Spatiotemporal history of fluid-fault interaction in the Hurricane fault zone, western USA

The Hurricane Fault is a ~250-km-long, west-dipping normal fault located along the transition between the Colorado Plateau and Basin and Range tectonic provinces in the western U.S. Extensive evidence of fluid-fault interaction, including calcite mineralization and veining, occur in the footwall damage zone. Calcite vein carbon (13CVPDB) and oxygen (18OVPDB) stable isotope ratios range from -4.5 to 3.8 ‰ and -22.1 to -1.1 ‰, respectively. 10 Fluid inclusion microthermometry constrain paleofluid temperatures and salinities from 45–160 °C and 1.4–11.0 wt % as NaCl, respectively. These data identify mixing between two primary fluid sources including infiltrating meteoric groundwater (70 ± 10 °C ,~1.5 wt % NaCl, OSMOW ~-10 ‰) and sedimentary brine (100 ± 25 °C, ~11 wt % NaCl, OSMOW ~5 ‰). Interpreted carbon sources include crustalor magmatic-derived CO2, carbonate bedrock, and hydrocarbons. U-Th dates from 5 calcite vein samples indicates punctuated fluid-flow and fracture healing at 539 ± 15 10.8, 287.9 ± 5.8, 86.2 ± 1.7, and 86.0 ± 0.2 ka in the upper 300 m of the crust. Collectively, the data imply that the Hurricane Fault imparts a strong influence on regional flow of crustal fluids, and that the formation of veins in the shallow parts of the fault damage zone has important implications for the evolution of fault strength and permeability.


Introduction
Secondary mineralization, alteration products, and associated textures in fault rocks provide windows into the history 20 of past fluid-fault interaction in the crust. The fracture networks and associated sealing cements associated faults are widely recognized not only for their tectonic significance, but also for their impact on fluid movement and distribution in the crust of groundwater, hydrocarbons, and ore-deposits (Mozley and Goodwin, 1995;Benedicto et al., 2008;Caine and Minor, 2009;Eichhubl et al., 2009;Caine et al., 2010;Cao et al., 2010;Laubach et al., 2019). The rates, spatiotemporal patterns, and mineralogy of fracture sealing in fault zones control fault-zone strength, the buildup of 25 pore-pressures, location and frequency of failure events, and the overall fault system architecture through time (e.g., Caine et al., 1996). In order to constrain fluid-fault interaction during and after fault slip, we need information on the sources of fluids moving through the systems, their temperature and chemistry, and the age of fracture in-filling minerals that aid in their healing. Microscopy, geochronology, stable and radiogenic isotope geochemistry, bulk-rock and micro-scale geochemistry, and fluid inclusion analysis of diagenetic products in fault zones collectively inform 30 these processes.
Exhumed brittle faults and fault damage zones are excellent natural laboratories for interpreting the interaction between fluids and faults with implications for fault-zone permeability evolution, diagenesis, and the seismic cycle https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License. al., 2009). We present the first quantitative results on the spatiotemporal thermochemical history of paleofluid flow and fluid-rock interaction along the Hurricane fault zone using stable isotope geochemistry, fluid-inclusion microthermometry, and U-Th geochronology of calcite vein networks exposed in the footwall damage zone.
Geochemical data indicate that the paleofluids migrating along the fault were mixtures of deeply circulated meteoric 75 water and sedimentary brines with contributions from hydrocarbons and possibly recent magmatism. Textural and preliminary geochronological results from veins suggest punctuated fluid flow and fracture sealing events, possibly associated with fault slip along the Hurricane Fault.

Geological Setting of the Hurricane Fault
The Hurricane Fault strikes roughly north-south in the transition zone between the Colorado Plateau and the Basin 80 and Range tectonic provinces in southwest Utah and northwest Arizona (Fig. 1). Major tectonic events that have shaped the region include the Sevier orogeny, Laramide orogeny, and subsequent Basin and Range extension. The Sevier orogeny and associated fold and thrust belt initiated at ~125 Ma due to subduction and formation of a continental arc along the western margin of North America (Armstrong, 1968;Heller et al., 1986). The fold and thrust belt progressed eastward until shallowing of the subducting Farallon slab marked the onset of the Laramide orogeny 85 at ca. 75 Ma (Livaccari, 1991;Yonkee and Weil, 2015). The Laramide event is marked by basement-cored uplifts and formation of the Rocky Mountains. Hydration of the continental lithosphere during this time lab led to widespread magmatism following foundering of the Farallon slab (Humphreys et al., 2003). Basin and Range extension and widespread normal faulting in the western U.S. began in the late Eocene (Axen et al., 1993).

90
Normal faults of the Basin and Range broadly follow Proterozoic accretionary and Sevier-Laramide compressional structural fabrics to accommodate late Paleogene extension (Armstrong, 1968;Quigley et al., 2002). Extension along the eastern margin of the Basin and Range adjacent to the Colorado Plateau inititated ~ 15 Ma (Axen et al., 1993).
The Colorado Plateau province has remained largely un-deformed by Basin and Range extension, and the transition from the thick, strong crust of the Colorado Plateau to the relatively thin crust of the Basin and Range occurs over a ~ 95 100-km-wide interval (Zandt et al., 1995). The eastern margin of the transition zone is also coincident with the Intermountain Seismic Belt, (Fig. 1), with multiple seismically active normal faults including the Wasatch and Hurricane fault zones . Late Cenozoic volcanism along the margin between the two tectonic provinces is bimodal, indicative of high heat flow and partial melting of the mantle (Best and Brimhall, 1974).

100
The Hurricane Fault is a 250-km long, segmented, west dipping normal fault in southwestern Utah and northwestern Arizona with poorly constrained origins in the mid-Miocene to Pliocene (Lund et al., 2007;Biek et al., 2010). Fault activity occurred predominantly in the Pleistocene, including up to 550 m of its total 600-850 m of vertical displacement (Lund et al., 2007). Six segments of the Hurricane Fault are 30-40 km long and have been defined based on geometric and structural complexities at segmentation boundaries ( Fig. S1) (Pearthree et al., 1983;Stewart and 105 Taylor, 1996;Stenner and Pearthree, 1999). The Hurricane fault is recently active as evidenced by Quaternary scarps, and the magnitude ~5.8 earthquake occurring in 1992 east of St. George, Utah with a focus at ~15 km depth along the https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License. projected dip of the Hurricane fault surface (Stewart and Taylor, 1996). Long-term slip rates based on paleoseismic studies range from 0.44 to 0.57 mm/y (Lund et al., 2007).

110
Rock types juxtaposed by the fault include Paleozoic and Mesozoic sandstones, siltstones and mudstones, marine limestones, and evaporites (Biek, 2003;Biek et al., 2010). Exposures of hanging wall bedrock are broadly covered by Quaternary colluvium concealing Triassic units that are exposed in a few locations. Permian and Triassic units are well exposed in the footwall along the Hurricane cliffs, especially in canyons cutting the escarpment (Fig. S2) The Permian Hermit Formation includes fine-grained quartz-rich sandstones with minor hematite and calcite cements.
Where exposed along the fault, the Permian Queantoweap Sandstone is composed of quartz-rich sandstone, cemented by calcite and quartz. 120 Basaltic volcanism in the eastern Basin and Range in the transition zone to the Colorado Plateau began at ~15 Ma and has been most active within the last 2.5 My (Nelson and Tingey, 1997). Quaternary basaltic volcanic centers are spatially associated with the Hurricane fault (Fig. 2). Basalt flows are offset by the Hurricane faults, and these are used for constraining long-term slip rates (Lund et al., 2007). Volcanic eruptions are predominantly alkali-rich basalts 125 with lesser basaltic andesite. Neodymium isotope ratios of Quaternary basalts reflect primarily lithosphere sources along the northern half of the Hurricane Fault and asthenosphere/mixed source to the south (Crow et al., 2011). These periods of basaltic magmatism associated with Basin and Range extension may have created hydrothermal systems in the past that locally influenced groundwater chemistry and circulation in the Hurricane Fault.

130
Prior work on fluid movement associated with this fault is limited to geochemical and isotopic studies of modern spring systems at Pah Tempe hot spring, near La Verkin, UT, and the Travertine Grotto and Pumpkin warm springs in Grand Canyon. At Pah Tempe hot spring, deeply-circulated meteoric waters emerge as CO2-charged and saline fluids along the fault trace, and precipitation of calcite veins is evident in the exhumed fault rocks (Nelson et al., 2009).
Travertine Grotto and Pumpkin warm springs are attributed to meteoric water mixing with deeply-sourced fluids that 135 are flowing upwards along the basement-rooted Hurricane Fault (Crossey et al., 2006;Crossey et al., 2009). Analyses of volatiles exsolving from these springs identifies a predominantly deep (endogenic) source, with some modern contributions from mantle or magmatic sources. https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License.

Field Locations
Field investigations along the Hurricane Fault were conducted between Cedar City, UT and the fault's intersection with Grand Canyon in Arizona (Fig. 2). Studies were restricted to well-exposed areas of the fault zone, typically where drainages cross the fault. Due to colluvial cover on the hanging wall, this study focused on footwall exposures of the fault and damage zone. Twenty-three field stations ( Fig. 2) along Hurricane Fault were investigated, and representative 145 hand samples were chosen for subsequent microscopic and geochemical characterization of diagenetic alteration and secondary vein mineralization. Sampling criteria included vein morphology, cross-cutting vein relationships, varying vein/fracture orientations, and range of apparent diagenetic modification, including unaltered host rocks. Sample locations were recorded using a GarminTM GPS unit in decimal degrees relative to the WGS 1984 datum (Table S1).

Microscopy 150
Standard petrographic thin sections were made from 34 hand samples displaying a range of vein types and diagenetic alteration. Of these 34 samples, 15 petrographic doubly polished thick sections (150 μm) of calcite veins were prepared for fluid inclusion analyses. Thin section petrographic observations were made using Leica Z16 APO and Leica DM 2700P petrographic microscopes. Photomicrograph images were acquired with a Leica MC 170 HD camera and processed using the Leica Application Suite 4.6 software. 155

Fluid inclusion microthermometry
Fluid inclusions in secondary calcite mineralization were investigated using a Zeiss Universal transmitted light microscope with a Zeiss Epiplan 50x long-working distance objective. A USGS gas-flow heating and freezing stage was used to measure fluid inclusion homogenization and melting temperatures. The stage was calibrated to the critical point of water using a synthetic supercritical H2O inclusion (374.1 °C), the freezing point of a synthetic 25 mol % 160 CO2-H2O inclusion (-56.6 °C), and the freezing point of double-deionized water using an ice bath (0 °C). Using the 15 thick sections, 107 homogenization temperatures (Th) and 35 melting temperatures (Tm) were determined from two-phase fluid inclusions in calcite (Table 1).
Fluid inclusion were classified and homogenization and melting temperatures were determined using criteria and 165 procedures described by Goldstein and Reynolds (1994) and Goldstein (2001). After performing heating measurements, numerous 2-phase fluid inclusions with homogenization temperatures from 45 -85 °C became metastable 1-phase liquid inclusions (i.e., the bubble did not re-nucleate upon cooling). In order to re-nucleate the second phase to facilitate measuring the melting temperatures, these fluid inclusions in these samples were intentionally stretched by heating to 110 °C for 18 hours in a laboratory oven (Goldstein and Reynolds, 1994). For a 170 few of these treated inclusions, unreliable melting temperatures >0 °C were obtained, and these were omitted from the data set. No pressure correction was performed to convert Th measurements to trapping temperatures (Tt). Assuming vein formation at a maximum depth of 800 m equivalent to the maximum throw on the fault (Anderson and Mehnert, https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License. 1976), a maximum pressure using a lithostatic load (2675 kg m-3 rock density), and the maximum measured Th of 160 °C, the pressure correction is <10 °C (Fisher, 1976;Bakker, 2003) and considered insignificant for this study. Th 175 measurements in this study are considered representative of Tt.

Carbon and oxygen stable isotope analysis
A Dremel® tool was used to collect 290 powdered sub-samples from calcite veins, mineralized fracture surfaces, limestone host rock, and calcite-cemented sandstone host rock. Carbon and oxygen stable isotope analyses of these samples was performed in the Utah State University Department of Geosciences Stable Isotope Laboratory using a 180 Thermo Scientific Delta V Advantage Isotope Ratio Mass Spectrometer (IRMS) and a GasBench II using the carbonate-phosphoric acid digestion method (McCrea, 1950;Kim et al., 2015). Specifically, ~ 120-150 μg aliquots of relatively pure calcite samples and standards were placed into 12 ml Exetainer© vials and flushed with ultra-highpurity helium. Impure carbonate cements (e.g., calcite-cemented sandstone) required 300 to 8000 μg of sample to achieve acceptable peak amplitudes during analysis. After helium flushing, ~ 100 μL of anhydrous phosphoric acid 185 was added to each sample and allowed to react for two hours at 50 °C before analysis. Carbon and oxygen stable isotope ratios were calibrated and normalized to the VPDB scale using the NBS-19 and LSVEC, and NBS-19 and NBS-18 international standards, respectively (Kim et al., 2015). In house calcite standards were used to correct for drift and mass effects. Carbon and oxygen isotope ratios are reported using delta notation (δ13C, 18O) in per mil (‰) relative to VPDB. Based on repeat analyses of in-house calcite standards, errors on δ13C and 18O are <0.1 ‰. Oxygen 190 isotope ratios are also converted and reported relative to standard mean ocean water (SMOW) for fluid inclusion calculations using Eq. (1) (Sharp, 2007): Eq. (1)

Uranium-thorium (U-Th) dating
Pilot U-Th geochronology was conducted on 5 key calcite vein samples from two field locations (Table S2). These 195 include locations 1-2 and 1-4, where veins are hosted in limestone and sandstone strata, respectively (Fig. 2). Veins were slabbed with a rock saw and approximately 300 mg of calcite powder was collected from discrete veins or vein laminations using a Dremel® tool and submitted to the University of Utah ICP-MS laboratory for analyses. At location 1-2, one laminated vein was subsampled at two locations (one near vein wall and near the outer part of the vein) to capture the timing of vein growth. At location 1-4, 3 generations of veins, determined based on cross-cutting 200 relationships, were subsampled.
Chemical preparation and analyses were performed at the University of Utah following methods modified from Edwards et al. (1987) using a Thermo NEPTUNE Plus Multi-Collector-Inductively-Coupled-Mass-Spectrometer (MC-ICP-MS). Powdered carbonate samples were dissolved in 16 M HNO3 and equilibrated with a mixed 229Th-233U-205 236U spike and refluxed on heat for at least one hour to ensure total dissolution. Uranium and thorium sample fractions were separated for analyses by anion exchange column chemistry. Measured peak heights were corrected for abundance sensitivities and mass bias, dark noise, background (blank) intensities, hydride contributions, ion-counter https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License. yields, and spike contamination. The spike was calibrated against solutions of CRM 145 and HU1 uraninite.
Uncorrected age uncertainties are reported as one standard error and include measurement error and uncertainties of 210 activity. Details of the spike calibration and data treatment can be found in Quirk et al. (in review) .

Fault zone diagenesis and veins
Evidence for fluid-fault interaction along Hurricane fault zone exists at the macroscopic and microscopic scale.
Collectively referred to as "fault zone diagenesis" (Knipe, 1992), these observations form the foundation for 215 subsequent geochemical and geochronological work. Examination of the fault zone at the field sites (

Carbon and Oxygen stable isotope ratios
Stable isotope ratios of carbon and oxygen were determined for calcite veins and host rocks from the field sites (see 235 data repository, Newell and Koger, 2020). The δ13CVPDB and δ18OVPDB values for the entire data set range from -4.5 to 3.8 ‰ and -22.1 to -1.1 ‰ (8.1 to 29.8 ‰ vs SMOW), respectively (Fig. 3 a). In the host rock units, carbonate cements in siliciclastic units and bulk limestone host-rock were analysed adjacent to veins and at ~ 1-2 m away for comparison.
Away from fractures, host rock δ13CVPDB and δ18OVPDB values range from -2.0 to 3.8 ‰ and -8.5 to -1.1 ‰, 240 respectively. The 13C and δ18O values for calcite veins, breccia cements, mineralized fractures, and slip surface cements span a wide range of values with considerable scatter. For the purposes of presentation and discussion these https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License. data are divided into 4 "vein sets" based on common lithological associations, vein morphologic features, and 13C and δ18O data patterns (Fig. 3 a). Note that these 4 vein sets span multiple locations (Fig. S1) and show no correlation in C and O isotope ratios with location. 245 Vein set 1 calcite exhibits a positive correlation (slope = 1.6) between 13C and δ18O and is commonly intergrown with hematite when hosted in siliciclastic strata. Calcite in set 2 also displays a positive 13C and δ18O correlation (slope = 0.9) with δ13C shifted to lower values compared to vein Set 1. Set 3 has a wide range of isotopic values, showing no strong trends or patterns. Set 4 calcite 18O values that overlap with set 3 with δ13C values that trend to 250 significantly lower values. The majority of set 4 data are from location (1-2).

Fluid inclusion microthermometry
Of the 15 thick sections of calcite veins observed, 6 contain populations of two-phase fluid inclusions that yield homogenization (Th) and melting (Tm) temperatures (Table 1, Fig. 4). Homogenization temperatures are used to approximate the trapping temperature (Tt) and are representative of fluid temperatures during mineralization. Melting 255 temperatures depend on the nature and concentration of dissolved species and are used to estimate the salinity of paleofluids (Bodnar, 1993).
Fluid inclusion homogenization and melting temperature data is organized by the calcite vein set as described in  (Goldstein and 265 Reynolds, 1994;Goldstein, 2001). Ice melting temperatures from vein set 1 range from -3-0 °C, equating to a salinity of 0 to 5 wt% as NaCl (Fig. 4). Calcite set 3 yield melting temperatures from -11 to 0 °C, equating to 0 to 15 wt% NaCl. Since no initial melting was observed, NaCl dominated salinity is assumed and calculated via Eq. (2) where Tm is the measured melting temperature in °C (Bodnar, 1993).

Paleofluid sources in the Hurricane fault
The carbon and oxygen stable isotope ratios of the calcite veins can inform the groundwater composition, source, and processes at work during paleofluid circulation in the Hurricane Fault. The C and O equilibrium isotopic fractionation between CO2 and calcite (cc) and water and calcite (cc), respectively are temperature dependent, and assuming that isotopic equilibrium during mineralization is valid, additional information on the paleofluid temperature is needed to 285 proceed. Homogenization temperatures of primary 2-phase fluid inclusions in calcite, when present, are a reliable method to estimate temperature, and thus to calculate the paleofluid O and C isotopic composition using Eq. (3) (Deines et al., 1974) Eq. (4) where αx-y is the temperature dependent fractionation factor between water and calcite (H2O-cc), CO2 and calcite (CO2cc), and T is temperature in Kelvin. For the fractionation factor magnitudes expected for these two systems, the difference in delta values between the phases (i.e., 18OH2O -18Occ and 13CCO2 -13Ccc) is a good approximation for 295 1000lnα (Sharp, 2007). When fluid inclusion data are not available, temperatures may be estimated based on other constraints, such as estimates on mineralization depths and the geothermal gradient, but the resulting paleofluid isotopic estimates will be far more uncertain due to surface-ward advection of geotherms (Forster and Smith, 1989).
Clumped carbonate isotopic methods (47) can yield reliable temperature estimates from fault-zone calcite mineralization (Swanson et al., 2012;Hodson et al., 2016), but are not available for this study. In the absence of 300 these constraints, a range of temperatures or starting fluid isotopic compositions can be explored to provide some interpretations of the calcite stable isotope data, again resulting in considerable uncertainty.
For the 6 samples that hosted populations of two-phase fluid inclusions, microthermometry heating and freezing data are used to estimates of fluid trapping temperatures and salinities of the paleofluids present in the Hurricane 305 Fault. In combination with oxygen stable isotope ratios from the calcite hosting these fluid inclusions, the paleofluid  (Table 1). Similarly, we associate this range of oxygen isotope values to the mean and standard deviation of the paleofluid salinity as estimated from fluid inclusion melting temperatures. The paleofluid  18 O and salinity (wt % as NaCl) estimates for these samples show a strong positive correlation (r 2 = 0.8; Fig 6). We interpret this correlation as mixing between two endmember fluid types, and that over the history of We suggest that the endmember characterized by high  18 O and high salinity is consistent with sedimentary 320 formation water (brine) that originated from extensive meteoric water-rock interaction and oxygen isotope exchange with marine sedimentary sequences (e.g., Clayton et al., 1966). Assuming a 25 -30 °C geothermal gradient and the range fluid inclusion temperatures, circulation depths for these ground waters ranges from 2 to 6 km. This is adequate to infiltrate all of the Mesozoic and Paleozoic strata in the region, including thick sections of marine carbonate and evaporite bearing units (Biek, 2003;Dutson, 2005;Biek et al., 2010). Infiltration into these marine 325 units is a likely source for the salinity in these ground waters. The endmember characterized by relatively low-salinity and low  18 O is likely dominantly meteoric groundwater. Using the same geothermal gradient, these ground waters have circulated to ~3.5 km based on fluid inclusion constraints. For comparison, Pah Tempe hot springs (Nelson et al., 2009), and Pumpkin and Travertine Grotto springs (Crossey et al., 2009) emanate along the Hurricane fault and have similar oxygen isotope composition and salinity to this endmember (Fig. 6). Based on comparisons of Pah 330 Tempe hot spring  18 O and  2 H with other local and regional meteoric waters, Nelson et al. (2009) interpret the source of the hot spring water as meteoric water that infiltrated during the last glacial interval. Based on geochemical geothermometry estimates, and the observed shift in hot spring water to higher  18 O, Nelson et al. (2009) suggest that Pah Tempe thermal waters circulation depths of 3-5 km with temperatures of 70-150 °C. This approach has also been employed at other faults to interpret paleofluid compositions. For example, coupled fluid inclusion 335 microthermometry and stable isotope values from fault-hosted calcite along the Moab fault, UT, USA, point to a mixing process between upwelling basin brines with meteorically derived groundwater (Eichhubl et al., 2009).
In terms of the carbon sources in these two fluids, there are alternative ways to interpret the relatively narrow range of calcite  13 C values (0.35 to 1.73 ‰). First, using the average calcite formation temperatures from fluid inclusions, 340 we can estimate the carbon isotopic composition of the paleofluid dissolved CO2 from -6.1 to -4.3 ‰ (VPDB) using Eq. (4) ( Table 1). However, unlike the oxygen isotope system that most likely reflects the water composition, carbon composition can be reflective of a carbonate host rock. For example, dissolved carbonate in equilibrium with limestone bedrock (i.e., strongly buffered by the host rock) will result in calcite veins with a  13 C similar to the host limestone (e.g., Dietrich et al., 1983). In this case, calculating the carbon isotopic composition of and external CO2 345 source may not be appropriate, and the vein value is simply representative of the source carbon. In this study, the https://doi.org/10.5194/se-2020-69 Preprint. Discussion started: 25 May 2020 c Author(s) 2020. CC BY 4.0 License. host rock limestone  13 C values range from -2.7 to 3.8 ‰ with an average of 1.2 ‰, which is in the range of expected values from marine carbonates (e.g., Hoefs, 1987;Sharp, 2007). However, in parts of the fault that have higher water-rock ratios or are generally carbonate poor (e.g., siliciclastic host rock), the carbon isotopes of calcite veins can be representative of an external CO2 source dissolved and traveling in the groundwater. With these uncertainties 350 in mind, we interpret the endmember carbon sources for the calcite veins as external CO2 sources and local marine limestones. Based on the results from this study, there may be a weak association between the two carbon sources and the fluid endmembers based on oxygen isotopes and salinity. In some but not all cases, vein  13 C values that are similar to host limestone tend to be associated with the highest salinity fluids. Veins hosted in sandstone units and associated with an external CO2 source (~-6 ‰) are in most cases associated with the lower salinity fluids. These 355 carbon isotope values are similar to the observed  13 C of CO2 at Pah Tempe (-5.5 ‰) and Pumpkin (-6.1 ‰) springs (Crossey et al., 2009;Nelson et al., 2009). These values overlap with mantle CO2 values (Marty and Jambon, 1987), but are also are similar to values observed in many crustal fluids and continental hot springs globally (Sherwood Lollar et al., 1997;Ballentine et al., 2002;Newell et al., 2008;Newell et al., 2015). Based on helium and carbon isotopes, Crossey et al. (2009) andNelson et al. (2009) suggest that mantle CO2 could range from a just a few percent 360 to as high as ~40 % in the Hurricane fault hosted hot springs, depending on the mantle and crustal endmembers used. We do not have constraints on the helium isotope ratios of the paleofluids, so we cannot further evaluate the possibility of magmatic contributions.

Subsurface processes impacting isotopic values
As shown earlier, the  13 C and  18 O values from calcite veins and cements associated with the Hurricane fault display 365 a large range of values (Fig. 3 a). In addition to the binary mixing described in section 5.1, precipitation of calcite from fluids with a range of temperatures is occurring along flow paths. A fairly wide range of temperatures is evident from the fluid inclusion work on vein sets 1 and 3. For a given water  18 O and  13 CCO2, varying temperature in Eqs.
(3) and (4) results in trends in calcite  18 O and  13 C with a slope of ~2.3 (Fig 3 b). To explore the impacts of both temperature change and mixing, the calcite forming from the saline ( 18 OVSMOW = 5 ‰) and meteoric water 370 ( 18 OVSMOW = -10 ‰) endmembers, both with a  13 CVPDB = -6 ‰ are superimposed on the observed data (Fig. 3 b, shaded region). The vein set 1 pattern is fairly well matched by calcite forming over a the range of T consistent with the fluid inclusion measurements (90-45 °C) from the low-salinity meteoric end member ( 18 OVSMOW = -10 ‰ ;  13 CVPDB = -6 ‰). The scattered values observed for vein set 3 are encompassed by the calcite forming from mixed saline and meteoric fluids over the range of temperatures consistent with the fluid inclusion measurements (160-50 375 °C). Although not shown on the diagram, using a limestone-buffered  13 C (~1 ‰) predicts values that are not consistent with any of the observed data, and this suggests that an external source of CO2 may be most appropriate for veins in both the carbonate and siliclastic host rock. The vein set 4 isotopic values are best explained by mixing the saline end member at ~100 °C with a  13 C source of ~ -12 ‰ (Fig. 3 b). This low  13 C value is consistent with derivation from organic matter (Boles et al., 2004). Hydrocarbons are present regionally and in the strata that hosts the Hurricane fault (Bahr, 1963;Blakey, 1979). Mobilization and microbial oxidation of these hydrocarbons to form dissolved carbonate (Baedecker et al., 1993;Tuccillo et al., 1999) has been shown in other fault settings to form calcite veins with low  13 C values (e.g., Eichhubl et al., 2009). range in  13 C and  18 O values. In combination with U-Th geochronological constraints, these authors suggest that this system has been active periodically for >100 ky with a consistent paleofluid source and isotopic composition.

420
Focusing on vein set 1 and a range of temperatures observed from fluid inclusion, the computed trends cannot explain the observed data (Fig. 3 b). Rayleigh fractionation trends under equilibrium conditions for CO2 degassing and coupled CO2 degassing + calcite precipitation <85 °C yield positive correlations that are not similar to the observed slope or range of values in set 1 calcite veins. Based on this analysis, it is unlikely that open system processes such as these are major processes involved in the formation of calcite veins in the Hurricane fault zone. It 425 is important to note that these Rayleigh fractionation models assume isotopic equilibrium. Rapid degassing and calcite precipitation may result in disequilibrium and kinetic fractionation that cannot be quantitatively addressed.
To summarize these analyses and interpretations, most of the vein C and O isotopic compositions observed can be explained by a combination of the mixing of two primary fluid endmembers over a range of temperatures, with 430 second order impacts from open system processes such as degassing during calcite precipitation. Vein set 1 is best explained by formation over a range of temperatures from the low-salinity endmember. Most of the values in set 2 and 3 can be explained by a mixture of the low-salinity meteoric and the sedimentary brine end members and precipitation over a range of temperatures. Vein set 4 requires addition of a much lower  13 C component to the fluids responsible for vein set 3, likely derived from an organic source. 435

Implications of vein geochronology
Pilot U-Th geochronology on 5 samples indicates that calcite veins formed from 539 to 86 ka. These samples are from two different sample locations (1-2 and 1-4) separated by 13 km along strike (Fig. 2, 5), and from vein sets 1, 3, and 4. Specifically, at location 1-2, hosted in limestone, veins formed at 133 ka and 86 ka (set 4). At location 1-4, hosted in sandstone, veins formed at 539 and 288 ka (set 1) and 86 ka (set 3). As described in the results, the dates are 440 consistent with interpreted vein growth direction and cross-cutting relationships (Fig. 5).
Based on the stable isotope results and analyses, the 539 and 288 ka veins are likely associated with the low-salinity meteoric water endmember (Fig. S6) and formed at moderate temperatures (60-70 °C). The 113 ka and both 86 ka veins are best associated with ~100 °C saline groundwater, with varying contributions of a low 13C carbon source 445 (Fig. S6). The 86 ka sample from location 1-2 has the lowest 13C (-7 ‰) observed in this study, and as discussed in section 5.2 requires an organic carbon source.
Although this data set is small, it suggests punctuated vein forming events. Interestingly, these two fault locations appear to preserve the 86 ka event, both requiring similar composition and temperature fluids, suggesting that the fluid circulation events have continuity over ~13 km of fault zone strike. These dates can be used along with constraints on fault slip rate to estimate the maximum depth of vein formation. Using the published slip rate estimates of 0.44 to 0.57 mm/y (Lund et al., 2007), this equates to vein formation of depths of ~40 to 300 m. This assumes negligible exhumation of the footwall over the last 540 ka. Consistent with the findings at thermal springs along the Hurricane Fault (Crossey et al., 2006;Crossey et al., 2009;Nelson et al., 2009), this indicates that deeply-circulated thermal 455 fluids are moving up the fault zone, advecting deeper geotherms towards the surface.
Although not the primary objective of this paper, the calcite vein textures in context of the preliminary geochronological results warrant a brief discussion. The vein wall breccias and laminated calcite veins observed along the Hurricane Fault share similar characteristics to those in other major fault zones that have been attributed to co-460 seismic or post-seismic sealing (e.g., Nuriel et al., 2011;Nuriel et al., 2012). These fracture openings filled with laminated growth bands of fibrous calcite crystal are indicative of post-fracture opening sealing (crack-seal cycle) (Ramsay, 1980). These suggest fluid-pressurization and fluid flow cycles associated with periodic fracturing in the fault damage zone, possibly due to seismic activity (e.g., Sibson, 1994). Williams et al. (2017b) showed that detailed U-Th dating of these types of laminated veins inform the periodic nature of fracture opening and sealing via calcite 465 precipitation and argue these are associated with seismic events. Specifically, they documented 13 seismic events between 550 ka and 150 ka and use this to estimate long-term earthquake recurrence intervals. In addition to earthquakes, Pleistocene climatic cycles could influence groundwater flow and fluid pressure, and possibly be associated with vein forming events similar to what is observed at the Little Grand Wash and Salt Wash faults (e.g., Kampman et al., 2012). We have documented 4 such events in the last ~540 ka, suggesting that a similar high-470 resolution geochronological study could yield meaningful information about the long-term recurrence of fluid-flow triggering events along the Hurricane fault zone, whether triggered by seismicity or linked to climatic cycles.

Supplemental Information 500
Supplemental information, tables, figures are available at the following link: XXXXXX.

Author Contributions
Jace Koger conducted the field sample collection, sample preparation, and sample analyses as part of the requirement for his MSc in Geology from Utah State University (Koger, 2017). Dennis Newell was the thesis supervisor, provided assistance and mentorship on sampling and analytical techniques, guidance on data analysis, and is the corresponding 505 author for the preparation of this manuscript.

Competing Interests
None.

Acknowledgements
We thank Andrew Lonero (USU Geosciences) and Diego Fernandez (University of Utah) for their assistance with 510 stable isotopic and U-Th analyses, respectively. Funding for this research was provided by a Geological Society of America Student Research Grant to J. Koger

560
The strong positive correlation of 18O and salinity (R2 = 0.8) is interpreted as a mixing trend between and low salinity, meteoricaffinity groundwater and high salinity sedimentary brine endmember. For comparison, the composition of Pah Tempe, Pumpkin, and Travertine Grotto springs are included, are very similar in composition to the low salinity endmember.