Cave Spring

Cave spring is an upland spring located at the northern boundary of the NTC in the Avawatz Mountains, about 25 mi northeast of Irwin subbasin (fig. 2). The spring is near the intersection of the Garlock and Southern Death Valley Fault Zones in a narrow northward-draining canyon about 2 mi north of the main drainage divide of the Avawatz Mountains (figs. 2, 3, 4). Cave Spring was described by Mendenhall (1909) and used by travelers between Barstow (fig. 1) and Death Valley, California (not shown). Cave spring is on the east sidewall of the wash in fractured metamorphic rocks (fig. 4). Mendenhall (1909) reported that two springs, each about 5 ft across and 5 ft deep, existed in the early 1900s. Thompson (1929) described two caves or short tunnels about 40 ft apart that were dug 10 ft into the wall of the canyon. However, when these caves were visited in 2015, only the northern cave contained water. Land-surface altitude of Cave Spring is about 3,605 ft above NAVD 88.

Cave Spring is fed by water seeping from fractured metamorphic rocks into a pond inside the northern cave that covered an area of at least 5 ft across. Water also is present in a second shallow pond outside the cave wall; this pond is about 5 ft south of the northern cave entrance and adjacent to the east sidewall of the canyon, where it flowed a short distance and infiltrated into the thin wash deposits within the canyon (fig. 5). The pond in the wash presumably is fed by seepage through fractures along the cave wall. The main area of vegetation, primarily grasses, cattails, and a small tree, was around this second pond, which covers an area of about 5 ft and where the ground was moist.

Desert King Spring

Desert King Spring is on the north side of the Granite Mountains and along the southern edge of Leach Valley (fig. 2), about 20 mi north of Irwin subbasin. This spring is along an unnamed shear zone and may be part of the Garlock and Southern Death Valley Fault Zones. Desert King Spring lies in a north-northeasterly canyon, about 2 mi north of the main drainage divide of the Granite Mountains (fig. 2). The spring/seep is on the east side of the wash in fractured plutonic rocks. Land-surface altitude of the spring is about 2,900 ft above NAVD 88.

In 1917, Thompson (1929) described Desert King Spring as a well east of the wash, about 900 ft south of and 50 ft above an old cabin and stamp mill. Thompson observed a small pond of water near the well that was dug 15 ft into granite and had a depth to water of about 1 ft below land surface ft bls. Bowen (1943) reported locating a 5 by 5 ft, 12.5-ft-deep well with a depth to water of about 2 ft bls, excavated in granite on the west side of a pronounced north/south-trending shear zone about 50-ft wide and with a nearly vertical dip. Bowen (1943) measured discharge at the outlet of a 1-in. diameter pipe from the well at about 0.17 gallons per minute (gal/min). Based on Bowen’s (1943) suggestions, a well reportedly was dug along the south side of the wash and west side of the shear zone to further develop this spring. In 2015, neither a pond nor wells were observed. The present (2016) location of Desert King Spring (figs. 6, 7) is about 400 ft south of the site reported by Thompson (1929). This “present-day” spring is a small 1 by 1 ft wetted area, where water was present on the surface and supported grasses and melons, near an old bathtub (fig. 8). White deposits, presumably nitrate-bearing caliche, were observed upgradient from this location and appear to lie along or near the shear zone.

Devouge Spring

Devouge Spring is an upland spring in the southern part of Avawatz Valley groundwater basin. The spring is a small seep on an alluvial fan. The Drinkwater Lake Fault (Schermer and others, 1996) and Granite Mountains (fig. 9) are south of the spring. The Granite Mountains are comprised of felsic plutonic rocks (Miller and others, 2014) that generally are non-water bearing except where jointed or fractured (Densmore and Londquist, 1997). The spring is east of an outcrop of plutonic rocks that extends north from the main part of the Granite Mountains and indicates shallow bedrock is in this area (fig. 9). Land-surface altitude of the spring is about 3,720 ft above NAVD 88.

Devouge Spring covers a wetted area about 6 ft long (oriented downslope from the mountain front) and 5 ft wide (fig. 10); this is a main area of vegetation, primarily cattails, shrubs, and grasses, where the ground is moist (fig. 11). An open pit that appears to be a hand-dug well is about 150 ft in distance upslope from the seep (fig. 10). The open pit is approximately 5 ft square and about 14 ft deep, and it appeared to be dry. An old metal water tank is about 200 ft downslope from the wetted area (fig. 10). Old piping lies between the wetted area and tank, indicating that water (presumably from the open pit) may have been piped and stored in the tank and the flow from the spring was likely greater than flow observed in 2015. Because of the location of the wetted area relative to the old piping, it appears that Devouge Spring may actually be seepage from a break in the buried piping resulting in a “man-made” wetted area.

Thompson (1929) described visiting a spring called “Drinkwater Spring” in the general vicinity of the present day (2016) Devouge Spring. When Thompson visited in 1917, there was a cabin with a well dug in granite about 250 ft east of the cabin. Thompson (1929) reported the depth of the well was about 8 ft and the depth to water was 3 ft in 1917. In Thompson’s 1929 report, this well was mistaken for the spring, which was observed to be several hundred feet farther southeast, on the southeastern side of the granite ridge. Although not sampled by Thompson (1929), Drinkwater Spring reportedly had “good” quality water. Based on the location description relative to a playa in Avawatz Valley groundwater basin, Devouge Spring, visited in 2015, could be Thompson’s Drinkwater Spring. Bowen (1943) also reported visiting Drinkwater Spring, where he located the cabin and perhaps the former well, but not any spring. Follow-up communications with NTC personnel (Liana Ayers, U.S. Army, written commun., 2017) indicated that Drinkwater Spring is west of Devouge Spring and was reportedly dry.

Panther Spring

Panther Spring is in the Granite Mountains, south of Avawatz Valley groundwater basin (fig. 2). The spring lies in a narrow wash on the north side of the Granite Mountains. Land-surface altitude of the spring is 3,825 ft above NAVD 88. Panther Spring is between a series of fault splays of the Drinkwater Lake Fault (figs. 9, 12). The spring is in thin wash deposits among felsic plutonic rocks, indicating shallow bedrock in this area (fig. 9). The active channel of the wash is lined with a thin veneer of younger alluvium composed of decomposed granite (unconsolidated sand and gravels derived from the plutonic rocks).

Panther Spring appears as a series of intermittent shallow pools in an ephemeral wash that flows only after precipitation (fig. 12). The shallow pools can hold water for several months after precipitation but reportedly dry up because of lack of substantial, sustained groundwater discharge from the fractured plutonic rocks or wash deposits coupled with high evaporation rates (Liana Ayers, U.S. Army, oral commun., 2017). A man-made trough built to capture water was dry when visited in 2016. The main area of wetland vegetation covers about 6,465 ft² and consists of a few trees dominating the west side of the wash (fig. 13).

No Name Spring

No Name Spring is in the Granite Mountains, near the southwest side of the Avawatz Valley groundwater basin (fig. 2). The spring lies in a narrow wash on the eastern end of the Granite Mountains. Land-surface altitude of the spring is 4,520 ft above NAVD 88. No Name Spring is between splays of the Drinkwater Lake Fault (fig. 9). The spring is about 1.5 mi southwest of Devouge and Panther Springs. The spring is in thin wash deposits overlying felsic plutonic rocks, indicating shallow bedrock in this area. Scattered patches of younger alluvium line the active channel of the wash. These wash deposits are composed of coarse-grained decomposed granite.

No Name Spring appears as an area of vegetation, primarily grasses and shrubs, covering about 2,200 ft² (figs. 14, 15). A small 1-in. diameter buried pipe was observed at the downstream end of the vegetated area but was not followed to its termination. An old metal water tank is about 200 ft downstream from the vegetated area (fig. 14), indicating that the spring may have previously had more discharge for capture than was observed in the current study (2016). Bowen (1943) reported visiting a spring called “Taylor Spring.” Based on the description of the man-made features, it is believed that No Name Spring may be the same as Taylor Spring described by Bowen (1943). In 1943, discharge of the spring was measured at the outlet of the 1-in. diameter pipe at a rate of 0.5 gal/min. During 2016, discharge from No Name Spring was merely a slow drip from the 1-in.-diameter pipe and was insufficient for determining either a flow rate or collecting water-quality samples.

Garlic Spring

The spring lies at the base of the low hills that separate the Langford Valley groundwater basin from the Bicycle Valley and Cronise Valley groundwater basins to the east (fig. 2). Garlic Spring is between Irwin subbasin and Langford subbasin (fig. 2). Land-surface altitude of the spring is 2,312 ft above NAVD 88. Garlic Spring is on the northeast side of a dry wash that flows only during extreme precipitation through a narrow gap between low hills that connects Irwin and Langford subbasins (fig. 16). The hills are crossed by several faults and are composed of felsic and mafic plutonic, schistose, and carbonate rocks (Miller and others, 2014) that generally are non-water bearing except where jointed or fractured (Densmore and Londquist, 1997). Garlic Spring appears in a relatively narrow part of the wash and forms as a series of seepages along the trend of the Garlic Spring Fault (fig. 16) that discharge southwest toward the wash. The NTC wastewater-treatment facility (not shown) and disposal ponds also are about 1-mile northeast near this wash (fig. 16).

The spring appears as an area of wetland seeps for a distance of about 450 ft. There are three main areas of vegetation: (1) where groundwater discharges to the surface, (2) flows downgradient from the surface discharge, and (3) and infiltrates into the main channel of the dry wash (figs. 17, 18). The western area supports large trees, shrubs, and grasses; the middle area supports cattails and small shrubs and trees; the eastern area supports cattails, grasses, and an occasional small tree (fig. 18). The extent of vegetation shown in areal photographs of Garlic Spring in 2007 (fig. 17) is greater than vegetated extent near Garlic Spring indicated by areal photographs taken in 1995 (U.S Geological Survey National Aerial Photography Program, September 30, 1995, accessed November 15, 2016, at https://earthexplorer.usgs.gov/; not shown on fig. 17). Two seepage points along the eastern side of the spring are reported to have been developed by shallow excavation (Bowen, 1943). During 2015, removal of invasive species vegetation by NTC personnel uncovered one of these excavations, an approximate 5- by 5-ft square pit that is about 14 ft deep on the eastern most end of the seepages. Water-quality samples collected from this pit in 1993 are representative of Garlic Spring; the location of the excavation where water samples were collected in 1993 is about 320 ft southeast of the general location shown on figure 17 and near the temperature profile site GS-2 Trod (East-Up).

Bitter Spring

Bitter Spring, the largest spring by area at Fort Irwin, is in Cronise Valley (fig. 2). Bitter Spring is about 1 mi southeast of the Langford Lake Main Supply Route (fig. 19). Land-surface altitude of Bitter Spring ranges from about 1,360 ft above NAVD 88 at the northern extent of vegetation to about 1,330 ft above NAVD 88 at the southern extent. The spring lies in the Bicycle Lake Fault Zone (fig. 19).

Bitter Spring is in a sandy wash cutting through hills consisting of Tertiary, partly consolidated deposits and northeast of a black volcanic hill composed of basalt, which is a mafic volcanic rock (also known as “The Whale”; fig. 19). The wash generally drains from the northern part of Cronise Valley groundwater basin to the southeastern part where the West and East Cronise Lake (dry) playas are located (fig. 1). This wash flows during extreme precipitation. The partly consolidated deposits that form the hills (fig. 19) are composed of coarse-grained deposits, such as conglomerate and sandstone underlain by tight clays. The coarse-grained deposits dip northeasterly, with the underlying clays forming a barrier to southward groundwater flow (Bowen, 1943). Several faults, splays of the Bicycle Lake Fault Zone, cross this area and may have offset or tilted the tight clay so that the altitude of the top of the clay is higher south of the faults than it is north of the faults. Bowen (1943) mapped the deposits in the east wall of the wash as dipping approximately 60–65 degrees to the northeast. Bitter Spring formed by groundwater draining from Cronise Valley groundwater basin being forced to the surface by the fault, through the overlying coarser-grained deposits and above the saturated tight clay that is part of the aquifer system.

Bitter Spring supports an extensive area of vegetation, primarily small trees, shrubs, and grasses, within the wash for about 1,500 ft (figs. 20, 21). This area of vegetation indicates that groundwater is shallow in this part of the wash. Along the southeastern end of the vegetation are several wildlife watering holes that are the surface expression of the water table. Reportedly, several trenches were dug before the 1940s along the southeastern part of the spring to better develop this area (Bowen, 1943); however, no obvious evidence of these trenches was observed during this study. The trenches may have been eroded during flash flooding, which commonly happens in the wash during precipitation. In 2016, the wash and much of the vegetation at Bitter Spring was concentrated along the southeastern part of the wash. Comparison of historical imagery showed that the area of vegetation changes because of flash floods, several of which happened during this study. Several small ponds of water and surface flow are present in two main incised channels that were about 5 ft below the land surface where the vegetation is well established. It is possible that these ponds may be the remnants of the trenches described by Bowen (1943). A well was canvassed by USGS in 1965 but was not located in 2016.

A borehole log from nearby multiple-well monitoring site CRTH2 (site name 013N005E08B001–2S; Kjos and others, 2014; U.S. Geological Survey, 2017; Nawikas and others, 2019), located about 2.5 mi northwest of Bitter Spring (fig. 19), indicated sandy, gravelly alluvial deposits overlie a 660-ft-thick clay deposit from 200 to about 860 ft bls (Kjos and others, 2014) and was consistent with the lithology observed in the hills east of the spring. The top of the clay deposit at monitoring site CRTH2 is at an altitude of 1,230 ft above NAVD 88, which is about 100 ft below the altitude of about 1,330 ft where the clay layer is observed along the northern wall of the wash at Bitter Spring.

Jack Spring

Jack Spring is near the northwest end of the Coyote Lake Valley groundwater basin (fig. 2). The spring is about 0.3 mi east of Fort Irwin Road that connects the NTC to Barstow (figs. 1, 22). Land-surface altitude of Jack Spring ranges from about 2,390 ft above NAVD 88 along the western extent to about 2,400 ft above NAVD 88 along the eastern extent. The spring lies within a zone of small faults along the Coyote Lake Fault and between sandy washes that cut through low hills consisting of plutonic and metamorphic (schistose) rocks (Miller and others, 2014) that are offset by faulting (fig. 22). These washes flow during extreme precipitation.

The washes that bound Jack Spring drain surface flows from the low hills that separate Irwin subbasin, to the north, from Coyote Lake Valley to the south, where Coyote Lake (dry) playa is located (fig. 2). The plutonic and metamorphic (schistose) rocks weather into coarse-grained alluvial deposits. Several faults cross this area and are likely splays of the Coyote Lake Fault (figs. 22, 23). Jack Spring appears as three separate seeps defined as east, middle, and west by areas of vegetation that are within about 500 ft (fig. 24). In this report, the east, middle, and west seeps are referred to as “Jack Spring East,” “Jack Spring,” and “Jack Spring West,” respectively. Jack Spring East supports large trees, cattails, grasses, and small shrubs; Jack Spring supports grasses and small shrubs; and Jack Spring West is a mounded region that primarily supports grasses and some cattails (fig. 24). These three areas of extensive vegetation indicate that shallow groundwater is in this area where shallow bedrock forces and joints and fractures in the rocks provide conduits for groundwater to flow to the surface.

Garlic Spring

At Garlic Spring, the 2015 and 2017 ERT model comparisons (2017 resistivity–2015 resistivity; figs. 26A, 26B) showed larger inverted resistivity differences than the apparent resistivity analysis indicated, although both model and apparent resistivity comparisons indicated a lack of clear depth dependent change. The 2015 and 2017 models generally resolve the same geometry of the resistivity structure, and differences primarily are attributed to higher relative contrasts in the 2017 models, possibly resulting from model non-uniqueness in independent inversions in a complex, faulted geologic setting. However, discharge measurements of Garlic Spring taken in December 2015 and March 2016 (Mesmer and others, 2024) indicate that spring discharge may be higher in spring than in fall, which could support the interpretation that the more conductive areas in the 2017 ERT model results may indicate higher water content.

The ERT models showed two shallow conductive bodies (1–8 ohm-m; indicated by the darker blue areas) near Garlic Spring (figs. 26A, 26B). One conductive body lies just beneath the surface at the location of a shallow pond and wetland area (at about 82 ft), and one lies beneath the surface at the uphill end of the vegetated area (at about 5 m; figs. 26A, 26B). Both conductive bodies were bordered by highly resistive bodies (200–1000 ohm-m) representing low-permeability bedrock (figs. 26A, 26B). In 2017, an additional long profile was collected to include the mapped fault in the resistivity profile (fig. 27A). The long resistivity profile confirmed that the uphill conductive region coincides with the location of the Garlic Spring Fault. Together, the position of the conductive bodies relative to the Garlic Spring Fault, the pond and wetland, and the vegetated area strongly indicated these bodies are higher-porosity saturated zones or fault splays and serve as conduits for groundwater feeding this spring, whereas the highly resistive bodies bordering the fault splays are blocks of lower porosity, lower water-content bedrock.

Bitter Spring

At the two resistivity profiles downstream from Bitter Spring (figs. 28A, 28B, 29A, 29B), the ERT model comparison showed clear resistivity changes in the upper 0–9 ft in the wash. Model subtractions of 2015 and 2017 data at Bitter Spring 1 and 2 resistivity profiles (figs. 28C, 29C) showed a negative change in resistivity (more conductive, indicating wetter) in much of the upper 6–9 ft, with smaller areas showing a positive change in resistivity (more resistive, indicating drier). In both resistivity profiles, the larger, negative changes were measured in the most resistive regions of the resistivity profile, which correlate to sandy areas on the surface in the wash. These results were supported by the apparent resistivity patterns—apparent resistivity data became generally more conductive (less resistive) between 2015 and 2017 and Δρ_a was most strongly negative at small electrode spacing (figs. 25B, 25C), indicating the largest changes near the surface. These data indicate that much of the upper 0–9 ft of sandy material along the wash was wetter or more saline during the 2017 survey.

The subsurface beneath the Bitter Spring resistivity lines was a relatively simple 2-layer structure, with a moderately resistive layer (20–1,000 ohm-m), at approximately 9–15 ft depth, likely representing wash deposits sitting on top of a generally homogeneous, highly conductive region (1–8 ohm-m), presumably associated with the clay layer observed in the northern wall of the wash at Bitter Spring (see “Bitter Spring” subsection in the “Description of Study Areas” section). As previously stated, Bitter Spring was formed by groundwater draining from the Cronise Valley groundwater basin being forced to the surface by the Bicycle Lake Fault Zone and above the tight clay. The high conductivity of the deeper subsurface here could be attributed to elevated groundwater or soil salinity at depth, which is consistent with the relatively high TDS concentrations (greater than 3,000 milligrams per liter [mg/L]) measured in Bitter Spring (see the “Water Quality” section). The low resistivity values seen in the ERT models generally are inconsistent with dry, non-saline sand below 9 ft. Additional subsurface data such as borehole lithology would be helpful in determining the nature of the conductive unit at depth.

Jack Spring

Electrical resistivity tomography data were collected along three profiles at Jack Spring. The ERT surveys were numbered based on the order in which they were completed; Jack Spring 1 (JS1) was done at Jack Spring and Jack Spring 2 (JS2) and Jack Spring 2 Cross (JS2C) was done at the west seep (Jack Spring West). An ERT survey was not done at Jack Spring East.

Jack Spring 1 (JS1)

At Jack Spring 1 (JS1), ERT data strongly indicated that the electrode profiles were offset by about 16 ft between the 2015 and 2017 surveys (fig. 30). In the 2015 ERT model (fig. 30A), there was a conductive (low resistivity) vertical body at a lateral position about 125 ft that connected the conductive surface to a conductive body below 16 ft. In the 2017 model (fig. 30B), this conductive body appeared but was offset about 16 ft to the south, as was a conductive body at a lateral position of 39–52 ft. Although this is a tectonically active region, only a few small (less than magnitude 3) earthquakes were reported within a 15-mile radius during this study (U.S. Geological Survey, 2022). The magnitudes of these earthquakes were believed to not be strong enough to shift the faults and subsurface bodies 16 ft. It was more likely that the resistivity lines were offset by 16 ft. This kind of resistivity line positioning error could be caused by GPS accuracy differences between surveys or erroneous coordinates. The potential for lateral offset between profiles precluded subtracting models or direct comparisons between 2015 and 2017. Apparent resistivity patterns also were non-conclusive for the same reason.

A 16-ft offset was applied to the plots for the 2015 and 2017 models so notable features matched (fig. 30). Qualitative comparison of the two models showed relatively little change in resistivity. Both models indicate a complex structure: the southern half of the model generally was a moderately resistive body (12–80 ohm-m) with a conductive body sitting from 8- to 16-ft depth at a lateral position of roughly 26–52 ft, whereas the northern half of the model showed a three-layer structure, consisting of (from shallowest to deepest) a conductive surface (1–12 ohm-m), a moderately resistive slope-parallel layer from 9- to 16-ft depth, and a conductive subsurface beneath (1–5 ohm-m). A vertical body of conductive material connects the subsurface to the surface at a lateral position about 125 ft. We hypothesize that this vertical body represents a higher porosity saturated zone, likely a fault splay, that serves as a conduit to transfer groundwater to the surface. Although there was no obvious flowing groundwater discharge along this profile immediately downhill from this feature, the presence of vegetation, primarily grasses, indicate focused spring discharge. The highly conductive body at lateral position 66–88 ft in the 2015 model was suspect because of data-quality issues.

Jack Spring 2 (JS2)

At Jack Spring 2 (JS2), 2015 ERT data were collected on a profile that followed an elevation contour across a sandy, partially vegetated mound that forms the main geomorphic feature of the spring. Groundwater discharge was visible in puddles 7–26 ft along the southern edge of the mound and south of this profile. ERT models for the 2015 and 2017 surveys (figs. 31A, 31B) showed a three-layer structure: from shallowest to deepest, a conductive (9–30 ohm-m) layer in the upper 10–13 ft, a resistive (80–800 ohm-m) layer between 13- and 33-ft depth, and a conductive layer beneath that, with the exception of a conductive (1–3 ohm-m) body at lateral position 125–164 ft. Model subtraction (fig. 31C) showed that the most pronounced differences in resistivity appeared below 9 ft between 2015 and 2017. Similar to the differences seen between ERT surveys at Garlic Springs, the 2015 and 2017 surveys showed similar resistivity geometry and could be related to inversion differences in this more complex setting. Apparent resistivity and inversion results indicated that the shallow conditions above 9 ft near the Jack Spring West seep did not notably change between surveys.

The Jack Spring West seep is along a splay of the Coyote Lake Fault that runs approximately east–west through a shallow wash (fig. 23). The Jack Spring 2 resistivity profile (JS2) is parallel to the fault. To aid in understanding the three-dimensional hydrologic structure of Jack Spring West seep and how it relates to the Coyote Lake Fault, a supplemental cross profile (JS2C 2017), perpendicular to the existing resistivity profile, was created (fig. 32) to bisect the fault. The ERT model of this cross-resistivity profile (JS2C 2017) showed that the north and south sides of the Coyote Lake Fault have similar resistivity (20–90 ohm-m), whereas the surface on the hill above the spring is conductive (1–5 ohm-m), likely due to copious natural salts occurring on the hill surface. Notably, there was a conductive vertical region that coincided with the surface location of the fault and the groundwater discharge zone (fig. 32). The ERT data indicate that the fault acts as a conduit for groundwater to reach the surface. The fault geometry is unknown at depth but is hypothesized to be high-angle based on regional geology (Miller and others, 2014).

Garlic Spring

Qualitative analysis of the temperature profiles at each of the instrumented sites indicated that variably saturated conditions likely existed throughout the period of temperature profile collection rather than just during continuously saturated conditions. In this study, variably saturated conditions were defined as periods of high soil moisture but not full saturation of spring sediments at the collected depths. Overall, the temperature profiles showed that the diurnal signal dominated the temperature with depth (figs. 33A–D). The phase shift in the signal varied depending on the time of year. The temperatures in December and February showed a similar phase shift with depth, whereas temperatures in April showed a greater separation between the temperatures in the shallow logger and the temperatures in the deeper loggers when the air temperature and plant growth increased. The data indicate shallow preferential flow (that is, rapid movement of groundwater between the sensors) rather than advective upwelling zones (as evidenced by the lack of a distinct phase shift with depth in the temperature record and no damping with depth of the temperature signal; Briggs and others, 2014). We assumed that the upwelling zone of the spring sites is upgradient from or adjacent to the drive points at Garlic, Bitter, and Jack Springs and that groundwater flowed along shallow preferential flow paths and discharged in visible areas of developed vegetation (figs. 18, 21, 24). During winter 2015, for example, diffuse discharge was observed east of Garlic Spring West (fig. 17) after vegetation was cleared by the NTC personnel. The purpose of the vegetation removal was an attempt at removing invasive species. The vegetation removal revealed saturated conditions, with surface discharge happening downslope of the cleared area. A seepage meter measurement was collected at Garlic Spring (Mesmer and others, 2024) in the cleared area where water upwelled. The discharge from these clear areas was substantial but spread out over a large, diffuse area; however, by April 2016, the vegetation had grown back, and water upwelling could no longer be seen.

Total Dissolved Solids and Chloride Concentrations

Total dissolved solids and chloride concentrations were used to further describe the areal variation in the water quality of spring discharge. Total dissolved solids commonly are used as a qualitative measure of groundwater salinity, as defined by Robinove and others (1958). In this report, groundwater salinity is defined by five ranges of TDS concentration: (1) less than 1,000 mg/L, fresh water; (2) 1,000–3,000 mg/L, slightly saline water; (3) greater than 3,000–10,000 mg/L, moderately saline water; (4) greater than 10,000–35,000 mg/L, highly saline water; and (5) greater than 35,000 mg/L, briny water. Salinity of sampled spring water ranged from fresh in upland springs (Cave, Desert King, Devouge, and Panther Springs) to slightly saline in groundwater basin springs (Jack and Garlic Springs) and moderately saline in Bitter Spring (U.S. Geological Survey, 2017).

The TDS concentrations in water samples from upland springs ranged from 446 mg/L at Devouge Spring to 556 mg/L at Desert King Spring, whereas concentrations in samples from groundwater basin springs ranged from 584 mg/L at Jack Spring East to 3,340 mg/L at Bitter Spring. The chloride concentrations in water samples from upland springs ranged from 42.5 mg/L at Panther Spring to 85.6 mg/L at Desert King Spring, whereas concentrations in samples from groundwater basin springs ranged from 69.4 mg/L at Garlic Spring to 685 mg/L at Bitter Spring (U.S. Geological Survey, 2017).

The highest TDS and chloride concentrations were measured in a sample collected from Bitter Spring. The concentrations in 2016 greatly exceeded concentrations in a sample collected from Bitter Spring in 1918 (Thompson, 1929, p. 545). The sample collected on February 26, 1918, had a TDS concentration of 1,628 mg/L (3,340 mg/L in sample collected on April 19, 2016) and a chloride concentration of 246 mg/L (385 mg/L in sample collected on April 19, 2016; U.S. Geological Survey, 2017). The moderately saline (TDS=3,340 mg/L) water discharging at Bitter Spring likely originates from buried lacustrine deposits in the Cronise Valley groundwater basin. These deposits are present in nearby multiple-well monitoring site CRTH2 (fig. 2), where the shallow well CRTH2 #2 is perforated from 270 to 290 ft bls in a 660-ft-thick clay deposit and in the Tertiary, partly consolidated clay deposits at the base of the hills bordering the wash where Bitter Spring appears (Kjos and others, 2014). Groundwater quality likely indicates remobilization of buried salts from these lacustrine deposits. The lower salinity water (TDS=3,340 mg/L) from Bitter Spring in comparison to the highly saline groundwater (TDS=13,400 mg/L in 2014) from shallow well CRTH2 #2 and slightly saline groundwater (TDS=829 mg/L in 2014) from deep well CRTH2 #1 (perforated from 920 to 940 ft bls) indicates that water discharging from Bitter Spring is a mixture of deep and shallow groundwater and a small percentage of surface water from occasional storms. The increase in TDS and chloride concentrations at Bitter Spring since 1918, indicate that in 1918, a greater proportion of shallow groundwater was mixing with deep groundwater discharging from the aquifer. Increased evaporation at the spring is another possible explanation for the increasing TDS and chloride concentrations. Further evidence of the mixture of source waters is provided in the “Stable Isotopes of Oxygen and Hydrogen” and “Tritium and Carbon-14” sections.

The lowest TDS and chloride concentrations measured in this study were in water samples from upland springs (Panther, Devouge, and Cave Springs) that are fed primarily by precipitation recharge infiltrating through the fractured rock in the mountains and not by groundwater discharge from nearby groundwater basins that interacts with unconsolidated, coarse-grained deposits interbedded with clays and salts. Although the TDS concentration in the 2016 water sample from Desert King Spring also was relatively low (556 mg/L, U.S. Geological Survey, 2017), the chloride concentration (85.6 mg/L) was slightly higher than the concentrations at the other upland springs (range=42.5–61 mg/L).

Nitrate Concentrations, Redox Status, and Nitrogen and Oxygen Isotopes of Nitrate

Nitrate plus nitrite as nitrogen (NO₃+NO₂ as N) concentrations were measured for most water samples collected from springs during this study. However, nitrite typically is not detected in the water samples, less than laboratory reporting levels of 0.001 mg/L or concentrations are low relative to the total concentration of nitrate plus nitrite as N (less than 5 percent; U.S. Geological Survey, 2017), so the nitrate plus nitrite as N concentrations represent nitrate concentrations and hereinafter are referred to as just “nitrate.” Nitrate concentrations in water samples collected during 2016 varied from less than the reporting level of 0.040 mg/L at Jack, Garlic, and Bitter Springs to 17.3 mg/L at Desert King Spring. Nitrate concentration in a water sample from Desert King Spring in October 1917 was reported at 8.4 mg/L (Thompson, 1929; U.S. Geological Survey, 2017110). Nitrate concentrations commonly differed at springs with multiple sample collection sites (that is, different seeps at an individual spring). For example, nitrate concentrations in samples from Jack Spring East were 9.1 and 9.6 mg/L, whereas concentrations from Jack Spring and Jack Spring West were 0.4 mg/L and less than the reporting level of 0.040 mg/L, respectively (U.S. Geological Survey, 2017).

Nitrate derived naturally from precipitation, soils, and geologic sources generally occurs in groundwater at concentrations ranging from less than 1 to 3 mg/L (Mueller and Helsel, 1996; Dwivedi and others, 2007; Dubrovsky and others, 2010; Izbicki and others, 2015). Devouge, Desert King, and Jack Spring East had at least one sample with nitrate concentrations greater than 3 mg/L; the Desert King Spring sample (17.3 mg/L) exceeded the U.S. Environmental Protection Agency (EPA) maximum contaminant level (MCL; U.S. Environmental Protection Agency, 2023) of 10 mg/L for nitrate as N. Most springs at the NTC are remote, and nitrate concentrations greater than 3 mg/L could result from animal activity, most notably wild donkeys (Equus asinus), which are not native to the Mojave Desert. At the time of this study (2016), most springs were fenced for protection from animals, including Desert King and Devouge Springs, which had nitrate concentrations as high as 17.3 and 8.6 mg/L, respectively (U.S. Geological Survey, 2017). Garlic and Bitter Springs were accessible through breaks in the fence. Jack Spring East and Jack Spring West were not fenced. Evidence of wild donkeys was present at Garlic, Bitter, and Jack Springs.

Orthophosphate as phosphorous (henceforth, orthophosphate) concentrations, when elevated, are commonly associated with animal waste but were low in 2016 and ranged from 0.01 mg/L at Jack, Garlic, and Devouge Springs to 0.14 mg/L at Panther Spring, with all samples having detectable concentrations of orthophosphate (U.S. Geological Survey, 2017). The highest orthophosphate concentration (0.14 mg/L) was in water from Panther Spring. The spring is fenced to keep donkeys out, although evidence of donkeys was present upslope of the fenced area. Nitrate concentrations were not detectable in Panther Spring.

Nitrogen in animal waste can be oxidized to ammonium and then to nitrate in the presence of dissolved oxygen and bacteria. In water, the nitrate can be reduced to nitrogen gas (dissolved but removed from the system when in contact with the atmosphere) in the absence of dissolved oxygen. For the purposes of this study, the reporting limit for dissolved oxygen was 0.2 mg/L. Except for Jack Spring, most sampled springs at the NTC contained measurable dissolved oxygen. Dissolved oxygen concentrations among other sampled springs were as low as 0.8 mg/L in Bitter Spring, but concentrations varied widely, as seen in Panther Spring where dissolved oxygen concentrations ranged from 0.4 to 7.7 mg/L in the 2016 samples. Springs are complex environments that may be oxygenated near the surface and reduced (that is, lacking dissolved oxygen) at depth, with redox conditions changing throughout the day in response to spring discharge, productivity from biogeochemical reactions within the spring, and possibly other factors. Consequently, collecting representative samples from springs and seeps with small diffuse discharges can be difficult, and individual samples may not completely characterize the complex geochemical and redox environment at small springs and seeps.

McMahon and Chapelle (2007) used iron and manganese concentrations to evaluate redox conditions within groundwater in the absence of dissolved oxygen data. In their study, iron concentrations greater than 100 µg/L and manganese concentrations greater than 50 µg/L were evidence of reducing conditions (dissolved oxygen absent) within groundwater. Although other redox processes may occur in aquifer systems, constituents diagnostic of other redox processes are not routinely measured (McMahon and others, 2009). Therefore, in this study, iron and manganese concentrations were used to determine if spring water containing dissolved oxygen may have been exposed to reducing conditions elsewhere within the spring. Spring water that has dissolved-oxygen concentrations greater than 1 mg/L but iron concentrations greater than 100 µg/L or manganese concentrations greater than 50 µg/L in one or more samples were classified as “suboxic” or as “variably oxic” if not all samples were consistent with iron and manganese reducing conditions within the spring (McMahon and Chapelle, 2007; Jurgens and others, 2009).

During this study, iron concentrations ranged from less than the laboratory reporting level (LRL) of 4.0 µg/L in Desert King Spring and Cave Spring to 1,350 µg/L in Bitter Spring; iron concentrations greater than 100 µg/L indicated reduced conditions in one or more samples from Jack Spring West, Bitter Spring, and Panther Spring. Manganese concentrations ranged from less than the LRL of 0.04 µg/L in Cave Spring to 1,240 µg/L in Panther Spring; concentrations greater than 50 µg/L indicated reduced conditions in one or more samples from Jack Spring East, Jack Spring West, Garlic Spring East, Garlic Spring West, Bitter Spring, and Panther Spring. Suboxic or variably oxic conditions were present in samples from Jack Spring East, Jack Spring West, Bitter Spring, Garlic Spring East, Garlic Spring West, and Panther Springs. Dissolved-oxygen concentrations less than 0.2 mg/L and reduced conditions relative to iron and manganese were present in water in Jack Spring. Oxic conditions were consistently present in samples from Garlic, Devouge, Desert King, and Cave Springs.

In addition to chemical constituents, water samples from selected springs also were analyzed for the delta nitrogen-15 (δ¹⁵N) in nitrate (δ¹⁵N-NO₃) and the delta oxygen-18 (δ¹⁸O) in nitrate (δ¹⁸O-NO₃). Although nitrogen is biogeochemically reactive and the nitrogen cycle is complex, isotopes of nitrogen and oxygen can provide insight into potential sources of nitrate and biochemical processes that may have altered nitrate chemistry and isotopic composition in the hydrologic system. Nitrate concentrations above the natural range (0–3 mg/L) and δ¹⁵N-NO₃ compositions greater than about 6 parts per thousand (per mil) indicated that in some springs samples, human waste could be a potential source of nitrate (Hinkle and others, 2008), and δ¹⁵N-NO₃ compositions greater than 10 per mil indicate than animal waste could be a potential source of nitrate (Kendall, 1998).

A geochemical framework model incorporating changes in nitrogen concentration and isotope composition (fig. 36A) was used to evaluate potential sources of nitrate concentrations above the natural range in some of the springs. δ¹⁵N-NO₃ data were used to evaluate the environmental history of nitrogen with respect to microbially mediated denitrification and subsequent removal of nitrate. δ¹⁸O-NO₃ data were used to determine whether the source of nitrate was from precipitation or if nitrate was from the microbially mediated conversion of nitrogen to ammonia in animal waste (ammonification) or from the fixation from the atmosphere by algae or vegetation (such as cattails) within a spring. The primary limitation on collection and analysis of nitrogen and oxygen isotopes of nitrate was that the sample had to have at least 0.06 mg/L of nitrate as N (Coplen and others, 2012). None of the samples had enough ammonia to analyze for the δ¹⁵N composition of ammonium (δ¹⁵N-NH₄).

δ¹⁵N-NO₃ values in spring water ranged from 5.8 to 20.2 per mil (Devouge and Jack Springs, respectively; U.S. Geological Survey, 2017; fig. 36). Higher δ¹⁵N-NO₃ values were associated with reduced, suboxic, or variably oxic redox conditions, whereas lower values were associated with oxic conditions. Nitrate concentrations generally decreased with increasing δ¹⁵N-NO₃ values (fig. 36A). The δ¹⁵N-NO₃ composition of animal waste, presumably as ammonia or organic nitrogen, ranges from 0 to 35 per mil, with most values between 10 and 20 per mil (fig. 36A; Kendall, 1998). Data were not available for the δ¹⁵N-NO₃ composition of donkey waste at the NTC. Groundwater samples affected by wastewater in Irwin subbasin had δ¹⁵N-NO₃ values ranging from 10 to 20 per mil (Densmore and Böhlke, 2000). The range of nitrate directly measured in samples of human waste is narrow; the δ¹⁵N-NO₃ value for wastewater from human sources was about 7 per mil (Hinkle and others, 2008) and generally is representative of sewage. Higher values in the Irwin subbasin likely reflect a combination of denitrification and microbially mediated changes in nitrogen isotope composition of nitrate during treatment and infiltration from ponds.

δ¹⁵N-NO₃ values commonly increase during nitrification as a result of transformation of ammonia from animal waste. δ¹⁵N-NO₃ values in water from Jack Spring, Jack Spring East, and Garlic Spring East, all of which had evidence of donkey activity, were within the published ranges for animal waste (Kendall, 1998). However, because of its wide range in composition, δ¹⁵N-NO₃ generally is not definitively diagnostic of nitrogen sources when used alone. Additionally, water from Jack Spring, Jack Spring East, and Garlic Spring was reduced or suboxic and may have been partly denitrified, lowering nitrate concentrations and increasing δ¹⁵N-NO₃ values.

Decreases in NO₃ as N concentrations with increases in δ¹⁵N-NO₃ values in suboxic/variably oxic or reduced spring water are consistent with denitrification. Assuming a reductive fractionation factor of α=−29.4 derived from laboratory data (Mariotti and others, 1981), denitrification trend lines were calculated that show decreases in nitrate concentrations expected as δ¹⁵N-NO₃ values increased as a result of possible denitrification for selected springs (fig. 36A).

δ¹⁸O-NO₃ data were used to provide additional evidence of denitrification at selected springs, evaluate the initial δ¹⁵N-NO₃ composition, and evaluate the potential source of nitrate in spring water. δ¹⁸O-NO₃ values in spring water ranged from 1.65 to 11.16 per mil (for Devouge and Jack Springs, respectively; U.S. Geological Survey, 2017). The two sites with the highest values of δ¹⁸O-NO₃ (10.42–11.16 per mil at Jack Spring and Jack Spring East) had suboxic and reduced water; Devouge, Cave, Desert King Springs, and Garlic Spring East had lower values (1.65–3.74 per mil) and had oxic and suboxic water (fig. 36B). Denitrification happens along a trend having a slope (Δx to Δy) of 0.5 (1:2). The three lines shown on figure 36B represent potential denitrification through selected wells.

The δ¹⁸O composition of water (δ¹⁸O-H₂O) in the springs ranged from −11.7 to −9.4 per mil (Garlic and Jack Springs, respectively), with a median value of −10.9 per mil. Given that during nitrification one-third of the oxygen in nitrate is contributed from dissolved oxygen within the water and two-thirds of the oxygen is contributed from hydrolysis of the water molecule (Mayer and others, 2001) and given a δ¹⁸O composition of atmospheric oxygen of 23 per mil (Mayer and others, 2001), the initial δ¹⁸O-NO₃ composition set during nitrification of ammonia from animal waste would be 0.3 per mil (blue line on fig. 36B). Projecting measured δ¹⁵N-NO₃ values in Jack Spring along a denitrification trend line to the calculated initial δ¹⁸O-NO₃ composition (intercept of δ¹⁸O-NO₃ composition set during nitrification and denitrification trend line; fig. 36B), an initial δ¹⁵N-NO₃ composition of −1 per mil would be expected in Jack Spring. A value of −1 per mil is outside the literature reported range for δ¹⁵N-NO₃ in most animal waste (Kendall, 1998), and the data are not consistent with an animal waste origin for nitrate; the value is consistent with nitrate from desert nitrate deposits (Böhlke and others, 1997) observed at the site. In contrast, δ¹⁵N-NO₃ and δ¹⁸O-NO₃ data from Garlic Spring differ from data at Jack Spring (fig. 36B). Projecting measured δ¹⁵N-NO₃ values from Garlic Spring along a denitrification trend line to the calculated initial δ¹⁸O-NO₃ composition (fig. 36B), the initial δ¹⁵N-NO₃ composition in Garlic Spring would have been about 7 per mil. A value of 7 per mil is outside the literature reported range for δ¹⁵N-NO₃ in most animal waste (Kendall, 1998) but is consistent with a human origin for nitrate (Hinkle and others, 2008) and may be associated with treated wastewater discharges in the Irwin groundwater subbasin about 0.5 miles upstream (fig. 16). Increases in flow at Garlic Spring, expansion of areal extent of vegetated area, and changes in major-ion data through time also are consistent with treated wastewater from the Irwin subbasin that may have reached Garlic Spring.

Water from Desert King, Cave, and Devouge Springs is oxic and nitrate in spring water likely has not been extensively denitrified in the same manner as water from Jack or Garlic Springs (fig. 36). Measured δ¹⁵N-NO₃ values from Cave and Devouge Springs are within the range of compositions expected for nitrate fixed from algae (fig. 36B) and closely approximate the median composition from algal sources (3.4 per mil δ¹⁵N-NO₃) estimated by Swart and others (2014). However, the δ¹⁵N- NO₃ compositions of algae are poorly described in the literature and likely differ among species, and an algal nitrate source was not definitive for Cave and Devouge Springs because of the overlap with the range of desert nitrate deposits. Although nitrate concentrations are higher and δ¹⁵N-NO₃ and δ¹⁸O-NO₃ compositions differ from those in Cave and Devouge Springs, algal sources also are possible for nitrate in Desert King Spring (fig. 36). Desert King Spring is fenced to protect it from donkeys and exclude animal waste sources.

The geochemical framework model used to estimate nitrate sources and processes affecting nitrate concentrations within sampled springs is not definitive. The ranges in δ¹⁵N-NO₃ isotope compositions are often broad and, in many cases, especially for nitrate from algal sources, poorly defined. Although isotope composition of atmospheric oxygen is known, possible fractionation of oxygen as it dissolves in water and reacts within the spring system were not considered in the calculations presented. However, the data indicate that sources of nitrate other than animal waste are possible in reduced water from Jack and Garlic Springs, and Desert King Spring, the spring having the highest nitrate concentrations. Interpretation of isotope data, especially δ¹⁸O-NO₃ data (fig. 36B), indicate that additional management efforts intended to exclude donkeys from these springs may not yield the desired water-quality results.

Stable Isotopes of Oxygen and Hydrogen

More negative values of δ¹⁸O and δD represent enrichment in the lighter isotopes ¹⁶O and ¹H (or a depletion in the heavier isotopes ¹⁸O and D), and less negative values represent enrichment in the heavier isotope of ¹⁸O and D.

There was no further change in isotopic composition at low temperatures of most groundwater systems at Fort Irwin after recharged water had migrated below the depth of evaporation. If groundwater flows to a depth where geothermal heat has raised water temperatures sufficiently, isotopic exchange between the oxygen isotopes in water and the large reservoir of oxygen in rocks is possible, complicating the interpretation of ¹⁸O isotopes in samples. However, this process does not appear to affect the isotopic composition of the spring samples (Smith and others, 2002). Therefore, any subsequent differences in the isotopic composition of groundwater along a flow path between the points of recharge and discharge generally reflect mixing within the aquifer system or concentration by evaporation. As a consequence, the δ¹⁸O and δD composition of groundwater relative to the global meteoric water line (Craig, 1961) and the isotopic composition of other potential source waters can be an indicator of the source of groundwater to the sampled springs.

The δ¹⁸O and δD composition of 28 groundwater samples from 7 springs at the NTC ranged from −9.37 to −11.88 per mil and from −79.3 to −95.5 per mil, respectively (fig. 37A). The isotopic data of spring samples plotted to the right of the global meteoric water line (Craig, 1961) and indicated possible partial evaporation during precipitation (before recharge), possible evaporation at or below land surface before recharge or at the spring discharge location, or a “local” meteoric water line that is sub-parallel to and differs slightly from the global meteoric water line. The isotopic compositions of groundwater from sampled wells in the Irwin subbasin (Densmore and Londquist, 1997), representing groundwater samples with an evaporated signature, were plotted for comparison (fig. 37A). A volume-weighted sample of averaged desert winter precipitation (Izbicki, 2004), representing modern-day (2004) precipitation in the western Mojave Desert, plotted near the global meteoric water line (fig. 37A). A volume-weighted sample of local precipitation at Daggett, California (fig. 1; Friedman and others, 1992), plotted to the right of the global meteoric water line and to the upper right of the spring samples (fig. 37A). The Daggett sample, representing the isotopic composition of precipitation collected during summer 1985 through winter 1987, was notably different from the weighted average composition for winter precipitation in the Mojave Desert. A sample of local precipitation at Goldstone weather station (fig. 1), plotted to the right of the spring samples and evaporated groundwater (fig. 37A). The variation in isotopic compositions in the precipitation samples illustrate the seasonal variability of the isotopic composition of precipitation. More recent precipitation samples were not available for comparison.

The water samples from the springs plotted into three main groups (circled and numbered on fig. 37A): (1) water samples with lighter isotopic composition (more negative), representative of groundwater inflow from nearby groundwater basins; (2) water samples with heavier isotopic composition (more positive), representative of evaporated groundwater that plots along an evaporation trend line; and (3) water samples that represent young spring discharge that was recharged from local precipitation (fig. 37A). Water samples from Garlic and Bitter Springs and Jack Spring West plotted among group 1, which was consistent with a groundwater source in nearby unconsolidated deposits. Water samples from most wells in Irwin subbasin are isotopically light and plotted among group 1. Water samples from wells affected by artificial recharge from wastewater and irrigation-return flow (Densmore and Londquist, 1997) that have undergone evaporation were isotopically heavier and plotted in group 2. A water sample from Jack Spring, also groundwater sourced, plotted along an evaporation line representing group 2 from the middle site and group 3 from the east site. The water samples from Jack Spring East were a mixture of evaporated basin groundwater (evaporation line) and local precipitation (local meteoric water line). Water samples from Cave, Desert King, Devouge, and Panther Springs plotted among group 3, which was consistent with a local precipitation source. Overall, the δ¹⁸O and δD data show the isotopic compositions in upland springs were consistent with recharge from winter precipitation, and compositions in groundwater basin springs were consistent with groundwater sampled from nearby wells. Summer precipitation from monsoonal storms often flowed away quickly overland and seemed to contribute little water to discharge at the springs.

Water samples from Bitter Spring were isotopically light, plotted with groundwater samples, and were isotopically similar to water from deep well CRTH2 #2 at multiple-well monitoring site CRTH2 (fig. 37). The relation between specific conductance (as a surrogate for TDS) and δD has been used to determine if dissolution of soluble minerals or concentration by partial evaporation is the predominant cause of increased salinity in groundwater (Densmore and Londquist, 1997). If evaporation is the only process concentrating salinity, δD should increase more rapidly than with increasing specific conductance. If dissolution of evaporites is the only process increasing salinity, δD will not change as salinity increases along a mixing trend. If dissolution of evaporites is the dominant process increasing salinity, δD increases less rapidly than it does with evaporation and δD and specific conductance would not be strongly correlated. Comparison of specific conductance and δD indicated that water from Bitter Spring had higher specific conductance relative to δD than water from the deep well CRTH2 #1 and lower specific conductance relative to δD than water from shallow well CRTH2 #2 (fig. 37B). This difference is consistent with a mixture of water from shallow well CRTH2 #2 and deeper well CRTH2 #1. However, the mixture of shallow and deep groundwater discharging at Bitter Spring alone does not explain the complex story of the sources feeding discharge at Bitter Spring. Further evidence is provided in the “Tritium and Carbon-14” section.

Tritium and Carbon-14

The radioactive isotopes of hydrogen (tritium) and carbon (carbon-14 or ¹⁴C) were used to determine the age (time since recharge) of groundwater, locate sources of recharge, and identify geologic controls on the movement of groundwater that discharges at sampled springs. Tritium and ¹⁴C can be used to help determine the age and movement of groundwater discharging to these springs (that is, older and younger springs), which can provide insight into sustainability of the springs. Older groundwater at springs typically is related to regional groundwater flow systems; whereas younger groundwater at springs typically is locally recharged groundwater and can be more sensitive to long-term dry conditions. In general, all spring discharges measured during 2015–16 were less than spring discharges reported by Thompson (1929).

Tritium is a naturally occurring radioactive isotope of hydrogen with a half-life of 12.3 years (Lucas and Unterweger, 2000). The concentration of tritium typically is reported in tritium units (TU). About 760 pounds of tritium was released into the atmosphere between 1952 and 1962 because of the atmospheric testing of nuclear weapons, with atmospheric concentrations of tritium reaching a maximum of more than 1,000 TU in 1963–64 (Michel, 1976). As a result, tritium concentrations in precipitation and groundwater recharge increased during 1963–64. Tritium concentrations are not affected by chemical reactions other than radioactive decay because tritium is part of the water molecule. The process of radioactive decay means tritium can be used to trace the movement and relative age of water to about 70 years before present (pre-1952).

Tritium concentrations in modern-day (2012) precipitation in the part of the Mojave Desert that includes the NTC are about 5 TU (Eastoe and others, 2012). In this report, groundwater samples with ³H concentrations less than the detection limit of 0.20 TU were interpreted as water recharged before 1952, and groundwater samples with greater concentrations (measurable ³H) were interpreted as waters recharged after 1952, or recent recharge. Eastoe and others (2012) predict the threshold of ambiguity for using tritium alone as a tool for mapping aquifer recharge will occur between 2025 and 2030.

Carbon-14 is a naturally occurring radioactive isotope of carbon that has a half-life of about 5,730 years (Mook, 1980). Carbon-14 data are expressed as percent modern carbon (pmC) by comparing ¹⁴C activities to the specific activity of National Bureau of Standards oxalic acid: 13.56 disintegrations per minute per gram of carbon in the year 1950 equals 100 pmC (Kalin, 2000). Carbon-14 above natural levels was produced, as was tritium, by the atmospheric testing of nuclear weapons (Mook, 1980), and as a result, ¹⁴C activities may exceed 100 pmC in areas where groundwater contains tritium. Carbon-14 activities are used to determine the age of a groundwater sample on timescales ranging from recent to more than 20,000 years before present. Carbon-14 is not part of the water molecule; therefore, ¹⁴C activities in groundwater may be affected by chemical reactions that remove or add carbon to solution or by mixing of younger groundwater that has high ¹⁴C activity with older groundwater that has low ¹⁴C activity. Izbicki and Michel (2004) reported an initial ¹⁴C activity of 85 pmC for groundwaters in the Mojave Desert region before atmospheric testing of nuclear weapons.

Carbon-14 ages presented in this report did not account for changes in ¹⁴C activities resulting from chemical reactions that remove or add carbon to solution or mixing and, therefore, were considered uncorrected (or apparent) ages. These ages were interpolated and not individually calculated. In general, uncorrected ¹⁴C ages were older than the actual age of the associated water due to mixing with radiocarbon “dead” dissolved inorganic carbon. Izbicki and others (1995) estimated that uncorrected ¹⁴C ages were as much as 30-percent older than the actual ages of groundwater in the regional aquifer in the Mojave River groundwater basin near Victorville, California (not shown), which is about 60-mi southwest of the study area. In this report, groundwater samples from wells or springs with ¹⁴C activities less than 85 pmC were interpreted as being recharged before 1952, whereas samples with ¹⁴C activities greater than 85 pmC were interpreted as being recharged after 1952 (Izbicki and Michel, 2004).

Tritium concentrations in 10 water samples from 6 springs (and associated spring sites) ranged from less than 0.20 (Cave, Devouge, and Panther Springs; Garlic Spring and Garlic Spring East; Jack Spring) to 0.44 TU (Jack Spring East; fig. 38; U.S. Geological Survey, 2017). Tritium concentrations less than 0.20 TU indicated that water discharging at these spring sites (Cave, Devouge, and Panther Springs and the east sites at Garlic Spring) was recharged before 1952, and greater concentrations (Garlic Spring West, Bitter Spring, Jack Spring East, and Jack Spring West) indicated that water discharging at these spring sites was (or had some part) recharged since the early 1950s. Measurable ³H (0.2 TU or greater) in water from Garlic Spring West and increased discharge at Garlic Spring (based on aerial imagery and major-ion data previously described) indicated that discharges of treated wastewater from the Irwin subbasin along the Garlic Spring Fault may have reached Garlic Spring West. Measurable ³H (0.2 TU or greater) in water from Bitter Spring and Jack Spring East and West indicated younger groundwater (likely originating as surface-water runoff from intermittent storms in nearby washes) could be mixed with primarily older groundwater flowing and ultimately discharging to these springs (based on ¹⁴C data discussed later in the text). During an extended dry period, there may not be surface-water runoff, and the discharge from Bitter and Jack Springs could decrease.

Uncorrected ¹⁴C data indicated the waters of these springs had apparent ages of about 700–6,300 years old and were much younger in comparison to apparent ages of about 12,000–40,000 years for groundwater sampled in nearby groundwater basins (Densmore and Londquist, 1997; Voronin and others, 2013; Densmore and others, 2018). Measured ¹⁴C activities in water samples from sampling points at the seven springs studied ranged from 46.75 pmC at Garlic Spring to 91.87 pmC at Garlic Spring West drive point (fig. 38; table 1; U.S. Geological Survey, 2017). The highest ¹⁴C activity (91.87 pmC) was measured in water from Garlic Spring West, at the western end of the series of seeps along the Garlic Spring Fault (fig. 17) and nearest to well WC3-60 at the southeastern-most end of the ponds where treated wastewater infiltrates in Irwin subbasin (fig. 16). The uncorrected ¹⁴C data indicated that groundwater at Garlic Spring West and well WC3-60 was about 700 years old. Conversely, the lowest ¹⁴C activity (46.75 pmC) was measured in water from Garlic Spring, at the eastern end of the seeps, where the uncorrected ¹⁴C data indicated that groundwater in this site was about 6,276 years old. Carbon-14 activities in water samples from all other springs were between these outliers.

Measured ¹⁴C activities in water samples from upland springs (Cave, Desert King, Devouge, and Panther Springs) ranged from 63.95 and 86.51 pmC (fig. 38). The ¹⁴C data (unadjusted from reactions with aquifer material) indicated that water at these springs were from about 3,700 to 1,200 years old. Water from Cave, Devouge, and Panther Springs did not contain measurable ³H (³H data were not available for Desert King Spring). Water from Panther Spring had a ¹⁴C activity of greater than 85 pmC, indicating it may contain recent water, even though there was no measurable tritium present in the sample. Thus, water from Panther Spring was interpreted to be a mixture of water from the fractured bedrock at the spring and local precipitation, with each source containing different ³H and ¹⁴C concentrations.

Water from Garlic Spring, with the highest and lowest ¹⁴C activities (91.87 and 46.75 pmC) of any sampled spring sites, had corresponding tritium concentrations of 0.33 and less than 0.20 TU in the western- and the eastern-most samples, respectively (fig. 38). These ¹⁴C activities (unadjusted for reactions with aquifer material) indicated that water from these two Garlic Spring sites was about 6,300 and 700 years old (46.75 and 91.87 pmC, respectively). The ¹⁴C activity (91.87 pmC) measured at Garlic Spring West, at an activity greater than 85 pmC, combined with measurable tritium (0.33 TU), indicated that water at the western edge of Garlic Spring contained a higher fraction of young water than the other seeps; however, a ¹⁴C activity (61.38 pmC) of less than 85 pmC measured at Garlic Spring East, combined with no measurable tritium, indicated that groundwater at the eastern edge of Garlic Spring was older. Water from Garlic Spring West was a mixture of water from wastewater-infiltration ponds in Irwin subbasin (based on ¹⁴C activities in water from well WC3-60 at the southeastern-most end of the ponds) and old groundwater, with each containing different ³H concentrations and ¹⁴C activities. These ¹⁴C data provide further support that Garlic Spring Fault acts, in part, as a conduit for groundwater flow. The fact that splays of the Garlic Spring Fault are bordered by blocks of nonporous bedrock, with splay geometries controlled by the architecture of the fault, supports the interpretation of the observed differences in ³H concentrations and ¹⁴C activities.

Water from Bitter Spring had a ¹⁴C activity of 81.5 pmC and a ³H concentration of 0.31 TU (fig. 38). The ¹⁴C activity (unadjusted for reactions with aquifer material) indicated that water at this spring was less than 1,700 years old. Based on stable isotope data and TDS concentrations (previously described), water from Bitter Spring was a mixture of shallow and deep groundwater. Water from this spring also had measurable tritium (0.31 TU) and a relatively large ¹⁴C activity (81.5 pmC), indicating some percentage of recent recharge. Shallow and deep groundwater sampled from monitoring wells at multiple-well monitoring site CRTH2 did not contain measurable tritium (less than 0.2 TU) and had very low ¹⁴C (7.87 and 9.87 pmC), indicating old water. As previously described in the “Stable Isotopes of Oxygen and Hydrogen” section, a mixture of shallow and deep groundwater discharges at Bitter Spring. Thus, water at Bitter Spring may have contained some recent, focused recharge of surface-water runoff from intermittent storms based on the tritium and ¹⁴C data.

Water from the three seeps at Jack Spring had ¹⁴C activities ranging from 64 to 85.4 pmC and tritium concentrations ranging from less than 0.20 to 0.44 TU (fig. 38). These ¹⁴C activities (unadjusted for reactions with aquifer material) indicated that waters from these seeps were about 3,700–1,400 years old; based on ¹⁴C activities and tritium concentrations, the youngest water was at the west and east seeps, which lie along dry washes draining the nearby granitic hills (figs. 22, 23). Like Garlic Spring, water from the west and east seeps was a mixture of water from the fractured bedrock moving along the Coyote Lake Fault and recent, focused recharge by surface-water runoff from intermittent storms. These data supported the interpretation that Coyote Lake Fault also acts as a conduit for groundwater flow. Water from Jack Spring was representative of water discharging primarily from the fractured bedrock that had the least mixing with recent recharge.

Overall, most springs contained little detectable tritium (0.2 TU or greater) and appeared to be primarily older groundwater. The relatively high ¹⁴C activities and the variability in measurable ³H indicated that some of these springs have received some percentage of their water from recent recharge with measurable ³H sufficient to be detected in the groundwater samples collected for this study. Tritium was detected at concentrations greater than the study reporting level (0.2 TU or greater) in groundwater basin springs, which could be a result of occasional surface-water runoff from infrequent storms in nearby washes near the springs (Bitter and Jack Spring East and West) or from recharged wastewater (Garlic Spring). These groundwater basin springs may be susceptible to decreases in discharge during extended dry periods due to the loss of focused recharge through nearby washes near the springs. These data also indicated that some recharge since 1952 was present in shallower parts of the groundwater system, along faults, and where groundwater mixes with storm water runoff along washes, specifically at Bitter Spring, Garlic Spring West, and Jack Spring East and West. Recharge of local precipitation through the fractured bedrock in upland springs at Cave, Desert King, and Devouge Springs happened within the last several thousand years. Recharge of local precipitation in the fractured bedrock in the mountains at Panther Spring happened more recently (less than 1,000 years ago), but before 1952, where ages based on ¹⁴C data indicate water is younger but lacks measurable ³H.

The sampled springs may be sensitive to future groundwater development and pumping of nearby existing production wells (fig. 1). The groundwater basin springs are especially sensitive to water-level declines in nearby groundwater basins from which they are fed. The upland springs also may be sensitive to pumping if they are hydraulically connected to nearby groundwater basins. Data were not collected to assess hydraulic connections between springs at the NTC and nearby groundwater basins as part of this study because there were few wells in the nearby groundwater basins; the exception was hydraulic data from wells in Irwin subbasin near Garlic Spring and in Cronise Valley groundwater basin near Bitter Spring.