Identifying origins of and pathways for spring waters in a semiarid basin using He, Sr, and C isotopes: Cuatrociénegas Basin, Mexico

He, Sr, and C isotopes are used to identify the origins of and groundwater pathways for Cuatrociénegas Basin springs; the basin is an oasis in the Chihuahuan Desert region of northeastern Mexico (Fig. 1). Dozens of springs flank the 2300-m-high Sierra San Marcos anticline, which bisects the basin and sources critical flows to groundwater-dependent pools, streams, and marshes that are refugia for >70 endemic species (Hendrickson et al., 2008; Fig. 2). Springs have varied spatial distribution and geochemistry. Spring vents are often obscured by valley-fill alluvium, but high-discharge springs generally issue directly from fractures in Cretaceous carbonate rocks. Other springs are located at the base of alluvial fans (Wolaver and Diehl, 2011). CO₂-rich spring water with pCO₂ 1.5–3 orders of magnitude higher than atmosphere facilitates deposition of rare modern freshwater stromatolites (Dinger et al., 2006).

The study area is in the Chihuahuan Desert, where long-term (1942–2003) annual precipitation of 200 mm exceeds annual potential evaporation (1964–1988) of 1960 mm (Aldama et al., 2005). Irrigated agriculture since the mid-1990s in adjacent valleys caused groundwater declines of ∼1 m/yr, and the main springs that formed the initial basis for irrigated agriculture in the town of Cuatrociénegas no longer flow (Wolaver et al., 2008). In addition to supporting groundwater-dependent ecosystems with unique endemic species, the Cupido aquifer that flows to Cuatrociénegas springs supplies water to more than four million people in northeastern Mexico and is correlative to the prolific Texas Edwards-Trinity aquifer (Johannesson et al., 2004). Understanding the Cuatrociénegas Basin spring-water origins is important for both developing groundwater resources for human needs and preserving spring flow for groundwater-dependent ecosystems in similar settings.

This study tests the hypothesis that Cuatrociénegas Basin springs are divided into two classes based upon discharge and geochemical properties, which are inferred to be controlled by relationships between springs and fault conduits. The first class has characteristics that imply discharge primarily from a regional carbonate aquifer with some deeply sourced fluids along high-permeability fault-associated pathways. These springs have elevated discharge (≤550 L/s), elevated temperature (30–35 °C), an absence of ³H, and distinctive water chemistry (high total dissolved solids, TDS > 2000 mg/L, calcite saturation, high pCO₂, mantle-derived He). The second class reflects mixing of regional carbonate aquifer and locally recharged waters. These springs have lower discharge and lower temperatures (<30 °C) and detectable ³H; water chemistry shows lower TDS (500–1500 mg/L), calcite undersaturation, and less mantle-derived He.

This study also addresses spring origins and pathways using major and trace element water geochemistry and Sr and He isotope data to calculate mixing of three source waters that results in the two spring classes. The three sources are: (1) epigenic waters (dominantly younger, locally derived mountain-front precipitation); (2) mesogenic waters (older, regional carbonate aquifer groundwater); and (3) endogenic waters (inputs from deep fault systems). Isotopes of Sr in groundwater have been used to identify the source of Sr ions and to evaluate groundwater flow paths (Musgrove et al., 2010). Sr isotopes indicate that groundwater fluxes are dominated by mesogenic, regional carbonate aquifer flow. Water chemistry shows that regional carbonate aquifer water mixes with deeply derived fluids that flow along basement-penetrating faults and also mountain-front recharge, resulting in the two spring classes.

Cuatrociénegas Basin consists of Jurassic and Cretaceous marine and nonmarine siliciclastics overlying Precambrian to Paleozoic crystalline basement (Goldhammer, 1999; Lehmann et al., 1999). Cretaceous carbonate rocks, as thick as 800 m, form the Cupido and Aurora Formations that are the primary regional aquifer (Wolaver and Diehl, 2011; Wolaver et al., 2008). The region has a complex tectonic history with several stages of faulting (Fig. 1). During the Permian–Triassic, left-lateral motion along the Coahuila-Tamaulipas transform, associated with the opening of the Gulf of Mexico, faulted Precambrian and Paleozoic basement blocks of northeastern Mexico (Aranda-Gómez et al., 2005, 2007; Dickinson and Lawton, 2001; Goldhammer, 1999). The imprinted zones of structural weakness influenced Laramide and subsequent tectonic activity. First, extension along Neocomian normal faults (ca. 140–136 Ma) formed fault-bounded structural highs and lows, such as the Coahuila block and adjacent Sabinas Basin. Then, compression folded and faulted strata in the Paleogene (ca. 45–23 Ma; Chávez-Cabello et al., 2005). Two less intense periods of normal fault reactivation occurred in the Late Miocene to Quaternary (Aranda-Gómez et al., 2005; Chávez-Cabello et al., 2005).

The tectonic activity created faults that are critical to the area hydrogeology. Mesozoic normal faults associated with the opening of the Gulf of Mexico created conditions that permitted the deposition of Cretaceous carbonate rocks that now form a regional carbonate aquifer (Lehmann et al., 1999; Wolaver and Diehl, 2011; Wolaver et al., 2008). Paleogene reactivation of reverse faults generated the ≤3000 m anticlinal uplifts that serve as recharge areas and fractured the sedimentary rocks to create secondary permeability. Late Miocene–Quaternary normal faulting fractured the carbonate aquifer, reactivated older faulted structures, and formed foci for Pliocene–Pleistocene volcanism (Aranda-Gómez et al., 2007). Wolaver and Diehl (2011) investigated fault-associated fracture permeability influences on springs. However, previous studies did not assess the role of these faults as conduits for deeply sourced fluids.

The study area is near the boundary between the Precambrian-cored North American craton (Whitmeyer and Karlstrom, 2007) and Mesozoic accreted terranes of Mexico (Dickinson and Lawton, 2001). Northeastern Mexico is a southern continuation of a region of low mantle velocity (van der Lee and Nolet, 1997) related to Mesozoic and ongoing mantle modification due to asthenospheric upwelling following foundering of the Farallon plate (Schmandt and Humphreys, 2010). Regions of low mantle velocity in the western United States are associated with high ³He/⁴He isotopic ratios in thermal spring waters (Newell et al., 2005).

Influences on regional groundwater chemistry caused by mantle degassing in areas of structural weakness have been studied using He isotopes (Crossey et al., 2006; Kennedy and van Soest, 2007; Newell et al., 2005). In Mexico, He isotopes have been used to study volcanic and hydrothermal systems (Inguaggiato et al., 2005; Taran et al., 2002; Vidal et al., 1982; Welhan et al., 1979) and for economic geology to evaluate ore fluids (Camprubí et al., 2006; Nencetti et al., 2005), but no He isotopic data are available for study area springs and groundwater to confirm the presence of deep faults in the regional carbonate aquifer.

Analytical methods used for water chemistry, gas chemistry (e.g., He isotopes and other noble gases), C isotopes, and Sr geochemistry analyses are described in the following. Water samples were collected at springs (vents and pools), spring runs, and wells for water and gas chemistry analyses. Sr isotope geochemistry analyses were done on water samples collected from springs (vents and pools), spring runs, and travertine. Water and travertine samples were collected to provide representative samples basin-wide trends in geochemistry (Fig. 2).

Standard methods (described in Crossey et al., 2009) were used for water chemistry analyses (Table 1).

The ³He/⁴He ratio is evaluated to elucidate terrigenic He crust and mantle source components by analyzing water for dissolved gases and He isotopes. Most dissolved noble gas samples were collected using passive diffusion samplers (De Gregorio et al., 2005; Gardner and Solomon, 2009). Noble gas samples were processed at the University of Utah Dissolved Gas Laboratory (following Manning and Solomon, 2003). One copper tube sample was collected from La Becerra warm vent and analyzed for CO₂/³He (methods of Poreda and Farley, 1992). The He precision is ±0.5%–1% and ±1%–2% for all other gasses. Measured ³He/⁴He ratios are often expressed as R/R_A, where R = ³He/⁴He of the sample and R_A = ³He/⁴He of the atmosphere (1.384 × 10⁻⁶; Clarke et al., 1976). We report air-corrected values for total He [He]_c, and ³He/⁴He (R_C/R_A) for the non-air portion of gases dissolved in springs (using X the air-normalized He/Ne ratio correction by Hilton, 1996) by accounting for potential amounts of ⁴He contributed by air and air saturated water to calculate terrigenic He (i.e., ⁴He from crust and mantle sources; Solomon, 2000). The He isotopic ratios are presented compared to atmospheric concentration as R/R_A.

Spring water was analyzed for C isotopes using methods described by Hilton (1996). One sample (La Becerra warm vent; Table 2) was analyzed at the University of Rochester (New York), while other samples (as reported in Aldama et al., 2005) were analyzed at the University of Arizona. Methods for distinguishing CO₂ sources from water chemical data follow Chiodini et al. (2004, 2000) and Crossey et al. (2009), except an evaporite correction is used to estimate external C in high SO₄ waters (i.e., deeply derived C that flows to springs via basement-involved faults) and new methods are used to estimate dissolved inorganic C (DIC) using the computer program PHREEQC (Parkhurst, 1995).

The CO₂ sources include: (1) C_carb (from carbonate dissolution along flow paths); (2) C_org (C of organic origin); and (3) C_endo (endogenic CO₂ from crust and mantle). In a mass balance model, C_carb = Ca²⁺ + Mg²⁺ – SO₄^2– (Chiodini et al., 2000; Fontes and Garnier, 1979). Many Cuatrociénegas Basin springs are SO₄ rich, making this calculation problematic. In cases where SO₄^2– is greater than Ca²⁺ + Mg²⁺, C_carb is set to zero because there must be a sulfur source, such as H₂S. Remaining C is external carbon, where C_ext = DIC_total – C_carb. DIC_total is computed from measured DIC (as alkalinity) and pH by a speciation model (PHREEQC; Parkhurst, 1995). To estimate proportions of C_org and C_endo of C_ext, the measured δ¹³C is corrected for each sample by removing the isotopic proportion due to C_carb. Depending on regions being studied, workers have used average values of 0‰ (Crossey et al., 2009; Sharp, 2007) to +2‰ (Chiodini et al., 2000; Crossey et al., 2009). The C_ext calculation is not especially sensitive to choice of this value, but here we use δ¹³C = +0.5‰, a central value found for the range of values (–6‰ to +5‰) of late Paleozoic southern New Mexico limestone (Koch and Frank, 2012). C_ext is composed of both organic components (C_org) and deeply sourced endogenic components (C_endo). Data plot within a set of binary mixing curves defined empirically by our own data, with chosen end members of (1) a soil-respired CO₂ end member with δ¹³C_org = –28‰ that is compatible with δ¹³C_org = –15‰ to –30‰ from vegetated areas (Deines et al., 1974; Robinson and Scrimgeour, 1995; Sharp, 2007) and with variable C_ext = 0.0001–0.004 mol/L, and (2) an endogenic CO₂ end member of δ¹³C_endo = –3‰ and C_ext≥ 0.06 mol/L that is comparable to values of –5‰ (Crossey et al., 2009) to –3‰ used in other studies (Chiodini et al., 2000), and comparable to a mid-oceanic ridge basalt (MORB) mantle reference value of δ¹³C = –6‰ ± 2.5‰ (Sano and Marty, 1995).

The Sr isotopes in water and travertine are evaluated to estimate relative contributions of deeply sourced Sr ions from basement faults and Sr ions derived from the regional carbonate aquifer or groundwater interactions with volcanic rocks (Table 3). An Sr-specific resin separated Sr from water and travertine samples by ion exchange (University of Texas at Austin Department of Geological Sciences Isotope Clean Laboratory; Banner and Kaufman, 1994).

Water chemistry data (Table 1) are compiled from published data (Johannesson et al., 2004; Aldama et al., 2005). Evans (2005) also presented water chemistry data, but focused on sampling spring discharge at spring run locations downstream of orifices.

Cuatrociénegas springs exhibit two distinct classes (Fig. 3): (1) one class plots as a grouping of Ca-SO₄–type waters (e.g., La Becerra and Escobedo springs); (2) the second class exhibits Ca-HCO₃–type waters (e.g., Santa Tecla spring).

The first class commonly discharges in springs issuing from fractures in carbonate rocks and at elevated flow rates (to 550 L/s), representing ∼85% of total basin spring water (Wolaver et al., 2008). These waters have low ³H concentrations (generally below detection limits; Table 2), TDS of 2000–3200 mg/L, and elevated temperatures (30–35 °C).

The second class has springs commonly at the base of alluvial fans with lower discharge rate, representing ∼15% of total basin spring flow (Wolaver et al., 2008). These springs often have low, but measureable, ³H (i.e., Santa Tecla has 0.2 tritium units, TU) and lower TDS (i.e., 1,415 mg/L). Spring-water temperatures are lower, but still elevated (30 °C).

Most groundwater samples have high measured ⁴He concentrations between 10⁻⁸ and 10⁻⁶ cm³/g (Table 2). The Δ⁴He, the percentage of terrigenic ⁴He relative to ⁴He in the atmosphere in equilibrium with water (4.31 × 10⁻⁸ cm³/g; Giggenbach et al., 1993), ranges from 434% to 3125% with the exception of the lowest Δ⁴He value (19%, sampled in an open canal; Table 2).

Springs with X factors >4 and air-corrected R_C/R_A values ranging from 1.25 R_A to 1.85 R_A (Table 2) indicate that 16%–23% of He is from MORB-type mantle sources. The highest ³He/⁴He values with lowest air corrections are in Poza Azul (1.84 R_A; X = 19.8), Escobedo vent (1.84 R_A; X = 21.5), and La Becerra warm vent (1.76 R_A and 1.85 R_A by two methods, X = 34 and 18.3, respectively). Gas chemistry copper tube samples (Poreda and Farley, 1992) are similar to those from passive diffusion samples from the same spring (La Becerra warm vent; Table 2). All spring waters are significantly above crustal radiogenic production values of ³He/⁴He = 0.02 R_A (Craig et al., 1978).

Water chemistry, CO₂/³He ratios, and C isotope data elucidate different C sources contributing to CO₂-rich springs in the Cuatrociénegas Basin. C isotopes reported as δ¹³C range from –15.4‰ (La Becerra warm vent) to –2.9‰ (CNA-82; Table 1).

The CO₂/³He value for the copper tube sample (1.2 × 10¹⁰; Table 2) is within the range of crustal CO₂/³He (10⁸ to 10¹⁴; O’Nions and Oxburgh, 1988). For 117 total groundwater samples (Table 1), C_carb from carbonate dissolution varies spring to spring, but has a mean value of 23% ± 25%. C_ext (computed as DIC_total – C_carb) has a mean of 77% ± 25% for 117 springs. The δ¹³C_ext values calculated for each groundwater sample with δ¹³C relative to C_ext are described by binary mixing models (Fig. 4). For 9 water samples with C isotope values, C_carb = 30% ± 22%, C_org = 24% ± 16%; and C_endo = 46% ± 33%.

There are 17 spring-water samples that exhibit high homogeneity, with ⁸⁷Sr/⁸⁶Sr = 0.707429–0.707468 (Table 3; Fig. 5); two samples from a travertine quarry at the southwest flank of the Sierra San Marcos (Fig. 2) with an age estimated by Aldama et al. (2005) as Pleistocene (∼17,000 yr) have ⁸⁷Sr/⁸⁶Sr = 0.707448 and 0.707449. One modern travertine sample (Fig. 2) has ⁸⁷Sr/⁸⁶Sr = 0.707428. Results of Sr isotope analyses of rock samples from Cretaceous carbonate (and gypsum bearing) rocks of the regional aquifer have similar ⁸⁷Sr/⁸⁶Sr values of 0.7072–0.7076 (Fig. 5; Lehmann et al., 2000). In contrast, volcanic rocks of nearby Las Coloradas, Las Esperanzas, and Ocampo volcanic fields (Fig. 2) have much lower ⁸⁷Sr/⁸⁶Sr values, 0.70333–0.70359 (Aranda-Gómez et al., 2007; Chávez-Cabello, 2005).

The first water class is sourced from a regional carbonate aquifer (mesogenic fluids) mixed with endogenic fluids that ascend along basement-involved faults (e.g., La Becerra and Escobedo springs). The ³H analyses, as indicated by levels below the detection limit, indicate regional carbonate aquifer residence times in excess of 60 yr. Ca-SO₄–type waters suggest rock-water interactions of a limestone aquifer with concomitant gypsum dissolution in a regional flow system (Fig. 3). Gypsum dissolution in Cuatrociénegas groundwater is consistent with the presence of hundreds of meters of cyclic carbonates and evaporites (Lehmann et al., 2000; Lehmann et al., 1999).

The second water class is consistent with a regional aquifer (mesogenic fluids) source mixed with locally recharged mountain precipitation (epigenic fluids). These Ca-HCO₃–type waters have slightly higher ³H (i.e., 0.2 TU at Poza Santa Tecla; Fig. 2) and lower TDS, reflecting shorter residence time consistent with localized flow systems. Water chemistry also reflects mixing between meteoric recharge (epigenic waters) and an SO₄-rich end member (mesogenic), consistent with dissolution and near equilibration with gypsum (Fig. 3). Canyons in the Sierra San Marcos, associated with alluvial fans (located in Cenozoic alluvium 2–4 km south of the A–A′ cross-section line in Fig. 2), permit mixing of epigenic groundwater recently derived as mountain precipitation with mesogenic groundwater from the regional carbonate aquifer that discharges along an inferred normal fault.

PHREEQC modeling of the two groundwater classes (Parkhurst, 1995) shows equilibrium pCO₂ values as high as 10^−0.5 atm, indicating high CO₂ relative to meteoric waters. Alkalinity as HCO₃⁻ is also high (≤229 mg/L), suggesting that CO₂ degassing from endogenic fluids ascending along basement-involved faults influences basin water chemistry.

The Δ⁴He (434%–3125%; Table 2) indicates that ⁴He from terrigenic sources is greater than atmosphere-sourced ⁴He. Spring waters contain nonatmospheric gases from a deep endogenic fluid system (i.e., crust or mantle). However, the lowest Δ⁴He value (19%) is an air-contaminated sample approaching atmospheric equilibrium from an open canal that flows into Escobedo spring.

The presence of mantle-derived fluids in spring water is supported by ³He/⁴He = 0.02 R_A, significantly above crustal radiogenic production values (Craig et al., 1978). Tritiogenic ³He contribution to R/R_A values in Cuatrociénegas Basin springs is miniscule, as groundwater ³H concentration is ≤0.4 TU and generally less than the method detection limit (≤0.1 TU; Table 2). Spring distribution relative to faults (Fig. 2; e.g., Poza Churince, Poza Azul, and Poza Anteojo springs) indicates that fluids containing appreciable mantle gases flow to a near-surface groundwater system along basement-penetrating faults.

Cuatrociénegas Basin He isotopes (R_C/R_A = 0.89–1.85; Table 2) show that western United States regional mantle degassing (Newell et al., 2005) continues into northern Mexico south of the Proterozoic Laurentia basement edge (Whitmeyer and Karlstrom, 2007). Published Mexican He isotope data (R/R_A) for volcanic, hydrothermal, and economic geology studies are consistent with mantle He degassing: 0.3–0.6 R_A (Vidal et al., 1982), 2.36–6.33 R_A (Welhan et al., 1979), 0.66–7.5 R_A (Taran et al., 2002), 1.25–2.84 R_A (Inguaggiato et al., 2005), 0.5–2.0 R_A (Camprubí et al., 2006), and 0.57–1.03 R_A (Nencetti et al., 2005). Mantle degassing along the plate margin of western Mexico is not surprising, but is also taking place in northeastern Mexico because of the importance of basement-involved faults.

Waters in the Cuatrociénegas Basin have high endogenic CO₂ and mantle He isotopes that show crust and mantle fluid input along basement faults into the regional aquifer. He gas transport through the mantle and crust is partly coupled to movement of CO₂/H₂O supercritical fluids (Giggenbach et al., 1993). Gas transport from the mantle is also supported by the CO₂-rich character of Cuatrociénegas springs, which have two to three orders of magnitude higher pCO₂ than air and δ¹³C = –15.4‰ to –2.9‰.

Cuatrociénegas Basin water, travertine, and carbonate rock have similar Sr isotopic ratios. Volcanic and basement rocks do not appear to be significant Sr ion sources to groundwater (Table 3; Fig. 5). Spring-water and travertine ⁸⁷Sr/⁸⁶Sr data show that the regional carbonate aquifer is the Sr source. Rock-water interaction of volcanic rocks and deeply sourced waters recirculated through granitic basement (Kesler and Jones, 1980–1981) do not appear to contribute measurable Sr (we use ⁸⁷Sr/⁸⁶Sr > 0.72 for the Precambrian to Paleozoic crystalline rock Sr end member; Crossey et al., 2009). Assuming that the Pleistocene climate was cooler and wetter, homogeneous ⁸⁷Sr/⁸⁶Sr from travertine of modern and inferred Pleistocene age suggest that the regional carbonate aquifer has remained the spring-water source despite drying Holocene climate (Musgrove et al., 2001). A comparison of ⁸⁷Sr/⁸⁶Sr in Cuatrociénegas Basin spring water with the Cretaceous section (Lehmann et al., 2000) reveals matches with the Albian Aurora and Barremian–Aptian Cupido Formations. Because the Aurora Formation is relatively thin compared to the Cupido Formation (Lehmann et al., 1999), this suggests that the Cupido Formation is the most important regional aquifer.

Cuatrociénegas Basin springs exhibit He, Sr, and C isotopes that indicate mixing of groundwater from the regional carbonate Cupido aquifer groundwater with endogenic sources (e.g., Poza La Becerra) and younger meteoric mountain front recharge (e.g., Poza El Venado). He, Sr, and C isotope data indicate that basement-involved faulting (1) is a critical factor controlling spring vent locations, and (2) provides conduits for circulation of deeply derived gases charged with mantle-derived ³He and CO₂. Sr isotopes show that the Cupido aquifer is the primary Cuatrociénegas Basin spring source and the most important regional aquifer.

From the hydrogeologic conceptual model explaining groundwater, He, Sr, and CO₂ fluxes (Fig. 6), it is inferred that neotectonic extensional faults facilitate fluid ascent along reactivated, older, deeply penetrating Laramide reverse faults. The hydrogeologic conceptual model is based upon regional structural styles, hydrogeologic data, and geophysical surveys in Wolaver and Diehl (2011; refer to table 1 therein for a generalized hydrostratigraphic column).

Basin and Range–type normal faults along mountain fronts are inferred based on spring alignment on the east side of Sierra San Marcos and the presence of mantle-derived fluids in these springs. Normal faults have offsets below gravity survey resolution (Wolaver and Diehl, 2011). Normal faulting has reactivated older deeply penetrating reverse faults and provides endogenic fluid pathways. Deeper fluid conduit system components likely involve hydrothermal fluid transfer through the lithospheric mantle and lower crust.

Sr isotopes indicate that carbonate rock-water interactions dominate Cupido aquifer spring-water chemistry. Paleogene–Quaternary reactivation of basement-involved faults associated with the opening of the Gulf of Mexico generated fractures in the regional Cretaceous carbonate aquifer (Ca-SO₄ facies) that focus spring discharge (⁸⁷Sr/⁸⁶Sr = 0.707429–0.707468). Basement-involved faults also provide conduits for Pliocene–Pleistocene mafic volcanism, but spring waters show no detectable contribution of Sr ions sourced from volcanic rock-water interactions (with ⁸⁷Sr/⁸⁶Sr = 0.70333–0.70359). A comparison of Sr isotopes from recent and inferred Pleistocene aged travertine shows that the Sr ion source to the groundwater flow system has remained the same, despite long-term climatic fluctuations. The likely dominant source of Sr ions to springs is the regionally extensive Cupido aquifer.

He isotopes show that basement-involved faults create pathways for mantle-derived gases to escape and discharge into springs (³He/⁴He = 0.89–1.85 R_A). Mantle degassing is consistent with regional tomographic images that show low-velocity mantle at shallow sublithospheric depths. Similar to the western U.S. (Bethke and Johnson, 2008; Crossey et al., 2009), northern Mexico is undergoing regionally pervasive mantle degassing through CO₂-rich cool and warm springs located along faults.

Spring-water C isotopes have high dissolved CO₂ (pCO₂ = 10⁻¹ to 10^−0.3 compared to atmospheric concentrations of 10^−3.5 atm) and alkalinity ≤230 mg/L (as HCO₃⁻) and are similar to Colorado Plateau travertine-depositing springs (Crossey et al., 2006). Water chemistry analyses indicate that CO₂ is (1) 33% ± 15% C_org, derived from soil gas and other organic sources, (2) 30% ± 22% C_carb, derived from dissolution of carbonate in the aquifer, and (3) 37% ± 29% C_endo, derived from endogenic (i.e., deep, geologic) sources. Faults act as conduits for elevated CO₂ flux with concomitant elevated groundwater alkalinity, as observed in other North American travertine-forming springs (Bethke and Johnson, 2008; Crossey et al., 2006, 2009).

These findings on Cuatrociénegas Basin springs have implications for managing groundwater resources and endemic species. The spring waters that irrigate farms and sustain groundwater-dependent ecosystems are primarily mesogenic fluids that discharge from the Cupido aquifer with residence times in excess of 60 yr (with minor endogenic fluids). Springs also discharge a mixture of mesogenic and epigenic fluids, a combination of Cupido aquifer and mountain recharge waters. High-discharge springs in the Cuatrociénegas Basin (e.g., Azul, Churince, Escobedo, La Becerra) have geochemical characteristics of mesogenic fluids discharging from the Cupido aquifer.

This study was partially funded by the the University of Texas, the Geological Society of America, BHP Billiton, the Gulf Coast Association of Geological Societies, the Houston Geological Society, and the Tinker Foundation. Larry Mack and John Lansdown provided Sr isotope and cation analyses expertise. Alan Rigby provided analytical expertise with dissolved noble gas analyses. Robert Poreda provided He isotope analyses of one of the samples. Dean Hendrickson and Suzanne Pierce collected water samples for Sr analyses. Gareth Cross, Peyton Gardner, Dean Hendrickson, Randall Marrett, Juan Manuel Rodriguez, and Vsevolod Yutsis provided helpful conversations. Laura Merner and Dawn Saepia of the Environmental Science Institute’s National Science Foundation Research Experience for Undergraduates program (NSF, EAR-0552940) assisted with field work. NSF grant EAR-0838575 supported Crossey and Karlstrom’s work. Publication is authorized by the Director, Bureau of Economic Geology, Jackson School of Geosciences, the University of Texas at Austin.