Groundwater Quality Defined Relative to Water-Quality Benchmarks

The GAMA-PBP uses established benchmarks for public drinking water to contextualize groundwater quality from domestic wells, which are not regulated in California. Groundwater-quality data are reported here as relative concentrations (RCs), which are defined as the ratio between the concentration measured in a groundwater sample to the concentration of a constituent’s regulatory or non-regulatory benchmark, following approaches by previous studies to understand concentrations in a toxicological context (Toccalino and Norman, 2006; Rowe and others, 2007). This report focuses primarily on water-quality constituents with regulatory maximum contaminant level (MCL) benchmarks (Faulkner and Dupuy, 2023). Constituents with aesthetic-based, secondary maximum contaminant level (SMCL) benchmarks that affect groundwater color, taste, or odor are discussed to support the water quality assessment. Other constituents of interest with non-regulatory, health-based benchmarks (such as California State response levels or Federal health advisory levels) are discussed on a more limited basis. The comparison benchmarks for all constituents detected in samples collected from domestic wells are listed in Faulkner and Dupuy (2023). Regulatory and non-regulatory benchmarks used in the study are established by the U.S. Environmental Protection Agency (EPA), the SWRCB-Division of Drinking Water (SWRCB-DDW), and the USGS, and were selected in the following order of priority:

Concentrations of a given water-quality constituent exceeding its respective benchmark were considered “high.” Concentrations between the benchmark and one-half the benchmark (for inorganic constituents) or one-tenth the benchmark (for organic and PFAS constituents) were considered “moderate.” Concentrations smaller than the moderate threshold were considered “low.” The moderate classification indicates elevated concentrations below the benchmark and can help identify constituents of emerging concern in aquifer systems. The one-tenth threshold was set for organic constituents (such as pesticides, VOCs, and PFAS) because they are typically not naturally present in groundwater systems, and a lower threshold effectively highlights parts of the aquifer area vulnerable to anthropogenic contamination as opposed to inorganic constituents that are often present naturally at low levels in aquifer systems from water-rock interactions (Fram and Belitz, 2014).

Data Collected for Domestic-Supply Assessment

Groundwater-quality data used for the domestic-supply assessment came from sites sampled by the USGS for the GAMA-PBP Kern County subbasin study unit (fig. 1). The Kern County domestic-supply aquifer study unit was defined as the extent of the San Joaquin Valley Kern subbasin (basin number 5-022.14) delineated in the 2020 update of the California Department of Water Resources' Bulletin 118 used for domestic resources (California Department of Water Resources, 2020). The extent of the domestic groundwater resources was drawn using 1-km buffers around areas with at least 1 well determined from well density coverage calculated in Johnson and Belitz (2019). This method resulted in a discontinuous study area (fig. 1). The areal extent of the domestic groundwater resources was divided into 40 equal-area grid cells with a slight rotation, each having an area of about 85 km² (fig. 1). From February to April 2022, one or two domestic groundwater wells were randomly selected for 29 of the grid cells to represent the groundwater resources used for domestic supply. No wells were available, or received permission to sample, in the remaining 11 grid cells. Sample sites for each grid cell were selected from lists of domestic-supply wells compiled using information obtained from the California Department of Water Resources well completion report database (Borkovich and others, 2019; California Department of Water Resources, 2024). Grid cells are non-continuous, and wells 9, 15, 16, 18, 29 fall outside the cell boundaries, but were assigned to the closest cell that did not contain a grid well as described in a USGS data release (Harkness, 2026). Additional information about the methods used to identify sites for sampling and assign site identifiers is provided in Faulkner and Dupuy (2023).

Groundwater samples were taken as close to the wellhead as possible to reflect the quality of ambient aquifer water before household plumbing or treatment systems. Sampling methods followed USGS protocols for groundwater quality sampling as described in the USGS National Field Manual (U.S. Geological Survey, variously dated) with modifications as described in Shelton and Fram (2017) and Burow and others (2024). The median and interquartile range of sampled well depths were 151 meters (m) and 105–197 m, respectively (Faulkner and Dupuy, 2023). Perforated or open intervals were mostly completed above the Corcoran Clay in locations where the clay was present, but seven wells were perforated through or below that confining layer.

The quality assurance and quality control protocols and results for the water-quality data collected for the Kern County subbasin domestic-supply study are described in Faulkner and Dupuy (2023) and are summarized here. Quality assurance evaluations were based on data from quality control samples collected for the Kern County subbasin study and on results from quality control samples collected for other GAMA-PBP study units or for the USGS National Water Quality Networks. Quality control samples collected for this study included three field replicate samples, three field blanks, and one source-solution blank (Faulkner and Dupuy, 2023). Blank detections include four VOCs (dimethoxymethane, benzene, carbon disulfide, and 1,1-difluoroethane) and five inorganic constituents (sulfate, copper, lead, zinc, and hexavalent chromium). Concentrations of trace elements below the study reporting levels (SRLs) defined by Bennett (2020) or below the concentrations detected in blanks collected for the Kern County subbasin study were reported as nondetections relative to a raised reporting level. Of the results originally reported by the laboratory as detections of low concentrations of inorganic constituents in groundwater samples, 46 were reclassified as nondetections, 15 with concentrations less than the highest blank detection, and 31 with concentrations less than SRLs defined by Bennett (2020) and Faulkner and Dupuy (2023). Results originally reported by the laboratory as detections of low concentrations of benzene and carbon disulfide were reclassified as nondetections based on detections of benzene in blanks (Faulkner and Dupuy, 2023) and an SRL established for carbon disulfide (Bexfield and others, 2022). A total of three replicate samples were collected, and all replicate pairs met the acceptance criteria with the following two exceptions: one replicate pair for enterococci consisted of a detection and a nondetection, and one replicate pair for diuron consisted of a detection at a concentration three times the method detection level and a nondetection. Faulkner and Dupuy (2023) also reported that replicate pairs for potassium and nickel did not meet acceptance criteria; however, in both cases, the replicate pair in question consisted of a detection with concentration at the method detection level and a nondetection.

The gridded sampling design used in GAMA-PBP studies allows results to be put in terms of the proportion of aquifer area with water quality in a defined range (for example, exceeding the MCL, or “high”). Wells sampled from individual grid cells represent a random sampling within a defined unit of the total aquifer area used for domestic groundwater resources (Belitz and others, 2010, 2015). The grid-based statistical approach was used to calculate the areal proportions of the domestic groundwater resources in the Kern County subbasin with high, moderate, and low RCs of constituents (eq. 1; Belitz and others, 2010):

Moderate- and low-RC aquifer-scale proportions for the study unit were calculated in the same manner as high-RC aquifer-scale proportions. In addition to the aquifer-scale proportions of high, moderate, and low RCs, the detection frequencies for organic constituents were calculated as area-weighted detection frequencies. The detection frequency was calculated by using equation 1 with $W_c^{high}$ replaced by the number of samples with detections.

For ease of discussion, these proportions are referred to as “high,” “moderate,” and “low” aquifer-scale proportions. Aquifer-scale proportions were calculated for individual constituents, for groups of related individual constituents, referred to as “constituent classes,” and for “all constituents.” Aquifer-scale proportions for constituent classes were calculated using the maximum RC for any constituent in the class to represent the class for each site. For example, a site having a high RC for nitrate, moderate RC for 1,2,3-trichloropropane (1,2,3-TCP), and low RCs for other trace elements and organics would be counted as having a high RC when evaluated for the “all constituents” class.

Explanatory Variables

Characteristics of the aquifer system, geochemical conditions, and land cover are described using explanatory-variable data compiled for the 33 sites sampled for the study unit. Data generated during this study are available as a USGS data release (Harkness, 2026). The methods used for assigning values for each of the explanatory variables are described in the two USGS data releases: Faulkner and Jurgens (2025) and Harkness (2026). Summaries of statistical methods to evaluate the explanatory variables and modeling of noble gases and age tracers are described in this section.

Noble Gas and Residence Time Modeling

The distribution of groundwater age, or time traveled in the aquifer system since entering as recharge, in a sample can be used to assess the vulnerabilities of domestic groundwater resources. Modern water recharged post-1950 can contain anthropogenic contaminants (Shelton and others, 2001; Zogorski and others, 2006; Jurgens and others, 2010; Lindsey and others, 2017). Noble gases and age tracers, tritium, carbon-14, and sulfur hexafluoride, can be used to calibrate a groundwater age model (Jurgens and others, 2020). Noble gas modeling estimates excess air as excess neon concentrations (ΔNe), the temperature at the time of recharge, called the noble gas recharge temperature (NGRT), terrigenic helium-4 concentrations (⁴He_terr), tritogenic helium-3 (³He), and atmospheric helium isotopic ratio ([³He/⁴He]/[³He/⁴He]_air or R/Ra). Terrigenic helium in groundwater is derived from radioactive decay of uranium and thorium in aquifer sediments, and higher terrigenic helium is associated with extensive water-rock interactions over hundreds to thousands of years. Tritogenic helium-3 is derived from radioactive decay of tritium in groundwater, and the atmosphere has a constant ³He/⁴He (Ra) ratio that can be identified in young, more recently recharged water. The R/Ra ratio changes due to increases in terrigenic helium with age and water-rock interactions (R/Ra<1) or rapid recharge that enriches tritogenic helium in groundwater (R/Ra>1). We used the program Dissolved Gas Modeling and Environmental Tracer Analysis (DGMETA; Jurgens and others, 2020) to model the recharge gas parameters as described in the USGS data release (Faulkner and Jurgens, 2025).

Groundwater age was estimated through calibration of environmental age tracers (sulfur hexafluoride, tritium, and carbon-14) to lumped parameter models (LPMs) using the computer program TracerLPM (Jurgens and others, 2012; Faulkner and Jurgens, 2025). The samples were calibrated to a dispersion model (DM) or a binary mixture of two DMs. The DM is an LPM that assumes that the age distribution in the well is due to dispersion in the aquifer. The DM has two free parameters (age and dispersion parameter), or five free parameters (age of the younger and older components, the dispersion parameters of the younger and older components, and the mixing fraction of the two components) in a binary mixture. The dispersion parameter can be adjusted to approximate other LPMs. Relative environmental tracer concentrations guided the calibration of LPMs to tracer concentrations; a DM was used to compute a best-fit mean age when environmental tracers indicated a single age, and a binary mixture of two DMs was used to simulate the mixing of modern and premodern groundwater (Jurgens and others, 2012). The mean age of the groundwater and the proportion of modern water determined from the resulting groundwater age distributions were used to examine correlations to groundwater quality. Groundwater age results and detailed processing steps are available as a USGS data release (Faulkner and Jurgens, 2025).

Explanatory Factors

The explanatory factors described here provide insight into the processes potentially affecting groundwater quality. Factors that show relations to water-quality constituents that exceed thresholds in the domestic groundwater resources in the Kern County subbasin include groundwater age classification, stable isotopic composition of water, reduction-oxidation (redox) conditions, and land use.

Groundwater Age

Groundwater age is a measure of the time that water has resided in an aquifer or traveled from a recharge source to a sample point, such as a well. A single sample represents a distribution of ages through the depth of the aquifer captured by the well. Modern groundwater recharged after about 1950, is characterized by higher concentrations of tritium and carbon-14 scavenged from the atmosphere by rain before infiltrating into the aquifer system (Lindsey and others, 2019). Premodern groundwater, recharged before about 1950 and sometimes up to thousands of years old, has lower concentrations of tritium and carbon-14. Wells often contain mixtures of modern and premodern groundwater, and lumped parameter modeling using the tritium and carbon-14 concentrations was used to calculate mean ages and fraction of modern water in wells with binary age mixtures (Jurgens and others, 2020; Faulkner and Jurgens, 2025). Modeled mean ages in domestic wells in the study unit ranged from 12 to 36,000 years, but mean age alone does not provide enough information to determine if modern and premodern groundwater are mixed in a sample. Groundwater age classification was determined from modeled fraction of modern water and tritium concentrations. Wells with modeled modern fractions less than 5 percent were classified as premodern (18 percent). Wells with tritium concentrations greater than 1 tritium unit (TU) and the modern fraction greater than 75 percent were classified as modern (45 percent), and wells with tritium concentrations below 1 TU and more than 5 percent modern fraction were classified as mixed (37 percent; table 1; fig. 6A; Harkness, 2026). Groundwater ages increased with distance from the Kern River (fig. 7A) and wells with modern-aged groundwater were located near the Kern River, local rivers, canals, and recharge ponds (fig. 6A).

Water Isotopes

Stable isotopes of water, delta oxygen-18 (δ¹⁸O) and delta hydrogen-2 (δ²H) provide a tracer of recharge source and can be used to identify the relevant contributions of river water to groundwater wells throughout the Central Valley and other parts of California (Visser and others, 2018; Wright and others, 2019; Burow and others, 2024; McMahon and others, 2024). In the domestic well samples, we see clustering of lighter, more negative oxygen stable isotope ratios in waters with δ¹⁸O values between −11 and −14 per mil and heavier, less negative δ¹⁸O values between −10 and −7.5 per mil (figs. 6B; Faulkner and Dupuy, 2023). The lighter isotope ratios were present in groundwater closer to the Kern River and recharge areas around Bakersfield (fig. 6B). Wells with heavier isotope ratios were further away from the Kern River, with the heaviest values in wells in the northern part of the study area (fig. 6B). Isotope ratios increased significantly with distance from the Kern River (p-value<0.001; table 1), consistent with the classification of groundwater with ratios lighter than −12 per mil as recharged with river water (fig. 7B; Visser and others, 2018).

Lechler and Niemi (2012) calculated a local meteoric water line (LMWL) for the southern San Joaquin Valley of δ²H=7.0×δ¹⁸O−5.2 (fig. 8). Local precipitation in the San Joaquin Valley has an δ¹⁸O value of −8.2±0.2 per mil (Visser and others, 2018). The Kern River has a lighter, more negative δ¹⁸O value of −13.34 (Kendall and Coplen, 2001) per mil that intersects the LMWL consistent with water sources from high-elevation Sierra Nevada headwaters (fig. 8; Wright and others, 2019). In addition to the Kern River, managed recharge projects receive water from the State Water Project and the Central Valley Project. These sources have higher δ²H and δ¹⁸O values compared to the Kern River, which could explain the number of wells with mixed source stable isotope ratios (Warden and others, 2024a).

Most domestic well water samples containing modern water had lighter stable isotope ratios (river water group), whereas most mixed and premodern samples had heavier stable isotope ratios (precipitation and mixed group; fig. 8). Wells in the river water group fell to the right of the LMWL which is consistent with the stable isotopes ratios being affected by higher rates of evaporation in managed recharge ponds or agricultural fields (fig. 8). Wells in the precipitation and mixed group tended to fall on or between the LMWL and global meteoric water line from Craig (1961; fig. 8). A subset of samples (S15-KERBS-13, S15-KERBS-15, S15-KERBS-23) close to the northeastern boundary of the study area plotted to the right of the LMWL, consistent with a more evaporative source (fig. 8).

Redox Classification

The oxygen content of groundwater affects the mobilization and degradation of some constituents in groundwater and appears to be a driver in the presence of nutrients, trace elements, and VOCs in domestic wells in Kern County subbasin. Wells sampled in the study unit were categorized as oxic or anoxic (fig. 6C; Harkness, 2026). Samples with DO greater than 1 mg/L were classified as oxic (76 percent of samples), and samples with DO concentrations below 1 mg/L were classified as anoxic (24 percent of samples). Anoxic wells were generally reported in the outer edges of the study area and the eastern extent of the Kern River, in wells with premodern groundwater (fig. 6C).

Groundwater from wells in fine-grained alluvial sediments, especially those at the far end of groundwater flow paths, are more likely to be anoxic because of the geochemical evolution of groundwater as it moves through the system (McMahon and Chapelle, 2008). Median DO concentrations were significantly higher (p-value=0.045) in modern (5.3 mg/L) and mixed (4.25 mg/L) groundwater than premodern groundwater (0.325 mg/L; fig. 9).

Land Use

Most domestic wells (80 percent) were in areas defined by agricultural land use, with an overall average agricultural land use of 77 percent within a 500-m radius of wells (fig. 2; Harkness, 2026). Three of the domestic wells (9.1 percent) were in urban areas, defined here as greater than 50 percent urban land use within a 500-m radius of the well. These wells are near Bakersfield, the only large metropolitan area in the study unit. The average proportion of urban land use in the 500-m buffers around all 33 wells was 9.6 percent, and the proportion of urban land decreased with distance from the Kern River (p<0.001; table 1; fig. 10). Three wells (9 percent) were in low-use natural areas, defined here as greater than 50 percent natural land use within a 500-m radius of the well, and the average natural land proportion for all 33 wells was 13.4 percent within a 500-m radius (Harkness, 2026).

Land use is related to groundwater age and recharge sources in domestic wells. The median age of groundwater in urban wells is 43 years, which is lower than the median age of 65 years in agricultural wells and significantly lower (p-value=0.012) than the median of 21,000 years in wells in natural areas (fig. 11). The fraction of modern water increased with proportion of urban land use and decreased with natural land use and δ¹⁸O decreased with proportion of urban land use due to recharge of young river water in the Bakersfield area and increased with proportion of natural lands (table 1; fig. 12). The proportion of agricultural land use was not correlated with age or stable water isotope ratios likely due to mixed sources of water used for irrigation that is recharging the domestic groundwater resources in agricultural areas.

Noble Gas Modeling

Noble gases in groundwater provide information about the recharge environment of the water source before it migrates to drinking water wells. Estimated NGRTs reflect the air temperature at the time and location of recharge. The NGRT in domestic wells are from 9.3 to 25 degrees Celsius (°C), reflecting a range from temperatures at high elevation in the Sierra Nevada to summer temperatures in the valley. Recharge of Kern River water in the Bakersfield area is associated with significantly lower NGRT (14.4 °C) than the mean annual air temperature (19.2 °C) due to higher elevation sources in the Sierra Nevada and rapid infiltration riverbed and recharge ponds, limiting equilibration of the colder river water with warmer atmospheric temperatures in the valley (Visser and others, 2014, 2018). The NGRT increased with the proportion of agricultural area around domestic wells and distance of domestic wells from the Kern River for wells with mixed and modern groundwater ages (table 1; fig. 13). Higher NGRTs in agricultural areas is consistent with recharge of summer irrigation water and lower NGRTs closer to the Kern River is consistent with recharge of river water derived from winter precipitation at higher elevation (fig. 6D; Cey and others, 2008; Visser and others 2018; Castaldo and others, 2021).

During rapid recharge, excess air bubbles are trapped or dissolved into groundwater, and ΔNe is a measure of that excess air (Cey and others, 2008). The ΔNe in domestic wells ranged from 6.2 to 283 percent. High excess air is associated with enhanced recharge, particularly due to MAR. The higher ΔNe in domestic wells in the Kern County study unit, particularly above 100 percent, are derived from enhanced recharge in the Kern River and MAR operations (Cey and others, 2008; Visser and others, 2014, 2018).

Principal Component Analysis

Multivariate analysis using PCA helps to evaluate the relations between groundwater age, recharge source indicators, redox conditions, and land use in a highly complex system like the Kern County subbasin. Variables that plot close to each other are more correlated than variables that plot further away. The length of the vector, or arrows, shows the amount of variability contributed by the variable in the first two principal components. Principal components (PC) 1 and 2 explain more than 60 percent of the variability in the explanatory variables (fig. 14). Mean age and proportion of natural land cover are strongly associated with PC 1 and increase in the positive direction, whereas proportion of agricultural land use is strongly associated with PC 1 in the negative direction (fig. 14). Dissolved oxygen is also associated with PC 1 in the negative direction but contributes a smaller amount to total variability. Proportion of urban land use is strongly associated with PC 2 in the positive direction, whereas water level, well depth, and NGRT are strongly associated with PC 2 and increase in the negative direction (fig. 14). The fraction of modern water, ΔNe, stable isotopes of water, and aridity are associated with both PC1 and PC2. Aridity and water isotopes increase in the positive PC1 direction and negative PC2 direction, whereas the modern fraction and ΔNe increase in the negative PC1 direction and positive PC2 direction. ΔNe shows a weaker association with PC1 and PC2 than that of other variables.

The PCA results indicate the relation between groundwater age and land use, with age increasing with increasing proportion of natural land cover, and the fraction of modern water increasing with agricultural and urban land use. The presence of modern water is related to DO and ΔNe variations, with a greater effect of urban land use on ΔNe and agricultural use on DO. The fraction of modern water and ΔNe vary inversely to the stable isotope ratios of water, likely due to young, river water recharging at MAR operations having lighter, more negative ratios compared to local precipitation. Less negative, heavier stable isotopes of water close to the local precipitation ratio of −8 per mil (Visser and others, 2018) are associated with older water in natural areas and deeper wells with deeper water levels. Well depth and water levels vary inversely with the fraction of modern water and the proportion of urban land use, suggesting that wells near the Kern River and MAR operations around Bakersfield have shallow water levels and younger water related to the river water recharge. The NGRT is closely related to water level and well depth, as well as agricultural land use, suggesting that summer irrigation water is a larger fraction of groundwater in deeper wells in agricultural areas further from the Kern River and MAR.

Recharge Sources

Combining geochemical, isotopic, and geospatial analyses provides a framework for delineating multiple sources and processes affecting water quality in the domestic groundwater resources of the Kern County subbasin. The explanatory variables helped build a conceptual model of the groundwater system and geochemical environment in the parts of the aquifer system used by domestic wells. Although the Kern River used to be a source of groundwater discharge in the basin, pumping and river diversions have led to the river becoming a primary source of recharge (Faunt, 2009; Visser and others, 2018; Wright and others, 2019). In addition, MAR projects at the Kern Water Bank and in the eastern part of Kern County, enhance recharge of surface water from the Kern River, State Water Project, and Friant-Kern Canal into the aquifer system (fig. 1). Wright and others (2019) reported modern water from the Kern River as a primary source of recharge in public-supply wells in the Fruitvale Oil Field, just north of Bakersfield near the Kern River and recharge areas. They reported a positive correlation between the fraction of modern water and distance to the Kern River in their sample set, but mean age was not correlated with distance. The samples collected by Wright and others (2019) were more densely located in an area within 6 km of the Kern River, whereas the samples collected for the GAMA-PBP were distributed throughout the Kern County subbasin and include samples collected in natural areas up to 52 km away from the river (fig. 1). A study by McMahon and others (2024) also reported that the Kern River, Poso Creek, and adjacent engineered recharge facilities were sources of modern recharge in the Poso Creek Oil Field northeast of Bakersfield. As noted in previous studies (Cey and others, 2008; Visser and others, 2018; Wright and others, 2019; Castaldo and others, 2021), a plume of isotopically light (δ¹⁸O<−12 per mil) groundwater extends out from the Kern River (fig. 6B). Wells in the plume contain modern and oxic water (DO concentrations between 4.6 and 16 mg/L; figs. 6A, 6C). Additionally, noble gas modeling also supports recently recharged water from the Kern River and adjacent MAR operations moving into the groundwater resource; NGRT are significantly correlated with distance from the Kern River and below 15 °C in domestic wells close to recharge areas (table 1; fig. 6D).

In addition to direct recharge of surface water in MAR, river water is used for irrigation throughout the extensive agricultural areas in the study unit. The NGRT and ΔNe can help distinguish summer irrigation recharge of river water delivered from the extensive canal system throughout the subbasin. Irrigation water is generally a mixture of sources, surface water from the Kern River and other northern Central Valley rivers delivered via canals and groundwater pumped locally. During drought years, up to 50 percent of irrigation water is from groundwater (Faunt and others, 2024). Wells in agricultural areas have intermediate δ¹⁸O between −9 and −12 per mil, consistent with mixing of river water with groundwater sourced from precipitation (fig. 6C). Irrigation generally happens in the summer months; therefore, NGRT in domestic wells recharged with irrigation water will have higher NGRT, regardless of water source (Visser and others, 2014). Recharge temperatures are higher in areas with a higher proportion of agricultural use, and NGRTs are greater than 14 °C, and as much as 23 °C (figs. 6D, 13; Faulkner and Jurgens, 2025). A positive correlation between NGRT and proportion of agricultural land use and distance from the Kern River supports irrigation as the source of recharge in agricultural areas (table 1; fig. 13).

Anthropogenic Constituents

Most inorganic constituents are naturally present in groundwater at low concentrations, but human activities can be sources of elevated concentrations. Volatile organic compounds can be naturally present in groundwater associated with hydrocarbon (natural gas and oil) deposits, but their presence in groundwater in most areas outside oil and gas fields is related to anthropogenic sources; however, pesticides are entirely derived from anthropogenic applications at the surface (Zogorski and others, 2006; Rowe and others, 2007; Toccalino and others, 2014). In the Kern County subbasin, nitrate, the fumigants 1,2,3-TCP and DBCP, and PFAS were detected at high relative concentrations in the domestic groundwater resources.

Nitrate

Nitrate is present naturally in groundwater at low concentrations, but concentrations above 2 mg/L are typically attributed to anthropogenic sources such as fertilizer applications and septic systems (Nolan and others, 2002; Dubrovsky and others, 2010). Nitrate persists in oxic groundwater with DO concentrations more than 1 mg/L. Under low oxygen or anoxic environments, denitrification, or the bacterial-mediated consumption of nitrate, removes nitrate from groundwater. Nitrate concentrations exceeded the MCL of 10 mg/L in five domestic wells and high RCs were only reported in wells in agricultural areas (fig. 15A). Nitrate concentrations were positively correlated with total number of VOC and pesticide detections, and with concentrations of 1,2,3-TCP, DBCP, and chloroform, which were the three most commonly detected VOCs in the study (tables 2, 3). Nitrate was significantly higher in oxic domestic wells (fig. 16), and nitrate was positively correlated with DO concentrations (p-value<0.001) and NGRT (p-value=0.049), indicating higher nitrate concentrations in wells receiving recharge of high oxygen, irrigation water during summer months (table 3; fig. 17). High and moderate concentrations of nitrate co-occurring with detections of fumigants and pesticides further supports irrigation recharge as a source of high nitrate to domestic wells in the Kern County subbasin.

Table 2.    

Results of Spearman's rank correlation tests for relations between water-quality constituents in the San Joaquin Kern County subbasin study unit, 2022, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project (Faulkner and Dupuy, 2023; Harkness, 2026).

[Spearman's rho (ρ) are shown for tests in which the variables were determined to be significantly correlated based on p-values (not shown) being less than the critical level (α) of 0.05. Negative numbers indicate a negative correlation. Abbreviations: —, Spearman’s test indicates no significant correlation between factors]

Table 2.    Results of Spearman's rank correlation tests for relations between water-quality constituents in the San Joaquin Kern County subbasin study unit, 2022, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project (Faulkner and Dupuy, 2023; Harkness, 2026).

ρ	Calcium	Magnesium	Potassium	Sodium	Bicarbonate	Bromide	Chloride	Sulfate	Total dissolved solids (TDS)	Nitrate	Arsenic	Boron	Hexavalent chromium	Manganese	Molybdenum	Selenium	Vanadium	Uranium	Perchlorate	1,2,3-Trichloro-propane (1,2,3-TCP)	1,2-Dibromo- 3-chloro-propane (DBCP)	1,2-Dibromo-ethane (EDB)	Chloroform	2-Chloro-4,6-diamino-s-triazine (CAAT)	4-Hydroxy-chloro-thalonil	Bromacil	Metolachlor oxanilic acid	Metolachlor sulfonic acid	Methane
Calcium		0.68	—	—	—	0.50	0.50	0.63	0.71	0.42	—	—	—	—	—	0.54	—	0.51	—	—	—	—	—	—	—	—	—	—	—
Magnesium			0.63	—	0.56	—	—	0.45	0.52	—	—	0.46	—	0.36	0.42	—	—	0.41	—	—	—	—	—	—	—	—	—	—	—
Potassium				—	0.53	—	—	—	—	—	—	0.49	—	—	—	—	−0.35	—	—	0.38	—	—	—	—	—	—	—	—	—
Sodium					—	0.76	0.77	0.78	0.79	—	—	—	—	0.50	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Bicarbonate						—	—	—	—	—	—	0.86	−0.42	—	—	—	—	0.48	—	—	—	—	—	—	—	—	—	—	—
Bromide							0.91	0.72	0.81	—	—	—	—	0.48	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Chloride								0.71	0.79	—	—	—	—	0.48	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Sulfate									0.92	—	—	—	—	0.53	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—
Total dissolved solids (TDS)										—	—	—	—	0.55	—	—	—	—	0.35	—	—	—	—	—	—	—	—	—	—
Nitrate											—	—	0.58	—	−0.34	0.74	—	—	0.83	0.59	0.52	0.43	0.61	0.49	0.42	0.49	—	—	−0.59
Arsenic												—	—	—	0.57	—	0.47	—	—	—	—	—	−0.38	—	—	—	—	—	—
Boron													−0.47	—	—	—	—	0.37	—	—	—	—	—	—	0.35	0.35	—	—	—
Hexavalent chromium														−0.39	−0.45	0.66	—	—	0.74	0.36	—	—	0.40	—	—	—	—	—	−0.42
Manganese															0.59	—	−0.52	—	—	—	—	—	—	—	—	—	—	—	0.66
Molybdenum																—	—	—	−0.37	—	—	—	—	—	—	—	—	—	—
Selenium																	—	0.48	0.68	0.41	—	—	0.39	—	—	—	—	—	−0.44
Vanadium																		—	—	—	—	—	—	—	—	—	—	—	−0.43
Uranium																			—	—	—	—	—	—	0.35	0.36	0.36	0.42	—
Perchlorate																				0.48	0.48	0.37	0.49	0.36	—	0.43	—	—	−0.44
1,2,3-Trichloropropane (1,2,3-TCP)																					0.59	0.59	0.78	0.43	—	0.59	—	—	−0.53
1,2-Dibromo-3-chloropropane (DBCP)																						0.69	0.51	0.38	—	0.59	—	—	—
1,2-Dibromoethane (EDB)																							0.55	—	—	0.51	—	—	—
Tetrachloroethene (PCE)																							0.48	0.44	0.53	—	0.37	0.40	—
Chloroform																								0.38	—	0.46	—	—	−0.46
2-Chloro-4,6-diamino-s-triazine (CAAT)																									0.87	0.50	—	—	—
4-Hydroxychlorothalonil																										0.52	—	—	—
Bromacil																											—	—	—
Metolachlor oxanilic acid																												0.58	—
Metolachlor sulfonic acid																													—
Methane

Table 3.    

Results of Spearman's rank correlation tests for between water quality constituents and explanatory variables in domestic wells in the San Joaquin Kern County subbasin study unit, 2022, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project (Faulkner and Dupuy, 2023; Harkness, 2026).

[Spearman's rho (ρ) are shown for tests in which the variables were determined to be significantly correlated based on p-values (not shown) being less than the critical level (α) of 0.05. Negative numbers indicate a negative correlation. Abbreviations: δ²H, delta deuterium in water; δ¹⁸O, delta oxygen-18 in water; m, meters; VOC, volatile organic compound; —, Spearman’s test indicates no significant correlation between factors]

Table 3.    Results of Spearman's rank correlation tests for between water quality constituents and explanatory variables in domestic wells in the San Joaquin Kern County subbasin study unit, 2022, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project (Faulkner and Dupuy, 2023; Harkness, 2026).

ρ	Depth to top of perforation	Dissolved oxygen (DO)	pH	δ2H	δ18O	Noble gas modeled recharge temperature (NGRT)	Tritium	Excess neon (ΔNe)	Terrigenic helium	Adjusted carbon-14	Mean age	Modern fraction	Percentage of land use within a 500-m radius			Aridity index	Distance to Kern River	Total VOC detections	Total pesticides detections
													Agricultural	Natural	Urban
Calcium	—	—	−0.60	—	—	—	—	—	—	—	—	—	—	—	—	—	0.46	—	0.52
Magnesium	—	—	−0.85	—	—	—	—	−0.39	—	—	—	—	—	—	—	—	—	—	—
Potassium	—	—	−0.52	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.38	—
Sodium	—	−0.39	—	0.59	0.65	—	−0.52	—	0.56	−0.53	0.59	−0.48	—	0.52	−0.47	0.44	0.47	—	—
Bicarbonate	—	—	−0.59	—	—	—	—	—	—	—	—	—	—	—	—	—	−0.58	—	0.39
Bromide	—	—	—	0.58	0.64	—	−0.51	—	0.54	−0.54	0.55	−0.44	—	0.34	—	0.43	0.55	—	—
Chloride	—	—	—	0.63	0.66	—	−0.45	—	0.38	−0.53	0.58	−0.45	—	0.36	−0.40	0.45	0.65	—	—
Sulfate	—	—	—	0.59	0.67	—	—	—	0.39	−0.35	0.34	—	—	0.37	—	—	0.52	—	—
Total dissolved solids (TDS)	—	—	—	0.61	0.69	—	—	—	0.40	−0.38	—	—	—	0.43	—	0.35	0.52	—	—
Nitrate	—	0.77	−0.38	—	—	0.37	—	—	—	—	—	—	—	—	—	—	—	0.46	0.35
Arsenic	—	−0.47	—	—	—	—	—	—	—	−0.35	—	—	—	—	−0.37	—	—	—	—
Boron	−0.39	—	−0.48	—	—	—	—	—	—	—	—	—	—	—	—	—	−0.55	—	—
Hexavalent chromium	—	0.57	—	—	—	0.42	—	—	—	—	—	—	—	—	—	—	0.38	—	—
Manganese	—	−0.43	—	0.38	0.41	—	—	−0.39	0.36	0.35	—	—	—	—	—	—	—	—	—
Molybdenum	—	−0.54	—	—	0.38	—	−0.38	—	0.44	—	—	—	—	—	—	—	—	—	—
Selenium	—	0.57	−0.42	—	—	0.48	—	—	—	—	—	—	—	—	—	—	—	—	—
Uranium	−0.37	—	−0.67	—	—	—	0.40	—	—	0.51	−0.62	0.63	—	—	—	−0.61	—	—	0.73
Perchlorate	—	0.60	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.40	—
1,2,3-Trichloropropane (1,2,3-TCP)	—	0.52	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.87	0.34
1,2-Dibromo-3-chloropropane (DBCP)	—	0.45	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.55	—
1,2-Dibromoethane (EDB)	—	0.52	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.62	—
Tetrachloroethene (PCE)	—	—	—	—	—	—	—	—	—	0.38	—	—	—	—	0.45	—	—	—	—
Chloroform	—	0.54	—	—	—	—	—	—	—	—	—	—	—	—	0.45	—	—	0.86	0.44
2-Chloro-4,6-diamino-s-triazine (CAAT)	—	0.36	—	—	—	—	—	—	—	—	−0.37	—	—	—	—	—	—	—	0.59
4-Hydroxychlorothalonil	—	−0.37	—	—	—	—	—	—	—	—	0.47	0.36	—	—	—	—	—	—	0.50
Bromacil	—	−0.39	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.50	0.43
Metolachlor oxanilic acid	—	−0.39	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	—	0.38
Metolachlor sulfonic acid	—	−0.35	—	—	—	—	—	—	—	—	−0.37	0.37	—	—	—	−0.43	—	—	0.49
Methane	—	−0.45	—	—	—	—	—	—	0.47	−0.36	—	—	—	—	—	—	—	−0.44	—

Previous studies in the Central Valley have indicated that river water recharge is an important factor in controlling nitrate concentrations in domestic wells (Castaldo and others, 2021). They hypothesized that dilution with river water is the driving factor for lower nitrate in domestic wells near rivers or MAR operations, and irrigation run-off is a source of elevated nitrate to groundwater recharge in agricultural areas. Nitrate concentrations in the major rivers that are the sources of water for irrigation and MAR are generally low. For example, average nitrate concentration in the San Joaquin River at Vernalis (USGS-11303500) in 2020–24 was 0.9 mg/L as N (U.S. Geological Survey, 2023). In 2017, a sample from the Kern Water Bank had a nitrate concentration below 0.04 mg/L as N (Warden and others, 2024b). Ambient precipitation in California generally has nitrate concentrations less than 0.2 mg/L N (Keresztesi and others, 2020). Therefore, the recharged river water and precipitation water themselves are not the source of the elevated nitrate concentrations in domestic wells.

Nitrate is applied on the land surface as fertilizer or accumulated in underground septic systems and swept into the groundwater system by recharge. The recharge from the river water is also oxic, and these oxic conditions prevent removal of nitrate by denitrification. Anoxic aquifer conditions conducive to denitrification are not widely present in the San Joaquin Valley and generally confined to the old discharge zone in the axial trough of the Valley (Landon and others, 2011; Rosecrans and others, 2017, 2018). In the Kern County subbasin, anoxic conditions are reported in wells in the northwestern and southwestern edges of the study area (fig. 6C).

Organic Compounds

The presence of organic compounds in the domestic-supply wells indicates anthropogenic surface sources mixing into the groundwater resources. Organic compounds were observed in domestic wells in agricultural and urban areas (fig. 18). The fumigants 1,2,3-TCP and DBCP, which are present at concentrations exceeding their MCLs in some domestic wells, are associated with agricultural applications. Wells with 1,2,3-TCP detections were distributed throughout the study area but mostly located in the central part of the subbasin, north of the Kern River, and the eastern region, south of the Kern River (fig. 15B). 1,2,3-TCP concentrations were correlated with other fumigants (DCBP and EDB) and the pesticides bromacil and CAAT (table 2). The extensive agricultural industry in the subbasin, the history of fumigant use for crops in the Central Valley, and oxic conditions explain the large proportion of high RCs of 1,2,3-TCP. The persistence of 1,2,3-TCP and other fumigants has been reported in other groundwater quality studies in the Central Valley (Burow and others, 2019, 2024).

Non-fumigant VOCs and PFAS were also detected in some domestic-supply wells. The proportion of urban land use was positively correlated with total VOC detections and concentrations of chloroform and tetrachloroethene (PCE; table 3). These chemicals are associated with industrial and disinfection processes, and concentrations do not exceed any MCLs. S15-KERBS-32, the only sample in Bakersfield (fig. 1), has a high concentration of PFOA and moderate concentrations of four other PFAS in addition to detections of chloroform, PCE, and 1,4-dioxane, all associated with an urban source of VOCs (Burow and others, 2024). Sample S15-KERBS-32 contained modern river water recharge with low TDS and ΔNe more than 200 percent. Sample S15-KERBS-32 was in an urban area with the highest density of septic tanks and the shortest distance to a leaking underground storage tank (Harkness, 2026). We cannot determine from this study whether the recharge water itself is a source of anthropogenic contaminants or if there is subsurface mixing of leakage from a septic system or underground storage tank.

Overall, 35 percent of the domestic groundwater resources have high (17 percent) or moderate (18 percent) nitrate concentrations from agricultural sources. However, in areas close to recharge, most notably the Kern River, nitrate concentrations are low, likely due to dilution by low-nitrate river water. However, the river water recharge does not necessarily reduce the presence of organic constituents. Persistent oxic conditions in agricultural and urban areas support the hypothesis that the high detections of fumigants are from agricultural applications, and the detections of VOCs and PFAS are from industrial applications in domestic wells throughout the study area.

Geogenic Constituents

Most inorganic constituents are naturally present in groundwater, but human activities can affect their concentrations (Hem, 1985). Uranium and arsenic were the two constituents with regulatory, health-based benchmarks that were present in domestic groundwater resources at high relative concentrations.

Uranium

Uranium concentrations exceed the MCL of 30 µg/L in wells located in agricultural areas (fig. 15C). Uranium concentrations are positively correlated with the fraction of modern water (p<0.001), bicarbonate (p=0.008), total pesticide detections (p<0.001) and concentrations of several pesticides and pesticide degradates and negatively correlated with pH (p<0.001; tables 2, 3; fig. 19). Uranium concentrations decreased with depth to the top of the perforated interval (p=0.046) and uranium concentrations were significantly higher in wells screened above the Corcoran Clay (mean=24.4 µg/L) than below (3.9 µg/L) or outside the Corcoran Clay extent (6.1 µg/L; fig. 20). A relation between elevated uranium and bicarbonate due to geochemical changes in groundwater driven by irrigation recharge has been reported in previous studies in the San Joaquin Valley and elsewhere (Jurgens and others, 2010; Coyte and others, 2018; Rosen and others, 2019). The introduction of young, alkaline irrigation recharge into shallow domestic wells is contributing to the mobilization of uranium.

Arsenic

Arsenic is the other geogenic trace element reported at high RCs in the study unit, and high concentrations were reported in wells with anoxic or alkaline groundwater (fig. 15D). Arsenic concentrations were negatively correlated with DO and RCs above the MCL mostly occurred in groundwater with pH between 8 and 9.5 (table 3; fig. 21). Dissolution of ion and manganese oxides in aquifer sediments under anoxic conditions can release arsenic into groundwater and arsenic is more mobile under alkaline conditions (pH>8), even in oxic conditions (Smedley and Kinniburgh, 2002; Belitz and others, 2003). Arsenic concentrations were not correlated with mean groundwater age, although they were inversely correlated with carbon-14, indicating that arsenic concentrations generally are higher in older groundwater (table 1). All samples with arsenic concentrations greater than the MCL had mean ages of 75 to 20,000 years and fractions of modern groundwater less than 25 percent (Faulkner and Dupuy, 2023; Faulkner and Jurgens, 2025).

The longer groundwater resides in the aquifer system, DO levels decrease and pH increases, resulting in conditions that favor high arsenic concentrations (Smedley and Kinniburgh, 2002). Arsenic concentrations were negatively correlated with the proportion of urban land use, suggesting the high-oxygen river water recharged along the Kern River in Bakersfield is protective against arsenic mobilization in domestic resources (figs. 15D). However, mixed or premodern anoxic or alkaline groundwaters in domestic wells outside of the Bakersfield area can have elevated arsenic concentrations.

Total Dissolved Solids

Total dissolved solids can be used as a proxy for groundwater salinity. Sources of TDS in the San Joaquin Valley are minerals dissolved from aquifer sediments, particularly marine sediments from the coastal ranges, concentration of solutes during evapotranspiration in shallow groundwater, mixing with hydrothermal fluids, and anthropogenic sources such as irrigation water recharged in agricultural areas and wastewater (Burton and others, 2012; Fram, 2017; Burow and others, 2024). A previous USGS-GAMA study of the groundwater resources used by public-supply in the Kern County subbasin reported that high TDS concentrations were related to both natural and agricultural sources (Burton and others, 2012). Total dissolved solid concentrations in domestic wells range from 186 to 1,660 mg/L, and 12 percent of the domestic groundwater resources have high TDS concentrations (greater than the SMCL of 1,000 mg/L). Total dissolved solid concentrations increased with distance from recharge around the Kern River and high relative concentrations are in wells in the north and southmost extent of the study area (table 3, fig. 15E). Kern River and recharge water have lower TDS concentrations, higher DO, and lower pH than most groundwater in the Kern County subbasin (Warden and others, 2024b; California Department of Water Resources, 2025). Changes in major ion concentrations also changed as TDS increased away from the Kern River, with calcium, sodium, bromide, chloride, and sulfate all increasing with distance from the Kern River and bicarbonate decreasing with distance from the Kern River (table 3). Total dissolved solid concentrations are positively correlated with δ¹⁸O, because of increasing TDS concentrations with decreasing proportions of river water recharge. Sodium (p-value<0.001), chloride (p-value<0.001), and sulfate (p-value<0.001) were also positively correlated with stable isotope ratios of water (table 3). However, TDS was not correlated to groundwater age despite a relationship with river water recharge.

Agricultural irrigation also is a source of recharge in the Kern County subbasin and has higher concentrations of TDS than the recharge along the Kern River or at MAR operations. Multiple sources of recharge with different salinity may explain why TDS is not correlated with groundwater age. We identified two types of wells in agricultural areas: Type A wells have low TDS concentrations below 500 mg/L, and Type B wells have moderate to high TDS concentrations over 500 mg/L, indicating varying degrees of mixing between irrigation water and river water (fig. 22). Previous studies have noted the increase in TDS with distance from the Kern River and hypothesized that water-rock interactions and enhanced mineral dissolution along groundwater flow paths in agricultural areas control salinity and dissolved solid concentrations (Hansen and others, 2018; Wright and others, 2019; McMahon and others, 2024). Marine sediments deposited from the Coast Ranges to the west and the San Emigdio Mountains in the southwest could be contributing TDS to groundwater in the more distal parts of the study area (Burton and others, 2012; Fram, 2017). Disposal of water produced from oil and gas operations may also contribute bicarbonate and TDS to irrigation water in the study area (Tisherman and others, 2023). Total dissolved solids in domestic wells are positively correlated with terrigenic helium concentrations, and wells in agricultural areas with moderate to high TDS (Type B) have lower neon-to-helium ratios, suggesting either accumulation of helium due to longer residence times or mixing with old water as sources of TDS (fig. 22).

Old Water in Domestic Wells

Kern County is one of California’s most densely developed areas for oil and gas extraction (Davis and others, 2018). Water quality effects on groundwater, including domestic wells, in Kern County oil and gas basins have been closely evaluated in other studies (Davis and others, 2018; McMahon and others, 2019, 2024; Wright and others, 2019; DiGiulio and others, 2021; Karolytė and others, 2021; Warden and others, 2024a). Although this study was not designed to provide the same level of in-depth analysis, there is evidence of old, crustal fluids mixing into the domestic resources in Kern County. There are seven samples in either undeveloped areas or withing a kilometer of an oil and gas basin that have a substantial source of non-atmospheric helium indicated by low R/Ra ratios and high terrigenic helium concentrations (fig. 22). Crustal fluids in groundwater wells near faults or oil and gas basins have been observed in previous studies in Kern County (Wright and others, 2019; Karolytė and others, 2021; McMahon and others, 2024). These fluids could be sourced from an oil and gas basin or shallow surface disposal of produced fluids, or groundwater with long residence times that have accumulated terrigenic helium (McMahon and others, 2021).

The geochemistry of these samples varies, and there are likely multiple processes that result in the mixing of old water into domestic wells. Samples S15-KERBS-02, S15-KERBS-33, and S15-KERBS-11 are mixed with modern-aged groundwater with high terrigenic helium concentrations and low R/Ra ratios (less than 0.5), but also high excess air and ΔNe over 1 (fig. 22). These results could suggest either mixing of deep brines into wells receiving MAR water or oil and gas fluids migrating from surface storage ponds. S15-KERBS-33 is within the West Bellevue Oil Field and near the Kern Water Bank and the Kern Fan Groundwater Storage Project in the Rosedale-Rio Bravo Water Storage District. Samples S15-KERBS-2 and S15-KERBS-11 are outside the Edison Oil Field and located along Caliente Creek in the southeastern part of the study area. Those samples also had moderate and high arsenic concentrations, respectively (fig. 15D). However, the methods used in this study cannot determine whether these fluids are derived from oil and gas production or storage of produced waters. The groundwater geochemistry is consistent with previous studies that have identified naturally occurring crustal fluids in drinking water wells in the Kern County subbasin (Wright and others, 2019; McMahon and others, 2021; Warden and others, 2024a).

Old, crustal fluids contain higher concentrations of arsenic, TDS, and methane in the domestic resources in the Kern County subbasin. Although arsenic and TDS can be associated with produced fluids from oil and gas development (McMahon and others, 2024), there is no indication that this is the source of high concentrations in domestic wells in the study area. The relations of arsenic and TDS concentrations and the residence times and geochemistry of domestic resources in the study area were discussed in the previous sections. Methane gas in groundwater is not a health concern, but it can be a useful tracer of groundwater migrated from oil and gas basins through either natural pathways or because of oil and gas production (Warden and others, 2024a). Methane concentrations in the domestic resources are low, but S15-KERBS-2, S15-KERBS-12, and S15-KERBS-17 have elevated concentrations (greater than 0.1 cubic centimeters per liter [cm³/L]) compared to the other samples, and ethane was also detected in S15-KERBS-2 and S15-KERBS-17. Methane produced during thermal maturation of hydrocarbons will co-occur with heavier hydrocarbons such as ethane and will have a heavier carbon isotope ratio (δ¹³C-CH₄>−50 per mil; Warden and others, 2024a). The stable isotopes of carbon in methane in the domestic resources do not conclusively show that the elevated hydrocarbons are derived from deep thermal sources, but the ratios in S15-KERBS-2 (δ¹³C-CH₄=−57 per mil and C₁/C₂+=104) and S15-KERBS-12 (δ¹³C-CH₄=−51 per mil) fall in a range that indicates there could be mixing between thermogenic gases and microbial methane (Warden and others, 2024a). Methane can migrate naturally from a thermogenic source through subsurface pathways or during the production and disposal of hydrocarbon fluids. However, the concentrations are low compared to other studies of gases released from production wells, and there is no other geochemical evidence suggesting they have migrated from the oil and gas development process (Warden and others, 2024a).