health-based screening levels (HBSLs; tables 4 and 5; U.S. Geological Survey, 2025b) to document the quality of groundwater that may serve as an alternate supply source for the City of Gary. These criteria provide the city, county, and State water-resources managers with valuable information on constituents that may be problematic for use as a supply source because of elevated concentrations of certain constituents that may require additional treatment in addition to the treatment costs already incurred by the City of Gary. A central premise of this study was the goal of locating an alternate supply of water that could augment the City of Gary’s current water supply source and, by blending water from the two supply sources, could reduce treatment costs incurred by the treatment facility.

Of the 203 constituents analyzed, only a few exceeded applicable EPA drinking water standards (table 4; U.S. Environmental Protection Agency, 2025). Table 6 summarizes the maximum, minimum, mean, and median concentrations and the standard deviation for all the significant constituents analyzed in the water samples collected for the seven mine sources. Iron exceeded the 300 µg/L secondary maximum contaminant level (SMCL) at only one of the seven sites (14.3 percent) sampled. Iron concentrations ranged from a minimum of <5 µg/L to a maximum of 724 µg/L with a median concentration of 7.62 µg/L. Manganese exceeded the 50 µg/L SMCL at two of the seven sites (28.6 percent) sampled. Manganese concentrations ranged from a minimum of 1.93 µg/L to a maximum of 271 µg/L with a median concentration of 4.03 µg/L. For arsenic, none of the sites sampled (0 percent) exceeded the maximum contaminant level (MCL) of 10 µg/L. Arsenic concentrations ranged from a minimum of <0.100 µg/L to a maximum of 2.35 µg/L with a median arsenic concentration of 0.20 µg/L. None of the seven sites sampled (0 percent) for selenium exceeded the EPA MCL of 50 µg/L. Selenium concentrations ranged from a minimum of <0.05 µg/L to a maximum of 5.26 µg/L with a median concentration of 3.21 µg/L. All seven sites were sampled for volatile organic compounds (VOCs), SVOCs, and polychlorinated biphenyls (PCBs), but most were not detected. Of the 10 PCB compounds analyzed for the seven sites sampled, none (0 percent) contained detectable concentrations of PCBs or Aroclor compounds. Of the 44 SVOCs analyzed from the seven sites sampled, only 1 SVOC, acenaphthene, was detected at a concentration of 0.020 µg/L. Of the 96 volatile organic compounds analyzed from the seven sites sampled, only two were found at detectable concentrations. Trichloromethane was detected at a concentration of 0.027 µg/L but did not exceed the 0.08 µg/L MCL for total trihalomethanes. Benzene was detected at four of the seven sites (57.1 percent) sampled, at concentrations of 0.021, 0.028, 0.029, and 0.035 µg/L, but none exceeded the EPA MCL for benzene of 5 µg/L.

Total coliform bacteria are ubiquitous in the environment, and their presence only suggests the potential for contamination by near-surface processes. Escherichia coli (E. Coli) bacteria, however, are derived from either human or animal fecal material and can be an indicator of potential contamination by pathogenic bacteria or viruses. Total coliform bacteria were detected at all 7 (100 percent) sites sampled with concentrations ranging from 17.5 to greater than (>) 2,420 most probable number per 100 mL of sample (MPN/100 mL), and a median total coliform concentration of 1,553 MPN/100 mL. Escherichia coli (E. Coli) bacteria were detected at 4 of the 7 (57.1 percent) sites sampled at concentrations ranging from <1 to 11.9 MPN/100 mL of sample, and a median E. coli concentration of 5.1 MPN/100 mL.

Geochemical Controls on Groundwater Quality

Overall, the water from five of the seven coal mine sources was of a calcium sulfate type (fig. 5), and the water from the two remaining sites was of a sodium and potassium sulfate type. Sulfate was by far the dominant anion for all seven sites sampled; calcium, magnesium, and sodium were dominant cations for the seven sites sampled. For one site, site Mcd-0232, the discharge from two older abandoned public supply wells at Havaco, W. Va., the dominant anion in the sample was sodium, not calcium and magnesium as was typical for the remaining six sites sampled (fig. 5). Boxplots showing the distribution of overall constituent concentrations for the more common ions are shown in figure 6. The boxplots also show that nitrate plus nitrite is found in very low concentrations within the study area, which is typical for above drainage underground coal mine aquifers of the southern West Virginia coalfields, corresponding to a lack of agriculture or other sources of nitrate within the region (Kozar and others, 2012; 2020).

Abandoned underground coal-mine aquifers in the coalfields of southern West Virginia typically have low concentrations of iron, and this is shown in a boxplot of trace metal concentrations for the seven sites sampled (fig. 7). The EPA SMCL for iron is 300 µg/L, and only site Mcd-0232, the discharge from two public supply wells at Havaco, W. Va., had a concentration (724 µg/L) exceeding the SMCL. Site Mcd-0232 emanates from within the abandoned underground mines in the Pocahontas No. 3 and No. 4 coal seams. The water emanating from this source is typified by low-oxygen-reducing conditions. Thus, high concentrations of iron are to be expected because the water emanating from the source lacks sufficient oxygen to precipitate metals, resulting in elevated concentrations of iron and manganese. Water samples from this site were collected downstream from the discharge site, so the dissolved oxygen value for the sample locations is not indicative of the low-oxygen environment present within the abandoned underground coal mines from which the source emanates. Aeration of the water discharging from the wells causes iron to precipitate in the water discharging to the Tug Fork. The aerated water exceeds the 300 µg/L SMCL and causes iron staining of the stream bed and the well pad adjacent to the Norfolk Southern railroad tracks, which are immediately adjacent to the wells.

The same redox processes controlling iron concentrations in the water from the seven sites sampled also control concentrations of manganese (fig. 7). Two of the seven sites sampled contained concentrations of manganese that exceeded the 50 µg/L SMCL. As was the case for iron, site Mcd-0232, discharge from the two public supply wells at Havaco, also contained the highest concentration of manganese of the seven sites sampled (271 µg/L). The tributary below multiple mine discharges at Havaco, W. Va., also contained a manganese concentration (185 µg/L) that exceeded the 50 µg/L EPA secondary maximum contaminant level (SMCL) drinking water standard (U.S. Environmental Protection Agency, 2025). Both site Mcd-0232, the discharge from two public supply wells at Havaco, W. Va., and the tributary below multiple mine discharges at Havaco, W. Va., are in areas where the Pocahontas No. 3 and No. 4 coal seams are at or below the base level of tributary streams in the area, thus representing below drainage mine sites, whereas the other five sites sampled are representative of above drainage mine discharge sources. The difference between above and below drainage underground coal mine sources is a controlling factor with regard to dissolved oxygen and associated redox processes for the abandoned underground coal mine aquifers in the Gary area, McDowell County, W. Va.

Another controlling factor is carbonate buffering capacity. Bedrock in the Pocahontas and New River Formations tends to have high carbonate content, which may buffer the water to some extent (Kozar and others, 2012, 2020). Median alkalinity for the mine sources sampled was 194 mg/L. Even though iron and sulfate concentrations can be elevated (mean concentrations of 123 µg/L and 197 mg/L, respectively), mean pH for the seven mine sources sampled was 7.32 standard units, indicating a water that is not all acidic but somewhat alkaline (fig. 8). As a result, acid mine drainage in the coalfields of southern West Virginia is not common, and water emanating from the Pocahontas No. 3 and No. 4 coal seams is commonly used as a source of water for public and residential supply in southern West Virginia.

Total dissolved solids concentrations for six of the seven mine sources sampled were moderate, median TDS concentration of 515 mg/L, and site (Mcd-0232), the site emanating from the two abandoned public supply wells at Havaco, had a moderate TDS concentration of 494 mg/L. Site Mcd-0232 differs from the other six sites sampled in one important aspect: geochemistry. Site Mcd-0232 emanates from deep abandoned underground coal mines where oxygen content is typically low, thus resulting in reducing conditions that can liberate minerals from source rock or coal seams, resulting in high concentrations of dissolved constituents (Kozar and others, 2012; 2020). Concentrations of iron, manganese, arsenic, silica, fluoride, chloride, and sodium were all much higher for site Mcd-0232 than the other six mine sources sampled. Sulfate concentrations were correspondingly the lowest at site Mcd-0232 of the seven mine sources sampled, also likely due to reducing conditions in the abandoned underground coal mine from which site Mcd-0232 is derived.

Nutrients such as nitrate, nitrite, total ammonia, total nitrogen, and orthophosphate were low in all seven mine sources sampled, generally less than the laboratory method detection limit (app. 1, table 1.1). The median nitrate + nitrite concentration was 0.080 mg/L because there is no agricultural activity in the study area, and the seven coal-mine sources were not in proximity or in connection with potentially sewage-affected discharge present in the Tug Fork (river).

The remaining factor to consider is the distance from the existing plant to the various potential supply sources. While the discharge and quality of water for each of the potential supply sources are important factors to consider, distance to the supply sources is also an important factor to consider, given the increased cost for installing pipelines from potential supply sources to the existing water plant. The mine shaft at Sawmill Hollow at Wilcoe, W. Va. (Mcd-0231) is closest to the water treatment plant, which is approximately 0.50 mi from the plant. For the remaining six potential supply sources, distances vary from 1.11 to 5.28 miles from the existing water plant for the City of Gary, W. Va. A long-term aquifer test would be needed to assess the viability of the mine shaft site (Mcd-0231) as an alternate source of water to augment or replace the existing Grapevine well the City of Gary currently uses as its sole supply source.