Hydrology

The Santa Clara River watershed has a Mediterranean climate, with warm dry summers and cool wet winters (Fleming and others, 2020). The Elizabeth Lake subwatershed covers 7.85 square miles (mi²), with the Elizabeth Lake drainage area accounting for 6.71 mi² of the area and an elevation that ranges from approximately 3,380 to 4,000 feet (ft) above sea level (ASL; North American Vertical Datum of 1988 [NAVD 88]; Tetra Tech, 2016; Los Angeles Regional Water Quality Control Board, 2016; U.S. Geological Survey, 2020). Annual precipitation is estimated to be 13.6 inches per year (in/yr) using north and south bounding National Oceanic and Atmospheric Administration (NOAA) stations (USC00042941 and USC00048014) located 3 and 6 mi away from Elizabeth Lake, respectively, with most of the precipitation falling between November and March (National Oceanic and Atmospheric Administration, 2020). The Elizabeth Lake subwatershed is considered a closed system with the exception of infrequent wet El Niño years, which deliver excessive amounts of precipitation causing intermittent channels to transport water from Elizabeth Lake northwest toward Munz Lakes and Hughes Lake and farther downstream toward the Castaic Lake (fig. 1) drinking water reservoir (Tetra Tech, 2016). Most residences receive imported water from northern California for consumption, and groundwater pumping is believed to be minimal and not taken into consideration for this study. Historically, Elizabeth Lake was previously used for recreational use and stocked with trout for fishing, and a picnic area is located on the western part of the lake.

Elizabeth Lake water levels are dependent on annual rainfall and groundwater discharging into the lake. Elizabeth Lake is primarily recharged from annual precipitation and stormwater runoff from the surrounding highlands during rainfall, with suspected minimal recharge of 38 acre-ft/yr from groundwater flow (Tetra Tech, 2016). Groundwater is assumed to recharge by infiltration of Elizabeth Lake surface water during dry years; runoff infiltration from uplands, including OWTS discharge; precipitation infiltration; and underflow from surrounding consolidated rocks, which are most likely fractured because of the proximity to the San Andreas Fault Zone. During years of severe drought with little to no recharge, the three lakes have been known to periodically dry up, which was the case during an initial 2018 field visit.

Groundwater levels can vary substantially in environments of extensive faulting (Haneberg, 1995). This variability can be observed within the Elizabeth Lake subwatershed based on groundwater levels recorded on well completion reports from the Department of Water Resources (DWR) Online System for Well Completion Reports (OSWCR) database (California Department of Water Resources, 2021). Historical water levels in wells measured by the USGS and well completion reports (WCRs) in the valley of the Elizabeth Lake subwatershed were about 8–16 feet below land surface (ft bls) during the late 1960s and early 1970s. Water levels in wells assumed to be screened within the unconfined aquifer (shallower than 200 ft bls and not containing a screened interval within hard rock) were documented on WCRs and ranged from 6 to 24 ft bls in lower parts and as much as 80 ft bls in the surrounding uplands of the Elizabeth Lake subwatershed. Groundwater flows from high to low hydraulic head in unconfined aquifers with no-to-minimal groundwater pumping and is typically assumed to follow topography; thus, we inferred that groundwater flows from the surrounding uplands toward Elizabeth Lake and discharges into Elizabeth Lake when water levels are high.

Geology

The San Andreas Fault Zone is an active transform fault distinguished throughout parts of California by troughs, ridges, sag ponds, and offset channels and extends from Mexico (not shown) to the Mendocino triple junction off the coast of northern California (not shown; Arrowsmith and Zielke, 2009). The study area lies along the San Andreas Fault Zone, which is shown on figure 4. Areas along the San Andreas Fault Zone, such as the Elizabeth Lake subwatershed, are heavy faulted and contain fault splays.

According to local drillers’ logs, depth to bedrock on the western boundary of the Elizabeth Lake subwatershed was measured at approximately 100 ft bls (California Department of Water Resources, 2021). Granite outcrops are present on the north side of Elizabeth Lake and bounding the northern part of residences. Lake sediment composed of Quaternary alluvium overlies this granite. The depth of the sediment is approximately 50 ft bls adjacent to the lakebed. On the other side of the lake, metamorphic rocks are present, and the alluvium thickness is approximately 70–150 ft.

According to a California Geological Survey (CGS) geologic map (Hernandez, 2011), the Elizabeth Lake basin primarily consists of Holocene lake deposits and modern alluvial fans sourced from mouths of stream canyons, older Pleistocene alluvium and alluvial fans, some outcrops of the Tertiary Anaverde Formation, shale and arkose sandstones, and artificial fill from human construction (from debris catchment basins, reservoirs, or road alignment). The area is bounded by Late Cretaceous granodiorite to the north and a quartzo-feldspathic and amphibolite gneiss (Early Cretaceous to Proterozoic) to the south. A simplified version of geologic units is shown on figure 4 (Hernandez, 2011).

Electrical Resistivity Tomography

To estimate depth to groundwater and bedrock, two ERT surveys were completed during a 3-day period, March 15–17, 2019, at the southern edge of Elizabeth Lake and on the overlying sediments (fig. 5). Resistivity measurements were taken by injecting current into the ground using two “transmitter” current electrodes and measuring the resulting voltage difference between two “receiver” electrodes (Minsley and others, 2010). Each electrode can work as either a transmitter or receiver electrode as the combination of electrode pairs used to make measurements is translated along the line (Groover and others, 2017). Electrical resistivity in ohm-meters (ohm-m) is an innate material property that is affected by (1) the amount and salinity of pore water; (2) the proportion of coarse or fine-grained material, such as clay; and (3) the fraction of metallic minerals, such as magnetite (Hubbard and Rubin, 2005; Minsley and others, 2010). The combination of these factors results in some ambiguity in interpreting data from techniques such as ERT; therefore, other data, such as borehole lithology, groundwater levels, or porewater salinity, are required to constrain processing and interpretation of ERT (Groover and others, 2017). ERT is relatively cost-efficient and can distinguish sharp contrasts in materials, such as basement to unconsolidated sediment, without detailed knowledge of the site parameters. This technique was used to determine the proper drilling method and to help plan locations for the installation of monitoring wells.

ERT data were acquired using an 8-channel SuperSting R8 resistivity/induced polarization meter (Advanced Geosciences Inc., 2011) with a maximum of 56 electrodes and a passive electrical cable system (Groover and others, 2021). Stainless steel electrode stakes, 18 inch (in.) long and 0.4-in. diameter, were installed to a minimum depth of 4 in. along the southwest side of Elizabeth Lake. Electrode and electrode stakes were positioned every 22.96 ft along Elizabeth Lake line 1 (A–A’), for a total length of 1,263 ft, and every 19.69 ft on Elizabeth Lake line 2 (B–B’) for a total length of 1,083 ft (fig. 5; Groover and others, 2021). Wider electrode spacings allow greater investigation at depth at the cost of poorer resolution of near-surface features (Binley and Kemna, 2005; Advanced Geosciences Inc., 20091; Minsley and others, 2010). In this study, electrode spacings were selected based on estimated depth to bedrock and field constraints on total survey line length. Line 1 (A–A’; fig. 5) was run from northeast to southwest paralleling the vegetated shoreline from a previous lake high stand and parallel to the approximate orientation of nearby housing developments. Line 2 (B–B'; fig. 5) was run from southwest to northeast, starting 100 ft from the road shown on figure 5 and extending toward the center of the lakebed. This line was slightly angled to avoid placing electrodes in deep standing water. Spatial data for all profiles were collected with the Trimble Geo7x differential global positioning system (GPS; Trimble, Inc., Sunnyvale, California), including latitude/longitude (referenced to World Geodetic System 1984 [WGS 84]) and projected into the North American Datum of 1983 (NAD 83) for the purposes of this report (fig. 5; Groover and others, 2021).

The ERT system used in this study allowed for eight measurements to be taken simultaneously along a single transect. Measurements were taken for a period of 1.2 seconds, equivalent to 0.83 Hertz, during which the polarity of the current electrodes was reversed (number of cycles set to 2) to minimize electrode polarization effects. The resistivity meter was powered by two 12-Volt marine batteries and injected as much as 2,000 milliamperes of current into the ground. Two separate array methods were used. The first method used the inverse Schlumberger array geometry noted for its measurement efficiency and good contrast between lateral and vertical resolution (Jamaluddin and Emi Prasetyawati Umar, 2018). The second method was dipole-dipole array geometry used to verify the inverse Schlumberger array. The dipole-dipole array has a poor signal-to-noise ratio but can resolve contrast among discrete horizontal features (Binley and Kemna, 2005).

ERT data were inverted using the robust, finite-element inversion method in Advanced Geosciences, Inc., EarthImager 2D software version 2.4.4, build 649 (Advanced Geosciences Inc., 2009). The method typically works well on datasets containing low-quality data and resolves resistivity boundaries well (Advanced Geosciences Inc., 2009). More technical information about processing techniques and parameters for ERT are available from Binley and Kemna (2005) and Minsley and others (2010). Spatial data were converted to “terrain” files and used in data processing to account for changes in elevation between electrodes. Inversion parameters were constrained by three assumptions: (1) sharp lateral contrasts in resistivity may be expected because of fault splays, (2) bedrock may be present within the depth resolution of the ERT surveys, and (3) surface moisture was likely highly saline. Geomorphic features of interest such as vegetation lineaments, coarse-grained flood channels, visible ponding of lake water, and differences in vegetation type were used to assess smoothing parameters for the inversions. Final inversions were selected to minimize the root-mean squared error (less than 5 percent) and least squared sum of errors (1.11 or less; Groover and others, 2017; Groover and others, 2021). Inverted data were compared to borehole geophysical logs in new monitoring wells to verify processing assumptions. The raw ERT data and inverted data are available in Groover and others (2021).

Monitoring Well Construction

In late 2019 and early 2020, the USGS Research Drilling Program drilled and constructed seven single-well monitoring sites around the perimeter of Elizabeth Lake as part of this study (fig. 3). These monitoring wells served two purposes: (1) characterizing the lithology and hydrogeology of the aquifer and (2) evaluating the quality of the groundwater. These seven sites were drilled using an 8-in. diameter auger drill rig and constructed of 2-in. threaded schedule 40 polyvinyl chloride (PVC) casing and screens. Wells ELLA-1 through ELLA-5 were installed on the southern perimeter, well ELLA-6 on the eastern perimeter, and well ELLA-8 on the northern perimeter of the lake. All wells were constructed with a slotted well screen in the bottom 20 ft of the well, except for well ELLA-8, which had a 10-ft screened interval. The annulus of the borehole in the screened interval was filled using a number 3 sand from the bottom of the borehole to approximately 10 ft above the screen. The sanitary seal was constructed using low permeability bentonite pellets and then sealed to land surface with a bentonite grout from the top of the sand pack to land surface. Additionally, three hand-augered shallow wells (ELLA-7, ELLA-9, and ELLA-10A) were installed using a 2-in diameter soil auger (table 1) and completed with 1-in. diameter PVC casing. The screened interval of the hand-augered wells was placed at the bottom of the PVC pipe and ranged in height from 0.2 to 2 ft long. The annulus of the hand-augered holes at these wells was backfilled with native soils. The purpose of these hand-augered wells was to provide additional monitoring locations for water-quality sampling and to assist in characterization of shallow (less than 10 ft) soils and groundwater surrounding the lake. Simplified versions of the well-construction diagrams for drilled and hand-augered sites are shown on the appendix figures 1.1–1.8. Subsamples of cutting materials are archived at the USGS California Water Science Center office in San Diego, California (not shown).

Borehole Geophysics

A suite of borehole geophysical logs was collected at the seven drilled wells to aid in the characterization of subsurface materials and to help conceptualize the subsurface geology. The geophysical logs were collected using logging tools that measured the electrical properties of the formation around the borehole (aquifer material) and the fluid in the formation. Electromagnetic (EM) induction logs (converted to resistivity) were collected to distinguish fine-grained clays and silt from coarser sand and gravel layers and to constrain ERT data. EM induction logs were completed using a Century model 9511 tool (Century Geophysical LLC, Tulsa, Oklahoma), in the upward direction at 5 ft per second. The EM logs were collected according to the manufacturer’s specifications (Century Geophysical LLC, Tulsa, Oklahoma) using a Century model 9511 tool (Century Geophysical LLC, Tulsa, Oklahoma). The tool uses an electromagnetic field to induce an electrical current in the surrounding formation (Williams and Johnson, 2004). The induced current sets up a secondary magnetic field that is measured, amplified, and then transmitted to the surface as a direct current. The magnitude of the direct current is proportional to the electrical conductivity of the formation, which is a function of lithology and pore-fluid conductivity (Keys, 1990). The volume of the material measured is a donut-shaped torus (Century Geophysical, LLC) with an inner diameter of 18 in. and an outer diameter of 50 in. Sand and gravel (or formations with low salinity porewater) tend to be more resistive than silts and clays (or formations with high salinity porewater), which tend to be more conductive (Keys, 1990). The natural-gamma logging tool measures the intensity of naturally occurring gamma-ray emissions, including material containing potassium-40, uranium-235, uranium-238, and thorium-232. Clays, volcanic material, and potassium-feldspar-rich gravel have higher intensity gamma-ray emissions (Schlumberger Limited, 1972; Hearst and Nelson, 1985; Driscoll, 1986). The natural-gamma sensors used in this study were present in the EM tool and the fluid tool; the data from the EM tool were used in all plots for simplicity. However, natural-gamma data were not used for interpretative purposes of this report.

Fluid temperature and the fluid resistivity measurements were collected with a Century Geophysical LLC model 9042 multiprobe. Data were collected in the downward direction at a consistent logging rate of 5 ft per minute to prevent any mixture of the fluid. Fluids with low salinity or low total dissolved solids (TDS) have a lower fluid resistivity response than more saline fluids (Keys, 1990).

Fluid resistivity and fluid temperature logs can be used to evaluate heterogeneity in fluid seepage rates in long-screened monitoring wells (Nawikas and others, 2016). Onsite wastewater treatment systems effluent tends to be more electrically conductive and less resistive than fresh water or rainwater and should provide a contrast with local groundwater depending on local salinity. EM induction, natural gamma, fluid resistivity, and fluid temperature data are available in the USGS Geolog Locator (U.S. Geological Survey, 2022) and may be retrieved using the USGS station number or station name in table 1.

Lithology, Hydraulic Conductivity Estimation, and Conceptual Model

Soil cuttings were collected from the auger drill rig at approximately 5-ft intervals or when a soil change was observed during installation of the seven monitoring wells. Care was taken to time the collection of the cutting sampling to the drilling rate and cutting travel time. Some discrepancies in the representativeness of drill cuttings may have occurred because of clay smear on the auger stem entrapping or contributing excess sediment or errors in the timing estimates of cuttings to travel up the auger stem. The cutting lithologies were initially characterized in the field by visually assessing grain-size distributions using a hand lens and characterizing colors using a soil color chart. The drill cuttings were later described in more detail using a binocular microscope following the Folk (1954) classification system and Wentworth grain size scale, as described by Kjos and others (2014). Grain rounding and sorting also were verified and described using a binocular microscope. Lithologies at all sites have been archived in the USGS GeoLog Locator (U.S. Geological Survey, 2022).

Hydraulic conductivity (K) and porosity (φ) were estimated for each site using lithologic descriptions of drill cuttings. This process was done following techniques and properties described by Heath (1984) and Morris and Johnson (1967). Additional information on subsurface properties around the lake, such as depth to bedrock and surrounding boundary conditions, was derived from California’s DWR OSWCR database (California Department of Water Resources, 2021).

The results from this study were combined and contextualized with historical data to develop a generalized conceptual hydrogeologic model narrative. This conceptual model includes information on boundary conditions, such as hydrostratigraphy and hydrogeologic properties, sources, sinks, and flow direction of water. Visual field observations, geologic maps, drillers’ logs, and surface resistivity were used to estimate depth to bedrock, historical water levels, and the type and location of other boundary conditions. Hydrostratigraphy and hydrogeological properties of the aquifer were characterized from descriptions of borehole lithologies. Estimated inputs and outputs and Elizabeth Lake subwatershed boundaries are based on previous works, as discussed earlier in the text.

Water Levels

The seven 2-in. drilled monitoring wells, ELLA-1 through ELLA-6 and ELLA-8, located along the southern, eastern, and northern perimeter of Elizabeth Lake were instrumented with vented continuous water-level sensors for water-level monitoring, which occurred from March 2020 to March 2021 (fig. 3). Before sampling, the monitoring wells depth-to-water were measured from a reference mark on the well casing with an electronic water-level meter to the nearest 0.01 ft and were recorded in the field using the USGS standard field application SVMobileAQ. After the depth-to-water was measured, the discrete water-level data were documented in SVMobileAQ. Once the water-level measurements were completed, the continuous water-level sensors were removed from the borehole before a Grundfos pump was deployed to purge the well. Total purge volumes were calculated by using the water-level and total casing depth before the collection of water-quality samples. All USGS water levels, discrete and continuous, can be accessed via the USGS National Water Information System (NWIS; U.S. Geological Survey, 2023a). Water-levels also were collected from the hand-augered wells before water-quality samples were collected from the wells.

Water Quality

Water-quality samples were collected from the seven 2-in. monitoring wells (ELLA-1 through ELLA-6 and ELLA-8) using a submersible Grundfos pump with Teflon tubing and from the 1-in. hand-augered wells (ELLA-7, ELLA-9, and ELLA-10A; table 1; fig. 3) using a peristaltic pump and Teflon tubing. Elizabeth Lake was sampled at the south shoreline (site ELS-01; table 1; fig. 3) using a peristaltic pump and Teflon tubing, and the well BW-PW (table 1; fig. 3) was sampled with the already equipped 5-horsepower submersible pump. Station number and type of sample (well, tank, Elizabeth Lake water, or a spring) are shown in table 1. Field collection and processing protocols as published in the USGS National Field Manual for Collection of Water-Quality Data (U.S. Geological Survey, variously dated) were followed but without the use of methanol during the cleaning process for field sampling equipment. Methanol was not used because of the potential interference for the DOC samples. These procedures were strictly followed during the collection process, with the exception of samples collected and processed from wells ELLA-3 and ELLA-4 due to the slow recovery of water in the wells and the inability to complete a borehole three-volume purge from the well. Due to the slow recovery of water from the wells, ELLA-3 and ELLA-4 were purged 2 days before sampling to allow time for the water levels to recover. During purging, water level was monitored with an electronic tape to prevent the water level from dropping below the top of the screened interval, which would compromise aquifer water quality and structural well integrity. At the time of sampling, due to the limited amount of water availability at wells ELLA-3 and ELLA-4 during sampling, water in the well was sampled 2 days after purging upon arrival to the sites, and field parameters were collected after the sample was collected from the well.

For each sampling event, samples were collected for analyses of SC, nutrients, pH, DOC, turbidity, and alkalinity at all sites. Stable water isotopes (δ¹⁸O and δD) and major-ion samples, along with selected trace elements, were collected once during the first sampling visit from each site. An additional stable water isotope sample was collected from Elizabeth Lake (ELS-01) during the July sampling trip. Nitrogen and oxygen isotopes in nitrate (δ¹⁵N-NO₃ and δ¹⁸O-NO₃) were collected during the July sampling event for all sites. Additionally, tritium and noble gas samples were collected during the July 27–29, 2020, sampling event for all groundwater sites, except at BW-PW, which shut off before the sample was collected.

Water-Quality Sample Processing, Laboratory Analytical Methods, and Statistical Methods

When the field parameters stabilized and were measured, the sample line was connected to the pump tubing for collection of unfiltered and filtered samples. Filtration was completed using a 0.45-micrometer (µm) capsule filter. All samples, except for water samples analyzed for stable isotopes, were chilled on ice until they could be refrigerated or shipped to laboratories for chemical analysis. Samples requiring low pH for preservation were acidified with concentrated nitric acid such that the sample pH was lowered to less than 2 prior to shipment to the laboratory. Details of the collection and processing of all sample types are given in the National Field Manual for the Collection of Water-quality Data (U.S. Geological Survey, variously dated). DOC samples were typically shipped to the lab within 48 hours of sampling and were not acidified in the field unless holding times were expected to be exceeded. A percentage of DOC samples collected in the field were preserved using sulfuric acid (H₂SO₄) due to extended hold time during the 2020 coronavirus-19 (COVID-19) pandemic (July samples) and again for quality control (QC) for comparison to unacidified DOC samples (September samples).

Samples were collected and analyzed for major ions and trace elements, nutrients, alkalinity, stable isotopes of water, nitrogen and oxygen isotopes in nitrate, noble gasses, tritium, and DOC, and total dissolved solids. Water samples were analyzed by the National Water Quality Laboratory (NWQL) in Lakewood, Colorado, for major ions, nutrients, and selected trace elements, using methods by Fishman and Friedman (1989), Fishman (1993), and Garbarino and others (2002, 2006), respectively. Total dissolved solids concentrations were measured by filtering water, drying the filter, and subsequently weighing the filter (Fishman and Friedman, 1989).

Analytical methods are further described in the following sections of this report. Samples were collected during different wet and dry periods to document potential variations in water quality relative to hydrologic conditions. Early February and late March were targets for the wet sampling periods. However, late March sampling was rescheduled to June due to interruptions of fieldwork because of the 2020 COVID-19 pandemic. Thus, February and June were sampled as wet conditions (closer to the rainy season of the water), and July and late September (toward the end of the water year) were sampled to represent dry conditions. All water-quality data collected are publicly available on NWIS (U.S. Geological Survey, 2023a).

Borehole Geophysics

Borehole geophysical logs were collected from the monitoring wells after drilling. Electrical resistivity is shown in figure 7 for the seven drilled monitoring wells (ELLA-1 through ELLA-6 and ELLA-8; fig. 7). Borehole resistivity, surface resistivity, and borehole lithology for wells ELLA-3, ELLA-4, and ELLA-5 are shown on figure 8. The logs for these three wells (ELLA-3, ELLA-4, and ELLA-5) displayed similar resistivity depth interval results relative to the surface resistivity transect for line 1 (line A–A’, fig. 5) that allowed a comparison to Elizabeth Lake surface resistivity results from line A–A’ and line B–B (fig. 5; Groover and others, 2021). Shallower materials surrounding the well casing had a resistivity of 0–35 ohm-m from 10 to 45 ft bls. Wells ELLA-2, ELLA-4, ELLA-5, and ELLA-6 borehole logs showed high-resistivity peaks (100 or greater ohm-m) from 35 to 40 ft bls, whereas high-resistivity peaks were deeper in wells ELLA-1 and ELLA-3 from approximately 50 to 55 ft bls. Well ELLA-8 was not drilled deep enough to include the high-resistivity peak, but the well depth was terminated at the upper extent of a high resistivity zone (approximately 30 ft bls). Wells deeper than the high-resistive peak transition back to lower resistivity of 15–30 ohm-m at deeper depths.

Lithology and Aquifer Properties

Lithologic descriptions were classified from drill cuttings collected at 5-ft intervals to the total depth of each well. The total depth of each well was dependent on an impassible highly consolidated blue clay layer observed during drilling. Because these wells were drilled into the surrounding lakebed, the geology was consistent with interbeds of sandy silt and gravelly sandy silt at wells ELLA-1, ELLA-2, ELLA-5, and ELLA-6 (fig. 3; table 1). In addition to interbeds of sandy silt and gravelly sandy silt, interbeds of primarily clayey-sediment were identified in the ELLA-3, ELLA-4, and ELLA-8 boreholes. Sandy silt and gravelly sandy silt were identified at all three shallow hand-augered monitoring wells: (1) ELLA-7, (2) ELLA-9, and (3) ELLA-10A. The borehole lithology for all wells can be seen in the “Borehole Lithologies” section of appendix 1 (figs. 1.1–1.8), which includes a summary of detailed notes documented by the onsite geologist during drilling and a simplified diagram of the borehole construction and lithology. During drilling, the onsite geologist noted the presence of perched water above the observed clay layer, which was likely the result of the recent storms and change in infiltration rates of sediments. This freshwater perched aquifer is likely the result of high resistivity at/near clay layers, where resistivity would be expected to have low resistivity.

The simplified lithologies were used to estimate aquifer properties. Hydraulic conductivity and porosity ranges were defined in relation to the 5-ft lithologies. Ranges of hydraulic conductivity applied to these materials were estimated from Heath (1984) and Morris and Johnson (1967). Estimated ranges of hydraulic conductivities were less than or equal to 3.3x10⁻³ feet per day (ft/d) for clay layers, from 3x10⁻³ to 16 ft/d for silt layers, from 16 to 32 ft/d for sand layers, and from 32 to 100 ft/d for poorly sorted gravel layers (appendix 1, figs. 1.1–1.8; Todd, 1959; Davis, 196925; Heath, 1984). Because the wells drilled during this study were assumed to be within an unconfined aquifer, hydraulic conductivity was specific but transmissivity was not because the thickness of the saturated zone varies with the water level. Thus, transmissivity was not estimated in this study. Porosity was estimated to be greater than 0.4 for clays and greater than 0.35 for fine gravel layers (Morris and Johnson, 1967).

Field Parameters

Field parameter results are shown in table 5. The median concentration of SC from groundwater sites was 918 µS/cm at 25 °C, whereas Elizabeth Lake (ELS-01) had a median SC of 9,720 µS/cm at 25 °C, a mean SC of 11,700 µS/cm at 25 °C, and a maximum SC of 21,300 µS/cm at 25 °C, which is consistent with brackish to saline lake waters (greater than 1,500 µS/cm at 25 °C; table 5; Stanton and others, 2017). The highest SC values for groundwater were measured from wells ELLA-4, ELLA-3, and ELLA-10A, at concentrations close to or more than 1,500 µS/cm at 25 °C. The lowest SCs (less than 750 µS/cm at 25 °C) were measured from wells BW-PW, ELLA-9, and ELLA-2. Specific conductance measurements from Tank-1 and Tank-2 were less than 500 µS/cm at 25 °C. For groundwater samples, the median concentration of TDS was 613 mg/L. The TDS concentration was measured by filtering a known amount of water, drying the filter at 180 °C, and weighing the dried filter.

Dissolved-oxygen concentrations in groundwater samples ranged from 0.3 to 5.7 mg/L, with a median concentration of 1.6 mg/L (table 5). The highest DO concentrations were measured from groundwater wells during February 2020. Wells with DO less than 1.0 mg/L were tested for sulfide using a field sulfide test. Dissolved sulfide was detected from wells ELLA-3, ELLA-4, and ELLA-2, at estimated concentrations of 0.03, 0.07, and 0.10 mg/L, respectively. The pH in the groundwater measured in all sites ranged from 6.75 to 8.27, with a median value of 7.44. The pH of Elizabeth Lake (ELS-01) samples ranged from 8.37 to 10.2. The temperature of groundwater from wells ranged from 16.7 to 23.5 °C, with a median value of 12.5 °C. Groundwater temperatures were coldest during February and September sampling.

Alkalinity, Major Ions, and Trace Elements

Alkalinities (expressed in mg/L as calcium carbonate) were measured in the field and were used to calculate bicarbonate and carbonate concentrations. Bicarbonate, major ions, and trace elements data from the groundwater had observed ranges of 121–921 mg/L of bicarbonate (ELLA-9 and ELLA-10a; table 5), 35.5–294 mg/L of sodium (ELLA-9 and ELLA-5; table 6), 12.1–265 mg/L of chloride (ELLA-9 and ELLA-4; table 6), 0.27–1.84 mg/L of fluoride (ELLA-8 and ELLA-10A; table 6), 30.4–97.9 mg/L of calcium (ELLA-3 and ELLA-8; table 6), 52.0–278 mg/L of sulfate (ELLA-7 and ELLA-10A), 1.82–447 µg/L of manganese (ELLA-7 and ELLA-1), and less than 10–410 µg/L of iron (BW-PW and ELLA-4; table 6). Median concentrations of major ions chemistry and trace elements in all groundwater samples can be summarized as 241 mg/L of bicarbonate_, 87 mg/L of sodium, 80 mg/L of chloride, 1.2 mg/L of fluoride, 61.0 mg/L of calcium, 127 mg/L of sulfate, 16.4 µg/L of manganese, and 10.6 µg/L of iron (table 6). Elizabeth Lake samples had a median concentration of 151 mg/L of bicarbonate, 1,570 mg/L of sodium, 1,640 mg/L of chloride, 3.2 mg/L of fluoride, 245 mg/L of calcium, 3,380 mg/L of sulfate, 118 µg/L of manganese, and less than 60 µg/L of iron (ELS-01; table 6). The median TDS concentration in groundwater was 614 mg/L, and the measured TDS concentration in Elizabeth Lake was 8,240 mg/L (ELS-01; table 6).

A trilinear diagram (fig. 11) shows the relative proportions of major ions and nitrate in trilinear diagrams. The interpretation of the trilinear diagrams is discussed later in this report.

Two trace elements, arsenic and uranium, were present with a concentration greater than the maximum contaminant levels (MCLs; 10 and 30 µg/L, respectively) in a few of the sites (U.S. Environmental Protection Agency, 2022). Arsenic concentrations of 42.9, 15.5, and 13.2 µg/L, were measured in the ELLA-4, ELLA-3, and Elizabeth Lake (ELS-01) samples, respectively. One concentration of uranium of 55.2 µg/L was measured in ELS-01.

Nutrients

Nutrient samples were collected and analyzed for multiple forms of phosphorus and nitrogen (table 7). Nitrite concentrations were generally low. Nitrate plus nitrite in groundwater samples (ELLA-1–10A and BW-PW) ranged from less than the detection limit of 0.04 to 27 mg/L as N with a median concentration of 4.7 mg/L as N. Nitrate was the primary form of nitrogen in 35 of the 47 nutrient samples that were collected in groundwater, but organic nitrogen was the dominant form of nitrogen in the Elizabeth Lake (ELS-01) samples (table 8). Nitrate plus nitrite concentrations were consistently below the detection limit in all ELLA-4 and ELS-01 samples and below detection in two of the four samples collected from ELLA-3 in February and September 2020. The highest concentrations of nitrate plus nitrite were detected in samples collected at ELLA-8 (table 7, fig. 12). Nitrate plus nitrite concentrations from all four sample periods collected from ELLA-8 ranged from 23.4 to 27.0 mg/L as N.

Nitrite concentrations ranged from less than the detection limit of 0.001–0.238 mg/L as N in all samples (table 7). The median concentration of nitrite detected in groundwater samples (ELLA-1–10A and BW-PW) was 0.007 mg/L as N. Nitrite concentrations in Elizabeth Lake (ELS-01) were below 0.003 mg/L as N or below the detection limit.

Ammonium as N concentrations in groundwater samples (ELLA-1–10A and BW-PW) ranged from below the detection limit of 0.01 to 0.25 mg/L as N with a median concentration of less than 0.01 mg/L. Ammonium concentrations generally were less than the detection limit, with a few exceptions. Concentrations generally were greater than 0.05 mg/L of ammonium as N in all samples collected at well ELLA-3, July and September samples collected at well ELLA-4, and February samples collected at well ELLA-1. Ammonium concentrations in Elizabeth Lake (ELS-01) ranged from 0.22 to 0.41 mg/L and from 1.90 to 4.05 mg/L of ammonium plus org-N as N.

Orthophosphate as P was the form of phosphorus measured in all wells, the Elizabeth Lake site (ELS-01), and the imported water tanks (Tank-1 and Tank-2). Total phosphorus as P was measured at the Elizabeth Lake site (ELS-01) in addition to orthophosphate as P because of the presence of particulate matter in the lake water. Total phosphorus was also measured in the Tank-1 water but was not measured in Tank-2. Total phosphorus should have been measured in the Tank-2 water, but unfortunately, only a sample for orthophosphate was submitted, so a direct comparison of all forms of phosphorus cannot be completed. However, the two tanks both contain water imported from northern California so the nutrient concentrations should be similar. Both tanks have water delivered to southern California from the Governor Edmund G. Brown California Aqueduct (not shown). Phosphorus concentrations in ELS-01 samples ranged from 0.045 to 0.387 mg/L. Concentrations of total phosphorus from the Tank-1 (0.007 mg/L) and orthophosphate from Tank-2 (0.008 mg/L) were similar, and both were measured during the July sampling. However, a direct comparison of the two forms of phosphorus is not possible because the two analytical methods are not the same. Total phosphorus is measured from an unfiltered sample, whereas orthophosphate is measured from a filtered sample. In groundwater samples, phosphorus was measured in the form of orthophosphate and ranged from 0.03 to 0.35 mg/L as P, with a median concentration of 0.08 mg/L as P.

Dissolved organic carbon was collected during each sampling for all sites (table 1) and ranged from 0.31 to 26 milligrams per liter as carbon (mg/L as C) in groundwater samples (ELLA-1–10A and BW-PW) and from 11 to 100 mg/L in Elizabeth Lake samples (ELS-01). The DOC concentrations in the water tanks (Tank-1 and Tank-2) were similar (1.9 and 2.0 mg/L, respectively). The median DOC concentration in groundwater was 1.9 mg/L, and the mean DOC concentration in groundwater was 4.1 mg/L. The lowest concentration of DOC in groundwater (0.31 mg/L) was measured from well ELLA-5, and the highest concentration of DOC in groundwater (26 mg/L) was measured from well ELLA-3 (table 7). Concentrations of DOC in Elizabeth Lake samples (ELS-01) ranged from 9.9 mg/L during the February sampling to 100 mg/L during the late September sampling.

Isotopes

The water isotopic values ranged from −9.90 to −1.93 per mil δ¹⁸O and −74.6 to −28.9 per mil δD (table 8; fig. 13). Samples were collected for δ¹⁸O and δD from Elizabeth Lake (ELS-01) on March 19, 2019, February 13, 2020, and June 7, 2020. The δ¹⁸O and δD data of these samples ranged from −6.3 to −1.93 and −43.1 to −28.9 per mil, respectively. Samples from the water tanks had values that ranged from −9.9 to −9.6 δ¹⁸O per mil and from −74.7 to −72.9 δD per mil, respectively. The δD values from the Tank-2 results were consistent with the δD value of −74 per mil expected in imported water reported by Izbicki (2004). The δ¹⁸O and δD data in samples from the two water tanks were the most negative (isotopically light), whereas ELS-01 samples were more positive (isotopically heavy) and deviated from the GMWL, indicating that surface waters were affected by evaporation (fig. 13). The δ¹⁸O and δD data in groundwater from wells surrounding the perimeter of Elizabeth Lake ranged from −9.8 to −8.5 per mil and from −70.0 to −58.0 per mil, respectively, and had median values of −8.8 and −61.2 per mil, respectively.

Elizabeth Lake δ¹⁸O and δD values plotted to the right of the GMWL, with an evaporative trend line having a slope of 3.2, which was consistent with the expected slope from evaporation (Gibson and others, 2008; fig. 13). Evaporation of Elizabeth Lake resulted in enriched δ¹⁸O and δD values in the lake water; the slope of the evaporative trend line intersects the GMWL at approximately −7 per mil δ¹⁸O and −47 per mil δD. The δ¹⁸O and δD data were most negative in water from tanks and were consistent with the expected lighter δ¹⁸O and δD of imported water. The water tanks δ¹⁸O and δD values also plotted to the right of the GMWL, indicating that the water may have been partly evaporated prior to storage. Slightly lighter isotopes were measured in Tank-1 in comparison to Tank-2, which contained a mixture that was mostly imported water mixed with small amount of BW-PW groundwater.

The δ¹⁸O and δD values were most negative in groundwater from well ELLA-4, indicating a higher percentage of imported water. The δD was most positive in groundwater from well ELLA-5, whereas the δ¹⁸O was most positive in groundwater from well ELLA-6. The calculated means of δ¹⁸O and δD measured in samples (−8.9 and −61.8 per mil, respectively) were consistent with δD for the precipitation in this region, which has been estimated at approximately −61 δD per mil (Gleason and others, 1994; Izbicki, 200458). The δD data indicated that groundwater in most of the study area was slightly isotopically heavier than water collected from the water tanks and therefore contained lower percentages of imported waters.

Using the mean δD of −61 per mil for groundwater and the δD of −47 per mil from the intersecting evaporative trend line for Elizabeth Lake samples (ELS-01; fig. 13), the mean δD end member of native groundwater was −54 per mil. The δD for imported water was estimated to be −75 per mil and was consistent with results detected in water in Tank-1 and estimated imported water values determined from a previous study (−41.1 to −101 per mil for δD; Izbicki, 2004). By using a simple two-component mixing equation (see eq. 3 in the “Isotopic Analysis” section), the percentage of imported water was calculated. The highest estimated percentages of imported waters were 71, 57, and 50 percent from wells ELLA-4, ELLA-8, and ELLA-3, respectively. The lowest percentages of imported waters were 15, 16, and 18 percent from wells ELLA-5, ELLA-2, and ELLA-1, respectively.

The nitrogen isotopes and oxygen isotopes of nitrate (δ¹⁵N-NO₃ and δ¹⁸O-NO₃) were only able to be determined for samples with nitrate concentrations greater than 0.06 mg/L as N that were only collected during the June sampling period (table 8). The δ¹⁵N-NO₃ and δ¹⁸O-NO₃ in groundwater (ELLA-1–10A and BW-PW) ranged from 3.74 to 33.5 per mil and −2.11 to 17.0 per mil, respectively. The maximum values for δ¹⁵N-NO₃ and δ¹⁸O-NO₃ were measured in the sample collected from well ELLA-7, the minimum value for δ¹⁸O-NO₃ was measured in the sample collected from well ELLA-3, and the minimum value for δ¹⁵N-NO₃ was measured in the sample collected from well ELLA-9. The median values of δ¹⁵N-NO₃ and δ¹⁸O-NO₃ in groundwater were 12.35 and 4.38 per mil, respectively. The potential sources of nitrate were determined with these isotopic N values. Values δ¹⁵N-NO₃ of total soil N range from +2 to 5 per mil, and δ¹⁵N-NO₃ of effluent is about 7 per mil and higher (Kendall and McDonnell, 1998; Hinkle and others, 2008). Additionally, δ¹⁵N-NO₃ data were used to evaluate the environmental history of nitrate with respect to microbially mediated denitrification and attenuation of nitrate. Results of the δ¹⁵N-NO₃ data were used to further identify the processes that control ammonium and nitrate concentrations.

The transformation of the forms of nitrogen in water occurs through different microbial processes. Nitrification can occur in the unsaturated zone and in shallow oxic groundwater. The extent of these reactions were evaluated using δ¹⁵N-NO₃ and δ¹⁸O-NO₃ ratios. Values of δ ¹⁵N-NO₃ and δ¹⁸O-NO₃ of nitrate were significantly correlated with a slope of 0.5 (coefficient of determination, R²=0.54; p-value=0.009), which is consistent with denitrification (figs. 14A, 14B; Kendall and McDonnell, 1998). Isotopic ratios of nitrogen and oxygen in nitrate change as denitrification proceeds in the sub-surface, and because this change is predictable, measurements of these isotopes provide clues to potential sources (Kendall and McDonnell, 1998). Additionally, dominant sources of N as a function of δ¹⁵N-NO₃ and δ¹⁸O-NO₃ were determined based on δ¹⁵N and δ¹⁸O values. Sites with detectable isotopic measurements as nitrate either indicate the presence of OWTS effluent or are affected by denitrification (fig. 14).

Noble Gases and Tritium

As previously stated, samples of dissolved noble gases were modeled in DGMETA (Jurgens and others, 2020) using the CE model and the unfractionated excess air (UA) model to determine recharge temperatures and to parameterize tritium-helium ages (Jurgens and others, 2020). The CE model and UA model results were compared on a case-by-case basis for each sample (fig. 15). The CE model generally predicted higher recharge temperatures, balanced helium concentrations, better fits, and predicted more incorporation of excess air (air bubbles trapped below the water table) than the UA model (fig. 15). Additionally, the CE modeled samples more accurately because the water table fluctuated more than a 4.33 ft annually, consequentially trapping large volumes of air in the water during recharge and causing the noble gases in groundwater to become fractionated (Phillips and others, 2007; Jurgens and others, 2008). Thus, results generated from the CE model were chosen as the best fit to describe the set of gas concentrations.

Measured values of helium-4 (⁴He), Ne, Ar, Kr, Xe, and N₂ gas concentrations were used as inputs into the DGMETA program to calculate excess air, excess N₂, recharge temperatures, and ³He_trit values (U.S. Geological Survey, 2023a). Concentrations of ⁴He ranged from 5.88x10⁻⁸ to 1.59x10⁻⁶ cubic centimeters at standard temperature and pressure per gram of water (cm³ at STP/g H₂O), and minimum and maximum concentrations were measured in samples collected from wells ELLA-6 and ELLA-3, respectively. Ne and Ar concentrations ranged from 2.09x10⁻⁷ to 3.40x10⁻⁷ and from 3.31x10⁻⁴ to 5.28x10⁻⁴ cm³ at STP/g H₂O, and median concentrations were 2.36x10⁻⁷ and 3.94x10⁻⁴ cm³ at STP/g H₂O, respectively. Minimum and maximum concentrations of Ne were detected in samples collected from wells ELLA-6 and ELLA-7, respectively. Krypton concentrations ranged from 7.02x10⁻⁸ to 1.09x10⁻⁷ cm³ at STP/g H₂O, with a median of 8.68x10⁻⁸ cm³ at STP/g H₂O; Xe concentrations ranged from 9.59x10⁻⁹ to 1.43x10⁻⁸ cm³ at STP/g H₂O, with a median of 1.16x10⁻⁸ cm³ at STP/g H₂O; and N₂ concentrations ranged from 0.01 to 0.02 cm³ at STP/g H₂O, with a median of 0.02 cm³ at STP/g H₂O. Minimum and maximum concentrations were consistently detected for Ar, Kr, Xe, and N₂ in samples collected from wells ELLA-8 and ELLA-3, respectively. Calculated recharge temperatures ranged from 10.6 to 17.0 °C, with a median temperature of 13.7 °C.

Measured N₂ was calculated from groundwater recharge temperatures and excess air concentrations. Excess nitrogen was calculated in samples collected from wells ELLA-1, ELLA-2, ELLA-3, ELLA-4, and ELLA-5, which indicated isotopic evidence of denitrification. Calculated concentrations of excess N₂ gas (differences between measured and modeled values for samples with higher N₂ than modeled values) ranged from 0.24 to 2.78 mg/L as N (U.S. Geological Survey, 2023a), with a median value of 0.86 mg/L as N. The maximum calculated concentration of N₂ was present in well ELLA-3. Groundwater at wells where excess N₂ was not present (wells ELLA-7, ELLA-8, ELLA-9, and ELLA-10A) had oxic conditions (DO greater than 0.50 mg/L) and were located on the north side of Elizabeth Lake. Where measured N₂ values were greater than modeled values, another source of N₂ is likely present, indicating excess groundwater N₂ is being added to groundwater after recharge by denitrification (Heaton and Vogel, 1981).

Measured tritium activities ranged from 0.77 to 4.89 TU, with a median concentration of 1.65 TU (table 8). Groundwater samples from wells ELLA-2, ELLA-7, ELLA-9, and ELLA-10A had measured tritium activities greater than 2.6 TU and were classified as classified as “modern” (recharged after 1950s; Lindsey and others, 2019). Groundwater samples from wells ELLA-1, ELLA-3, ELLA-4, ELLA-5, ELLA-6, ELLA-8, and BW-PW had tritium activities between 0.34 and 2.6 TU and were classified as “mixed age” (containing a mixture of groundwater recharged before and after the 1950s; Lindsey and others, 2019). Groundwater samples from the study area did not indicate evidence of pre-modern samples (less than 0.34 TU; Lindsey and others, 2019).

An “apparent” age was computed for most groundwater samples based on the ratio of ³He_trit to ³H using noble gas inputs into DGMETA. Apparent ages were not computed for groundwater samples from wells ELLA-8, ELLA-10A, and BW-PW because noble gas data could not be used to predict ³He_trit concentrations for these wells because the DGMETA model returned unrealistic values, such as unrealistic recharge temperatures. ³He_trit concentrations ranged from 1.20 to 99.49 TU (table 8), with a median concentration of 10.5 TU. Computed ages ranged from 6 to 86 years (table 8), indicating that some groundwater samples would be classified as “pre-modern” based on the age classifications derived from measured tritium activities. The median computed age of samples from wells with mixed groundwaters (excluding ELLA-8 and BW-PW) was 44 years, and the median computed age of groundwater from wells classified as “modern” (excluding ELLA-10A) was 7.5 years.

Water-Quality Discussion

The groundwater of the study area is fresh to slightly brackish, as indicated by measurements of specific conductance and TDS concentrations. In contrast, Elizabeth Lake is slightly saline and has TDS concentrations greater than 1,500 mg/L (Swenson and Baldwin, 1965; Stanton and others, 2017). DO concentrations, along with oxidation-reduction potential (redox) indicators in groundwater, can be used to interpret the chemical and microbial reactions that can occur. Samples with DO concentrations less than 0.5 mg/L typically indicate reducing conditions. The EPA (U.S. Environmental Protection Agency, 2023b) suggests that for healthy systems, groundwater pH values should be between 6 and 8.5, and surface-water pH values should be between 6.5 and 8.5 (fig. 16). Groundwater samples (ELLA-1–10A and BW-PW) collected in this study had a narrow range of pH values (7–8), with variable DO (0.3 to 10.2 mg/L), but samples from Elizabeth Lake (ELS-01) had consistently higher pH values (greater than 8; range was from 8.4 to 10.2) and DO concentrations (greater than 6 mg/L; fig. 16). Elizabeth Lake water was collected during the day and during an algal bloom, so elevated pH was expected. We also expected that pH values and DO concentrations would drop after daylight hours. All but one sample from ELS-01 exceeded the suggested pH thresholds for surface water (6.5–8.5) by more than 1.3 pH unit; the exception was the sample collected during the February 2020 sampling event. Higher pH values in ELS-01 were attributed to algal productivity and CO₂ uptake.

According to visual observations made from local residents and field visits during the dry season before the study, Elizabeth Lake periodically dries up and leaves behind a salt efflorescence (white appearing residue), which dissolves as the lake fills with water in wet years and contributes to the TDS in the lake. Elizabeth Lake has a different molar ratio of calcium to magnesium (0.26) relative to those in domestic source water (that is, imported water) used by homes surrounding Elizabeth Lake (1.05) and to groundwater in the wells surrounding the lake (about 2.00). Major element chemistry, including ratio of calcium to magnesium, is affected by water-rock interactions. Elizabeth Lake water is supersaturated with respect to calcite (as indicated by saturation indices exceeding 1, based on thermodynamic modeling), indicating active precipitation of calcite. Active precipitation will increase the ratio of calcium to magnesium (Hem, 1985). Isotopic analysis (δD and δ¹⁸O) of groundwater from wells (ELLA-1–10A and BW-PW) and surface water from Elizabeth Lake (ELS-01) indicated an evaporative signature for Elizabeth Lake. Evaporation of Elizabeth Lake water will also drive calcite precipitation because of evaporation.

Two major sources of water to the groundwater system immediately downgradient from the developed area are local precipitation and water from OWTS. The region is arid, with an average annual rainfall of 13 in/yr (National Oceanic and Atmospheric Administration, 2020). Water used for domestic consumption in the community surrounding Elizabeth Lake is imported mainly from the Governor Edmund G. Brown California Aqueduct and has the lowest TDS concentrations (243–278 mg/L). Water used for domestic purposes eventually recharges the aquifer as OWTS effluent. Imported water was also used for minimal residential lawn and gardening irrigation and infiltrated into the groundwater. Once discharged from the OWTS, the water provides a source of recharge to the local groundwater system. Isotopic analysis of δD and δ¹⁸O indicated the domestic (imported) water was isotopically light relative to the wells (ELLA-1–10A and BW-PW) and Elizabeth Lake (ELS-01; fig. 16). Isotopes measured in groundwater from wells were clearly separated from the more highly evaporated signature of ELS-01 samples. The enriched δD data indicated that wells ELLA-3, ELLA-4, and ELLA-8 were more closely related to Tank-1 samples and are thus more affected by imported waters from OWTS effluent into groundwater upgradient from the wells than the remainder of the wells.

Relative proportions of major ions in the imported water, wells (ELLA-1–10A and BW-PW), and Elizabeth Lake (ELS-01) varied considerably on a trilinear diagram (fig. 11). Groundwater in two wells, ELLA-3 and ELLA-4, had the most sodium plus potassium relative to calcium and magnesium, and the isotopic signatures of groundwater in these wells were similar to the imported water. One likely explanation for this is that the domestic (imported) water was treated with a water softener which takes the doubly positive ions (calcium and magnesium) out of the water and replaces them with singly charged ions, such as Na. We interpret this result as the groundwater chemistry in wells ELLA-3 and ELLA-4 having been altered by OWTS. Wells ELLA-1, ELLA-3, ELLA-5, ELLA-6, ELLA-7, and ELLA-10A were supersaturated with respect to calcite, further indicating alteration of the source (imported water or recharge). Well ELLA-9, the farthest from the developed area, was undersaturated with respect to calcite and is probably the most representative of groundwater recharged mainly by local precipitation. Groundwater in six wells, ELLA-1, ELLA-2, ELLA-5, ELLA-6, ELLA-10A, and BW-PW, have similar water quality, as shown on the trilinear diagram; observed isotopic similarities indicate a similar evolution along the groundwater flow path relative to water, rock, and atmospheric interactions (figs. 13, 16). The isotopic compositions of samples from well ELLA-8 differed substantially from groundwater in all other wells, were isotopically more similar to the domestic (imported) water, and had the highest nitrate concentrations and some degree of evaporative concentration.

Trace elements, such as iron and manganese, measured in the samples from the wells indicated mostly sub-oxic conditions (less than or equal to 0.05 mg/L of iron and manganese) consistent with low-measured DO concentrations (less than or equal to 1.0 mg/L), except for well ELLA-8 (Jurgens and others, 2008; McMahon and Chapelle, 2008; McMahon and others, 2009). The method of Jurgens and others (2008) was used to identify redox conditions in the water samples. Well ELLA-8 had the highest DO concentrations, undetectable iron, and high nitrate concentrations, indicating that conditions were oxic. The oxidation-reduction conditions in groundwater in most of the wells, except for well ELLA-8, would allow denitrification reactions to happen. Most of the groundwater sampled in this system was characterized as oxic to sub-oxic based on concentrations of DO, nitrate, manganese, iron, and sulfate collected from each site after the first sampling event (February 2020). Samples from wells ELLA-1, ELLA-2, ELLA-3, and ELLA-4 were characterized as having an oxic-anoxic mixture. Wells ELLA-1 and ELLA-2 were characterized as having an oxygen-manganese reduction process (greater than or equal to 0.5 mg/L of DO and greater than or equal to 0.05 mg/L of manganese), whereas groundwater samples from wells ELLA-3 and ELLA-4 were characterized as having an oxygen-iron/sulfate reduction process (greater than or equal to 0.5 mg/L of DO, 0.5 mg/L of sulfate, greater than or equal to 0.01 mg/L of iron, and less than 0.5 mg/L of nitrate). Wells ELLA-5, ELLA-6, ELLA-7, ELLA-8, ELLA-9, and ELLA-10A were characterized as oxic. Additionally, samples from wells ELLA-3 and ELLA-4 and from Elizabeth Lake (ELS-01) contained arsenic greater than the maximum contaminant level of 10 µg/L, indicating that redox conditions or other geochemical conditions may be causing mobilization of arsenic. Mobilization of arsenic in water is frequently the result of desorption from surfaces following reduction of manganese and iron oxyhydroxides by heterotrophic microorganisms, subsequent reduction of binding sites, desorption driven by competing ions such as phosphate, or oxidation and dissolution of arsenic-bearing sulfide minerals (Neil and others, 2012; Haugen and others, 2021).

DOC concentrations likely increased as a product of the decomposition of DOM in addition to DOC concentrations pre-existing in groundwater. DOC concentrations in natural groundwater are typically below 2.0 mg/L (Thurman, 1985); however, groundwater samples collected from wells ELLA-4, ELLA-3, ELLA-10A, and ELLA-7 had measured DOC concentrations that consistently exceeded 2.0 mg/L. Groundwater samples collected from wells ELLA-2, ELLA-1, and ELLA-9 sometimes had measured DOC concentrations that exceeded 2.0 mg/L (figs. 12, 17A, 17B), potentially indicating effects of recharge from OWTS. DOC concentrations relative to nitrate and ammonium concentration are shown on figure 17. Dissolved organic carbon concentrations and nitrate concentrations showed no linear relation (fig. 17A), whereas ammonium and dissolved organic carbon have a weak linear relation (R² = 0.5924). Sites in this study that contain low nitrate concentrations of 0.05 mg/L or less typically have DOC concentrations greater than 2.0 mg/L, which is the threshold of natural DOC in groundwater (table 7; Thurman, 1985), which indicates that OWTS may contribute DOC into groundwater (fig. 17). DOC concentrations were less than 2.0 mg/L in all groundwater samples collected from wells ELLA-5, ELLA-6, ELLA-8, and BW-PW.

Groundwater samples with high DOC concentrations generally had elevated ammonium concentrations (figs. 12, 17B). High DOC from OWTS drives heterotrophic oxidation by microbes that deplete DO and cause waters to be anoxic, which promotes the attenuation of nitrate through denitrification. Breakdown of DOC by microbes also will liberate org-N and convert the org-N to ammonium. Once DO is low enough, the ammonium will not convert to nitrate. Lakes with high algal productivity typically have 2.0–5.0 mg/L of DOC; these concentrations were consistently exceeded in Elizabeth Lake (ELS-01) samples (figs. 12, 17A; Thurman, 1985). Ammonium in Elizabeth Lake may originate from internal lake recycling or a waste product of the numerous aquatic bird species that are present on the lake, or biological fixation of gaseous nitrogen.

Low concentrations of nitrate measured in groundwater indicate that removal through denitrification is occurring due to reducing conditions in parts of the aquifer where DOC is elevated (fig. 17A). Thus, the remaining ammonium is less able to oxidize to nitrate during these reducing conditions, and increased DOC in the aquifer may be decomposing and increasing ammonium concentrations. The presence of ammonium is a potential indicator for effluent or nitrification of org-N upgradient in the flow path that was not lost to sorption or volatilization. Water containing elevated DOC, ammonium, and org-N concentrations likely discharged into shallow groundwater from OWTS. Onsite wastewater treatment system discharges likely alter the redox condition in the aquifer by increasing DOC concentrations that fuel microbial productivity, decrease DO levels, and increase the potential for nitrogen removal through denitrification (Seitzinger, 1994). Sites where nitrate concentrations in groundwater are low and ammonium is present also contain detectable nitrite concentrations (fig. 12). Nitrite is a transient species in groundwater and is part of the process of the bacterial transformation of nitrate to nitrogen gas or oxidation of ammonium to nitrate. Groundwater prior to infiltration of OWTS effluent was likely oxic and favored the transformation of org-N and ammonium through nitrification. Additional evidence for the occurrence of denitrification in groundwater at wells ELLA-3 and ELLA-4 came from the September 2020 sampling, in which the presence of sulfide indicated that sulfate-reducing bacteria were present.

Nitrate in groundwater in wells ELLA-8, ELLA-9, and BW-PW may have also resulted from soil-derived nitrogen that overlaps with isotope values of the nitrogen derived from manure and septic waste (OWTS effluent) boundaries (fig. 14). Soil-derived nitrogen was likely the source in groundwater from wells ELLA-9 and BW-PW, which contained nitrate concentrations less than 10 mg/L and were not suspected of being downgradient from any OWTS; for the duration of the study, groundwater in well ELLA-8 consistently contained nitrate concentrations exceeding 20 mg/L, DOC concentrations less than 2.0 mg/L (fig. 12), and evidence of oxic waters (DO greater than 0.05 mg/L; tables 5, 7).

Orthophosphate concentrations in groundwater were variable across all the wells but were highest in wells with low DO and higher in groundwater than in the imported water (figs. 12, 18). At individual wells, orthophosphate concentrations showed little to no variability throughout the study (fig. 18). Nitrate concentrations also did not indicate a seasonal variability at individual well sites (fig. 18).

Orthophosphate in aquifers is frequently sorbed onto oxide surfaces of iron or manganese during oxic conditions; however, sorption depends on pH, and low pH conditions favor sorption. During alkaline pH conditions, oxides may have been supersaturated with respect to orthophosphate (Domagalski and Johnson, 2011). The observed concentrations of orthophosphate in groundwater can most likely be attributed to variations in oxidation/reduction throughout the study area, redox conditions, and oxide surface availability.

The measured δD of water from Tank-1 was −75 per mil, which is consistent with previous measurements Izbicki (2004). Because of the effects of imported water on groundwater, δD values in native groundwater were estimated using precipitation values reported in Gleason and others (1994) and the intersection of the evaporation trend line of Elizabeth Lake water with the GMWL. Groundwaters that have a larger imported water ratio contain a more negative δD value than groundwaters that contain a more positive δD value. However, δD and imported ratio values within each well may vary seasonally, depending on the quantity of imported water used. The change in demand also results in variations in the groundwater flow direction (fig. 10A).

In this study area, water imported from northern California discharges into the groundwater system as recharge from OWTS. There is little landscape irrigation, and most of the water recharged is suspected to subsequently be sourced as OWTS effluent. Further evidence to support this interpretation is based on the estimated percentage of imported water of groundwater samples in wells ELLA-4, ELLA-8, and ELLA-3, which were estimated to be composed of greater than or equal to 49-percent imported water (table 8). Estimated imported water percentages in Elizabeth Lake were unable to be calculated because the lake is a mixture of recharge from direct precipitation and potential groundwater discharge to the lake. Additionally, the lake has undergone evaporative enrichment of δD and δ¹⁸O values. Of the three wells with increased imported water, well ELLA-8 was the only well with elevated concentrations of nitrate.

One possible explanation for the elevated nitrate concentrations in well ELLA-8 is OWTS-derived nitrogen, as indicated by similar δD and δ¹⁸O ratios to imported water. Because redox conditions were oxic, denitrification and attenuation of nitrate could not occur. Organic-N and ammonium from effluent were likely oxidized to nitrite and nitrate during the oxic conditions in groundwater. Vero and others (2017) noted a time lag for nitrogen moving through aquifers and denitrification. A similar time lag may have been occurring at well ELLA-8, where nitrate concentrations were increasing but denitrification had not started. However, the nitrogen and oxygen isotopes in nitrate for well ELLA-8 plot in the soil derived region (fig. 14A). Thus, the origin of the nitrate could not be determined from the available data.

Groundwater in the area contained relatively high concentrations of entrapped air; therefore, the best model to interpret dissolved gases was the CE model in DGMETA (Jurgens and others, 2020). The median average air temperatures, assumed to be the temperature at recharge in the unsaturated zone, were similar to the mean annual recharge air temperatures during spring and winter when the bulk of groundwater recharge occurs (National Oceanic and Atmospheric Administration, 2020). However, this interpretation of recharge temperature may not be correct where the water table is within approximately 6.5 ft of the surface. High recharge temperatures were determined in samples from the shallow wells (less than 10 ft deep). Of the 12 noble gas samples analyzed, the recharge temperature was 15 °C or greater in wells ELLA-7, ELLA-9, and ELLA-10A. Additionally, groundwater samples from wells on the north side of Elizabeth Lake contained higher noble gas measured concentrations of ⁴He than wells on the south side of Elizabeth Lake, which is likely a result of the wells to the north of the lake being shallower.

Lower excess air concentrations were calculated at wells on the north side of Elizabeth Lake, with a mean amount of 20 cm³ at STP/kg H₂O, compared to those on the south side, with a mean amount of 72 cm³ at STP/kg H₂O. In this arid environment, intermittent recharge events cause rapid rises of the water table (fig. 9) and can favor entrapment of air bubbles. Geogenic helium (⁴He) was detected in shallow groundwater samples, indicating that groundwater samples on the north side of Elizabeth Lake, such as ELLA-7 and ELLA-9, may be slightly affected by the upwelling of a mixture of mantle- or crustal-derived He (fig. 19; Kulongoski and others, 2003; Jurgens and others, 2020). Excess nitrogen gas produced from denitrification in reducing waters indicates that denitrification occurred in groundwater at all wells on the south side of Elizabeth Lake and well ELLA-6 on the east side of Elizabeth Lake. The most oxic groundwater during the time of sampling for noble gasses was obtained from wells ELLA-8, ELLA-7, and ELLA-10A. The presence of excess gas at wells with reducing redox conditions (ELLA-6, ELLA-4, ELLA-3, ELLA-5, ELLA-2, and ELLA-1) provides further evidence that denitrification is occurring.

Generally, groundwater ages in wells on the northern part of Elizabeth Lake contain much younger ages than groundwater on the southern part of the lake (fig. 3), with the exception of well ELLA-8. This difference in groundwater ages could potentially be an effect of well depths, which are generally deeper on the south side of the lake and are farther from recharge along runoff drainage channels.