Groundwater Quality Defined as Relative Concentrations

Groundwater quality was categorized using relative concentrations, which are defined as the ratio of a constituent’s measured concentration in groundwater to the concentration of a regulatory or non-regulatory water-quality benchmark used to evaluate drinking-water quality. Relative concentrations can only be computed for constituents with water-quality benchmarks; therefore, water-quality constituents without benchmarks were not included in the status assessment. Using relative concentrations allows evaluation and inter-comparison of a wide array of organic and inorganic constituents at concentrations that often range several orders of magnitude (Toccalino and others, 2004; Rowe and others, 2007; Toccalino and Hopple, 2010).

Regulatory and non-regulatory benchmarks typically are used to evaluate treated drinking water distributed by public-supply systems. The use of water-quality benchmarks developed to meet the health- and aesthetic-based standards for public-supply sources provides context to evaluate domestic-supply sources for the purposes of this study. The water-quality constituents measured for this study were compared to benchmarks established by the U.S. Environmental Protection Agency (EPA), the California State Water Resources Control Board Division of Drinking Water (SWRCB-DDW), and the USGS. The benchmarks used for each constituent in this study were selected in the following order of priority:

Relative concentrations were classified as low, moderate, or high categories for calculation of aquifer-scale proportions. Relative concentration values greater than 1.0 (constituent concentration or value greater than a benchmark) were defined as “high” for all constituents. For inorganic constituents (trace elements, nutrients, radiological constituents, and inorganic constituents with SMCL benchmarks), relative concentration values greater than 0.5 and less than or equal to 1.0 (constituent concentration or value is greater than one-half of the benchmark but less than the benchmark) were defined as “moderate” and relative concentration values less than or equal to 0.5 (constituent concentration or value is less than one-half the benchmark) were defined as “low.” For organic and special-interest constituents, relative concentration values greater than 0.1 and less than or equal to 1.0 were defined as “moderate,” and relative concentration values less than or equal to 0.1 were defined as “low.” Low relative concentrations of inorganic, organic, and special-interest constituents included non-detections and values less than moderate concentrations.

The “special-interest” class of constituents in GAMA-PBP studies has historically included constituents that the State of California was actively considering for an MCL-CA at the beginning of GAMA-PBP in 2003 (Belitz and others, 2003). Perchlorate is a trace inorganic compound and received an MCL-CA in 2007 (California State Water Resources Control Board, 2017). However, perchlorate is still classified as a constituent of special interest for the purposes of this study and is evaluated in a manner similar to the organic constituents for consistency with previous reports.

The SWRCB-DDW notification level (NL-CA) is a non-regulatory, health-based advisory level that is associated with the RL-CA and functions as an early warning indicator for certain contaminants without regulatory benchmarks (California State Water Resources Control Board Division of Drinking Water, 2022c). The NL-CA can range from 6 to 100 times less than the RL-CA and has a similar function to that of the low-to-moderate boundary in the relative concentration classification system described previously; therefore, if a constituent has an NL-CA, then the value of the NL-CA is used as the low-to-moderate threshold for the relative concentration classification instead of the benchmark multiplied by 0.1 (for organic constituents) or 0.5 (for inorganic constituents).

In this study, if the measured constituent concentration was greater than the NL-CA and less than or equal to a corresponding non-regulatory, health-based primary benchmark value (the RL-CA or HAL-US, whichever is lower), the constituent was considered present at a moderate relative concentration. Boron and vanadium were the only detected constituents with NL-CA values for which relative concentration thresholds were affected by this modification to the relative concentration classification system; for example, although the primary benchmark for boron is the HAL-US of 5,000 µg/L, the low-to-moderate concentration boundary is the NL-CA of 1,000 µg/L and not 2,500 µg/L (one-half the HAL-US).

Data Collected for Domestic-Supply Assessment

This report section describes methods used to select wells for the MOBS domestic-supply assessment, the scope of the water-quality data collected for the study, quality-control methods used for the water-quality results, and methods used for evaluation of groundwater-age tracer results. The selection of constituents for discussion in the status assessment is also described.

Age-Dating and Geochemical Tracers

The radiological isotopes of hydrogen (tritium) and carbon (carbon-14) can be used to determine the age (time since recharge or isolation from the atmosphere) of groundwater, help locate sources of recharge, and help identify geologic controls on the movement of groundwater (Clark and Fritz, 1997). Tritium and carbon-14 are naturally occurring but also were produced in large quantities to the atmosphere beginning in 1952 because of the atmospheric testing of nuclear weapons. In this report, tritium is used to identify presence of modern groundwater, water recharged since approximately 1952, and carbon-14 is used to estimate the age of pre-modern groundwater.

Use of tritium as a tracer of young groundwater is complicated because the Mojave River did not flow during much of the 1950s and 1960s (see hydrograph for USGS streamgage 10262500; U.S. Geological Survey, 2022), which indicates that water with the large tritium excess from nuclear testing is largely missing from the aquifer (Izbicki and Michel, 2004; Warden and others, 2023). Nevertheless, tritium does provide some indicator of the age of groundwater pumped by domestic and public-supply wells. Lindsey and others (2019) used tritium alone to classify groundwater samples into three age categories: (1) pre-modern, (2) mixed, and (3) modern. Following the method of Lindsey and others (2019), samples with tritium activities less than 0.13 tritium units (TU) for samples collected in 2018 and less than 0.23 TU for samples collected in 2008 were classified as pre-modern, samples with tritium activities greater than 0.86 TU for samples collected in 2018 and greater than 1.52 TU for samples collected in 2008 are classified as modern, and all other samples are classified as mixed. However, samples classified as pre-modern may contain a small fraction of modern water, and samples classified as modern may contain a small fraction of pre-modern water.

Carbon-14 activities are used to determine the age of groundwater ranging from recent to 30,000 years before present (Clark and Fritz, 1997). For this study, carbon-14 activities were not corrected for geochemical reactions or mixing that may have removed or added carbon as groundwater moved from the point of recharge to the aquifer locations where the water was sampled. Uncorrected carbon-14 ages are useful for qualitative comparisons between samples (Wright and others, 2015). Izbicki and Michel (2004) found that corrected carbon-14 ages for samples from the Mojave River groundwater basin were 0 to 2,500 years younger than the uncorrected carbon-14 ages. Uncorrected carbon-14 ages were estimated using the standard equation for radioactive decay (Clark and Fritz, 1997).$$An initial carbon-14 activity ($A_0$) value of 84 percent modern carbon (pmC) was assumed following Izbicki and Michel (2004). Samples with carbon-14 activities greater than 84 pmC were assumed to be dominated by modern recharge, and no attempt was made to estimate carbon-14 age. Unadjusted carbon-14 ages in this report (table 5) are used for relative comparison of groundwater ages between wells.

Table 5.    

Selected water-quality data, geochemical and age-dating tracer results, septic tank densities, and inferred aquifer sources for samples from grid and understanding wells, Mojave Basin Domestic Aquifer study unit, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project, January–May 2018

[Water-quality data and measured tracer data available from Groover and others (2019) and U.S. Geological Survey (2022). Aquifer sources: Classification into floodplain or regional aquifer source is based on well location (fig. 5) and water stable isotope data. Regional aquifer source is separated into north and south at approximately the latitude of El Mirage (dry) Lake (fig. 5). Wells located in the regional study area, but with stable isotope compositions consistent with the floodplain aquifer are classified as floodplain, inferred. Wells in Lower Mojave River Valley (fig. 5) with aquifer source of floodplain or floodplain, inferred are at the end of the floodplain in the study unit. Understanding wells considered to be affected by nitrate source endmembers are labeled with those sources: wastewater treatment plant (wwtp) discharge, septic tank (septic) discharge, recharge beneath field with dairy waste application (dairy waste), or agricultural irrigation (ag) recharge. Water-type classification based on Piper (1944) and Landon and others (2010), reported as "major cations : major anions": Ca, calcium; Na, sodium; Mg, magnesium; HCO₃, bicarbonate; SO₄, sulfate; Cl, chloride; mixed, either no dominant cation or no dominant anion. Other abbreviations: >, less than; TU, tritium units; pmC, percent modern Carbon, mg/L, milligrams per liter; tanks/km², density of septic tanks in tanks per square kilometer within a 500-meter radius of the site; na, not available]

Table 5.    Selected water-quality data, geochemical and age-dating tracer results, septic tank densities, and inferred aquifer sources for samples from grid and understanding wells, Mojave Basin Domestic Aquifer study unit, California Groundwater Ambient Monitoring and Assessment Program Priority Basin Project, January–May 2018

Local identification number	Delta oxygen-18 of water, in per mil	Delta deuterium of water, in per mil	Aquifer source	Strontium-87/ Strontium-86 atom ratio	1Tritium, in TU	2Uncorrected Carbon-14, as pmC	3Uncorrected carbon-14 age, in years before present	Total dissolved solids, in mg/L	Water-type classification	pH, in standard units	Dissolved oxygen, in mg/L	Nitrate, in mg/L as nitrogen	4Septic density, in tanks/km2
Floodplain study-area grid wells
F01	−8.76	−62.3	Floodplain	0.71125	1.12	108	Modern	187	Ca-Na : HCO3	7.16	6.05	0.36	3.1
F02	−7.96	−57.9	Floodplain	0.71103	0.71	121	Modern	1,354	Ca-Na : mixed	7.00	0.35	0.51	1.1
F03	−9.17	−63.3	Floodplain	0.70957	<0.11	78	650	320	Na : HCO3	7.81	4.06	0.19	1.1
F04	−8.36	−59.1	Floodplain	0.71072	0.39	98	Modern	563	Ca-Na : mixed	7.58	7.24	6.05	5.1
F05	−9.47	−69.2	Floodplain	0.71089	1.60	107	Modern	272	Ca-Na : mixed	7.35	8.32	1.48	5.7
F06	−8.98	−65.3	Floodplain	0.71071	0.73	87	Modern	726	Ca-Na : mixed	7.45	8.04	3.49	20
F07	−9.27	−63.3	Floodplain	0.71315	1.16	106	Modern	151	mixed : HCO3	7.54	8.53	1.64	20
F08	−8.62	−59.9	Floodplain (end)	0.71036	<0.11	88	Modern	269	Ca-Na : HCO3	7.94	4.64	0.33	3.8
F09	−8.47	−59.5	Floodplain (end)	0.71082	<0.11	86	Modern	286	Ca-Na : HCO3	7.85	4.78	0.62	3.8
F10	−8.49	−60.2	Floodplain (end)	0.71072	<0.11	84	Modern	235	Ca-Na : HCO3	7.91	4.94	0.33	3.8
F11	−8.55	−60.2	Floodplain (end)	0.71082	<0.11	92	Modern	272	Ca-Na : HCO3	7.87	6.27	0.88	3.8
F12	−8.73	−60.6	Floodplain (end)	0.71033	0.71	87	Modern	313	Ca-Na : HCO3	7.52	4.95	1.27	3.8
F13	−7.32	−55.2	Floodplain (end)	0.71037	0.12	68	1,800	388	Ca-Na : mixed	7.76	7.80	0.47	0.3
F14	−5.06	−47.4	Floodplain (end)	0.71038	0.26	81	320	610	Ca-Na : HCO3-SO4	7.24	2.50	0.11	0.3
F15	−8.54	−61.4	Floodplain (end)	0.70917	<0.11	41	6,000	375	Na : mixed	7.90	0.20	<0.04	0.3
Regional study-area grid wells
R01	−10.68	−79.7	Regional (south)	0.71588	<0.11	26	9,600	281	Na : HCO3	8.02	2.14	0.64	15
R02	−11.57	−82.5	Regional (south)	0.71272	<0.11	22	11,000	315	mixed : SO4	7.81	2.83	0.57	5.5
R03	−11.73	−83.4	Regional (south)	0.71240	<0.11	42	5,700	331	mixed : SO4	7.76	1.50	0.42	1.4
R04	−12.10	−86.8	Regional (south)	0.71112	<0.11	19	12,400	310	Na : HCO3-SO4	8.88	3.48	0.62	1.4
R06	−10.74	−85.0	Regional (north)	na	<0.11	11	16,700	2151	Ca : SO4	7.09	0.77	<0.04	0.3
R08	−12.47	−93.0	Regional (north)	0.70950	<0.11	9	18,400	353	Na : HCO3	7.82	5.16	0.55	0.6
R09	−10.59	−78.9	Regional (south)	0.71111	0.15	87	modern	502	Na : mixed	7.79	8.62	3.05	0.6
R10	−12.40	−90.2	Regional (south)	0.71117	<0.11	2	31,600	215	Na : HCO3-SO4	8.88	0.16	0.40	1.4
R11	−11.47	−82.9	Regional (south)	0.71103	<0.11	24	10,300	302	Ca-Na : SO4	7.94	0.46	1.24	7.7
R12	−9.13	−64.4	Regional (south)	0.71124	1.45	100	modern	284	Ca : HCO3	6.87	7.50	11.06	4.9
R13	−9.06	−64.9	Regional (south)	0.71277	2.03	86	modern	135	Ca-Na : HCO3	7.17	8.44	8.65	3.1
R14	−9.20	−63.1	Regional (south)	0.71096	<0.11	62	2,500	140	Na : HCO3	8.33	6.42	2.41	128
R15	−9.74	−68.3	Floodplain, inferred	0.70933	<0.11	67	1,800	426	Na : mixed	7.98	1.52	0.26	1.1
R16	−8.65	−66.9	Floodplain, inferred	0.70858	0.47	92	modern	2285	Ca-Na : SO4-Cl	7.04	5.29	6.85	0.2
R17	−12.40	−98.6	Regional (north)	0.70839	<0.11	7	21,000	834	Na : mixed	7.62	3.53	1.33	0.0
R18	−9.40	−77.0	Regional (north)	0.70810	<0.11	43	5,500	1522	Na : SO4-Cl	7.84	4.93	0.08	0.0
R19	−8.59	−61.3	Floodplain, inferred	0.71046	<0.11	87	modern	313	Ca-Na : mixed	7.75	6.07	0.42	5.7
R20	−11.73	−93.1	Regional (north)	0.71046	0.35	9	18,300	535	Na : HCO3-SO4	7.70	2.10	0.24	0.4
R21	−11.52	−91.2	Regional (north)	0.71054	<0.11	17	13,000	554	Ca-Na : mixed	7.23	0.68	1.37	1.8
R22	−12.40	−93.7	Regional (south)	0.71133	0.18	18	12,800	356	Na : HCO3-SO4	8.18	2.00	1.34	66
R23	−9.39	−68.7	Floodplain, inferred	0.71144	1.17	115	modern	200	Ca-Na : mixed	6.94	7.79	0.66	2.4
R24	−11.87	−86.3	Regional (south)	0.71145	<0.11	48	4,700	285	mixed : HCO3-SO4	7.37	5.97	1.24	7.6
R25	−11.55	−84.0	Regional (south)	0.71042	0.29	34	7,600	1131	Ca-Na : Cl	7.61	4.56	0.95	7.0
R26	−12.42	−93.2	Regional (south)	0.71020	<0.11	30	8,400	341	Ca-Na : HCO3-SO4	7.76	5.85	1.37	1.8
R27	−11.17	−89.0	Regional (north)	0.71059	<0.11	71	1,400	821	Ca-Na : mixed	7.34	7.75	20.69	0.4
R28	−11.55	−93.0	Regional (north)	0.70934	<0.11	1	37,300	948	Na : mixed	8.59	0.46	0.34	8.8
R29	−8.71	−62.1	Floodplain, inferred	0.70991	<0.11	61	2,600	530	Na : mixed	8.12	5.35	3.92	3.9
R30	−8.40	−61.9	Floodplain (end), inferred	0.70988	0.15	40	6,200	407	Ca-Na : mixed	7.91	1.53	2.75	0.3
R31	−9.36	−67.8	Floodplain (end), inferred	0.70917	<0.11	5	23,700	558	Na : HCO3-SO4	9.10	1.44	<0.04	3.8
R32	−4.62	−46.1	Floodplain (end), inferred	0.71037	0.40	78	590	681	Ca-Na : mixed	7.49	2.25	0.09	0.3
R33	−8.64	−63.0	Floodplain (end), inferred	0.70922	<0.11	27	9,200	314	Na : mixed	8.04	0.42	<0.04	0.3
R34	−8.87	−64.9	Floodplain (end), inferred	0.70921	<0.11	39	6,300	392	Na : mixed	8.11	5.87	0.28	0.3
R35	−12.29	−87.8	Regional (south)	0.71147	<0.11	50	4,200	213	Ca-Mg : HCO3	7.83	8.63	1.12	7.6
Floodplain study-area understanding wells
U01	−8.22	−63.5	Floodplain (wwtp affected)	0.71054	na	93	modern	1,475	Ca-Na : SO4-Cl	7.21	0.59	12.40	5.2
U02	−8.89	−63.6	Floodplain (ag field affected)	0.71089	na	105	modern	697	Ca-Na : mixed	7.00	3.38	1.98	14
U04	−9.49	−67.1	Floodplain	0.71135	na	105	modern	178	Ca-Na : HCO3-Cl	6.89	9.79	0.95	2.4
MOJO-34	−8.66	−62.3	Floodplain	na	1.47	117	modern	160	na	6.83	7.79	na	na
Regional study-area understanding wells
U03	−9.92	−76.3	Regional (dairy waste affected)	0.71207	na	103	modern	4,909	mixed : SO4	6.75	0.32	98.14	0.6
U05	−12.62	−92.9	Regional (south)	0.70940	<0.11	10	17,500	249	Na : HCO3	8.54	5.27	1.01	0.6
U06	−11.98	−88.0	Regional (south)	0.71010	<0.11	14	14,900	340	Na : HCO3	8.07	5.52	0.83	0.6
U07	−12.89	−94.7	Regional (south)	0.70843	0.23	2	30,700	241	Na : HCO3	9.43	0.32	<0.04	0.6
U08	−10.85	−78.8	Regional (septic affected)	0.71095	na	52	4,000	514	Ca-Na : mixed	7.75	5.11	1.42	108
U09	−10.46	−76.6	Regional (septic affected)	0.71098	na	74	1,000	507	Ca-Na : mixed	7.67	6.43	4.57	108
U10	−10.55	−77.0	Regional (septic affected)	0.71078	na	68	1,700	435	Ca-Na : mixed	7.68	7.17	3.33	108
U11	−11.90	−86.3	Regional (south)	na	<0.11	na	na	1,480	Ca-Na : SO4-Cl	7.37	6.21	na	0.6

¹

Tritium activities reported with an "R" remark code in Groover and others (2019) are reported as nondetections less than the maximum value reported with an "R" code. Tritium activities in pCi/L in Groover and others (2019) have been converted to tritium activities in TU.

²

Carbon-14 activities reported in Groover and others (2019) in percent modern have been converted to Carbon-14 activities in percent modern Carbon (Plummer and others, 2004).

³

Uncorrected carbon-14 ages were computed using the standard equation for radioactive decay (Clark and Fritz, 1997), assuming an initial carbon-14 activity of 84 pmC (Izbicki and Michel, 2004). Ages are not computed for samples with uncorrected carbon-14 activities greater than 84 pmC; these samples are defined as "Modern" age.

⁴

Mean septic tank density within a 500-meter radius of the well computed from U.S. Census Bureau (1992).

Isotope tracers used in this report included the stable isotopic ratios of oxygen (δ¹⁸O-H₂O) and hydrogen (deuterium, or δD) in water, and the isotopic ratio of strontium-87 to strontium-86 (⁸⁷Sr/⁸⁶Sr) in dissolved strontium. Stable isotope ratios in water are report in Groover and others (2019), and strontium isotope date are available in NWIS (U.S. Geological Survey, 2022); both are also listed in table 5 for convenience.

The major-ion compositions of MOBS samples were categorized into different water types on a Piper diagram (Piper, 1944; Landon and others, 2010; table 5). The Piper diagram was plotted using the Piper plot macro in SigmaPlot 14.5 (Systat Software, Inc., 2022).

Calculation of Aquifer-Scale Proportions for the Domestic Supply Assessment

A grid-based statistical approach was used to calculate the areal proportion of the floodplain and regional aquifer study areas in MOBS having high, moderate, and low relative concentrations for selected water-quality constituents or classes of constituents (“aquifer-scale proportions”; Belitz and others, 2010). Non-detections were included in the low relative concentration class. Aquifer-scale proportions were calculated for individual water-quality constituents, and for classes of constituents. The computations for classes of constituents used the highest relative concentration for any constituent in the class to represent the class. Aquifer-scale proportions for high relative concentrations were calculated as the percentage of grid wells in a study area having high relative concentrations for a given constituent (equation [eq.] 1):

Aquifer-scale proportions for moderate relative concentrations, detections of organic constituents at any concentration, and presence of microbial indicators were calculated similarly by replacing terms using the superscript “high” in eq. 1 with terms using the superscripts “moderate,” “detection,” or “present,” respectively.

Compilation of Data and Calculation of Aquifer-Scale Proportions for Public-Supply Wells

The results from the GAMA-PBP public-supply study unit in the Mojave region in 2008 (MOJO; Dawson and Belitz, 2012) could not be compared directly to the results from this study; therefore, groundwater quality in public-supply wells was reassessed for this study. The 2008 study did not include inorganic, isotopic, and age-dating constituents for all sampled wells, and it was designed with one grid-cell network (Mathany and Belitz, 2009; Dawson and Belitz, 2012) rather than the two grid-cell networks used in this study (floodplain and regional study areas; fig. 5). For MOJO grid cells in which the well sampled by the GAMA-PBP was not analyzed for inorganic constituents, Dawson and Belitz (2012) randomly selected one public-supply well with data in the SWRCB-DDW database for samples collected during 2005–08 to represent each grid cell. They used the resulting dataset of one well per cell to calculate aquifer-scale proportions for the groundwater resource used for public supply with moderate and high relative constituent concentrations and constituent classes.

For this study, water quality in groundwater resources used for public supply in the floodplain and regional study areas was reassessed using the same grid-cell networks as used for the MOBS study and using a larger dataset to represent the public-supply wells. Water-quality data for public-supply wells were compiled from the following sources: (1) wells in the SWRCB-DDW database located in MOBS grid cells and having data for samples collected between 2008 and 2018 (California State Water Resources Control Board, 2019b); (2) GAMA-PBP data collected between 2008 and 2018; and (3) all public-supply wells sampled by USGS projects other than the GAMA-PBP between 2008 and 2018 (U.S. Geological Survey, 2022). The GAMA-PBP data included samples collected during the original assessment in 2008 for MOJO grid wells and MOJO understanding wells MOJOU-01 and MOJOU-06 (Mathany and Belitz, 2009), samples collected from a subset of MOJO grid wells in 2011 and 2018 for monitoring of groundwater-quality trends (Jurgens and others, 2018; U.S. Geological Survey, 2022), and samples collected from well LUBU-03 (Wright and others, 2015). Reassessment with the larger dataset of public-supply wells was necessary because the MOJO grid wells were located in only 10 of 15 floodplain study area grid cells and 18 of 33 regional study area grid cells, and this coverage was considered insufficient for representation of the entire study area. Expanding the dataset to include wells sampled by other USGS projects and wells with data in the SWRCB-DDW dataset increased the coverage to up to 13 of 15 floodplain study area grid cells and up to 24 of 33 regional study area grid cells.

Datasets were checked for duplicate wells by comparing USGS site identifiers between the NWIS (U.S. Geological Survey, 2022) and GAMA-PBP (Jurgens and others, 2018) databases and by comparing public-supply well codes compiled in the USGS datasets (U.S. Geological Survey, 2022; Jurgens and others, 2018; field not publicly available) to those in the SWRCB-DDW dataset (California State Water Resources Control Board, 2019b). Public-supply wells were assigned to MOBS grid cells based on the latitudes and longitudes provided with the data; no additional verification of locations was done for this study. If wells had location information from USGS and SWRCB-DDW, the USGS location was used.

Many of the 322 public-supply wells located in the MOBS grid cells were sampled more than once for some constituents during 2008–18. For each constituent at each well, the data from the sample collected closest to the midpoint of the sampling period for the corresponding domestic-supply assessment (that is, the midpoint of sampling from January–May 2018 for the MOBS study) was selected. This method prioritized selection of samples measured as close in time as possible to the domestic-supply assessment, without excluding wells for which only considerably older or more recent data were available.

Some adjustments to inorganic constituent parameters from the SWRCB-DDW dataset (California State Water Resources Control Board, 2019b) were done to compare the data with those collected by the USGS (U.S. Geological Survey, 2022; Jurgens and others, 2018). These adjustments included the following: converting nitrate concentrations to nitrate as nitrogen concentrations, calculating adjusted gross alpha-particle activity (following procedures described in Groover and others, 2019), converting specific conductance values to TDS concentrations for wells without TDS data, and reviewing the data for inconsistencies in reported units and removing results with uncertain reporting units. For all inorganic constituents, except the special-interest constituent perchlorate, the reporting levels used in the SWRCB-DDW dataset had concentrations below the boundary between concentrations classified as moderate and low relative concentrations; thus, aquifer-scale proportions computed for the combined public-supply well dataset were directly comparable to those computed from the MOBS grid well dataset. The reporting level for perchlorate in the SWRCB-DDW dataset is 4 µg/L; therefore, the aquifer-scale proportions for moderate relative concentrations of perchlorate in public-supply wells are minimum values.

The reporting levels for organic constituents in the SWRCB-DDW dataset are generally several orders of magnitude higher than the reporting levels for the same constituents in the GAMA-PBP dataset (Landon and others, 2010). The combined SWRCB-DDW and GAMA-PBP dataset was re-censored to the most common SWRCB-DDW reporting level for each constituent. Only organic constituents analyzed for both the MOJO and MOBS GAMA-PBP studies, and for which data are commonly compiled in the SWRCB-DDW dataset, were included in the compilation. Because of the large number of VOCs analyzed but not detected, the compilation for VOCs was further limited to only those reported as detected in at least one sample in any of the data sources. The only pesticides detected in MOJO wells at concentrations greater than a SWRCB-DDW reporting level were metolachlor and dieldrin (Mathany and Belitz, 2009), but neither were analyzed during the MOBS study; therefore, metolachlor and dieldrin were not included in the combined public-supply well dataset. A total of 208 public-supply wells had data for either simazine or atrazine or both, and all results were nondetections relative to the SWRCB-DDW reporting levels Simazine and atrazine were reported as detected at low concentrations by Mathany and Belitz (2009), but the concentrations were below the SWRCB-DDW reporting levels.

The number of public-supply wells per cell varied from 1 to 32, and to reduce spatial bias caused by this uneven distribution of wells per cell, aquifer-scale proportions for the public-supply data were computed for the selected constituents and constituent classes using spatially weighted calculations to decluster the data (eq. 2; Isaaks and Srivastava, 1989; Belitz and others, 2010) in each study area. The spatially weighted method for calculating aquifer-scale proportions was used because the objective was to determine proportions on an areal basis for comparison with the results from the domestic well assessment. If spatial weighting were not used, cells with greater numbers of wells would contribute more than cells with fewer wells to the computed proportions.

This approach calculates the proportion of wells in each grid cell with high relative concentration groundwater for a given water-quality constituent and then averages these proportions across all cells with data in the study-area grid network. Spatially weighted aquifer-scale proportions for moderate relative concentrations were calculated similarly. Spatially weighted aquifer-scale proportions for low relative concentrations of constituents or constituent classes were not calculated for the public-supply dataset because of the differences in reporting levels between data obtained from USGS (U.S. Geological Survey, 2022) and SWRCB-DDW (California State Water Resources Control Board, 2019b).

Sources and Age of Groundwater used for Domestic Drinking Supplies

Previous studies in the area have shown the primary source of groundwater recharge to the floodplain aquifer is from infiltration of streamflow from the Mojave River that originated as winter precipitation near Cajon Pass (fig. 2; Lines, 1996; Stamos and others, 2001; Izbicki and Michel, 2004). This source of recharge is identifiable using stable isotopes of water (Izbicki, 2004). Air masses that contribute precipitation near Cajon Pass have not been uplifted over the higher mountains; therefore, precipitation condenses at lower altitude and warmer temperature than precipitation in the surrounding San Gabriel and San Bernardino Mountains, creating a distinctive isotopically heavy signature with δ¹⁸O-H₂O of approximately −10 to −8 per mil (fig. 6; Izbicki, 2004). Other sources of recharge to the floodplain aquifer are (1) discharge from regional wastewater treatment plants serving the Victorville and Barstow areas, (2) septic discharge and irrigation return in rural and agricultural area along the floodplain aquifer, and (3) managed recharge of imported water from the Governor Edmund G Brown California Aqueduct in spreading areas located on the floodplain (Lines, 1996; Stamos and others, 2001; Izbicki and Michel, 2004).

The regional aquifer surrounds and underlies the floodplain aquifer. The regional aquifer is primarily recharged by infiltration of intermittent streamflow from precipitation in the San Gabriel and San Bernardino Mountains on the south side of the study area and in the desert mountain ranges on the west, north, and east sides of the study area (Izbicki, 2004). Recharge to the regional aquifer is less than recharge to the floodplain aquifer during present-day (2020) climatic conditions. Recharge from intermittent streams other than the Mojave River is distinguishable from recharge from the Mojave River using water stable isotopes. In general, groundwater in the regional aquifer has lighter (more negative) stable isotopic compositions than groundwater in the floodplain aquifer because the runoff is from precipitation that condensed at higher elevations in the mountains than the elevation of Cajon Pass, or because the recharge is from precipitation in the past when the climate was considerably wetter and cooler than the current climate (Izbicki, 2004; Izbicki and Michel, 2004). Much of the groundwater in the regional aquifer corresponding to the MOBS regional study area is old (recharged more than 10,000 years before present; Kulongoski and others, 2003; Izbicki, 2004; Izbicki and Michel, 2004; Wright and others, 2015; fig. 6). Groundwater pumping in the regional aquifer exceeds groundwater recharge resulting in water level decline (Hardt, 1971).

Stable isotope data and location were used to identify whether wells were likely pumping groundwater from the floodplain or regional aquifer. This classification by geochemical characteristics may not be the same as the original assignment of the wells to the floodplain or regional study area. The boundary between the floodplain and regional study areas was based on the estimated extent of the floodplain aquifer from Stamos and others (2001) and may not correspond exactly to the hydrologic boundaries at the locations and depths of individual wells. All floodplain study area wells had δ¹⁸O-H₂O values greater than −10 per mil, consistent with pumping groundwater from the floodplain aquifer that receives recharge from the Mojave River (table 5; fig. 6; Izbicki, 2004). Five regional study area wells with δ¹⁸O-H₂O values greater than −10 per mil and locations less than 400 meters from the boundary between the floodplain and regional study areas (R15, R16, R19, R23, and R29) were assumed to be pumping groundwater from the floodplain aquifer. Five regional study area wells near the eastern end of the Lower Mojave groundwater basin that had δ¹⁸O-H₂O greater than −10 per mil (R30, R31, R32, R33, R34) were assumed to be pumping groundwater recharged to the regional aquifer from the floodplain aquifer (table 5; Hardt, 1971; Stamos and others, 2001; Izbicki, 2004). Three regional study area wells with δ¹⁸O-H₂O values greater than −10 per mil and locations near the San Gabriel Mountains and more than 6 kilometers west of the Mojave River (R12, R13, and R14) were assumed to be pumping groundwater from the regional aquifer, despite their stable isotope compositions, because of recharge from intermittent streams on the flanks of the San Bernardino and San Gabriel Mountains (Izbicki, 2007).

The stable isotopes of hydrogen and oxygen in water and groundwater ages from domestic wells sampled for MOBS (fig. 6) show patterns similar to those found in previous studies of groundwater in the Mojave groundwater basins (Izbicki, 2004; Izbicki and Michel, 2004; Metzger and others, 2015). Groundwater in the floodplain aquifer mostly has modern carbon-14 ages (blue circles in fig. 6). A few wells in the floodplain aquifer between Victorville and Barstow have uncorrected carbon-14 ages up to 2,600 years old (F03, R15, R29). Groundwater from the floodplain aquifer that has recharged into the regional aquifer in the Lower Mojave River Valley groundwater basin at the eastern end of the study unit ranges in age from modern to greater than 20,000 years old, demonstrating the long flow paths and low amounts of recharge in this area (Izbicki, 2004; Izbicki and Michel, 2004). Groundwater from three wells (F13, F14, R32) at the eastern end of the Lower Mojave River Valley groundwater basin have stable isotope compositions on an evaporation line extending from the global meteoric water line near the composition of unevaporated groundwater from Lower Mojave River Valley groundwater basin (Izbicki, 2004). Izbicki (2004) used the stable isotopes to also distinguish groundwater derived from wastewater discharge and from imported water, but the MOBS dataset has too few samples to distinguish contributions from those endmembers to the floodplain aquifer groundwater samples.

Groundwater in the regional aquifer has two distinct stable isotopic trends (fig. 6). Most samples have stable isotopic compositions on the global meteoric water line (Clark and Fritz, 1997) between δ¹⁸O-H₂O of −13 to −10 per mil, and a subset of samples form a parallel line offset to the right of the global line. Samples on the global meteoric water line are located in the southern part of the study unit and are recharged by runoff derived from precipitation on the San Gabriel and San Bernardino Mountains (Izbicki, 2004). Samples on the offset line are located farther away from the San Gabriel and San Bernardino Mountains and are likely recharged by runoff from the mountainous areas on the north, east, and west sides of the study unit (Izbicki, 2004). This offset line defines a local meteoric water line for precipitation formed under different climatic conditions that precipitation that lies along the global meteoric water line (Clark and Fritz, 1997; Izbicki, 2004). Most of the wells in the regional aquifer have groundwater with uncorrected carbon-14 ages of 5,000 to greater than 20,000 years. Most of the wells with younger groundwater are located close to the primary source of recharge, the San Gabriel and San Bernardino Mountains. Groundwater in the regional aquifer with δ¹⁸O-H₂O lighter than −11 per mil is generally more than 10,000 years old (fig. 6). These light isotopic values are lighter than most current precipitation in the mountains around the study unit and indicate recharge from precipitation in a climate regime that was wetter and colder than the current climate (Izbicki, 2004). Well R18 near Hinkley has stable isotopic composition between the Mojave River and the local meteoric water line and may indicate evaporation of ancestral Mojave River water rather than locally derived recharge (Izbicki and others, 2023b).

Inorganic and Special-Interest Constituents

Of the 42 inorganic and special-interest constituents analyzed by the GAMA-PBP, 38 were detected in samples from one or more MOBS grid wells (tables 3, 4; Groover and others, 2019). Of the 38 detected constituents, 12 constituents with health-based benchmarks and 5 constituents with secondary maximum contaminant level benchmarks were detected at moderate to high relative concentrations (table 3; fig. 7). The remaining 21 inorganic constituents detected in grid wells (Groover and others, 2019) were either present only at low relative concentrations or did not have established benchmarks for comparison (table 4).

High relative concentrations of inorganic constituents with health-based benchmarks were more prevalent in the regional study area than in the floodplain study area: 58 percent of the regional study area had high concentrations of one or more constituent compared to 13 percent of the floodplain study area (table 6). The inorganic constituents with MCL benchmarks detected at high relative concentrations in the regional study area were arsenic, chromium, fluoride, adjusted gross alpha, uranium, and nitrate; only arsenic was also detected at high relative concentrations in the floodplain study area (table 7). The inorganic constituents with non-regulatory health-based benchmarks detected at high concentrations in the regional study area were hexavalent chromium, molybdenum, and strontium; none were detected at high relative concentrations in the floodplain study area.

One or more inorganic constituents with SMCL benchmarks were present at high concentrations in 15 and 6.7 percent of the regional and floodplain study areas, respectively, and moderate concentrations of one or more SMCL constituents were present in 27 percent of both study areas (table 6). The inorganic constituents with SMCL benchmarks detected at high relative concentrations in the regional study area were TDS, chloride, sulfate, and iron; only TDS and sulfate were detected at high relative concentrations in the floodplain study area (table 7).

The study-area scale aquifer-scale proportion results presented in table 6 cannot be combined to yield the study-unit scale results presented in Groover and Goldrath (2019). The study-unit level results in Groover and Goldrath (2019) were computed as detection frequencies of high and moderate relative concentrations of constituents or classes of constituents in the MOBS grid well dataset, without regard to study area. However, detection frequency in the 48 grid wells is not equivalent to aquifer-scale proportion at the study unit level because the grid cells in the floodplain and regional study areas are not the same size. The aquifer-scale proportion for high relative concentrations of a constituent or class of constituents at the study-unit scale would be computed as the area-weighted combination of the aquifer-scale proportions at the study-area scale (see appendix B of Fram and Belitz, 2012). Study-unit scale aquifer-scale proportion results are not presented in this report.

Arsenic

Arsenic was the constituent most commonly detected at concentrations above benchmark concentration (table 6). Arsenic was detected at concentrations above the MCL-US of 10 µg/L in 36 percent of the regional study area and in 13 percent of the floodplain study area (table 6; fig. 8A). Concentrations in MOBS grid wells ranged from less the reporting limit of 0.05–172 µg/L (Groover and others, 2019). Most domestic wells with high or moderate relative concentrations of arsenic were distributed in the regional aquifer north of the approximate latitude of Victorville (R04, R08, R10, R17, R18, R20, R22, R28; fig. 8A). These wells generally have uncorrected carbon-14 ages greater than 10,000 years, pH values greater than 7.8, and oxic or suboxic oxidation-reduction (redox) conditions (table 5; Groover and others, 2019). Elevated arsenic concentrations in oxic, alkaline, old groundwater in the regional aquifer indicate increased solubility of arsenic in alkaline conditions resulting from silicate weathering along long groundwater flow paths (Welch and others, 1988, 2000; Izbicki and Michel, 2004; Bowell and others, 2014; Wright and others, 2015; Izbicki and others, 2023a). Some domestic wells with high or moderate relative concentrations of arsenic were located near the terminus of the floodplain aquifer east of Barstow (F13, F15, R30, R31, R33, R34; fig. 8A). These wells generally have uncorrected carbon-14 ages between 5,000 and 10,000 years, pH greater than 7.8, and alkaline suboxic or oxic conditions. The stable isotopic compositions of these groundwater samples are similar to the stable isotopic composition of modern groundwater in the floodplain aquifer (table 5; Izbicki, 2004). Low rates of groundwater recharge and depletion of the modern groundwater in the floodplain aquifer by pumping may result in domestic wells extracting groundwater that is older than normally considered representative of the floodplain aquifer.

A few wells with high or moderate relative concentrations of arsenic were located in or adjacent to the floodplain aquifer between Adelanto and Helendale (F03, R15, R16; fig. 8A). These wells generally had younger ages (modern to 1,800 years) compared to other wells with high or moderate arsenic concentrations (table 5). Izbicki and Michel (2004) and Izbicki and others (2023a) also identified greater prevalence of elevated arsenic concentrations in groundwater containing modern recharge in this segment of the floodplain aquifer and suggested that redox changes related to recharge of treated wastewater in this segment may increase arsenic solubility.

Total Dissolved Solids

The SMCL constituent most commonly present at moderate or high relative concentrations was TDS (table 7). TDS was detected at high and moderate relative concentrations in 12 and 27 percent of the regional study area, respectively. TDS was detected at high and moderate relative concentrations in 6.7 and 20 percent of the floodplain study area, respectively. All wells with high or moderate relative concentrations of chloride or sulfate also had high or moderate relative concentrations of TDS (Groover and others, 2019). Most grid wells with high or moderate relative concentrations of TDS were located north of the approximate latitude of Victorville (fig. 8I).

Unlike most other inorganic constituents commonly detected at moderate or high relative concentrations, TDS was present at high and moderate relative concentrations in wells with modern-age groundwater and in wells with groundwater having uncorrected carbon-14 ages greater than 10,000 years (high: F02, R06, R16, R18, R25; moderate: F04, F06, F14, R09, R17, R20, R21, R27, R28, R29, R31, R32; table 5). TDS concentrations in grid-well samples also were not related in a simple manner to the major ion composition. The major-ion composition of groundwater from MOBS grid wells is summarized with a Piper diagram (Piper, 1944, Hem, 1985; fig. 9). The MOBS wells display a wide range of major ion compositions similar to the range observed in Mojave public-supply wells (Wright and others, 2015). The anion composition of MOBS grid wells generally had increasing proportions of sulfate or sulfate and chloride relative to the proportion of bicarbonate as the concentration of TDS increased (fig. 9).

As groundwater travels through aquifer materials, cation exchange, mineral dissolution and precipitation, and oxidation-reduction reactions can cause changes in the dissolved ion composition of the groundwater. One common pattern of evolution of groundwater chemistry along long aquifer flow paths (travel times of thousands of years) is the cation composition evolving from a mixture of sodium, calcium, and magnesium toward the sodium apex of the Piper cation ternary (Hem, 1985; Drever, 1997). This pattern is observed in groundwater from Mojave and other desert basins (Wright and others, 2015), and in the MOBS dataset, older groundwater generally is characterized as a sodium-type groundwater (table 5). However, there was no systematic relation between TDS and the cation composition of MOBS grid wells (fig. 9). The lack of relation between groundwater evolution and TDS demonstrates the complexity of the geology in MOBS and indicates that multiple sources of salinity may contribute to groundwater used by domestic wells. The geologic composition of aquifers supplying water to domestic wells in the regional aquifer is highly varied; composition of this aquifer most likely consists of alluvium from surrounding rock outcrops and Miocene and older river and lake deposits (Cox and others, 2003; Stamos and others, 2003; Izbicki, 2008; Miller and others, 2020; Izbicki and others, 2023a). The regional aquifer is interrupted by multiple strike-slip and normal fault zones that juxtapose geologic units with different rock compositions and chemistry next to each other (fig. 3; Glazner and others, 2002; Stamos and others, 2003; Langenheim and others, 2019; Miller and others, 2020).

The importance of local geology as a source of TDS to groundwater is supported by strontium isotope data. The isotopic composition of dissolved strontium (⁸⁷Sr/⁸⁶Sr) can be used as a tracer of the sources of dissolved inorganic constituents in groundwater (Faure and Powell, 1972; Clark and Fritz, 1997; Bataille and Bowen, 2012). Rocks and geologic materials with higher ⁸⁷Sr/⁸⁶Sr ratios are referred to as “radiogenic” because of the accumulation of radiogenic ⁸⁷Sr, whereas rocks and geologic material with lower ⁸⁷Sr/⁸⁶Sr ratios, closer to mantle values, are referred to as “non-radiogenic.” Strontium is an exchangeable cation similar to calcium, and strontium weathered from rock and minerals may exchange rapidly with strontium in groundwater. Consequently, the ⁸⁷Sr/⁸⁶Sr composition of water from wells is indicative of the ⁸⁷Sr/⁸⁶Sr composition of rocks and geologic materials the water has interacted with and the rates of dissolution of strontium-bearing minerals in those materials (Faure and Powell, 1972; Nimz, 1995; Johnson and DePaolo, 1997; McNutt, 2000; Bataille and Bowen, 2012).

In general, groundwater from MOBS wells south of the approximate latitude of El Mirage (dry) Lake has more radiogenic ⁸⁷Sr/⁸⁶Sr ratios than groundwater from MOBS wells to the north (fig. 10). This pattern follows the isotopic composition of the sources of alluvium: the San Gabriel and San Bernardino Mountains contain rocks that generally have more radiogenic strontium isotopic compositions than most rocks in the desert mountain ranges in the northern half of the study unit (DePaolo, 1981; Glazner and O’Neil, 1989; Miller and others, 1996; Bataille and Bowen, 2012; Cecil and others, 2019; Stone and others, 2021). Strontium isotopic compositions in the wells in the floodplain aquifer are most radiogenic near the San Gabriel and San Bernardino Mountains, and are less radiogenic downstream, indicating dissolution of, or equilibration with, strontium-bearing minerals from the local alluvium as water moves along flowpaths of increasing distance from the San Gabriel and San Bernadino Mountains.

Organic and Microbial Indicator Constituents

Neither VOCs nor pesticides were detected at high or moderate relative concentrations in MOBS study unit grid wells (fig. 7). VOCs were detected at low relative concentrations in 21 percent of the regional study area and in 27 percent of the floodplain study area (table 6; fig. 11A). VOCs classified as trihalomethanes were the most commonly detected organic constituent class, followed by solvents (table 6). Nine different VOCs were detected in grid wells (tables 3, 4; fig. 7).

Trihalomethanes are formed by reactions between chemicals used for disinfection of drinking water and wastewater and organic matter present in water or solid materials (Zogorski and others, 2006). The most commonly detected trihalomethane was trichloromethane, and 3 of the 7 grid wells with trichloromethane detections also had detections of one to three brominated trihalomethanes (Groover and others, 2019). Trichloromethane was detected in 12 percent of the regional study area and 20 percent of the floodplain study area (table 7). Well R35 had detections at low relative concentrations of all four trihalomethanes (trichloromethane, bromodichloromethane, dibromochloromethane, and tribromomethane) and also had the highest ratio of bromide to chloride measured in MOBS samples (Groover and others, 2019). Disinfection of water containing bromide results in formation of brominated trihalomethanes (Ivahnenko and Barbash, 2004). Trihalomethanes may be detected in groundwater either because of recharge of treated water or wastewater from septic systems, leakage from treated water distribution systems, landscape use of treated water, or wastewater discharge (Ivahnenko and Barbash, 2004; Zogorski and others, 2006), or due to disinfection of wells. Shock chlorination (often carried out by pouring bleach down a well) is a recommended procedure for treating bacterial contamination and odor problems in domestic drinking-water supply wells (U.S. Centers for Disease Control and Prevention, 2006).

One or more solvents were detected at low relative concentrations in 15 percent of the regional study area and were not detected in the floodplain study area (table 6; fig. 11C). The most frequently detected solvent was tetrachloroethene (PCE; Groover and others, 2019). Well R09 had detections of four solvents at low relative concentrations and is in an area that may have been settled since the 1940s (based on historical aerial photographs, not shown); therefore, the solvents detected in the groundwater may be from legacy activities. Well R09 is about 5 km from the George Air Force Base (historical; fig. 11C) where there has been extensive remediation of groundwater contaminated by the solvent trichloroethene (Air Force Civil Engineer Center, 2014). However, water-level maps show that well R09 is upgradient from the remediation site (Dick and Kjos, 2017; U.S. Geological Survey, 2023), and the mapped contaminated plume does not extend in the direction of the well (Air Force Civil Engineer Center, 2014).

Five pesticides with water-quality benchmarks were detected at low concentrations in 20 percent of the floodplain study area and 9.1 percent of the regional study area (table 6, fig. 11D). Pesticides were not detected at moderate or high relative concentrations (fig. 7), and no individual pesticide constituents were detected in more than 10 percent of grid wells (table 4; Groover and others, 2019). All five detected pesticides were herbicides.

Grid wells in both study areas were sampled for the presence or absence of microbial indicators, including total coliform bacteria, Escherichia coli (E. coli), and Enterococci. Repeat sampling of wells required to evaluate microbial indicator results with benchmarks (California State Water Resources Control Board Division of Drinking Water, 2022a) was not done for this study. These microbial indicator constituents are used as indicators of potential contamination of water supplies from human or animal waste (California State Water Resources Control Board, 2019c). Coliform bacteria are ubiquitous in soils and surface water but are also present in the digestive tracts of warm-blooded animals. Enterococci and E.coli are more specific indicators of potential fecal contamination than are total coliform bacteria (California State Water Resources Control Board, 2019c). E. coli was not detected in the MOBS grid wells. Total coliform bacteria were detected in 27 percent of the floodplain study area grid wells and there were no Enterococci detections (table 7; fig. 11E). Total coliform bacteria and Enterococci were detected in 9 and 6 percent of grid wells in the regional study area, respectively (table 7; fig. 11E).

Well Depth and Groundwater Age

The depths of public-supply and domestic wells in the study unit were characterized using well-depth data compiled for the domestic wells sampled for MOBS (table 2) and 51 public-supply wells sampled for MOJO (Mathany and Belitz, 2009) or with well-depth data available from the USGS NWIS (U.S. Geological Survey, 2022). Public-supply wells in the floodplain study area, with a median well depth of 374 feet (ft) and a maximum well depth of 650 ft below land surface, generally are deeper than domestic wells sampled for MOBS, which have a median well depth of 210 ft below land surface and a maximum well depth of 332 ft below land surface (fig. 13). Like domestic wells in MOBS, public-supply wells are deeper in the regional study area than in the floodplain study area, reflecting the generally greater depths to water in the regional aquifer. Public-supply wells in the regional study area have a median well depth of 530 ft and a maximum well depth of 1,130 ft below land surface compared to domestic wells sampled for MOBS, which have a median well depth of 300 ft below land surface and a maximum well depth of 620 ft below land surface (fig. 13).

The difference in depths between public-supply and domestic wells in the floodplain and regional study areas indicates that there may be a difference in the age of the groundwater pumped by the two types of wells because groundwater age generally increases with depth in the aquifer system. Because many wells in MOBS (fig. 6) and MOJO (Wright and others, 2015) have uncorrected carbon-14 ages that are between 1,000 and 40,000 years, groundwater age comparisons would ideally be made using uncorrected carbon-14 ages. However, only half of the MOJO wells were sampled for carbon-14. The available carbon-14 data indicate minimal difference between domestic and public-supply wells. In the floodplain study area, both well types have median carbon-14 values of about 85 pmC and in the regional study area, both well types have median carbon-14 values of about 37 pmC (fig. 14).

All MOJO and MOBS wells were sampled for tritium (table 5; Mathany and Belitz, 2009). Following the method of Lindsey and others (2019), samples were classified as modern, mixed or premodern (see “Methods” section). Wells in the regional study area have a greater proportion of pre-modern ages than wells in floodplain study area, and there is minimal difference in age classifications between the public-supply and domestic wells (fig. 14).

Similar groundwater ages between domestic and public-supply wells were not expected given that domestic wells generally are shallower than public-supply wells (fig. 13). Because public-supply wells are deeper and have longer well-screen intervals than domestic wells sampled for MOBS, the public-supply wells may pump groundwater from a mixture of aquifer sediment types compared to shallower domestic wells. Older sediment at depth in the floodplain aquifer generally is finer grained (Huff and others, 2002; Cox and others, 2003; Stamos and others, 2003; Clark and others, 2009) than shallower alluvium; consequently, in wells screened throughout long intervals, shallower aquifer lithologic units may contribute proportionally more water to the well than deeper aquifer lithologic units. Thus, the age of groundwater extracted by public-supply wells in the floodplain aquifer may be more similar to the age of groundwater extracted by the shallower domestic wells.

The relation between well depth and groundwater age in the regional aquifer may have spatial patterns related to proximity to the floodplain aquifer. Domestic wells on the margins of the regional study area further from the floodplain aquifer may be pumping groundwater that is greater than 10,000 years old that is present at shallower levels in the aquifer system. In contrast, most of the public-supply wells that are in the regional aquifer are located in the more heavily used areas closer to the floodplain aquifer.

Water Quality

Heterogeneity in aquifer sediments (lithology) and pumping may result in differences in water quality between public-supply wells with long well screens and domestic wells with short well screens. Groundwater from finer-grained sediment horizons may have higher concentrations of some naturally occurring trace elements leaching into groundwater due to desorption from secondary mineral surfaces (Groover and Izbicki, 2018b) and weathering of primary mineral grains. Interbedded coarse-grained intervals and fine-grained intervals (fig. 15) may result in redox boundaries at interval contacts, resulting in higher concentrations of some trace elements dissolving into groundwater due to changes in water chemistry. Additionally, large-capacity pumps installed in public-supply wells are designed to pump large volumes of water and, as a result, may withdraw more water from coarse-grained intervals in heterogeneous aquifers compared to domestic wells. Domestic well placement and construction generally are less optimized for high pumping rates, and lower pumping rates may result in these wells obtaining a greater proportion of groundwater from finer-grained layers with different water chemistry to sustain water levels.

Stable isotopes of oxygen and hydrogen in water were available for public-supply wells sampled for the GAMA-PBP MOJO study unit (Mathany and Belitz, 2009; U.S. Geological Survey, 2022). The MOJO public-supply wells (not shown) have similar ranges in isotope values to those observed in the MOBS domestic wells (fig. 6) and in previous studies of the stable isotope systematics of groundwater in the Mojave groundwater basins (Izbicki, 2004). The primary difference is that the local meteoric water line to the right of the global meteoric water line is not as well-defined in the public-supply well dataset (not shown; Mathany and Belitz, 2009; U.S. Geological Survey, 2022). In the domestic well dataset, the local meteoric water line to the right of the global meteoric water line is formed by wells from the northern part of the study unit (fig. 6; table 5; Groover and others, 2019). Many of the parts of the regional aquifer on the northern side of the study unit did not have public-supply wells available to be sampled by the GAMA-PBP public-supply assessment study (fig. 12).

Comparison of inorganic data by constituent class and relative concentrations indicates that the large contrast between groundwater quality in the floodplain and regional study areas detected in MOBS domestic wells also is present in the public-supply wells (table 6; figs. 16A, B). There are two notable differences between the water quality in the domestic and public-supply wells. In the floodplain study area, the public-supply wells have greater prevalence of high or moderate relative concentrations of iron or manganese compared to the domestic wells. In the regional study area, the domestic wells have greater prevalence of high or moderate relative concentrations of trace elements.

The greater prevalence of high and moderate relative concentrations of iron and manganese in public-supply wells in the floodplain study area may be an artifact of sample collection protocols or may indicate a difference in the composition of the groundwater pumped by domestic and public-supply wells.

Most samples from public-supply and domestic wells in the floodplain aquifer have dissolved-oxygen concentrations greater than 0.5 mg/L (Mathany and Belitz, 2009; Groover and others, 2019), and iron and manganese have low solubilities in oxic conditions. Groundwater with dissolved-oxygen concentrations less than 0.5 mg/L was found in two locations: in some public-supply wells and some domestic wells in the Lower Mojave River Valley groundwater basin at the eastern end of the study unit, and in some public-supply wells between Victorville and Helendale (locations on fig. 2; data not shown; Mathany and Belitz, 2009; Wright and others, 2015; Groover and others, 2019; table 5).

All the public-supply wells sampled by USGS that had dissolved-oxygen concentrations less than 0.5 mg/L also had low relative concentrations of iron and manganese (Mathany and Belitz, 2009; U.S. Geological Survey, 2022); these anoxic conditions would favor dissolution of iron and manganese oxides present in the aquifer sediments. Many of the SWRCB-DDW public-supply wells with moderate or high relative concentrations of iron or manganese were located in the floodplain aquifer between Victorville and Helendale, where public-supply wells sampled by USGS had dissolved-oxygen concentrations less than 0.5 mg/L (locations on fig. 2; data not shown; Mathany and Belitz, 2009; Wright and others, 2015). However, dissolved-oxygen concentrations were not available in the SWRCB-DDW dataset (California State Water Resources Control Board, 2019b).

Trace element data from the USGS used in this report (U.S. Geological Survey, 2022) are from filtered water samples, whereas trace element concentration data compiled in the SWRCB-DDW dataset (California State Water Resources Control Board, 2019b) may be from filtered or unfiltered water samples. Comparison of data for samples from wells sampled by USGS and with data in the SWRCB-DDW dataset indicates iron and manganese concentrations commonly are greater in the SWRCB-DDW samples, which likely demonstrates the contribution of particulates to the total concentration (Levy and Fram, 2021). Of the 75 public-supply wells in the floodplain aquifer with manganese data in the SWRCB-DDW dataset, 27 percent had high or moderate relative concentrations of manganese (California State Water Resources Control Board, 2019b); of the 12 public-supply wells in the floodplain aquifer with manganese data in the USGS dataset, 8.3 percent had high or moderate relative concentrations of manganese (U.S. Geological Survey, 2022). Based on the available data, it is not possible to determine whether the difference in frequency of high and moderate relative concentrations of manganese was due to contributions from particulates to samples in the SWRCB-DDW dataset or due to the smaller number of wells sampled by the USGS not including zones of the public-supply aquifer represented in the larger set of wells in the SWRCB-DDW dataset.

If the greater prevalence of high and moderate relative concentrations of iron and manganese in the public-supply aquifer of the floodplain study area compared to the domestic-supply aquifer is not an artifact of differences in sample collection protocols between data from the USGS and SWRCB-DDW datasets, then we hypothesize that the difference in prevalence of high and moderate relative concentrations of iron and manganese could be related to differences in the depths of public-supply and domestic-supply wells. Public-supply wells in the floodplain aquifer are screened to greater depths than domestic wells (fig. 13), and consequently, may be screened in finer-grained alluvium that lies beneath the coarse sand near the surface in which many of the domestic wells are screened (fig. 4; Stamos and others, 2001; Huff and others, 2002). This finer-grained alluvium at depth has higher concentrations of iron and manganese oxides (fig. 15), and is more weathered (Groover and others, 2023), which may allow for greater mobility of trace elements in groundwater (Groover and Izbicki, 2019). Alternative explanations could include higher concentrations of organic carbon in aquifer sediments associated with ancestral lake and overbank deposits; however, data from Morrison and others (2018) indicated even seemingly reduced blue or black silt and clay in the floodplain aquifer has low concentrations of organic carbon (less than 1 percent).

The greater prevalence of moderate and high relative concentrations of trace elements in domestic wells compared to public-supply wells in the regional study area is mostly caused by greater prevalence of high and moderate relative concentrations of arsenic, fluoride, and molybdenum (table 7). This greater prevalence of high relative concentrations of trace elements in domestic wells may partially be due to well location; some of the grid cells represented by a domestic well with high relative concentration of arsenic, fluoride, or molybdenum did not contain any public-supply wells with data for those trace elements (compare fig. 5 to fig. 12). The greater prevalence of high relative concentrations of trace elements in domestic wells compared to public-supply wells may also partially be caused by differences in the characteristics of domestic and public-supply wells. Because domestic wells commonly pump at lower rates than public-supply wells, they can be screened in less productive aquifer materials, for example, finer-grained sediments that may have groundwater with higher concentrations of trace elements.

The greater prevalence of moderate and high concentrations of trace elements in domestic wells in the regional study area also could be due to regulation of public-supply wells, which may be taken offline if constituents exceed health-based benchmarks such as MCLs, whereas domestic well owners may be unaware of water-quality concerns in their well or unable to switch to a different source of water. MOBS wells in five of the eight regional study-area grid cells with no public-supply wells exceeded benchmarks for one or more inorganic constituents.

Organic constituents were not widely detected in either MOBS grid wells or in the public-supply dataset. There were no organic constituents detected at high relative concentrations in either dataset. No organic constituents were detected at moderate concentrations in the MOBS grid wells, and VOCs were detected at moderate concentrations in 2.0 and 0.5 percent of the public-supply regional and floodplain study areas, respectively (table 6; fig. 16). The organic constituents detected at moderate concentrations in the public-supply regional study area were the solvent PCE and the trihalomethane trichloromethane; only PCE was detected at moderate concentration in the public-supply floodplain study area. Detection frequencies at any concentration were not calculated for the public-supply dataset because of the difference between the reporting limits in the SWRCB-DDW and GAMA-PBP datasets would have required re-computing the MOBS results for comparison.