We assessed tectonic controls on the spatial and temporal distribution of fault zone flow pathways in the Rio Grande rift (New Mexico, USA) by using fault zone calcite cements as a geochemical record of syntectonic fluid flow. Cement δ18O, δ13C, and 87Sr/86Sr values indicate that older, large-displacement master and basin-margin faults were cemented by more isotopically evolved basinal brines than younger intrabasin faults. These data suggest that diagenetic fluids in basin-bounding faults equilibrated predominantly with downdip Paleozoic carbonates. In contrast, intrabasin faults transmitted fluids from shallow stratigraphic sources. This pattern of flow pathways is linked to the systematic distribution of sediments and faults that record rift evolution, which dictated spatial and temporal variations in fault zone architecture and permeability structure. Our results indicate that the depths from which fluids can be transported in active rift basins ultimately depend on both tectonically mediated variations in the grain size of syntectonic sediments entrained in fault damage zones and fault displacement magnitude.
Understanding basin-scale subsurface fluid flow is critical to petroleum exploration (e.g., Bethke et al., 1991; Gong et al., 2011), geological CO2 sequestration (e.g., Birkholzer et al., 2009; Person et al., 2010), and contaminant transport modeling (e.g., Carle et al., 2006). In rift basins, factors affecting regional flow patterns include variations in sediment type (Person and Garven, 1994; Mailloux et al., 1999), basin subsidence and/or compaction (Cathles and Smith, 1983; Cartwright, 1994), and hydrothermal circulation (Person and Garven, 1992; Simms and Garven, 2004). Faults included in regional flow models are often treated as discrete boundaries with no intrinsic permeability (Mailloux et al., 1999). Few previous studies have explicitly considered fault zone fluid transport (cf. Simms and Garven, 2004; Guillou-Frottier et al., 2013), despite the fact that fault zones can be distinct hydrologic units with measurable permeability anisotropy (Rawling et al., 2001; Bense and Person, 2006), capable in some cases of transporting fluids from great depth to the surface (Crossey et al., 2006; Williams et al., 2013).
The systematic distribution of sediments and faults established during extension in rift basins (Gawthorpe and Leeder, 2000; Fig. 1A) suggests that fault zone permeability structure will vary with basin position, as the sediment or rock type cut by a fault exerts a primary control on fault zone architecture (Rawling et al., 2001). If correct, this hypothesis will help predict which faults serve as conduits for fluids from differing stratigraphic levels. We tested this hypothesis by field, microstructural, and geochemical studies of fault zone calcite cements (a geochemical record of fluid source) in different structural positions in the Rio Grande rift, New Mexico (USA) (Fig. 1B). We show that syntectonic fluid source depths varied systematically with fault zone structural position, providing, for the first time, an opportunity to examine the role of tectonic processes in determining the spatial and temporal distribution of fault zone flow pathways during basin development.
Basins of the Rio Grande rift are bound by large normal faults (Fig. 1A; master and basin-margin faults) that accommodated the majority of extension since rifting began in the middle Oligocene (Chapin and Cather, 1994). In the Late Miocene, faulting near basin margins slowed, and subsidence was increasingly accommodated by numerous, relatively small displacement (<200 m) intrabasin faults that formed nearer basin axes. Middle Oligocene to Miocene basin fill (Fig. 1A; Lower Santa Fe Group, LSF) records closed-basin deposition, and grades from coarse-grained, high-permeability alluvial fan and fluvial deposits near uplifted basin margins to fine-grained, low-permeability lacustrine and playa deposits nearer basin centers (Cather et al., 1994; Hawley et al., 1995). Pliocene–Pleistocene basin fill (Upper Santa Fe Group, USF) records the establishment of throughgoing axial drainage, and grades from basin-margin alluvial fan to eolian and fluvial sediments nearer basin centers. Most faults in the Rio Grande rift exhibit low-permeability clay-rich cores flanked by damage zones (Minor and Hudson, 2006). Where faults cut poorly lithified Santa Fe Group sediments, damage zone structures record particulate flow (mixed zones of Rawling et al., 2001; Rawling and Goodwin, 2006). Where faults cut fully lithified pre-rift units, damage zones are dominated by fracture networks (cf. Caine et al., 1996).
This spatial distribution of sediments and normal faults allows us to define three end-member fault types (Fig. 1A): (1) master faults, the largest displacement, basin-bounding faults, which juxtapose coarse-grained, high-permeability alluvial fans and lithified pre-rift units; (2) basin-margin faults, which have relatively large downdip displacements but juxtapose relatively coarse grained USF and LSF sediments near the surface; and (3) relatively small displacement, intrabasin faults that juxtapose fine-grained, low-permeability lacustrine and playa deposits of the LSF throughout much of their downdip extent. High-permeability hanging-wall damage zones of sheared, rift-margin sediments and footwall damage zones with a significant extent of fractured pre-rift units suggest that master and basin-margin faults may act as conduits for deep basinal brines (Fig. 1A). Conversely, damage zones of sheared, low-permeability LSF lacustrine and playa deposits should block flow between basement and USF damage zones, implying that intrabasin faults transport fluids between shallow fluvial and eolian aquifers of the USF. Calcite cement is common in these faults and oriented concretions locally record fault-parallel flow (Mozley and Goodwin, 1995; Minor and Hudson, 2006). Locally, cements are cut by fractures and slickenlines are evident between cemented damage zone and core, recording syntectonic cementation (Heynekamp et al., 1999). Determining the sources of fluids that precipitated these syntectonic fault cements is key to testing our hypothesis.
We studied five representative fault types in the Albuquerque Basin (northern Albuquerque and Santo Domingo subbasins) and Socorro Basin (Fig. 1B). All have USF sediments in cemented hanging-wall damage zones, indicating that fluids were transported to similar stratigraphic levels. Fault names and locations, footwall characteristics, and additional details of hanging-wall sediments follow. (1) The La Bajada fault, a master fault of the Santo Domingo subbasin, has a footwall of Mesozoic sedimentary rocks. (2) The Loma Blanca fault, a Socorro basin-margin fault, juxtaposes Pleistocene alluvium and Pliocene fluvial sands. (3) An intrabasin transfer fault, here referred to as the Santa Ana transfer, juxtaposes USF fluvial and eolian sands. (4) The informally named Arroyo Sediento intrabasin fault juxtaposes USF fluvial sands. (5) The Sand Hill basin-margin fault has a footwall of LSF sediments. The latter three faults are located in the northern Albuquerque subbasin.
The distribution and character of fault zone structures, lithologic units, and mesoscale cement paragenesis were documented at each site. Samples were collected from well-cemented portions of fault zones, which in these faults are restricted to hanging-wall damage zones. Modal mineralogy was determined by point counting. We selected a subset of samples from each fault for electron microprobe, oxygen, carbon, and strontium isotope analyses. Details of methods and tabulated data are given in the GSA Data Repository1.
Cements were precipitated in immature sands (Table DR1 in the Data Repository). Pore-filling calcite cement accounts for 20–60 vol% of the rocks. Sections cut perpendicular to fault strike display a well-defined shape-preferred orientation of elongate grains subparallel to each fault (Fig. DR1). In contrast, elongate grains outside of the fault zone are aligned parallel to bedding (Rawling and Goodwin, 2006). No evidence of multiple generations of calcite cement was found in either petrographic or backscattered electron images of samples. Cements are nonluminescent, and record no evidence of post-cementation deformation.
Fault cements are nearly pure calcite; Mg and Sr are the only trace elements detected by electron microprobe. Master fault cements are enriched in Mg and Sr relative to basin-margin and intrabasin fault cements (Figs. 1C and 2E). Intrabasin faults have the lowest Mg and Sr concentrations. Concentrations of Mg and Sr decrease systematically with distance from the master fault, then increase near basin-margin faults on the opposite side of a basin.
Fault calcite cement 87Sr/86Sr, δ18O, and δ13C values also show systematic trends as a function of fault position (Fig. 2). Master and basin-margin faults show relatively high 87Sr/86Sr, δ18O, and δ13C values compared to intrabasin faults. Master and basin-margin faults also show a relatively narrow range in 87Sr/86Sr, δ18O, and δ13C values, whereas intrabasin faults have nearly constant δ18O and a wide range of 87Sr/86Sr and δ13C values (Figs. 2A–2C).
Fault cement chemistry varies spatially in the Rio Grande rift (Figs. 1C and 2). We consider these variations in terms of fluid source to test our hypothesis that fault zone permeability structure varies systematically with basin position.
Precipitation from meteoric fluids is suggested by the comparatively low δ18O, low δ13C (organic carbon), and Mg and Sr concentrations of intrabasin fault cements (Figs. 2A and 2E; cf. Plummer et al., 2004; Williams et al., 2013). Calcite precipitation temperatures of 4–22 °C calculated from cement δ18O values assuming equilibrium fractionation (Kim and O’Neil, 1997) and modern meteoric water values for the study area (−10‰ to −14‰ Vienna standard mean ocean water, VSMOW; Kendall and Coplen, 2001) support this interpretation. Relatively low 87Sr/86Sr ratios of intrabasin fault cements can also be explained by precipitation from a fluid comparable in 87Sr/86Sr ratio to the Rio Grande river (headwater values of ∼0.7096; Hogan et al., 2007), possibly following minor interaction with relatively young rift volcanic rocks (∼0.7020–0.7080; McMillan, 1998).
Cementation by deep basinal brines is consistent with the higher 87Sr/86Sr, δ18O, and δ13C values in master and basin-margin fault cements. As meteoric fluids contain negligible Sr2+ concentrations and the Rio Grande headwaters have 87Sr/86Sr ratios of ∼0.7096, cement 87Sr/86Sr > ∼0.7100 in master and basin-margin faults indicate fluid input from deeper (i.e., older) crustal sources (Banner, 1995; Crossey et al., 2006; Hogan et al., 2007; Williams et al., 2013). Carbon isotope data for these faults are also consistent with cementation by high δ13C, deep basinal fluids (Fig. 2A), an interpretation supported by high concentrations of Mg and Sr (Figs. 1C and 2E). Furthermore, precipitation temperatures calculated for master and basin-margin fault cements using meteoric δ18O values are improbably low (−8 to 9 °C), indicating an isotopically evolved fluid source. Deeply derived fluids associated with modern hot springs in the Rio Grande rift have δ18O between −2‰ and 6‰ VSMOW (Williams et al., 2013), although some of these values may reflect evaporative enrichment.
The correlation between Mg and Sr in master and basin-margin fault cements and their 87Sr/86Sr, δ18O, and δ13C values are consistent with cementation by a fluid that equilibrated mainly with Paleozoic carbonate units (Fig. 1A). Williams et al. (2013) showed that dissolved inorganic carbon in Rio Grande rift waters was sourced in part from these units. This interpretation explains the positive correlation between cement 87Sr/86Sr, Mg and Sr concentration, and δ13C, as marine limestones can provide substantial Mg2+, Sr2+, and near-zero δ13C values to solution. Paleozoic carbonates are likely to have 87Sr/86Sr ratios of 0.7070–0.7084 (McArthur et al., 2001). Relatively radiogenic values documented here suggest minor fluid interaction with interbedded calcareous shales of the Madera Formation (>0.7100; Mukhopadhyay and Brookins, 1976; Williams et al., 2013) and/or underlying crystalline basement rocks (>0.7481; Taggart and Brookins, 1975). Observed trends in cement geochemistry likely record mixing between end-member meteoric and deep basinal fluids. We propose that the specific contributions of end-member fluids to the overall geochemical signal recorded in a given fault cement are ultimately tectonically controlled. Specifically, variations in grain size of the syntectonic sediments sheared in fault damage zones and fault displacement magnitude were the primary factors determining whether deep basinal fluids could be transported to the near surface (Fig. 1A).
Our data demonstrate that fault zone flow pathways varied systematically with structural position in the Rio Grande rift. In master and basin-margin faults, proximity to uplifted sediment sources resulted in hanging-wall damage zones of sheared, high-permeability alluvial fan and/or fluvial sediments; large displacements produced footwall damage zones with a significant extent of fractured pre-rift units. This fault zone architecture allowed transport of fluids from relatively deep stratigraphic levels to cement precipitation sites in high-permeability hanging walls. In contrast, intrabasin fault zones served as barriers to flow through LSF low-permeability lacustrine and playa deposits. However, they were conduits for meteoric fluids within high-permeability damage zones in the USF, where they likely shuttled water either upward or downward between adjacent aquifers (cf. Haneberg, 1995).
The temporal distribution of different fault permeability structures was also tectonically mediated during basin development. Master and basin-margin faults initiated with rifting; intrabasin faults formed later in rift history. We conclude that the extension history of the Rio Grande rift resulted in a predictable spatial and temporal distribution of fault zone flow pathways, which transmitted fluids from different stratigraphic levels depending on slip magnitude and basin position. All of the studied fault types extend from crystalline basement into USF basin-fill sediments (Russell and Snelson, 1994); therefore we conclude that downdip fault extent was not a primary control on fluid source depths. Our results provide a fundamental first step toward predicting regional flow patterns in extensional settings by improving our understanding of tectonic controls on the distribution of syntectonic fault zone flow pathways in sedimentary basins.
S. Minor, D. Newell, and an anonymous reviewer provided constructive reviews. We thank the Santa Ana and Zia Pueblos for access to field sites. Work was supported by grants from the GDL Foundation, the Geological Society of America, the American Association of Petroleum Geologists, and the Wisconsin Alumni Research Foundation. We thank B. Link for field assistance, and S. Minor for outcrop locations.