Approximately 44 million people in the U.S. (about 13% of U.S. households) rely on domestic wells for drinking water (U.S. Geological Survey, 2016a, b). The Environmental Protection Agency (EPA) regulates water from public sources but it does not regulate private wells. Various contaminants, particularly those that occur naturally in groundwater, may be found in private wells. The U.S. Geological Survey's (USGS) National Water-Quality Assessment Program assessed the water quality of 2100 domestic wells within 48 states and reported that more than one in five of the sampled wells contained one or more contaminants at a concentration greater than either EPA's maximum contaminant levels (MCLs) or USGS' health-based screening levels (HBSLs) (DeSimone, 2009). Testing, treatment, and maintenance of these private wells may prevent contamination and thus protect the health of the well users.

The 2010 U.S. Census estimated the rural population of West Virginia at 51.3%, the second highest rural population in the U.S. (U.S. Census Bureau, 2016). An estimated 23% of West Virginians obtain their water from private wells (Atkins, 2007). However, a comprehensive assessment of constituents in drinking water derived from private wells in West Virginia does not exist.

USGS has conducted several studies on basic groundwater quality in West Virginia (Kozar and Brown, 1995; Kozar and Mathes, 2001; Kozar et al., 2001; Larsen and Mann, 2005; Mathes et al., 1998). Most notably, USGS tested groundwater across various rock strata in West Virginia from 1993 to 2008, and testing indicated that the quality of these public drinking water sources before treatment varied highly across various aquifer geologies (Chambers, 2008; Shiber, 2005).

Because of its geology, the majority of West Virginia falls into Radon Zones 1 and 2 as categorized by the EPA, demonstrating moderate to high concerns for radon exposure (Fig. 1) (West Virginia -EPA Map of Radon Zones, 2016). Radon Zones are classified as Zones 1, 2, or 3 with Zone 1 representing counties with predicted average indoor radon screening levels greater than four picocuries per liter (pCi/L), Zone 2 for counties with levels between two and four pCi/L and Zone 3 for counties with levels lower than two pCi/L. The EPA has set the indoor radon screening level of 4 pCi/L as the action level where homeowners are recommended to take corrective measures to reduce exposure to radon gas. Radon-222 is a radioactive gas that can seep from groundwater into plumbing and accumulate in indoor air as it exits the spigot. Though studies indicate only a small proportion (<5%) of the total radon in indoor air is attributable to radon in water, as the water radon level increases, so can the level of indoor air radon (West Virginia -EPA Map of Radon Zones, 2016).

One of the strongest determinants of groundwater quality is the geology where the groundwater resides (Chambers, 2008). For example, higher arsenic concentrations have been found to be closely associated with sandstone mudstone geologic formations found in certain geologic ages (Abbott, 2002; Schlottmann, 2016). Each geologic age has particular geologic content and rock strata which affect the quality of the groundwater residing there. West Virginia lies predominantly across five geologic strata which range in age from 251 and 444 million years old. The central region of West Virginia and most of the western region is mostly sedimentary rock formed between 318 and 251 million years ago during the Pennsylvanian and Permian ages. Southern West Virginia comprises sedimentary rock formed between 318 and 359 million years ago, predominantly during the Mississippian age. Eastern West Virginia comprises shale, siltstone, and sandstone formed between 359 and 444 million years ago during the Devonian and Silurian periods (Chambers, 2008). Aquifers in the Appalachian Plateaus Province where West Virginia is situated are divided into Mississippian, Permian, and Pennsylvanian aquifers (USGS, 2017).

The Centers for Disease Control and Prevention (CDC), in collaboration with the West Virginia Bureau of Public Health (BPH), initiated an investigation to characterize private well water quality based on the age of the local geology. The objective of this study was to characterize the ground water serving private wells by testing household drinking-water samples and comparing them to EPA's primary and secondary MCLs (Environmental Protection Agency, 2015a, b, c). The study can provide information to state and local public health and local communities about constituent concentrations for planning, testing, and mitigation.

One hundred and thirty-nine residents throughout 10 rural counties from the three regions returned cards inviting them to be a part of our study. We sampled from 8 to 28 wells in each of the 10 counties for a total of 139 water samples. Of these, 40 (28.8%) were in the central region, 52 (37.4%) in the eastern region, and 47 (33.8%) in the southern region. By geologic age, we tested 39 (28.1%) samples from the Devonian age, 32 (23.0%) from the Permian age, 27 (19.4%) from the Mississippian age, 14 (10.1%) from the Pennsylvanian age, and 5 (3.0%) from the Silurian age. GPS coordinates were not provided for 22 (15.8%) samples because a signal was not available. Among wells with recorded drilling year (n = 97, 69.8%), the oldest well was drilled in 1973 and the newest well was drilled in 2010; Thirty six (37.1%) of the wells tested in this study were drilled in 2009. Among wells with recorded depths (n = 90, 64.7%), the range was 40–800 ft.

We show basic water-quality measures by region in Table 1. The average temperature across all regions was 18.4 °C and the average pH was 7.4 (5.6–10.3) for all samples. The average pH was 7.2 (range 6.6–8.6) in the eastern region, 8.0 (range 6.9–10.0) in the central region, and 8.0 (range 5.6–10.3) in the southern region. The turbidity range in all water samples was 0.0–201.0 nephelometric turbidity units (NTU) (mean = 7.2); conductivity range was 0.04–5.1 microsiemens per centimeter (µS/cm) (mean = 0.5); dissolved oxygen range was 3.8–31.7 mg/L (mean = 16.0 mg/L); and salinity range was 0.0–3.7 g/L (mean= 0.3 g/L).

Previous studies of West Virginia groundwater demonstrated that most groundwater samples throughout the region did not exceed current MCLs (Kozar and Brown, 1995; Kozar and Mathes, 2001; Kozar et al., 2001; Larsen and Mann, 2005;Mathes et al., 1998). This study reiterates those findings; only 3 of the 13 constituents tested had at least one sample that exceeded their respective MCL. The most notable were the ten samples that exceeded the arsenic MCL (0.01 mg/L); the maximum arsenic concentration in a sample was more than three times the MCL. Even though these concentrations are lower than any levels that may cause acute arsenic toxicity, many studies have suggested that prolonged use of water with arsenic concentrations higher than the MCL may cause chronic adverse health effects, including cardiovascular disease, diabetes mellitus, and lung, bladder, liver, renal, and skin cancer (Meliker, 2007; Kurttio, 1999).

The arsenic concentrations of samples differed greatly by the age of the surrounding geology; most of the exceedances were from samples located in bedrock aquifers from the Permian age, which comprise primarily sandstone. Studies have shown that arsenic occurs naturally in certain deposits from the Permian age due to many factors, including the oxidized nature of the aquifer and the distribution of finer-grained rock formations such as sandstone and mudstone in the aquifer (Abbott, 2002). Studies in Oklahoma have also reported dissolved arsenic concentrations higher than the MCLs in aquifers from the Permian age (Schlottmann, 2016; Abbott, 2002).

We found that iron and manganese were the only two SMCL constituents where samples had high concentrations that exceeded SMCLs. While SMCLs are not enforceable and adverse health effects are unlikely at the concentrations found in our study, water with high concentrations of iron or manganese may be aesthetically unpleasant and affect the perceived quality of drinking water. Moreover, iron and manganese may increase the corrosiveness of the water, which can stain household fixtures and reduce water flow in distribution pipes (Environmental Protection Agency, 2015a, b, c). Samples with results for iron and manganese that exceeded the SMCL were reported in all regions and geological ages in this study. The geometric mean concentrations for iron were higher in bedrock aquifers from the Devonian age, a finding supported by previous studies in West Virginia (Kozar and Mathes, 2001; Kozar et al., 2001; Chambers, 2008).

We found levels of radon above the PMCL in a large percentage (34.5%) of the water samples we tested, with most of the exceedances occurring in aquifers from Permian-aged rock. Our findings are consistent with other studies suggesting higher groundwater levels of radon exist in the northwestern West Virginia (Chambers, 2008). Private well owners should be aware that groundwater radon concentrations may exceed the PMCL, particularly in aquifers from Permian-aged geologies. Private well owners should also be aware that the majority of West Virginia falls within EPA Radon Zones 1 and 2, which indicate a moderate to high radon potential in the indoor air. Though studies indicate only a small proportion (<5%) of the total radon in indoor air is attributable to radon in water, as the water radon level increases, so can the level of indoor air radon (West Virginia -EPA Map of Radon Zones, 2016). Lung cancer is associated with inhalational radon exposure and is considered the second leading cause of lung cancer after cigarette smoking (Environmental Protection Agency, 2013; National Research Council, 1999). Well owners, particularly those who obtain their water from Permian-aged aquifers or in Radon Zones 1 and 2, may be advised to routinely test their homes for radon to ensure that the indoor air levels are not of any health concern.

Our study had several limitations. West Virginia county health departments for the counties included in this study conducted the recruitment process so we could not determine the overall response rate for the study. Respondents may have had significantly different well characteristics and constituent levels than persons who did not reply to the recruitment card. Future studies using a random selection process from well permits may produce results that are more generalizable.

While we outlined water collection protocols prior to sample collection, there may have been uncontrollable deviations in the sample collection process that produced aberrant results. For example, testing for dissolved oxygen in an open bucket and in higher elevation could produce higher dissolved oxygen levels thus skewing the mean and range above expected levels.

We did not determine the frequency of water use, drinking water treatment, or alternative water sources in this study so we could not assess whether the constituent concentrations might lead to significant exposure. For future studies, researchers could assess water use and knowledge, attitudes, and practices of private-well owners at the time of testing to assess the extent of exposure to potential constituents found in the water samples.

While geologic age can be useful for characterizing differences in water quality and has been shown to be a predictor of potentially high constituent concentrations in private well water, it is important to note that the geochemistry of the aquifer and groundwater are the main drivers of constituent concentrations relating to geologic age. Geologic age has been shown to be a useful surrogate for the actual features of the system that ultimately determines private well water quality. Moreover, constituent concentrations may be associated with factors besides geologic age of the rock surrounding aquifer. Other variables not measured in this analysis, such as proximity to other sources of contamination (e.g., coal power plants and mining facilities) were not accounted for in this analysis.

When an individual sample exceeded the allowable levels for total arsenic or radon, or indicated the presence of E. coli or coliform, BPH provided an information fact sheet in the results packet to the household. The fact sheets aimed to help reduce risks from contaminated water by explaining how to identify potential sources of contamination, have tap water tested, and reduce contamination in drinking water. When arsenic concentrations exceeded the MCL, the fact sheet explained the potential sources of arsenic in private well water, the adverse health effects of ingesting arsenic, and ways to remove arsenic from drinking water. When radon exceeded the PMCL, the fact sheet explained radon sources, the possible health effects of radon in groundwater, and ways to avoid groundwater radon from being released in the air (e.g., by installing an aerator for indoor faucets or granular activated carbon filters). BPH also provided a free radon air sampling kit to the individuals for future testing by BPH. When E. coli or coliform were present in the sample, the fact sheet explained symptoms related to E. coli or coliform exposure and ways to ensure that drinking water was safe before ingesting it. BPH included in the packets a total of 11 fact sheets for arsenic, 48 fact sheets for radon-222, and 39 fact sheets for E. coli or coliform. Public health messaging following testing of private wells may be a cost-efficient method to inform private-well owners of ways to mitigate potential contamination of their wells. State and local public health authorities may consider standardized public health messaging for preemptive notification to private well owners who drill in aquifers that have potentially harmful constituent concentrations.

We found arsenic and radon concentrations were higher in water samples collected in areas mapped to Permian-aged aquifers compared to those sampled from aquifers mapped to other geologic ages. Homeowners with private wells in areas with Permian aged aquifers could benefit from targeted public health messaging about potentially harmful contaminant concentrations in the well water. This may help ensure proper testing and maintenance of private wells to ensure water quality.