Both the sources and pathways of fluid circulation are key factors to understanding the evolution of low-angle normal fault (LANF) systems and the distribution of mineral deposits in the upper crust. In recent years, several reports have shown the presence of meteoric waters in mylonitic LANF systems at mid-crustal conditions. However, a mechanism for meteoric water infiltration to these mid-crustal depths is not well understood. Here we report paired δ18O and δ2H isotopic values from dated, neoformed clays in fault gouge in major detachments of the southwest United States. These isotopic values demonstrate that brittle fault rocks formed from exchange with pristine to weakly evolved meteoric waters at multiple depths along the detachment. 40Ar/39Ar dating of these same neoformed clays constrains the Pliocene ages of fault-gouge formation in the Death Valley area. The infiltration of ancient meteoric fluids to multiple depths in LANFs indicates that crustal-scale normal fault systems are highly permeable on geologic timescales and that they are conduits for efficient, coupled flow of surface fluids to depths of the brittle-plastic transition.


Fluid flow in both individual faults and sets of faults in a given tectonic regime has been the subject of considerable interest for the past 30 years (e.g., Kerrich et al., 1984; McCaig, 1997; Gébelin et al., 2012; Menzies et al., 2014). Fluids in middle- and upper-crustal normal faults show a strong influence of variably evolved, meteoric-derived fluids (e.g., Fricke et al., 1992; Mulch et al., 2004; Swanson et al., 2012; Hetzel et al., 2015). These observations would require downward circulation of surface waters into the mid-crust, a physiomechanical process that is poorly understood (Connolly and Podladchikov, 2004; Person et al., 2007; Lyubetskaya and Ague, 2009). In addition to fluid pathways, fault zone minerals with a meteoric fluid origin can be used to make inferences about regional paleoelevations (e.g., Mulch et al., 2004; Gébelin et al., 2012, 2013). By contrast with normal fault systems, it is thought that fluids in thrust faults are dominated by upward circulation of deep basinal fluids, with minor contributions from evolved meteoric fluids in late stages of orogeny (e.g., McCaig et al., 1995; Trave et al., 2007; Sample, 2010).

Fluid-flow models have outlined a set of narrow permeability and topography conditions by which downward flow of meteoric-water–dominated waters might still occur (Person et al., 2007). Key to testing the feasibility of these fluid-flow models for LANFs is isotopic data from the upper and middle reaches of LANF systems. Isotopic studies of the upper and middle reaches of LANF systems, which extend from the surface to the middle crust, are few relative to the now data-rich mylonitic rocks. Stable isotopes have been employed in faults in carbonate-dominated sequences (e.g., Losh, 1997; Losh et al., 2005; Swanson et al., 2012), but relatively few LANFs occur in carbonate-dominated sequences relative to those in silicate-dominated upper-crustal sections.

Neoformed clay-rich fault gouges are a common feature of LANFs (e.g., Haines and van der Pluijm, 2012) and have been recognized to both dramatically reduce the frictional strength of fault zones (e.g., Carpenter et al., 2011; Haines et al., 2014) and document the age at which fault gouge formation occurred (Solum et al., 2005; Haines and van der Pluijm, 2008). The clay minerals that are neoformed in LANF gouge thus have a major influence on fault behavior, but their potential as recorders of upper-crustal fluid circulation in LANFs has not been broadly examined to date. Phyllosilicates are unusual silicate minerals in that they contain structural hydrogen in addition to the oxygen that is found in all silicates and thus permit analysis of both δ18O and δ2H on a single mineral phase, allowing for a more complete characterization of the exchanging fluid. The δ2H value of the clay minerals preserves the initial source of the fluid until water-rock ratios become very low (water/rock <0.001; Menzies et al., 2014). By contrast, because all silicate minerals contain oxygen, the δ18O value is strongly sensitive to the degree of wall-rock–fluid interaction (Sharp, 2005). An analysis of both isotopic ratios from the same mineral separate allows for an evaluation of both the initial source and any degree of wall-rock–fluid interaction of fluids exchanging with that mineral phase. Fluids with δ2H <−80‰ and δ18O <0‰ are generally interpreted to be of meteoric origin, while fluids with δ18O >+5‰ and δ2H >−80‰ are interpreted to be of metamorphic or igneous origin (Sheppard, 1986).

Illitic clays are common to many clay-rich gouges; they contain K, and retain Ar, permitting dating of clay growth in gouge by 40Ar/39Ar methods. We use the illite age analysis method, which utilizes 40Ar/39Ar dating in conjunction with quantitative X-ray diffraction (XRD) to determine the age of authigenic and detrital (cataclastically derived) clay mineral populations.

To investigate the fluid-flow system of LANFs in the U.S. Basin and Range province as a class of fault, we utilized a suite of samples that were characterized as part of a companion study of clay gouge mineralogy (Haines and van der Pluijm, 2012), which identified systematic patterns of clay mineral transformations in clay-rich fault gouges. These faults range from shallow-rooted structures (such as the Panamint Range–Front detachment) to LANFs that reached mid-crustal depths (e.g., the Ruby Mountains detachment). We isolated authigenic phyllosilicate minerals from both upper-crustal clay-rich fault gouges and mid-crustal metasomatic, chlorite-rich breccias from a suite of faults (Fig. 1) and analyzed δ18O and δ2H values of neoformed phases in order to investigate the composition of fluids from which they grew. A subset of these samples was analyzed to determine the age of neoformed clays, and thus, the timing of fluid infiltration.


Low-angle normal faults (LANFs) are a special class of normal fault, first noted in the American Cordillera (Anderson, 1971; Wernicke, 1981) and now recognized globally (e.g., Collettini, 2011). These faults are unusual in that they have accommodated normal displacements of tens of kilometers and many slipped at dips below those predicted from conventional rock friction arguments (Axen, 2004; Haines and van der Pluijm, 2010). Many exposures of exhumed shallow-crustal LANFs have well-developed cm-thick to m-thick, clay-rich fault gouges that are dominated by neoformed clay minerals, predominantly illite, illite-smectite, and smectite. These neomineralized clays in fault gouge comprise the uppermost part of a suite of distinctive fault-related rocks in metamorphic core complexes (MCCs) that record progressive exhumation of footwall lithologies, often from pre-faulting mid-crustal depths. Many (but not all) LANF footwall exposures have clay gouges in direct contact with a distinctive greenschist-facies epidote + chlorite alteration of footwall metamorphic or igneous lithologies (Fig. 2A). This distinctive epidote + chlorite alteration can extend for tens of meters into the footwall, and, where brecciated, these rocks are lithified cataclasites, sometimes called “chlorite microbreccias” (Phillips, 1982; Selverstone et al., 2012). The fault rocks that are inferred to form at greatest depths are commonly quartzofeldspathic mylonites (formed at temperatures from 400 to 550 °C; Anderson, 1988; Mulch et al., 2007).

Fluid flow in LANF systems has been examined with numerous isotopic studies of the mid-crustal mylonitic fault rocks using both δ18O and δ2H on minerals and fluid inclusions (Lee et al., 1984; Wickham and Peters, 1990; Fricke et al., 1992; Wickham et al., 1993; Peters and Wickham, 1995; Mulch et al., 2004, 2007; Gébelin et al., 2011, 2012, 2015; Gottardi et al., 2011). The greenschist-facies microbreccias have also been studied, often in conjunction with the higher-temperature mylonites (Kerrich and Hyndman, 1986; Kerrich and Rehrig, 1987; Kerrich, 1988; Smith et al., 1991; Morrison, 1994; Nesbitt and Muehlenbachs, 1995; Morrison and Anderson, 1998). While many of these studies used δ18O and δ2H analyses, few performed both analyses on the same phase. Fluid circulation in LANFs in carbonate-dominated successions has been studied using carbonate veins formed in lower-temperature (30–300 °C) fault rocks (Losh, 1997; Losh et al., 2005; Swanson et al., 2012). The majority of these studies have documented low-δ2H/low-δ18O fluids, inferred to be of meteoric origin, although some (Smith et al., 1991) have documented predominantly igneous-dominated fluids, or the interaction of two (metamorphic and meteoric) fluid sources (Kerrich, 1988).

In recent years, δ2H isotopic studies of neoformed micas in LANF mylonites have yielded very depleted (δ2H <−100‰) values, interpreted to be indicative of (1) high-altitude meteoric fluid, (2) high-latitude meteoric fluid, or (3) a paleo-rain shadow (Mulch et al., 2004, 2007; Gottardi et al., 2011; Gébelin et al., 2011, 2012, 2015). The processes and pathways by which meteoric fluids of surface origin reach the mid-crust are controversial. Fluids migrating down a fault system will encounter unfavorable thermal and density gradients, and the buoyancy of hot waters at higher pressures is greater than that of colder waters, inhibiting downward flow (Connolly and Podladchikov, 2004; Lyubetskaya and Ague, 2009). In addition, mid-crustal rocks are widely assumed to lack the porosity and permeability to permit fluid flow at rates sufficient to prevent the very low water-rock ratios that would obscure the initial source of the fluid. Some studies, therefore, have suggested that isotopic evidence for meteoric fluids in mid-crustal lithologies is instead evidence of burial of pre-metamorphic fault rocks to mid-crustal depths and not actual incursion of meteoric fluids to mid-crustal shear zones (Clark et al., 2006; Raimondo et al., 2011, 2013). Person et al. (2007) presented a numerical model that suggested that a metamorphic core complex with a fracture-dominated flow system with a relatively narrow range of effective fault zone permeabilities (10−15 to 10−16 m2) and a crystalline basement wall-rock permeability <10−17 m2 could explain the observed isotopic depletion of micas in fault zone mylonites at the Sushwap metamorphic core complex. Our data test the hypothesis that upper-crustal, brittle faults in LANF systems act as pathways by which 2H- and 18O-depleted meteoric fluids can reach the middle crust. To test this hypothesis, we obtained paired stable isotope measurements of oxygen and hydrogen from authigenic clay minerals (illite and smectite) and authigenic chlorite from a suite of eight LANFs (Fig. 1). A subset of clay gouges from the Death Valley region and the Ruby Mountains of Nevada was also dated to test the ancient origin of fluids that were responsible for clay neomineralization.

2.1 Neoformed Clay-Rich Gouges

Fault gouges and breccias, which are common in Cordilleran LANFs, form in the brittle regime (<300 °C), and are commonly assumed to be predominantly the result of physical processes, such as cataclasis (e.g., Sibson, 1977; Holland et al., 2006). In recent years, it has been recognized that many “brittle” fault zones are also chemically very reactive environments (Vrolijk and van der Pluijm, 1999; Solum et al., 2005) and that significant mineral transformations occur in fault gouge (Haines and van der Pluijm, 2012), affecting the frictional strength and permeability structure of these rocks. Neoformed clay minerals form in fault gouges at a temperature range of 50–180 °C and in faults with a variety of wall-rock compositions. The common clays forming in gouges are illite, formed from the alteration of fragmental mica or feldspar in gouge, and smectite. Smectite in fault gouge mostly forms by two discrete pathways—tri-octahedral smectite (saponite) forms by the alteration of cataclastically derived chlorite in gouge, whereas di-octahedral smectite (montmorillonite) forms from the alteration of fragmental acid volcanics and tuffs in wall rocks. A predictable relationship has been observed between wall-rock lithology, temperature, and clay mineral formation in clay-rich gouge (Haines and van der Pluijm, 2012). Because these clays are authigenic hydrous phyllosilicates, they exchange with both the oxygen and hydrogen in the infiltrating fluids, providing information on fluid sources and pathways.

2.2 Chlorite “Microbreccias”

Chlorite metasomatic alteration and brecciation of the footwall extending for meters to tens to hundreds of meters below the detachment fault surface are common features of low-angle normal faults associated with metamorphic core complexes (Crittenden et al., 1980; Kerrich, 1988). Chlorite alteration is found at all of the detachments sampled in this study that are thought to have evolved from mid-crustal shear zones (Badwater and Mormon Point turtlebacks, Buckskin-Rawhide detachment, Chemehuevi detachment; Fig. 1). The chlorite metasomatic alteration is nearly always developed in footwall rocks that are dioritic to granitic in composition. Mylonitic marbles that are locally present in the footwalls of the Black Mountains and that are intercalated with extensively chloritized gneisses are visually unaltered. The breccias contain a distinctive assemblage of chlorite ± epidote ± (titanite or rutile) ± feldspar ± calcite ± Fe-oxide that overprints mylonitic fabrics and imparts a distinctive greenish color to the rocks (Selverstone et al., 2012). Isotopic and fluid inclusion studies indicate the alteration results from the breakdown of biotite, amphibole, or anorthitic feldspar at greenschist-facies metamorphic conditions (300–350 °C, Kerrich, 1988; 350–520 °C, Morrison and Anderson, 1998; and 380–420 °C, Selverstone et al., 2012) by an influx of Fe-, Mg-, and Mn-rich meteoric (Morrison, 1994; Morrison and Anderson, 1998) or igneous fluids (Smith et al., 1991). The “microbreccias” are commonly cataclastically reworked as brittle, unconsolidated fault gouges. While at some exposures of LANFs, alteration of the metasomatic chlorite to lower-temperature tri-octahedral clays is significant, other outcrops locally preserve a cataclastically derived, chlorite-dominated gouge without evidence of significant alteration (e.g., Figs. 1B and 1D; Haines and van der Pluijm, 2012). These unconsolidated gouges contain abundant chlorite cataclastically separated from the wall rock, permitting the chlorite to be separated by centrifugation and then analyzed.

2.3 Detachments Sampled

We sampled gouges from five suites of Cordilleran LANFs, comprising eight separate detachments: (1) The Ruby Mountains detachment in northern Nevada; (2) three detachments in Death Valley, California (Badwater detachment, Mormon Point detachment, and Amargosa detachment); (3) two detachments in the Panamint Mountains, west of Death Valley (Panamint Range–Front LANF and Mosaic Canyon detachment); (4) the Buckskin-Rawhide detachment in NW Arizona; and (5) the Waterman Hills detachment in southern California (Figs. 1 and 2). Further description of sampled outcrops is given in Data Repository File DR1; geospatial data in .kmz format are found in Data Repository File DR2. Gouges from these faults were all mineralogically characterized as part of a previous study that identified systematic patterns of clay mineral neomineralization in clay-rich fault gouges (Haines and van der Pluijm, 2012), and the samples analyzed in this study are all a subsample of samples from that study.


Fault gouges are mixtures of fragmental wall-rock material derived from one or both sides of a fault zone and authigenic (neoformed) clay minerals growing in the gouge. Isolating the neoformed clay component of clay-rich gouges is therefore a challenging process, because the clay crystallites are very small (<<2.0 μm). Gouges can also contain fragmental phyllosilicates that are superficially similar to the authigenic phases but would contaminate the isotopic value without careful characterization. Our sampling approach is shown in Figure 3. We use gravity settling in water to isolate the <2.0 μm (Stokes equivalent diameter) size fraction, followed by high-speed centrifugation to separate the clay-size fraction into three or four size fractions, coarse (2.0–0.2 μm), medium (0.02–0.05 μm), and fine (<0.05 μm). Each fraction is then characterized by XRD, using both oriented mounts (with and without ethylene glycol solvation) to identify principal clay phases and random powder mounts to accentuate the non-(00l) peaks characteristic of clay polytypes (which can be used to identify authigenic clay minerals in gouges). Additional site information and mineralogical description of these samples are found in Data Repository Files DR1 and DR2 and Haines and van der Pluijm (2012).

3.1 Sampling Clay-Rich Gouges

We analyzed only gouge clay samples that were well characterized in previous studies (Haines and van der Pluijm, 2010, 2012) for this study with >90% authigenic material based on XRD. X-ray diffraction patterns of all analyzed materials are given in Figure 4. Isotopic measurements were made on splits from the same material described in Haines and van der Pluijm (2010, 2012). We note that three of the 14 samples contain near–detection-limit quantities of one or two other mineral phases: quartz in ASH-1, a 10-A phase (illite, muscovite, or biotite/phlogopite) in MOR-3 and A-BOMB-3, and calcite in MOR-3. Although all illitic material contained some interlayered smectite as discernable by XRD, for this study we only used illitic clays that were >80% illite in illite/smectite, and most were >90% illite in illite/smectite.

3.2 Sampling Gouges Derived from Epidote/Chlorite Microbreccias

For this study, we only sampled chlorite-rich gouges where no other Mg-rich phyllosilicates (tri-octahedral clays such as saponite, chlorite/smectite, corrensite, talc/stevonsonite, or vermiculite-like phases, or other phyllosilicates such as sepiolite or palygorskite) were detectable by XRD. These samples also are free of other phases (e.g., epidote, feldspar, and calcite) to the level of XRD detection (Fig. 4C). In these gouges, chlorite in the fault gouge is structurally and compositionally indistinguishable from that found in chlorite-epidote alteration zones in the fault footwall as determined by XRD (Haines and van der Pluijm, 2012). These purely cataclastic gouges have effectively disaggregated the footwall lithologies, allowing footwall-derived chlorite grains to be efficiently separated by settling in water and subsequent centrifugation, similar to the authigenic clays in clay-rich gouge (see above).


4.1 δ18O Isotopic and δ2H Measurements

Oxygen isotopic analysis of clay separates was completed in the University of Wisconsin Stable Isotope Laboratory by laser fluorination using BrF5 (Valley et al., 1995) and an airlock sample chamber that prevented pre-fluorination (Spicuzza et al., 1998). Hydrogen isotope measurements were made by continuous-flow mass spectrometry at the Stable Isotope Laboratory at Leibniz Universität Hannover, except for samples WH68-1 (F) and WH68-3 (MF) that were analyzed at the U.S. Geological Survey in Denver. All isotopic ratios are reported relative to Vienna standard mean ocean water (VSMOW), and methods are detailed in Data Repository File DR3.

4.2 Illite Age Analysis (IAA)

40Ar/39Ar ages of samples were obtained by vacuum encapsulation (Dong et al., 1995) to address Ar loss during sample irradiation (“Ar recoil”). Samples were packaged into fused silica vials and sealed prior to irradiation (van der Pluijm et al., 2001). Thus, the 39Ar expelled from the crystallites during irradiation is retained for analysis (see van der Pluijm and Hall, 2015, for a full description of the method). The sample vials were broken open, the initial gas was analyzed, and the vials were then step-heated under a defocused laser until sample fusion occurred. Note that the total gas age obtained from the vacuum-encapsulated sample is functionally equivalent to a conventional K-Ar age (Dong et al., 1995).


Stable isotope values for LANF neoformed illite, smectite, and chlorite are shown in Table 1 and Figures 5A, 5C, and 5E. Individual illite δ18O isotope values range from −2.0‰ SMOW (Ruby Mountains, SEC 4-2) to +11.5‰ (Badwater), and illite δ2H values range from −142‰ (Ruby Mountains, SEC 1-2 and SEC 4-2) to −107‰ (ASH-1 [F]). Smectite δ18O isotope values are +3.6‰ (Ruby Mountains, SEC 4-3) and +17.9‰ (Waterman Hills, WH68 [<2 μm]), while smectite δ2H values are −147‰ (Ruby Mountains) and −95‰ (Waterman Hills WH-68-1 [F]). Both the Ruby Mountains main detachment illite (SEC 4-2) and illite from a hanging-wall normal fault (SEC 1-2) that formed coevally with the main detachment at 11–13 Ma (Haines and van der Pluijm, 2010) have isotopic values for δ18O of −1.8‰ and −2.0‰, respectively, and δ2H of −142‰ for both.

Values of δ18O chlorite range from +0.58‰ to +8.1‰, and δ2H values fall in a relatively narrow range from −97‰ to −113‰. The Mormon Point detachment samples—MOR-2 (M), MOR-2 (F), and MOR-3 (M)—all have relatively low δ18O values ranging from +0.58‰ to +3.1‰ and δ2H values from −99‰ to −108‰. The Chemehuevi detachment (LOBECK-3 [M]) and Badwater detachment (BAD-1 [G] [M]) samples have δ18O values of +2.5‰ and +4.84‰, respectively, and δ2H values of −106‰ and −113‰, respectively. The Buckskin-Rawhide detachment chlorite sample (A-BOMB-3) shows the highest δ values, with δ18O of +8.1‰ and δ2H of −97‰.

Ages of neoformed clay in selected gouge samples are listed in Table 2, and Ar degassing spectra for each grain-size fraction are included in Data Repository File DR4. We illustrate our results with a sample from the Badwater detachment (Fig. 6). Four size fractions show decreasing percentages of detrital illite with smaller grain sizes (Fig. 6A). Corresponding Ar ages for these samples are systematically younger with decreasing detrital illite, which we analyze in an IAA plot (van der Pluijm et al., 2001; Fig. 6B). Using linear York regression (Mahon, 1996) of percentage detrital illite versus e(λ.t) −1 (where λ is decay constant and t is age) produces extrapolated authigenic and detrital intercept ages of 3.3 ± 0.4 Ma and 12.2 ± 1.9 Ma, respectively. Note that this particular regression analysis treats both parameters as independent, resulting in age errors that primarily reflect the 2%–3% error in mineralogic quantification, while individual Ar ages have much smaller errors, on the order of 0.2–0.5 Ma.


Our results from LANF gouge illites show that they are significantly depleted in both 18O and 2H compared to previously published results from illite formed in three fault gouges from strike-slip and normal fault environments (Fig. 5A; Solum, 2005; Tonguç Uysal et al., 2006; Isik et al., 2014). Because paired oxygen and hydrogen isotopic data from either illite or smectite taken from within fault zones are rare, we compiled published δ18O and δ2H data for neoformed illite, smectite, and chlorite from several different geological settings to place our results in a broader context: (1) sedimentary basins, (2) active and fossil hydrothermal systems, (3) sedimentary basins that experienced meteoric water flushing (illite only, e.g., Whitney and Northrup, 1987), (4) bentonites (smectite only), (5) bentonites where significant post-formation alteration caused 2H and 18O exchange to became decoupled (smectite only, e.g., Cadrin et al., 1996; Horton and Chamberlain, 2006), (6) metamorphic rocks (chlorite only), and (7) fault zones (illite and chlorite only). Compiled literature data are shown in Figures 5A, 5C, and 5E. Compiled data are found in Data Repository File DR5, and supporting references in Data Repository File DR6.

6.1 Illite/Smectite Gouge Samples

Our gouge illite samples are isotopically depleted relative to illite that forms in sedimentary basins, and they are also depleted in 2H with respect to illite that formed in hydrothermal systems. Some gouge illites (Waterman Hills, Amargosa, and Badwater) have oxygen and hydrogen isotopic compositions that are similar to illites from sedimentary basins interpreted to have been formed during basinal flushing with meteoric water (“meteorically reset”) (Fig. 5A; Glasmann et al., 1989) or hydrothermal systems; but other gouge illites (Ruby Mountains and Panamint) have δ18O and δ2H values far lower than any reported from sedimentary basins. The Ruby Mountains illite samples preserve hydrogen and oxygen isotope values lower than any illite measurements yet reported (δ18O = −1.8‰ and −2.0‰, δ2H = −142‰ for both). Gouge smectite samples have isotopic compositions that are very similar to illite results, with the Waterman Hills sample (δ18O = +17.9‰, δ2H = −95‰) similar to smectites in sedimentary basins or smectites from bentonites (Fig. 5C) and the extremely isotopically depleted Ruby Mountains smectite (δ18O = +3.6‰, δ2H = −147‰), which is the most depleted smectite isotopic measurement with respect to both oxygen and hydrogen yet reported.

6.2 Chlorite Microbreccia Samples

The chlorite isotopic data are similar to the most isotopically depleted chlorites found in hydrothermal systems (Fig. 5E), with the BAD-1 (G) (M) sample being the lowest δ2H value yet reported (δ2H = −147‰). Overall, our chlorite samples are very isotopically depleted, especially with respect to hydrogen (all δ2H = −97‰ to −113‰), relative to those found in metamorphic terranes or in sedimentary basins (Fig. 5E and references in Data Repository File DR6). δ18O values for the chlorite samples are more variable, ranging from = +0.6‰ to +11.5‰, likely reflecting variable amounts of fluid–wall-rock interaction.

6.3 Equivalent Fluid Compositions

Interpreting stable isotopic values of phyllosilicate minerals and using them to estimate the composition of the fluid with which they exchanged requires constraints on the temperature at which neoformed minerals grew and the associated fractionation between mineral and fluid. While clay gouges lack fluid inclusions that permit direct estimation of the temperature of formational fluids, the clay mineral assemblages found in these gouges place constraints on temperature at their time of formation. Previous studies of clay gouge mineralogy with reliable thermal constraints indicate that neoformed illite in fault gouges from a range of fault settings form at temperatures 80 °C to 180 °C, and perhaps as low as 50 °C (Haines and van der Pluijm, 2012). Because the neoformed illite is the low-temperature 1Md polytype for all samples and XRD analysis indicates that all samples contain some interlayered smectite, the likely temperature of formation is no more than ∼120 °C for both illite and smectite in LANF gouge. From measurements of δ18O and δ2H and estimates of a plausible clay-formation temperature range, the isotopic composition of the fluid that exchanged with the clay can be calculated. Using published δ18O fractionation equations for illite and smectite (Sheppard and Gilg, 1996), we determine δ18O fluid compositions in equilibrium with the clay phases measured (Figs. 5B and 5D). Similarly, published water-mineral δ2H fractionation equations for illite and smectite (Capuano, 1992) permit calculation of the fluid composition exchanging with the neoformed clays in LANF fault gouge. Based on this analysis, we find that the compositions of the fluids with which clays exchanged range from nearly pristine meteoric water to weakly isotopically enriched meteoric water. Calculated end-member water compositions are compatible with prior estimates of Middle Miocene (Ruby Mountains and Waterman Hills) and Pliocene (Armargosa, Panamint, and Badwater) Basin and Range meteoric waters (Poage and Chamberlain, 2002; Gébelin et al., 2012, 2015). Only the fluid exchanging with the Badwater gouge illite (Bad-1) shows significant deviation of oxygen enrichment from the field of isotopic values of fluids found in sedimentary basins with increasing depth (Fig. 5B), possibly reflecting oxygen exchange with silicate minerals in the fault zone prior to illite growth. Alternatively, the Death Valley area has been periodically evaporative since the Pliocene (Knott et al., 2005). Evaporative fluids are higher in δ18O than the meteoric water line (Holser, 1979), which might also explain the observed O enrichment of the Badwater sample relative to other samples.

6.4 Previous Fault Zone Isotopic Results

The sole previous oxygen and hydrogen analyses of illite from the gouge of a normal fault, the Moab fault in Utah, USA (δ18O = +7.9‰ and +8.6‰, δ2H = −114‰ and −116‰, respectively; Solum, 2005) did not report an equivalent fluid composition, but our calculations from the reported mineral values are consistent with a weakly heavy isotope–enriched meteoric fluid (δ18O = −4.0 to −6.5‰, δ2H = −83 to −93‰). These limited results support our interpretation of a link between kinematic environment and fluid source, with normal fault systems being dominated by fluids of meteoric origin, while reverse fault systems are dominated by fluids of basinal or metamorphic origin (e.g., Kerrich, 1988; McCaig, 1997). By contrast, data from deeply rooted strike systems suggest fluid sources are more variable in these systems. Data from the crustal-scale North Anatolian fault zone indicate fluid infiltration at various times by fluids of metamorphic or magmatic origin (Tonguç Uysal et al., 2006) and meteoric origin (Boles et al., 2015). Data from the subparallel but shallower-rooted Savcili strike-slip fault zone (Isik et al., 2014) suggest a deep basinal origin for circulating fluids.

The temperature of chlorite formation in epidote-chlorite breccias is less constrained than that for illitic gouges. Estimates range from 300 to 350 °C (Kerrich, 1988) to 350–520 °C (Morrison and Anderson, 1998) and to 380–420 °C (Selverstone et al., 2012). To capture this uncertainty, we use a temperature range of 340–440 °C. The variation in oxygen isotope fractionation over the full range of proposed temperatures (300–520 °C) is <1.3‰, far smaller than the observed range for illite or smectite, and thus the uncertainty in temperature has little effect on interpretation of the chlorite data. Hydrogen isotope fractionation between chlorite and water is poorly constrained at −30‰ to −40‰ but is thought not to change significantly with temperature over the range at which these breccias formed (Graham et al., 1987). Unlike illitic and smectitic clay minerals, chlorite in both brittle fault zones (<300 °C) and mylonitic greenschist- and amphibolite-faces shear zones has been extensively studied with stable isotopic methods. Previous studies of chlorites in fault zones include LANF (Picacho Mountains metamorphic core complex [MCC], Kerrich and Rehrig, 1987), as well as upper greenschist- and/or lower amphibolite-facies shear zones in the French Pyrenees (Leclere et al., 2014) and central Australia (Clark et al., 2006; Raimondo et al., 2011) and a Tertiary thrust fault in the Pyrenees active at ∼200 °C (Lacroix et al., 2012) (Fig. 5E). The results most germane to this study are chlorite samples from the Picacho Mountains MCC, which have δ18O values of +4.7‰ to +5.5‰ and δ2H values of −85‰ and −95‰. Our results are similar to these and suggest exchange with a fluid moderately enriched in δ18O but depleted in δ2H. Overall, chlorites from LANF systems have similar δ18O values to chlorites from amphibolite-facies shear zones inferred to have been infiltrated by meteoric fluids but have far lower δ2H and δ18O values than chlorites taken from brittle thrusts in compressional tectonic settings (Lacroix et al., 2012). Significantly, our samples all have δ2H that is ∼20‰ lighter than those observed in fault zones other than LANFs.

6.5 Meteoric Water Infiltration and Circulation

Isotopic exchange with fluids of meteoric origin has been increasingly documented associated with faults at mid-crustal depths (Morrison, 1994; Mulch et al., 2004; Gottardi et al., 2011; Gébelin et al., 2012; Mancktelow et al., 2015), but the mechanisms by which surface fluids reach these depths is not well understood (Roddy et al., 1988; Barentt et al., 1996; Losh et al., 2005; Hetzel et al., 2013). Our data from the upper brittle reaches of LANF systems show that meteoric water (with evidence of some wall-rock–fluid interaction) is the predominant fluid in deformed upper crust of LANF systems down to several kilometers depths. Our results showing meteoric fluid infiltration in the brittle portion of LANFs, together with observations of meteoric fluids at greater depths (i.e., chlorite breccias and mylonites) and model predictions, suggest that the drawdown of meteoric water along brittle faults is the dominant fluid circulation system in and near fault zones in extended crust. Convective flow up to balance the fluid-flow system must therefore occur either away from the fault zones or elsewhere up some other reach of the same fault system. Recent studies in the Dixie Valley hydrothermal field have suggested that in some cases, fluids travel updip along discrete sections of basin-bounding normal faults and resurface in hydrothermal springs, the location of which are transient over thousand- to ten-thousand–year timescales, as supported by geochronologic studies of hot spring deposits (Blackwell et al., 2007). Additionally, geothermal modeling of this region suggests that thermal activity and fluid flow along faults may vary according to permeability structure of the fault, with some portions of the fault favoring the upward flow of fluids, whereas other along-strike portions of the fault may behave in a hydraulically opposite sense, allowing fluids to flow downdip (McKenna and Blackwell, 2004; Wanner et al., 2014). Variations in geothermal gradient in the basins also suggest that fluids may flow basinward away from faults through permeable sedimentary layers and layers with favorably oriented fracture networks (Blackwell et al., 2007). Our study of neoformed clays offers novel documentation that supports previous assertions that hanging-wall rocks of evolving LANFs experienced extensive infiltration of surface fluid at least along some, if not all, portions of transient fault and fracture systems to depths of as much as 10 km over time periods of millions of years. This upper-crustal plumbing system provides a pathway for meteoric fluids to mid-crustal depths and formation of mineral deposits by mixing of meteoric fluids with deeper-sourced, metal-enriched fluids (Spencer and Welty, 1986; Roddy et al., 1988).

6.6 Evaluation of Post-Faulting Isotopic Exchange

A concern with stable isotopic analysis of clay minerals is the possibility that the measured isotopic values record late isotopic exchange at near-surface conditions and that the measured values do not reflect the conditions at the time of clay formation at temperatures of 60–180 °C for illite and smectite, or the greenschist-facies conditions at which the cataclastically reworked chlorites originally formed. To address this concern, we: (1) compared our calculated paleofluid compositions to present-day meteoric-water compositions near the faults we sampled, and (2) dated the sample material we used for the stable isotopic measurements by 40Ar/39Ar methods.

The stable isotopic composition of modern precipitation across the western United States has been investigated extensively and was recently reviewed by Lechler and Niemi (2011). The δ18O values of precipitation at sites closest to our sample sites range from −15.6‰ to −8.3‰ with a general trend toward more negative values toward the north and northwest (Friedman et al., 1992, 2002; Lechler and Niemi, 2012; Table 3 and Fig. 7). The δ2H of precipitation at sites closest to our sample sites range from −115‰ to –57‰, decreasing toward the north and northwest, generally correlating with δ18O and following the global meteoric water line across the Great Basin. The δ2H values of precipitation do deviate slightly from the meteoric water line during the summer months when evaporative fractionation effects are strongest (Friedman et al., 2002). Our calculated paleofluid compositions record a similar trend in that more isotopically depleted paleofluid compositions are also found at faults where present-day precipitation is strongly isotopically depleted. However, two lines of evidence suggest our calculated paleofluid compositions reflect ancient fluids and not late alteration or mixing with present-day fluids. (1) Calculated fluid compositions are sometimes isotopically heavier with respect to both oxygen and hydrogen (e.g., BAD-1) or lighter (ASH-1, S-PARK-1, WH68-1, and WH68-3) than present-day precipitation (Fig. 7), suggesting that there is not a direct relationship between calculated paleofluid composition and observed present-day precipitation. Where calculated paleofluid compositions are similar to present-day precipitation compositions (SEC 1-2 and SEC 4-2), the required 100–120 °C temperatures are inconsistent with vadose-zone interaction with current precipitation and instead consistent with higher-temperature interaction with an even more isotopically depleted fluid. (2) Dating of the authigenic clays also excludes the concern that mineral isotopic signals are indicative of late, near-surface low-temperature exchange after faulting. All of the 40Ar/39Ar ages for the samples listed in Table 2 are geologically consistent with clay growth while the sampled faults were active. The Ruby Mountains ages (reported in Haines and van der Pluijm, 2010, on splits from the samples used in this study) document the last major period of slip and fluid activity on the detachment at ca. 12 Ma, consistent with previous thermochronometer work (Colgan et al., 2010). The Panamint detachment gouge age of 3.6 ± 0.2 Ma is consistent with an inferred Pliocene time of slip (Andrew and Walker, 2009), while the mid-Miocene age for the Tucki Mountain gouge (16.9 ± 2.4 Ma) is also geologically plausible (Hodges et al., 1990). Likewise, Late Pliocene ages for gouge formation in Armargosa (3.2 ± 3.9 Ma), Mormon (2.8 ± 0.5 ma), and Badwater (3.3 ± 0.4 Ma) detachments of the Black Mountains record the last major pulse of motion on these LANFs (e.g., Knott et al., 2005; Norton, 2011). The dated fault rocks do not show evidence of significant postfaulting alteration, which demonstrates that meteoric fluid signatures preserved in neoformed clays are representative of ancient fluid circulation and not modern surface alteration.


Our study of neoformed clays and chlorites in exhumed shallow-crust to mid-crustal LANF systems shows that both LANF clay gouges and mid-crustal chlorite “microbreccias” exchanged isotopically with pristine to weakly evolved meteoric water. The presence of meteoric waters in LANF detachments at multiple crustal levels indicates these systems were hydrologically open for large parts of their history. Instead of recording lateral infiltration along major detachments or burial of pre-metamorphic fluids (e.g., Clark et al., 2006; Raimondo et al., 2011, 2013), we conclude that fluid circulation of crustal-scale LANF systems occurs by drawdown of meteoric waters through evolving fault and fracture networks that form and propagate in response to regional extension in the hanging wall, possibly aided by topography to drive fluid flow (Fig. 8).

Our dynamic scenario explains the observations of near-pristine meteoric water at upper- to mid-crustal levels in LANFs, with transient fault networks providing efficient pathways for significant quantities of meteoric water to reach into the crust. Our interpretation of a surface-to-depth plumbing system in LANFs and comparison with depth-to-surface fluids in thrust systems suggests that fluid dynamics of the upper crust is closely linked to the kinematic environment.

Research was supported by National Science Foundation (NSF) grant EAR-1118704 (to van der Pluijm) and the Turner and Wilson funds of the University of Michigan (to Haines and Lynch). We thank Monamie Bhadra and Kaajal Bhadra for assistance with fieldwork. We thank R.E. Anderson at the U.S. Geological Survey in Denver for pilot measurements that supported this study at the initial stages and Mike Spicuzza at University of Wisconsin–Madison for laser fluorination. We are grateful to staff at Death Valley National Park for permission to sample within the park area. We thank Christian Teyssier and an anonymous reviewer for reviews, as well as Gary Axen, Vincent Famin, and another anonymous reviewer for comments on an earlier version. Mulch acknowledges support through Deutsche Forschungsgemeinschaft (DFG) grant Mu2845/2-1. The Stable Isotope laboratory at UW-Madison was supported by grants NSF-EAR-1144454 and DOE-BES DE-FG02-93ER14389.

1GSA Data Repository Item 2016208, which contains 6 Data Repository Files: DR1: Description of outcrops sampled and geological context; DR2: Geospatial data of outcrops as .kmz file for GoogleEarth; DR3: Methods for isotopic analysis; DR4: Ar degassing spectra for dated samples; DR5: Compiled global δ18O and δ2H data for phyllosilicate minerals (illite, smectite, and chlorite); and DR6: Reference list for data compiled in DR5, is available at www.geosociety.org/pubs/ft2016.htm, or on request from editing@geosociety.org.
Samuel Haines