Reconstructing the erosional and tectonic record of Laramide contraction to Rio Grande rift extension, southern Indio Mountains, western Texas, USA

The geology of southwestern United States has been dramatically shaped by the Laramide orogeny and subsequent Cenozoic extension. During the Late Cretaceous and Paleogene, contraction behind the Cordilleran arc extended eastward into the interior of the North American craton (e.g., Miller et al., 1992; Erslev, 1993; Saleeby, 2003; Liu et al., 2010; Yonkee and Weil, 2015; Thacker et al., 2023). From Nevada and New Mexico through Canada this deformation led to well-defined orogenic phases wherein deformation stepped farther into the craton through time (Armstrong, 1968). The Laramide orogeny is typically viewed as a foreland event that generated basement uplifts in what is now the U.S. Rocky Mountain region (e.g., Hamilton, 1988). Although it overlapped temporally with thin-skinned Sevier deformation to the west, it was distinct because it typically evolved through development of basement-cored faults, arches, and basins (e.g., Yonkee and Weil, 2015). Intracratonic Laramide deforma tion was driven by flat-slab sub-duction in association with subduction of oceanic plateau(s) in Late Cretaceous to Paleogene time (e.g., Saleeby, 2003; Dickinson, 2009; Humphreys, 2009; Liu et al., 2010). This Rocky Mountain–centric view of the process is, however, an oversimplification of the event because Laramide age contraction is a nearly Cordilleran-wide event that had different manifestations and postcontractional sedimentary, erosional, and volcanic histories (Cather, 2004; Copeland et al., 2017; Busby and Centeno-García, 2022; Thacker et al., 2023). For example, Laramide deformation merges to the south with the Mexican fold and thrust belt, which lies at the northern margin of deformation associated with the Mexican orogeny (Fitz-Díaz et al., 2018). The Mexican orogenic event is similar to Laramide deformation in that it formed through west-to-east contraction from the late Cretaceous through Eocene (e.g., Fitz-Díaz et al., 2018, and references therein). Contraction involved thin-skinned structures to the south that merge with thin- and thick-skinned deformation in the Mexican fold and thrust belt.

In addition to complexities resulting from Laramide deformation overlapping in space and time with Sevier and Mexican orogenic events, the details of how Laramide deformation transitioned (both in space and time) to ignimbrite volcanism and Neogene extension across the western United States remain incompletely understood (e.g., Chapin et al., 2004). Fragmentary sedimentary, tectonic, and thermal records of this tectonic shift result in a partial understanding that is compounded in many areas because subsequent extension either dissected Laramide features or buried them beneath deep rift basins (e.g., Seager, 2004). This is especially true for the southern Laramide and/or the northern Mexican orogen of western Texas, which lies outside of the classic Laramide terrane of northern New Mexico, Colorado, and Wyoming, and where extensional forces in the southern Rio Grande rift and adjacent Basin and Range have dramatically obscured older events. Here we refer to Cretaceous through Eocene contractional deformation of western Texas as “Laramide” due to its close proximity to Laramide structures of southern New Mexico while acknowledging that deformation in this region includes thin-skinned thrusting and that this region lies within a blurry transition with the Mexican fold and thrust belt to the south. We focus on the transition from Laramide shortening to Neogene extension as preserved in a paleovalley in the Indio Mountains of western Texas. We present new mapping results, field observations, structural and sedimentological analysis, and geo/thermochronologic data focused on constraining the post-Laramide geologic and tectonic framework. Sedimentary and volcanic rocks record incision of a Laramideage mountain range, creating a paleocanyon filled with sediments, mass-wasting deposits, and ash-flow tuffs. Geochronologic, thermochronologic, and structural data document the timing and kinematic evolution related to development of the Rio Grande rift. The Indio Mountains preserve a unique record of sedimentation and deformation in the southern Laramide orogen and Rio Grande rift, allowing for an understanding of deformation at the intersection of the Mexican fold and thrust belt, the southern Laramide orogen, and the Rio Grande rift.

The N-NW–trending Indio Mountains are located 50 km southwest of Van Horn, Texas (Fig. 1). The range is underlain by ~2100 m of Cretaceous strata (Underwood, 1962; Haenggi, 2002) that record a generally transgressive sequence originally deposited in the Mesozoic Chihuahua trough. The Chihuahua trough, one part of the larger Mexican Border rift system (Dickinson and Lawton, 2001; Stern and Dickinson, 2010; Lawton et al., 2020) is a NW-trending extensional structure that formed between the Middle Jurassic and the early Late Cretaceous (Haenggi, 2002; Haenggi and Muehlberger, 2005). The northern limit of the Mexican Border rift system trends eastwest through southern Arizona and southern New Mexico and bends toward the southeast near El Paso, Texas (Fig. 1). Proposed mechanisms for the initial opening of the trough include backarc extension associated with slab rollback of the Mezcalera plate (Lawton and McMillan, 1999) and dextral transtension along an intermittently reactivated Proterozoic structural fabric in the southwestern portion of the North American craton (Haenngi, 2002; Haenggi and Muehlberger, 2005). Both mechanisms, however, agree that the opening of the Chihuahua trough was related to the opening of the Gulf of Mexico (Haenggi, 2002). West-dipping, basin-bounding normal faults along the northeastern margin of the Chihuahua trough initially developed during the Jurassic (Uphoff, 1978) and are nearly coincident with the trace of the Rio Grande River, extending from Big Bend National Park to El Paso, Texas (Carciumaru and Ortega, 2008). These basin-bounding normal faults were active during Jurassic–Cenomanian sedimentation (Haenggi, 2002).

The Chihuahua trough underwent inversion and thrusting with the onset of contraction due to the arrival of the Laramide deformational front between 84 and 78 Ma (Denison et al., 1970; Lehman, 1991). Most researchers agree with a tectonic model wherein Laramide deformation was a consequence of flat-slab subduction of the Farallon plate beneath North America between 75 and 35 Ma (e.g., Coney and Reynolds, 1977; Bird, 1998; Saleeby, 2003), although collisional tectonic models have also been proposed (Tikoff and Maxson, 2001; Tikoff et al., 2022). In southern New Mexico and Arizona, western Texas, and the northern parts of Chihuahua, Sonora, and Coahuila states of Mexico, the typical Laramide contractional deformation overlapped in time and space with a thin-skinned thrust belt (part of the larger Mexican fold and thrust belt) that continues southward into eastern Mexico (Fitz-Díaz et al., 2018).

The Mexican fold and thrust belt as well as the segment that follows the United States–Mexico border inverted Mesozoic rift basins and generated east to northeast vergent, thin-skinned thrust systems that are not only distinct in structural style from the classic Rocky Mountain Laramide but also may have occurred from distinct mechanisms (Fitz-Díaz et al., 2018). Thin-skinned thrusting began in the Cenomanian and continued into the Paleocene (Seager and Mack, 1986). Deformation waned by 55 Ma, but folded margins of dikes and sills near El Paso, Texas, which are as young as 47 million years old (Hoffer, 1969), indicate that deformation in the fold and thrust belt continued to at least this time.

In southern New Mexico and in the Big Bend region of western Texas, typical Laramide oval basins formed from the latest Cretaceous and continuing to the late Eocene (Fig. 1) (Seager and Mack, 1986; Seager et al., 1997). Synorogenic deposition adjacent to major basement-involved thrust faults recorded Laramide deformation (Fig. 1) (Clinkscales and Lawton, 2015). One key element of the Eocene deformation along the United States–Mexico border was the association with abundant andesitic intrusions and volcanism that began by ~55 Ma and extended past the end of Laramide deformation (Seager and Mack, 1986; De los Santos et al., 2018) (Fig. 1).

During contraction within the southern Laramide orogen and Mexican fold and thrust belt, Cretaceous rocks of the Indio Mountains were thrust northeastward out of the Chihuahua trough and against the Diablo Plateau (Gries and Haenggi, 1970; Haenggi, 2002; Carciumaru and Ortega, 2008). Stratigraphic studies in the core of the thrust belt suggest this thrust system developed along an evaporite décollement (e.g., Haenggi, 2002), but in the foreland segment in west Texas, the thrusts are detached within the Cretaceous section forming imbricate stacks along the paleorift margin.

The end of Laramide contraction has been attributed to various factors, including buckling of the Farallon plate (Humphreys, 1995), steepening of the subduction angle (McMillan et al., 2000), loss of EW-directed tractional forces acting on the base of western North America (Bird, 1998), and loss of the gravitational potential energy stored in the overthickened crust as a result of Sevier-Laramide deformation (Coney and Harms, 1984). In western Texas and southern New Mexico, the close of Laramide contraction has been broadly correlated to the onset of post-Laramide volcanism at ca. 48 Ma, but paleostress analyses using dike orientations indicate contractional stresses may have continued until 32 Ma (Henry et al., 1991), overlapping with later tectonic events.

The emplacement of largescale ignimbrites in the southwestern United States during and following the Laramide orogeny suggests a major rise in the asthenosphere-lithosphere contact as the Farallon plate was removed from the base of the North America lithosphere (McMillan et al., 2000). The change from calc-alkaline, arc-like magma to more bimodal volcanism has been used to infer the collapse or foundering of the Farallon plate and replacement by asthenosphere, which provided the necessary heat source for production of lithospheric-derived melts (Humphreys, 1995; McMillan et al., 2000; Humphreys et al., 2003; Parker et al., 2012). Ignimbrite volcanism began in western Texas and southern New Mexico during the Eocene and propagated to the west toward northern Arizona and southern Nevada (e.g., Coney and Reynolds, 1977; Humphreys, 1995). Simultaneously, a separate trend of ignimbrite volcanism propagated southeast from Washington, beginning at ca. 55 Ma, until the two magmatic fronts nearly converged.

In the Trans-Pecos volcanic field of western Texas, magmatism occurred in three phases (Henry et al., 1986). An early phase from 47 to 39 Ma was composed of mostly mafic and minor silicic magma. The main phase of magmatism, from 38 to 28 Ma, was largely silicic in composition. The late phase ended at 17 Ma and was dominated by basaltic eruptions that temporally overlapped with evolution of the southern Rio Grande rift. Silicic magmatism began ca. 42 Ma in Big Bend, peaked at 38–32 Ma across much of western Texas, and progressed from the northeast to southwest through time (Parker and Henderson, 2021). Thus, in western Texas, Cenozoic volcanism appears to have overlapped with both Laramide contraction and rift extension.

Rio Grande rift extension overprinted all earlier episodes of tectonic deformation, sedimentation, and magmatism. The Rio Grande rift is a NS–trending alignment of extensional basins and uplifts that extends from northern Colorado to southern New Mexico. Southward, the rift trends NW-SE in western Texas and northern Chihuahua (Fig. 1). Initial rifting began in the Oligocene (Mack et al., 1994), and thermochronologic data collected from the central and northern segments of the Rio Grande rift suggest a main phase of extension from ca. 25–10 Ma (Kelley et al., 1992; Kelley and Chapin, 1997; House et al., 2003; Landman and Flowers, 2013; Ricketts et al., 2016; Abbey and Niemi, 2018; Biddle et al., 2018; Gavel et al., 2021; Ricketts et al., 2021a). In Colorado and northern New Mexico, the Rio Grande rift lies between the Colorado Plateau to the west and the Great Plains to the east. In southern New Mexico, it borders the highly extended Basin and Range Province along a well-defined vertical boundary along the southern edge of the Colorado Plateau (Ricketts et al., 2021a). While earlier studies contend that Rio Grande rift extension in southern New Mexico was broadly coincident with ignimbrite magmatism and initiated at ca. 36 Ma (Cather, 1990; Mack et al., 1994; McMillan et al., 2000), Gavel et al. (2021) relied on low-temperature thermochronologic constraints to suggest that extension was likely delayed until 27–25 Ma. In western Texas, Henry et al. (1991) bracketed the onset of extension between 31 and 28 Ma based on the orientations of dikes and fault kinematic data coincident with the change from calc-alkaline to bimodal volcanism.

The transition from Laramide contraction to Rio Grande rift extension may have varied temporally from north to south in the Trans-Pecos region. In the northern Trans-Pecos, the Indio Mountains contain a suite of unfolded Eocene ignimbrites dated at 38 Ma (Davidson, 2014). Normal faulting, and thus a regional change in stress, occurred sometime after the deposition of the youngest volcanic unit, dated at 36.57 Ma (Davidson, 2014). The absence of folds or thrust faults within this volcanic assemblage was interpreted by Davidson (2014) as evidence that contraction had ceased by 38 Ma; however, that interpretation is equivocal because of the potential for piggyback thrust basin development and contraction structures that bypassed the assemblage. For example, to the southeast of the Indio Mountains in Big Bend National Park, volcanic units of the Chisos Formation dating at ca. 32 Ma exhibit contractional structures that may be related to Laramide deformation (Davidson, 2014).

The Indio Mountains expose a succession of Mesozoic rocks that are unconformably overlain by Oligocene and younger volcanic, volcaniclastic, and clastic deposits. The Mesozoic stratigraphy in the Indio Mountains consists of Early Aptian to Late Cenomanian to Turonian marine and non-marine rocks (Underwood, 1962). Thicknesses of Mesozoic units vary significantly between the axis of the Chihuahua trough to the southwest and the Diablo Plateau to the northeast, which served as the basin margin during deposition (Underwood, 1980; Budhathoki, 2013). Unit thicknesses also vary significantly within the study area due to juxtaposition along thrust and normal fault systems (Plate 1).

General descriptions of rocks exposed in the Indio Mountains are provided in Underwood (1962), and complete stratigraphic descriptions are available for the upper member of the Yucca Formation (Kyu) (Page, 2011; Li, 2014), the Bluff Mesa Formation (Kbm) (Anderson, 2017), and the Cox Formation (Kc) (Budhathoki, 2013). Figure 2 summarizes the Cretaceous units and their general lithologies. The oldest exposed rocks belong to the Yucca Formation, which can be subdivided into upper (Kyu) and lower (Kyl) members based on an abrupt fining-upward sequence from conglomerate to sandstone, siltstone, shale, and limestone (Underwood, 1962; Page, 2011; Ramirez, 2018). Well-rounded clasts and cross beds on the order of 1 m in Kyl suggest deposition in a fluvial environment (Underwood, 1962), whereas Kyu consists of interstratified upward shallowing and coarsening packages of lacustrine and fluvial facies (Li, 2014). We divide the overlying Bluff Mesa Formation into four informal members (Kbm1, Kbm2, Kbm3, and Kbm4). Kbm1 is predominantly a carbonate unit that grades upward into quartz arenite (Kbm2), gray shale (Kbm3), and is capped by dark bluegray micrite (Kbm4) (Li, 2014). The Bluff Mesa Formation is overlain by cross-bedded quartz arenite with subordinate red and green shale of the Cox Formation (Underwood, 1962; Budhathoki, 2013). The overlying units consist of micritic limestone (Finlay Formation; Kf), shale capped by sandstone (Benevides Formation; Kbe), carbonate with minor shale and marl (Espy Formation; Ke), sandstone (Eagle Mountains Formation; Kem), and mictitic, nodular limestone (Buda Formation; Kb).

Oligocene ignimbrites belonging to the Garren Group have been mapped in the Wylie Mountains, Van Horn Mountains, Eagle Mountains, and Indio Mountains, extending north and east from the study area (Hay-Roe, 1957; Underwood, 1962). Chemical analyses of the Garren Group indicate they consist predominantly of calc-alkaline and minor alkaline mafic ignimbrite deposits (Davidson, 2014). They are compositionally correlative to deposits from the Van Horn Mountains caldera (Henry and Price, 1984), although targeted geochronologic and geochemical analyses are needed for robust correlation of the various units between mountain ranges (e.g., Barnes et al., 1979). In the Indio Mountains, the base of the Garren Group is an angular unconformity with Cretaceous rocks or pre-Garren Group sedimentary units described here. We describe the regional Garren Group units here with more detailed descriptions of local units in the Paleovalley Stratigraphy section.

The Hogeye tuff is the basal unit of the Garren Group (Teal, 1979) and has a reported thickness of 46–64 m (Teal and Hoffer, 1980). The unit is densely welded at its base but transitions to moderately to poorly welded up-section (Teal and Hoffer, 1980). It contains phenocrysts of K-feldspar (5%), oligoclase (3%–5%), hornblende (2%), lithic fragments (10%), and pumice in a devitrified microcrystalline matrix (Teal and Hoffer, 1980). The top of the Hogeye tuff contains a trachyte unit (Teal and Hoffer, 1980) that Underwood (1962) described as pale red, pink, gray, and light- to moderate-brown in color.

The Pantera trachyte (Tp) conformably overlies the Hogeye tuff and has a reported thickness of 27–91 m (Teal and Hoffer, 1980). The unit is a resistant, pale- to grayishred, densely welded trachytic crystal tuff with microeutaxitic texture overlying a basal vitrophyre (Underwood, 1962; Teal, 1979). Phenocrysts consist of andesine (15%–20%), K-feldspar (10%), augite (3%–5%), and small amounts of quartz, biotite, and hornblende in a devitrified matrix (Teal and Hoffer, 1980).

Barnes et al. (1979) mapped Flat Top Hill (Plate 1) as the Undivided Part of the Garren Group. From the base to top, the units consist of vitric crystal tuff, aphanitic trachyte that becomes more vesicular up-section, pale-colored tuff, aphanitic basalt, fine-grained tuff, porphyritic basalt, tuff and tuffaceous sandstone and conglomerate, and ledge-forming porphyritic andesite. The uppermost andesite at the top of Flat Top Hill was interpreted to be the Bell Valley andesite (Davidson, 2014). U-Pb dating on zircons from the unit yielded an age of 36.57 ± 0.80 Ma (Davidson, 2014).

Geologic mapping was completed using field computers running the Geographic Information System (GIS) program QGIS with a custom interface and data structure similar to that described in Pavlis et al. (2010). Mapping accuracy was aided by realtime Global Navigation Satellite System (GNSS) and external Garmin Bluetooth units using GPS and Glonas constellations. Mapping was conducted on layered base maps within the GIS, including a scanned topographic map and various generations of orthophotography. The result is a map with point locations accurate to the GNSS resolution of ~2 m and linework accuracy commensurate with the orthoimagery resolution (<1 m) subject to the interpretation level of the geologic mapping. The geologic map (Plate 1) is a composite work compiled from several generations of mapping by the authors as well as the work of Page (2011) and Ramirez (2018).

Kinematic analysis focused on collecting structural data to determine principal shortening (P) and extension (T) axes (Marrett and Allmendinger, 1990) associated with post-Laramide deformation. Kinematic analysis allows for the reconstruction of the principal strain axes that correspond to directions of maximum (S1) and minimum (S3) stretch and requires that fault-plane orientations, slip directions, and relative sense-of-slip are known (Marrett and Allmendinger, 1990). Strike- and- dip data from fault planes and associated slickenline rake measurements were coupled with shear-sense information from available localities. Shear-sense information was determined by orientations of chatter marks; R, R’, P, and T fractures; and asymmetric foliation structures that commonly develop in brittley deformed rocks (Petit, 1987).

Fault data were analyzed using FaultKin 7.5 (Marrett and Allmendinger, 1990; Allmendinger et al., 2012). Faultkin uses a graphical approach to solve for the infinitesimal P (maximum shortening) and T (maximum extension) axes. Fault planes and slickenlines were plotted onto lower-hemisphere projections, with arrows to indicate the hanging-wall movement direction. A movement plane was then calculated for each fault, which contains the slip vector, the pole to the fault, and the P-T axes that plot 45° from the pole to the fault plane. The location of the P-T axes relative to the pole of the fault were determined using the shear-sense information. After creating a scatter plot of all P-T axes, Bingham statistics linked all P-T axes to solve for the kinematic axes. A fault-plane solution, the P-T dihedral, was then calculated to graphically portray the best-fit conjugate P-T fields using the calculated kinematic axes. Gray and white areas of the P-T dihedral represent extension and contraction, respectively. Fault planes the P-T axes, which do not plot within the appropriate P-T field, were deemed incompatible and were removed from the data set for further inspection.

Sedimentological analysis used clast counting and paleoflow measurements to help reconstruct the evolution of Paleogene–Neogene sedimentary deposits. At each clast count location, a minimum of 100 clasts were identified. A minimum clast size of 3 cm was applied to avoid misidentifying smaller clasts. Paleocurrent analysis involved measuring the long (A) and intermediate (B) planes of sedimentary clasts following the strategy of Cavazza (1986). At least 20 orientations from imbricated clasts were collected to two accessible outcrops of Paleogene conglomerate. Measurements were only collected from imbricated clasts with an A-axis >5 cm.

Apatite (AHe) and zircon (ZHe) (U-Th)/He techniques are sensitive to temperature windows of 30–90 °C (Flowers et al., 2009) and 50–240 °C (Guenthner et al., 2013), respectively. At higher temperatures, helium that is produced through radioactive decay can easily diffuse out of the crystal lattice, and within these temperature windows, helium is partially retained. Retention of helium within an individual apatite or zircon grain depends on the effective uranium concentration (eU), calculated as eU = [U] + 0.235 [Th] (Flowers et al., 2009). Flowers et al. (2009) and Guenthner et al. (2013) describe important relationships between eU and helium retention in the crystal lattice, which are commonly observed as positive, negative, or flat date-eU trends.

AHe and ZHe techniques were applied at a single location to constrain the most recent exhumational history related to development of the Rio Grande rift. Two separate samples were collected 0.5 m apart from the lowest exposed Lower Yucca Formation. This sample site is in the immediate footwall of the Red Mountain normal fault in an attempt to constrain movement along this structure. The two samples were collected from a sandstone layer and adjacent conglomerate layer to account for possible differences in apatite and zircon source. Apatite and zircon were separated using standard mineral separation techniques, including rock crushing, sieving, and heavy liquids. Euhedral and unbroken grains were selected under a polarizing microscope and packed into Nb tubes for analysis. A total of ten AHe and ten ZHe dates were collected from this location to look for possible spread in dates related to eU concentration. Data were collected at the Thermochronology Research and Instrumentation Laboratory at the University of Colorado Boulder (CU TRaIL). Modeling of thermo chronologic data was done using HeFTy v. 1.9.3 (Ketcham, 2005).

Detrital sanidine ⁴⁰Ar/³⁹Ar geochronologic methods were used to constrain ages of various Paleogene–Neogene volcanic and sedimentary deposits. Sanidine is a common mineral associated with explosive volcanism, and many sanidine-erupting volcanic centers existed throughout the southwestern United States during the Cenozoic. Following sample preparation methods of Cather et al. (2019), samples were prepared by disaggregation, sieving, and cleaning with a weak acid solution. Magnetic and heavy liquid separations were used to concentrate K-feldspar grains. Grains were immersed in wintergreen oil before inspection to identify plutonic textures and selection of grains beneath a polarizing microscope.

For this study, two volcanic rock samples were collected from ⁴⁰Ar/³⁹Ar geochronology (Plate 1). One sample was collected from a trachyte member of the Hogeye tuff near the base of the Oligocene section in order to assign a minimum age of Laramide deformation to the southern Indio Mountains. A second sample of basalt that is interbedded with conglomerate farther up-section was also collected for geochronology. A third sample was collected from the Neogene half-graben for detrital sanidine ⁴⁰Ar/³⁹Ar geochronology to constrain a maximum depositional age for this deposit. Sanidine ⁴⁰Ar/³⁹Ar geochronology was conducted at the New Mexico Geochronology Research Laboratory.

Rocks exposed in the Indio Mountains range in age from Aptian–Albian to Quaternary and are important records of continental-scale transition from Laramide deformation to Rio Grande rift development. Laramide contraction, post-Laramide erosion and volcanism, and Rio Grande rift extension have imprinted several important mapscale features in the southern Indio Mountains (Fig. 3; Plate 1); these features are briefly summarized here.

Laramide folds and thrust faults in the southern Indio Mountains exclusively deform Mesozoic rocks (Fig. 3) (Rohrbaugh, 2001; Page, 2011; Sahin, 2015; Ramirez, 2018). Major thrust faults include the Squaw Peak thrust in the eastern portion of the mapped area and the Purple Sage thrust to the west (Fig. 3; Plate 1). The Purple Sage thrust dips shallowly to the west and places Lower Yucca Forma tion structurally on Bluff Mesa, Cox, and Upper Yucca formations. The Squaw Peak thrust is subhorizontal to east dipping and places Lower and Upper Yucca Formation on Espy, Eagle Mountain Sandstone, and Buda formations in the mapped area (Plate 1) but is more complex to the north where a footwall duplex produced a window in the Squaw Peak thrust sheet (Page, 2011). Previous work suggested that the Squaw Peak and Purple Sage thrusts were the same structure that has since been dismembered by younger faults (Ramirez, 2018).

Post-Laramide erosion led to the development of a highly irregular, angular unconformity throughout the southern Indio Mountains (Fig. 3). This structure separates folded and faulted Mesozoic rocks from gently deformed Paleogene volcanic, volcaniclastic, and sedimentary rocks. Stratigraphic thicknesses of units vary considerably across the angular unconformity, with some of the older units pinching out along strike against the unconformity (Plate 1). We interpret this stratigraphic relationship as manifestations of a paleovalley that was filled prior to deposition of the bulk of the volcanic assemblage. The northwest and northeast portions of the map area preserve the thickest stratigraphic thickness of post-Laramide volcanic and sedimentary deposition, referred to here as the western and eastern paleovalleys (Fig. 3).

Extensional structures that formed during development of the Oligocene–Recent Rio Grande rift trend predominantly NW-SE through the entire mapped area (Fig. 3). The most prominent extensional structure is the Indio fault, a NW-striking, SW-dipping normal fault that largely juxtaposes Eocene volcanic rocks and Neogene basin fill against Cretaceous rocks (Fig. 3). This fault is paralleled by an anticlinal fold in the immediate footwall (Plate 1) that could have formed either during Laramide contraction as a duplex anticline (Page, 2011) or during development of the Indio fault as a footwall drag fold. Other notable extensional structures include the NW-striking Red Mountain and Borrega normal faults in the western portion of the map areas that bound Red Mountain (Guerrero, 2018; Ramirez 2018). These structures truncate the older Purple Sage thrust fault. Several EW-striking, near-vertical strike-slip faults in the northern part of the mapped area cut rocks as young as Eocene volcanic and clastic rocks. NW-striking normal faults are also similar to a younger structure, the West Indo Mountains fault outside the study area; the fault separates the range from the Red Light Draw basin to the southwest and is responsible for an unknown amount of northwestward tilt of the range (Collins and Raney, 1994).

A second angular unconformity separates the youngest volcanic unit from an overlying moderately to wellsorted Neogene conglomerate (Tc) in the central and southern parts of the mapped area (Fig. 3). This conglomerate fills a halfgraben adjacent to the Indio fault and was deposited on more steeply dipping volcanic flows of the Garren Group.

Detailed stratigraphic descriptions of volcanic, volcaniclastic, and clastic deposits in the southern Indio Mountains are generally lacking or incomplete in previous studies. In particular, previous studies tended to focus on the most prominent volcanic units that make up the Garren Group, whereas the accompanying sedimentary deposits are either only briefly mentioned or mapped as Quaternary deposits (e.g., Underwood, 1962). Here we provide more detailed unit descriptions and relate these observations to previous unit subdivisions. A combination of detailed geologic mapping and volcanic sample analysis reveals a complex volcanic and sedimentological stratigraphy that filled a paleovalley cut into Cretaceous units. Complete unit descriptions are provided in Text File S1¹.

Since some of the paleovalley units are local and interstratified with more widespread units, correlation of units across the Indio Mountains was aided with thin-section analysis and point counting (Figs. 4 and 5). This analysis focuses on the Hogeye tuff (Thtr) and Pantera trachyte (Tp), two prominent welded tuffs that are located in both paleovalleys.

Samples of Thtr from both paleovalleys contain crystal constituents that are predominantly K-feldspar with minor quartz and magnetite (Fig. 4). Phenocryst content is less than 5%, and most crystals are <1 mm in length. Pumice lapilli are present in small quantities and are partially to fully devitrified. Rock fragments are rare, but when observed, are quartz sandstone and mudstone clasts ~3–5 mm in diameter.

Additional samples of presumable Tp were collected from both paleovalleys for comparison. K-feldspar is a dominant mineral, along with minor quartz, hornblende, biotite, and magnetite set within a brownishred matrix (Fig. 4). Flattened cuspate and bubblewall shards can be seen flowing around crystal grains, giving the sample a microeutaxitic texture (Fig. 4).

Point counting of units Thtr and Tp also reveals that samples from the two units can be distinguished from each other, allowing correlation between the western and eastern paleovalleys (Fig. 5). Both units are similar in that they are essentially devoid of rock fragments and are composed mostly of pumice and glass (Fig. 5). Point-counting results classify both units as vitric tuff. Correlation of these two units between the two paleovalleys is supported by the observation that samples of each unit from both paleovalleys plot in similar locations on the triangular diagram. However, Thtr consistently contains crystal percentages ranging between 5%–15%, whereas Tp crystal percentages range from 19%–31% (Fig. 5). These results allow correlation of these key marker units.

The post-Laramide angular unconformity is a highly irregular erosional surface. Cenozoic units above the unconformity are lensoid in map view, pinching out to the north and south against the unconformity. This map relationship suggests a paleovalley with deposits thickest within the core of the paleovalley and thinning laterally against the erosional surface (Fig. 3; Plate 1), an interpretation further supported by the rock units themselves. One of the most prominent characteristics of these units is that many of them are confined to the western paleovalley. All of the Garren Group units that were originally described by Underwood (1962) crop out within the paleovalley, but they are interstratified with additional previously unrecognized units (Fig. 6). Three megabreccia units overlie the angular unconformity and are interstratified with the Hogeye tuff (MB1, MB2, and MB3 in Figs. 3 and 6A). These units are characterized by angular blocks of Kyl that are up to 5 m in diameter and lack any coherent bedding. MB3 also contains minor wellrounded limestone clasts of Kbm1. Thicknesses of these deposits vary dramatically.

Within the center of the paleovalley, the brown, cliff-forming, lower welded Hogeye trachyte member (Thtr) overlies megabreccia units and is separated from Hogeye upper tuff (Thtu) by megabreccias. Thtu is interstratified with several units and includes airfall tuffs that drape paleotopography and resedimented tuffs that formed layers on the paleovalley floor. Along the southern margin of the paleovalley, Thtu overlies the unconformity.

The white Hogeye ash (Thtu) apparently filled the western paleovalley. This was later incised, and conglomerates and fluvial sandstones, along with the basal, indurated member of the Pantera trachyte filled the newly incised valley (Fig. 6A). Clastic deposits include a moderately indurated, wellrounded conglomerate (Pcgl). This unit contains clasts of sandstone, conglomerate, and limestone derived from several units in the Cretaceous section. A prominent basalt flow (Tb) is interbedded with Pcgl (Plate 1), and in outcrop, the basalt is highly weathered and contains abundant calcitefilled fractures and voids. Both of these units thin and pinch out within the paleovalley (Figs. 3 and 6A; Plate 1). Poorly to moderately consolidated beds of white and gray laminated ash overlie the conglomerate (Plate 1). This unit, referred to as the Tertiary pond deposit (Tpd), thins dramatically to the north and south. It is mostly fine grained but also contains fluvial conglomerate lenses and interbedded layers of ash. The spatial extent of this unit indicates that it represents an area within the paleovalley that temporarily remained a topographic low that responded to damming of the paleovalley by Tb to create a local depositional environment.

These local units are overlain by a whitishgray rhyolitic ashfall deposit (Tra) in the western paleovalley. In contrast to all underlying units within the paleovalley, Tra is an extensive unit that was deposited along the entire NS length of the mapped area, and to the south, this unit was deposited across the post-Laramide angular unconformity (Fig. 6A; Plate 1). This observation indicates that the paleovalley was essentially filled by the time Tra was deposited. The stratigraphically highest units in the Garren Group are the rhyolitic ash flow tuff (Trt) and the Bell Valley andesite (Tbv), which, in contrast to underlying units, are relatively uniform in thickness and extend across the entire map area (Plate 1).

In the eastern part of the mapped area in the footwall of the Indio fault, dramatic thickening of units and pinching out of units to the north and south also indicate deposition within a paleovalley. The megabreccias seen at the base of the Paleogene section to the west are conspicuously absent in the eastern paleovalley. Instead, the stratigraphy is simpler, and the paleovalley is largely filled with the Hogeye tuff. The oldest volcanic unit unconformably overlying Cretaceous rocks is a gray, poorly indurated ash, referred to here as the Lower Hogeye tuff (Thtl) (Figs. 3 and 6B). This unit is not present in the western paleovalley but pinches out along strike against the unconformity consistent with its deposition in a paleovalley.

In the eastern paleovalley, Thtl is overlain by the trachyte unit of the Hogeye tuff (Thtr), the Upper Hogeye tuff (Thtu), and discontinuous lenses of the Pantera trachyte (Tp) with no interstratified units (Fig. 6B). The rhyolite ashfall deposit (Tra) overlies Tp and is widespread and more regionally extensive than many of the stratigraphically underlying units, similar to its outcrop appearance in the western paleovalley. This unit is interstratified with several volcanic and volcaniclastic units that are local and not present in the western paleovalley. A darkbrown lithic breccia interbedded with Tra matches the descriptions of Twiss (1959) and Teal (1979), who describe a similar unit in the Van Horn and Wylie Mountains to the east. The unit is clast-supported and contains angular to sub-rounded volcanic clasts ranging from 0.1 to 0.3 m in size. The unit is referred to here as the Tertiary Garren Mountain breccia (Tgm).

Additional volcanic units younger than the Garren Mountain breccia are also present in the eastern paleovalley and match descriptions of the Garren Group (Undivided) (Hay-Roe, 1957, 1958; Twiss, 1959). The Garren Group (Undivided) of Hay-Roe (1957) and Twiss (1959) consists of an alternating sequence of poorly indurated tuffs and basalts, rhyolite ignimbrite, and an andesite lava flow correlative to the Bell Valley andesite (Hay-Roe, 1957; Twiss, 1959; Teal, 1979; Teal and Hoffer, 1980). A similar alternating sequence of darkgray to black basalt flows (Tb; Plate 1) interbedded with poorly bedded white ash was observed overlying the Garren Mountain breccia in the eastern paleovalley (Fig. 6B).

The top of the eastern paleovalley stratigraphy is similar to units observed in the western paleovalley. Unit Tra grades up section into the overlying Tertiary rhyolite tuff (Trt), which is in turn capped by the Bell Valley andesite (Tbva; Plate 1). These two uppermost units are regionally extensive and maintain relatively consistent thicknesses, indicating that when they were deposited, the western and eastern paleovalleys were essentially filled.

In the central and southern parts of the mapped area, the Garren Group volcanic rocks are overlain by a conglomerate (Tc) along an angular unconformity (Plate 1). This unit overlies the Bell Valley andesite along its western edge and is faulted against Mesozoic units by the NW-trending Indio fault along its eastern border. Beds within the conglomerate generally dip east-northeast toward the Indio fault. Clasts within this unit are derived entirely from the underlying Mesozoic and Garren Group units, where the percentage of Garren Group clasts relative to Mesozoic clasts increases, as documented more carefully in the Sedimentological Analysis section below. These observations suggest that Tc represents deposition within a nascent internally drained half-graben that formed during southern Rio Grande rift development.

Paleogene conglomerates (unit Pcgl) restricted to the western paleovalley record fluvial deposition between Laramide contraction and Rio Grande rift extension. This unit provides an opportunity to investigate paleoflow direction and sediment sources related to infilling of the paleovalley.

Clast counts and paleoflow measurements were conducted at two locations within unit Pcgl (Fig. 7; Plate 1). Due to the difficulty in distinguishing clasts from individual Cretaceous units, we instead generalize and use basic lithologies for clast-counting purposes. Both clast count sites yield similar results. They both contain minor percentages of volcanic clasts (5.3%–5.6%) relative to Cretaceous clasts (94.4%–94.7%), and Cretaceous clasts are dominated by sandstone over limestone and conglomerate (Fig. 7A).

Paleoflow was estimated by measuring the A-B plane of oblate imbricated clasts that are at least 5 cm in length (Cavazza, 1986). At each location, multiple clasts were measured to account for any variability, and the orientation of bedding was also measured in order to rotate back to horizontal. Paleoflow location 1, located stratigraphically below Tb, shows a range of paleoflow orientations indicating transport to the east (Fig. 7B; Plate 1). At paleoflow location 2, located stratigraphically above Tb, paleoflow orientations are instead predominantly to the south.

Clast counts were conducted at nine locations in the Neogene half-graben adjacent to the Indio fault (Fig. 8). Approximately 100 clasts were identified and counted at each location, along with facies characterization of each site. Clast counts were restricted to diameters greater than 3 mm; smaller clasts were not included to decrease the likelihood of improper clast identification. The Neogene half-graben sediments were deposited on all the older Cretaceous lithologies and Paleogene volcanic units, and clast lithologies identified in outcrop include Cretaceous conglomerate (Kyu or Kyl), Cretaceous sandstone, basalt, ignimbrite, and ash (pumice). Clast counts were also conducted at the base, middle, and top of the Neogene half-graben section.

Clastcount results generally show a relative decrease in volcanic and volcaniclastic clasts up section—from 100% at the base of the unit (location CC4) to a minimum of 46% near the top adjacent to the Indio fault (location CC11) (Fig. 8A). The decrease in volcanic clasts occurs along with an increase in Cretaceous conglomerate and sandstone clasts, from 0% at location CC4 to 54% at location CC11. Figure 8B shows the relative upsection percentages of volcanic, Cretaceous sandstone, and Cretaceous conglomerate clasts. An increase in Cretaceous sandstone and minor increases (10%) in Cretaceous conglomerate relative to volcanic clasts indicate an unroofing trend, where the Neogene basin was progressively filled as the footwall uplift was dissected and eroded. In contrast, clast-count locations CC8 and CC9 do not show the same trend. Instead, these locations consist entirely of clasts of ignimbrite and ash, even though they are near the top of the unit (Fig. 8). This likely reflects a local sediment supply late in the basin filling that drained a yet-to-be-eroded exposure of Paleogene volcanic rock.

Fiftyone fault plane and slickenline orientation measurements were collected for fault kinematic analyses, focused on the post-Laramide tectonic framework. Measurements were collected from the southern portion of the Indio fault, the Red Mountain fault, two NW-SE to E-W–trending normal faults offsetting Kyl in the north central portion of the mapping area, and from minor structures offsetting Kyu and Kyl in the central portion of the mapping area (Fig. 9). The kinematic analyses also incorporate ten fault measurements from Rodriguez Gonzalez (2019) collected from the Indio fault. Figure 10 shows the range of fault strike and slickenline rake values corresponding to the structure they were collected from. Rake data from slickenlines on the Indio fault indicate predominantly normal to normal sinistral slip sense and minor occurrences of dextral normal, normal, and sinistral normal slip. NW-trending normal faults straddling Red Mountain are normal and normal dextral slip in the western portion of the study area (Fig. 10). Seven rake measurements were collected from extensional structures with less than 1 m of slip in the central portion of the study area, and these are dextral normal to sinistral normal slip. Faults in the northern mapping area exhibit a range of rake values but predominantly group within the dextral normal, normal dextral, and normal sinistral-slip fields (Fig. 10).

Kinematic analysis of fault data was conducted in a stepwise approach to investigate strain compatibility among various structures and identify separate fault populations. Figure 11A shows the initial calculated kinematic axes using all (n = 61) structural data collected in the southern Indio Mountains. P-T axes that plot outside of the appropriate area in the bestfit conjugate P-T fields (P-axes plotting in the T-field or T-axes plotting in the P-field) were deemed incompatible and removed. The remaining compatible P-T axes are shown as Population 1 (n = 51) (Fig. 11B). The calculated kinematic T-axis for population 1 has an orientation of 13, 237, suggesting SW-NE extension. From the population of incompatible faults (n = 10) (Fig. 11C), a total of nine faults are compatible to form Population 2 (Fig. 11D), which suggests NW-SE extension based on a T-axis with an orientation of 18, 336. The remaining single incompatible fault is not considered further. Kinematic data were also analyzed by which structure they were collected from to compare extensional directions among different faults (Fig. 12). Notably, all four groups of faults predict similar NE-SW–directed extension. This also includes the group of EW to WNW-ESE–oriented, near-vertical, strike-slip faults that are at a high angle to the main NW-SE fault population.

Samples for apatite (AHe) and zircon (ZHe) (U-Th)/ He thermochronology were collected from the lowest exposed section of Lower Yucca (Kyl) exposed in the footwall of the Red Mountain normal fault (Plate 1). A total of ten AHe and ten ZHe dates were collected from two samples. Five AHe and ZHe dates were collected from a sandstone interval and five AHe and ZHe dates were collected from a chert pebble conglomerate 0.5 m away. Although these samples have experienced an identical thermal history since deposition of Kyl, separate samples were collected to capture a wider range of eU values from apatite and zircon grains from possibly different sources.

Five AHe dates collected from the chert pebble conglomerate range from 40.9 ± 1.1–75.9 ± 2.3 Ma (Fig. 13A; Table 1). Apatite crystals range from 15.5 to 104.3 ppm eU and show a positive correlation with AHe date. Five samples collected from a sandstone interval range from 13.9 ± 1.2 Ma to 52.8 ± 1.8 Ma (Fig. 13A). Apatite crystals range from 29.6 ppm to 85.5 ppm eU and also show a positive correlation with age. Collectively, these samples yield AHe data that show a positive correlation with eU, suggesting prolonged residence time within the AHe partial retention zone. Individual apatite grains range from 33 µm to 49 μm in size (Fig. 13B) and do not show a relationship with AHe date. This observation suggests that crystal size does not exert a firstorder control on AHe date, and instead, the observed range in AHe dates is due to radiation damage (Flowers et al., 2009).

Six ZHe dates collected from the conglomerate range from 79.8 ± 2.7–134.3 ± 3.9 Ma and show a moderate spread in eU from 177 ppm to 452 ppm (Fig. 13A; Table 2). A total of four zircon grains were collected from the sandstone layer, and ZHe dates range from 76.2 ± 2.9–110.0 ± 4.0 Ma across eU values that range from 216 ppm to 367 ppm (Fig. 13A). These ZHe dates are both younger and older than the Aptian to early Albian depositional age of the Yucca Formation (Underwood, 1962), indicating that burial of the Yucca Formation did not result in full resetting of ZHe dates. Similar to the AHe dates, crystal size does not correlate with ZHe date, indicating that the resulting ZHe dates more likely result from radiation damage (Fig. 13B) (Guenthner et al., 2013).

AHe data were modeled in HeFTy v. 1.9.3 software (Ketcham, 2005) to constrain the recent cooling history of the Yucca Formation in the footwall of the Red Mountain fault. ZHe data were excluded from modeling because these grains were not completely reset after deposition of the Yucca Formation, and each grain therefore records a unique thermal history. In contrast, the observation that all AHe dates are younger than the depositional age across a range of eU values indicates that this sample was buried sufficiently to reset AHe dates. For modeling, the conglomerate and sandstone samples were combined into a single sample because they experienced an identical thermal postdepositional thermal history.

Inverse modeling uses AHe dates, U and Th concentrations, and grain size as inputs to constrain a thermal history. Since HeFTy only allows a total of seven input grains, a synthetic grain approach is used, similar to previous studies (Flowers et al., 2020; Reade et al., 2020; Ricketts et al., 2021b; Thurston et al., 2021). To create synthetic grains, eU values are binned, and in each bin, the AHe data are averaged. eU bins used are 0–40 ppm, 40–60 ppm, 60–80 ppm, 80–100 ppm, and >100 ppm to create a total of five synthetic grains used for modeling (Fig. 13A). These synthetic grains follow the same positive trend as individual AHe data in date-eU space. The inverse model begins at 1400–1100 Ma, representative of the age of Precambrian basement in western Texas, and at temperatures of 280–300 °C, higher than the AHe sensitivity range.

The final model is completely unconstrained prior to 125 Ma, consistent with the observation that AHe dates were reset after deposition of the Yucca Formation. However, good paths indicate the sample was subsequently buried to maximum temperatures of 80–90 °C (Fig. 13C). The model is not very sensitive to the timing of maximum reheating and is permissible of cooling coincident with Laramide deformation (Fig. 13D). The model is also not very sensitive to the initiation of cooling in the critical period of 40–30 Ma during the transition between Laramide contraction and Rio Grande rift extension. However, the model does show that rapid cooling was ongoing by 27 Ma and that this pulse brought the sample close to nearsurface temperatures (Fig. 13D). Due to the sample’s proximity to the Red Mountain fault, this pulse in cooling that had initiated by 27 Ma is attributed to extensional exhumation in the Indio Mountains related to development of the Rio Grande rift.

Detrital sanidine ⁴⁰Ar/³⁹Ar geochronologic methods were applied to one sample collected from the base of the Neogene conglomerate (Tc) in the half-graben. Two additional samples, the Trachyte member of the Hogeye tuff (Thtr) and the oldest basalt flow (Tb), were collected from the western paleovalley for ⁴⁰Ar/³⁹Ar geochronology to refine the times of deposition (see Plate 1 for sample locations).

A total of 234 sanidine grains were analyzed from the half-graben conglomerate (Fig. 14A). The grains are dominated by a peak at ca. 38 Ma, with several grains that are slightly older (maximum age 121 Ma) and several smaller peaks between 38 Ma and 30 Ma. A single grain with an ⁴⁰Ar/³⁹Ar age of 11.857 ± 0.011 Ma is interpreted to be the maximum depositional age of this unit.

Twenty-one sanidine grains analyzed from the trachyte member of the Hogeye tuff (Thtr) indicate an emplacement age of 38.166 ± 0.016 Ma and a mean square of weighted deviates (MSWD) = 4.49 (Fig. 14B). This age is within error of, yet slightly older than, the 38.02 ± 0.99 Ma age Davidson (2014) reported for the overlying tuff member of the Hogeye tuff (Thtu). Stepheating of the Tertiary basalt sample (unit Tb1), which stratigraphically overlies Thtu, shows a plateau at 37.54 ± 0.05 Ma and a MSWD = 2.14 (Fig. 14C).

The southern Indio Mountains preserve a unique window into the regionally important transition from Laramide contraction to Rio Grande rift extension. This area records local sedimentation and volcanism influenced by end-Laramide topography that was affected by more regionalscale patterns in faulting during early development of the Rio Grande rift. As is the case for many parts of the western United States, the timing and nature of the transition from Laramide contraction to subsequent extension in western Texas are poorly constrained, and different studies provide conflicting viewpoints. For example, Henry et al. (1991) relied on the orientations of dikes to suggest that Laramide contraction persisted across western Texas until 32 Ma. However, ages and geochemistry of volcanic rocks in southern New Mexico suggest that extension began at 36 Ma (Cather, 1990; Mack et al., 1994; McMillan et al., 2000). Here we discuss the regional importance of the Indio Mountains local geology within the context of this ambiguity and outline the most salient geologic observations in Figure 15.

The Cretaceous rocks of the Indio Mountains experienced tectonic inversion and general E-NE transport out of the adjacent Chihuahua trough and onto the Diablo plateau during the Laramide orogeny (Fig. 15A) (Hennings, 1994; Rohrbaugh, 2001; Carciumaru and Ortega, 2008; Page, 2011). Contractional structures preserved in the Indio Mountains are diverse, including thrust faults, duplex systems producing folded thrusts, strike-slip faults, and a variety of folds from open and upright to tight, overturned, and nearly recumbent (Page, 2011; Guerrero, 2018; Ramirez, 2018). These structures are manifestations of Laramide deformation into Trans-Pecos Texas and Mexico (Hennings, 1994). In terms of total deformation, Reaser (1982) estimated thrust displacement of ~48 km along the margin of the Chihuahua trough, and Page (2011) estimated 17–33 km of total fault displacement in the central Indio Mountains.

The age of Laramide deformation can be difficult to constrain in many locations, but it appears to have varied spatially and temporally in both the southwestern United States and Trans-Pecos region of western Texas (Drewes, 1978). Triassic rocks in Sierra Salamayuca (Lawton et al., 2018) yield mica K-Ar dates that indicate deformation was ongoing at ca. 80 Ma (Denison et al., 1970). Igneous rocks (83 Ma) from the Pemex Camello-1 well in the northern portion of the Chihuahua trough may be related to early Laramide igneous activity (López-Ramos, 1988). In the Big Bend region to the southeast, Lehman (1991) suggests an early phase of Laramide deformation at ca. 70 Ma recorded in the Tornillo foreland basin followed by a second phase that ended by 50 Ma. North of Chihuahua City, Mexico, K-Ar dates from the folded Cuervo Formation, which is unconformably overlain by the Nopal Formation, constrain Laramide deformation to 54–45 Ma in that area (Reyes-Cortés, 1997).

Within the Mexican fold and thrust belt, including the Indio Mountains, contractional deformation progressed from west to east. Earliest deformation along the western margin initiated during Aptian–Albian time, and the latest deformation along the eastern margin (where the Indio Mountains are located) was active from 64 Ma to 42 Ma (Fitz-Diaz and van der Pluijm, 2013; Fitz-Díaz et al., 2018). More recently, Davidson (2014) conducted a geochronologic study of undeformed volcanic units that overlie older rocks that have been variably deformed during presumptive Laramide deformation. Davidson (2014) suggested a regional trend where the cessation of Laramide deformation was oldest in the northwest in and around the Indio Mountains at ca. 38 Ma but persisted in the Big Bend region until 32–31 Ma.

Thermal history modeling of AHe data in this study does not require any major pulse of Laramide exhumation in the southern Indio Mountains, although it is permissive of ~25–30 °C of cooling from 80 Ma to 40 Ma (Fig. 13D). In the western paleovalley, Eocene volcanic rocks directly overlie faulted and folded Cretaceous strata above an angular unconformity (Plate 1). The oldest volcanic unit that overlies this unconformity at this location (Thtr) yielded a sanidine ⁴⁰Ar/³⁹Ar age of 38.166 ± 0.016 Ma. There are two models to explain these observations. In the first model, Laramide shortening in Trans-Pecos Texas had ceased by 38 Ma, in general agreement with the age of Laramide cessation in southern New Mexico (e.g., Seager, 1983) and the end of Laramide deformation in the Mexican fold and thrust belt (Fitz-Díaz and van der Pluijm, 2013; Fitz-Díaz et al., 2018) (Fig. 15B). Alternatively, rocks in the Indio Mountains could have been translated northeastward along an underlying detachment rooted in evaporite horizons until 32–31 Ma (Davidson, 2014). In this scenario, the Purple Sage thrust fault would have become inactive by 38 Ma, and 38 Ma allochthonous rocks in the Indio Mountains would not show signs of internal deformation. The presence of megabreccias along the angular unconformity potentially supports this model but is inconclusive because they could represent landslides developed on remnant topography, following cessation of thrusting. This model would predict the emergence of a thrust front located somewhere farther to the east (Fig. 15B). Data presented here do not distinguish between these two models. Although most research suggests that Laramidestyle contraction had long ceased by 32 Ma and the region was at the height of ignimbrite volcanism, there are several lines of reasoning that Laramide contraction may have persisted until 32 Ma in Trans-Pecos Texas.

First, Davidson’s (2014) interpretation is based on 31.73 ± 0.56 Ma folded and faulted rocks of the Chisos Group in Big Bend National Park that are overlain by horizontal rocks of the Burro Mesa Forma tion with a reported U-Pb age of 31.2 ± 0.73 Ma. Davidson (2014) notes that folds and faults in the Chisos Group rocks trend northwest, consistent with a Laramide origin. However, a detailed structural study of the Chisos Group is necessary to link this deformation more confidently to the Laramide event.

Second, plate reconstructions of the subducted conjugate Shatsky and Hess oceanic plateaus suggest they may have driven flat-slab Laramide deformation (Liu et al., 2010). Predicted positions of these two plateaus indicate that the northern Shatsky conjugate moved northeast in an arcuate path beneath North America from 84 Ma to 68 Ma and was a primary driving factor in classic Laramide deformation of Colorado and Wyoming (Liu et al., 2010). In contrast, the southern Hess conjugate is younger and traveled northeast below Mexico from ca. 65–55 Ma and differs from the northern Shatsky conjugate in that it did not overlap in time with deformation in most of Mexico (Fitz-Díaz et al., 2018). Hence, only the eastern margin of the Mexican fold and thrust belt might have been influenced by passage of the Hess conjugate. Despite this difference, the younger timeframe of subduction of the Hess conjugate could possibly indicate that Laramide deformation along the northeastern margin of the Mexican fold and thrust belt is younger than 38 Ma.

Third, several studies have suggested that Laramide contraction in southern New Mexico continued until 31.5–29.3 Ma based on broad NW-SE–trending folds in ignimbrite rocks (Copeland et al., 2011; Tomlinson et al., 2013). However, this interpretation is tenuous because ignimbrite flareup magmatism initiated at ca. 36 Ma in southern New Mexico (e.g., Chapin et al., 2004). In addition, Clinkscales and Lawton (2017) note that thickest ignimbrite sections blanket Laramide structures and syndepositional strata and are thickest in normal fault grabens, indicating they were deposited after the cessation of Laramide deformation. Alternatively, Clinkscales and Lawton (2017) suggest that tilting of these ignimbrites could have occurred during subsequent extension. In summary, additional research is needed to document post–38 Ma Laramide deformation more confidently in Trans-Pecos Texas; therefore, we favor a model where contraction had ceased by 38 Ma in the Indio Mountains.

The influence of paleotopography on the geographic extent of sedimentary and volcaniclastic deposits has been extensively documented in Colorado and Nevada (e.g., Epis and Chapin, 1975; Chapin and Lowell, 1979; Henry, 2008). Henry (2008) documented abrupt thickness changes in volcanic and sedimentary units both parallel and perpendicular to strike. Eocene paleovalley deposits in Nevada are also characterized by localized conglomerates, megabreccias, and lacustrine sediments that are restricted to paleovalley floors as a result of paleotopography (Henry, 2008).

The Paleogene volcanic and sedimentary rocks in the southern Indio Mountains are interpreted as the preserved fill of a paleovalley within a highland of significant topography. Based on the preserved sedimentary and volcanic rocks filling both paleovalleys, available paleoflow orientations, and cross-section reconstruction, these paleovalleys are interpreted to have been a single valley that has since been dissected by the Indio fault (Plates 1 and 2). Evidence that sedimentary and volcanic rocks were deposited within a paleovalley includes (1) deposits that are thickest in the middle of the inferred paleovalley and thin to the north and south, indicating that they both have a similar E-W orientation; (2) the presence of megabreccias and conglomerates in the western paleovalley; (3) thinning and abrupt pinching out of volcanic units, such as the Pantera trachyte (Tp) and the trachyte member of the Hogeye tuff (Thtr) against the underlying unconformity; and (4) thickening of lowparticle concentration ashflow tuffs (Tpd and Tra) in the western paleovalley. Since the oldest volcanic unit in the paleovalley is dated at 38.166 Ma, the topography responsible for these patterns in deposition was most likely remnant from Laramideage mountain building along the transition from the Chihuahua trough to the Diablo plateau (Fig. 15B).

Megabreccias (MB1, MB2, and MB3) were previously undocumented in the southern Indio Mountains and are confined to the western paleovalley (Plate 1). MB1 consists of angular clasts sourced from the Lower Yucca Formation that are up to 5 m in diameter. The diverse size and range of angular clasts with the lack of any observable bedding suggest that MB1 was eroded and deposited as debris flows or landslides sourced from an adjacent mountain with significant topographic relief. The Hogeye tuff (38.17 Ma; this study) directly overlies MB1 in the western paleovalley. MB1 was thus emplaced prior to the onset of extension and precludes any model where MB1 was eroded because of slip on the nearby Red Mountain normal fault or that it is a Quaternary debris flow (Underwood, 1962). MB2 is more extensive than MB1 and directly overlies the 38.17 Ma Hogeye tuff. This unit, which also contains angular pebbles and boulders sourced from the Lower Yucca Formation, indicates that debrisflow deposition was ongoing after initial volcanism had begun. Similarly, MB3 is interbedded with unit Thtu and indicates that the filling of the paleovalley was a gradual transition from very local debrisflow sources to more distal volcanic sources. The changes in depositional environment within the paleovalley suggest a large Laramide mountain range existed to the southwest, from which the megabreccias were sourced.

Paleogene conglomerate (Pcgl) and Paleogene Lower Yucca conglomerate (Pylcgl) are only present in the western paleovalley. Paleoflow orientations collected from the stratigraphically lowest location reveal W-to E-directed paleoflow, supporting a model where the western paleovalley is more proximal to the source, in agreement with the location of megabrecca units. Paleoflow orientations show a change to more southdirected flow after the emplacement of the 37.61 ± 0.07 Ma basalt (Tb), suggesting that this basalt flow impacted the drainage system enough to redirect paleoflow direction. The Tertiary Pond Deposit (Tpd) overlies the uppermost conglomerate beds (Pcgl), and abrupt lithology changes parallel and perpendicular to strike may indicate transient damming by Tb to create a localized depositional environment not observed outside of the western paleovalley. Tpd contains interbedded intervals of poorly indurated ash, conglomerate, and darkgray silt beds, which indicate localized interactions between fluvial and lacustrine depositional environments intermittently covered by ashflow and ashfall deposits. Lava damming and redirecting of fluvial systems is common in paleovalley settings (Wobus and Epis, 1978) and has been documented for larger river systems, such as the Colorado River in Grand Canyon (e.g., Hamblin, 1994; Fenton et al., 2002; Crow et al., 2015).

The upper tuff member of the Hogeye tuff (Thtu) and the Upper Tuff (Tra) increase from a cumulative thickness of 230 m in the eastern paleovalley to ~320 m in the western paleovalley, not including the additional 74 m of Tpd stratigraphically between Thtu and Tra not present in the eastern paleovalley. Increase of thickness in these low-density units supports the hypothesis of increased accommodation space in the western paleovalley. Filling of the paleovalley is recorded in widespread deposition of the stratigraphically higher volcanic units, including Tra, Trt, and Tbva, which blanketed the landscape (Fig. 15C).

Although volcanic units throughout the southern Indio Mountains lack demonstrable Laramide structures, the Pantera trachyte (Tp) and Hogeye Upper tuff (Thtu) preserve folding evidenced by the abrupt change in bedding orientations in the southwest portion of the western paleovalley. However, rather than an indication of continued Laramide shortening, this is likely a consequence of deposition of ash blankets on paleovalley walls and postdepositional compaction of ignimbrites. Chapin and Lowell (1979) documented postcompaction folding of ignimbrites emplaced in paleovalleys in central Colorado. Assuming paleovalley walls to be at a maximum angle of repose of 35°, postdepositional compaction of an ignimbrite unit can achieve dips of up to 22° toward the axis of the paleovalley (Chapin and Lowell, 1979). Bedding dips in Thtu and Tp (20°–30°) are consistent with their emplacement within a paleovalley.

Laramide deformation was likely a timetransgressive event (e.g., Perry and Flores, 1997; DeCelles, 2004; Carrapa et al., 2019; Thacker et al., 2023), and Laramide basins, arches, and uplifts have different geologic histories across the orogen (Galloway et al., 2011). By the Middle Eocene, Laramide contraction had largely ceased, yet residual highlands continued to shed detritus into existing basins (e.g., Cather and Johnson, 1984; Galloway et al., 2011; Lawton, 2019) (Fig. 16). This period of erosion is coincident with the development of an extensive erosional surface, known as the Rocky Mountain erosion surface, which eroded parts of the Laramide uplifts and is estimated to have occurred ca. 50 Ma in central Colorado (Chapin and Kelley, 1997). Evidence for widespread erosion from 42 Ma to 37 Ma in Wyoming, Colorado, and parts of Montana includes deep paleovalleys that incised into the Rocky Mountain erosion surface (Evanoff and Chapin, 1994; Chapin and Kelley, 1997; Cather et al., 2012). Cather et al. (2012) suggest that major Eocene erosion was restricted to the Colorado and Wyoming region and did not extend any farther south than northern New Mexico (Fig. 16). In western Texas, the Tornillo basin in the Big Bend region represents one of the southernmost Laramide basins of the North American Cordillera (Fig. 16). Pre-, syn-, and postorogenic sediments in the Tornillo basin record two major periods of high sedimentation rates from ca. 69–66 Ma and 61–51 Ma that likely reflect pulses in Laramide deformation (Lehman, 1991). Absence of alluvial fans suggested to Lehman (1991) that no significant topography existed along the western or northern margins of the Tornillo basin during its formation. Galloway et al. (2011) indicate that the Chihuahua tectonic belt existed as a highland until the early Eocene, at which time it was beveled by erosion and was absent by the middle Eocene (Fig. 16).

Incision of the paleovalley as documented in this study occurred prior to deposition of the 38.17 Ma Hogeye tuff (Thtr), which overlaps with the regional ca. 50–37 Ma incision of paleodrainages into the Rocky Mountain erosion surface in Colorado and Wyoming (Cather et al., 2012). In the southern Indio Mountains, the western and eastern paleovalleys have estimated depths of 200–375 m and 150–225 m, respectively. This is less than the maximum paleodrainage depths from Wyoming but is similar to paleodrainage sizes in central Colorado. The age of the oldest paleovalley fill in the Indio Mountains (38.17 Ma) is also very similar to the 37 Ma age of fill in central Colorado. The age and size of paleodrainages in the Indio Mountains suggest that Late Eocene erosion may have been more widespread than previously thought and extended into Chihuahua and western Texas, although here the extent is uncertain as Neogene extension has dissected the Laramide uplands. Megabreccia units preserved in the western paleovalley described in this study suggest very short transport distances and considerable topographic relief at 38 Ma within the Chihuahua tectonic belt northwest of the Tornillo basin. These observations argue for continued topographic relief within the Chihuahua tectonic belt through the Middle Eocene (Fig. 16). Although sedimentation within the Tornillo basin slowed after ca. 50 Ma, postorogenic sediments of the Canoe Formation and younger units contain abundant extrabasinal clasts (Lehman, 1991) that may have been derived from the relict Chihuahua tectonic belt along a highland that extended through the Indio Mountains.

Rio Grande rift deformation in the southern Indio Mountains occurred primarily along three subparallel NW-striking structures, including the Indio fault, Red Mountain fault, and Borrega fault (Plates 1 and 2). The Indio fault is the largest and extends for ~30 km along the Indio Mountains (Underwood, 1962). The topography of the range and erosion of Tertiary volcanics indicates that fault displacement decreases toward each end. The study area lies closer to the southern end, and slip is inferred to decrease between cross sections A–A’ and B–B’. Restorations along the fault confirm this. The locations of the angular unconformity in the footwall and hanging wall of the Indio fault were used as piercing points to restore slip and calculate the net slip for cross sections A–A’ and B–B’ (Plate 2). Restoring total slip along normal faults in A–A’ indicates a minimum of 1270 m of normalsense slip across the Indio fault and a minimum of 235 m slip along the Red Mountain fault. Restoration of B–B’, which lies closer to the southern termination of the Indio Ffault, indicates a minimum of 760 m of slip along the Indio fault and 135 m of slip along the Red Mountain fault. Estimated slip along the Red Mountain fault is slightly less than previous palinspastic restorations by Ramirez (2018), which show ~500 m of minimum net slip. Along the Indio fault, vertical displacements of 1685 m (Rohrbaugh, 2001), 1169 m (Sahin, 2015), and 1000–2000 m (Underwood, 1962) have been estimated, and Page (2011) indicates a minimum of 1550 m net slip. In total, restoration of extensional structures suggests ~10%–11% total extension in cross section A–A’ and 7%–8% extension along B–B’ over a horizontal distance of 7–8 km.

These estimates of extension are similar to the 8%–12% estimated from the northern part of the rift (Chapin and Cather, 1994; Kluth and Schaftenaar, 1994) and far less than the 17%–25% extension from central New Mexico and >50% extension for southern New Mexico (Russell and Snelson, 1994). However, total extension from the southern Indio Mountains is only estimated from a short spatial scale, and this does not include the largest and youngest extensional fault, the southwest-dipping West Indo Mountains fault that separates the range from the Red Light Draw basin (Collins and Raney, 1994). The Neogene gravels along the West Indo Mountains fault show similar rotation into the fault, as seen in the study area, and this fault is likely the cause of the overall eastward rotation of the units in the study area that remains after restoration along the Indio fault (Plate 2).

Thermal history modeling of AHe thermochronologic data indicates rapid cooling had initiated by 27 Ma (Fig. 13). The samples were collected from the lowest exposed portion of the footwall of Red Mountain fault, and the time of cooling postdates emplacement of ignimbrites in the southern Indio Mountains. Therefore, the rapid cooling is taken to represent fault-related exhumation in the footwall of the Red Mountain fault. Published apatite fission-track (AFT) and zircon fission-track (ZFT) dates (Kelley et al., 1992; Kelley and Chapin, 1997) and thermal history modeling of AHe, AFT, and ZHe data (House et al., 2003; Landman and Flowers, 2013; Ricketts et al., 2015, 2016; Abbey and Niemi, 2018, 2020; Gavel et al., 2021) from ranges in the southern, central, and northern portions of the Rio Grande rift in Colorado and New Mexico suggest synchronous extension from 25 Ma to 10 Ma. However, the onset of extension is less clear along the length of the rift, possibly due to the temporal and spatial overlap of ignimbrites and rift flank cooling between ca. 40–25 Ma (Ricketts et al., 2016). The pulse in cooling in the southern Indio Mountains (Fig. 13D) that is recorded by thermal history modeling falls within a generally accepted range of 31–28 Ma for onset of extension in the southern Rio Grande rift (Henry and Price, 1986; Morgan et al., 1986; Henry et al., 1991). More extensive thermochronologic investigations are needed to test and refine these estimates.

Early onset of extension in the southern Rio Grande rift is evidenced by an inferred half-graben associated with a linear belt of 35 Ma rhyolite domes near Las Cruces, New Mexico (Seager and Clemons, 1975; Mack, 2004). This is at odds with recent modeling of thermochronologic data, which indicate that extension was delayed in the southern rift until ca. 27–25 Ma (Gavel et al., 2021). The pulse in cooling in the Indio Mountains coincides with the 27–25 Ma time period of extension observed in thermal history models from southern New Mexico (Gavel et al., 2021). It agrees with the first phase of late Oligocene–early Miocene extension in southern New Mexico (Mack et al., 1994) and is only slightly older than the ca. 24 Ma age of onset of extension proposed by Henry et al. (1991) in the Tran-Pecos region of west Texas.

The youngest detrital ⁴⁰Ar/³⁹Ar sanidine date of 11.87 ± 0.016 Ma from the Neogene half-graben conglomerate overlaps with graben and half-graben formation in other portions of the rift (e.g., Morgan et al., 1986; Dickerson and Muehlberger, 1994; Mack et al., 1994; Langford et al., 1999). Langford et al. (1999) dated basal rift fill sediments to 12.4 Ma in the Eagle Flat basin to the northwest of the Indio Mountains. A younger but similar age of ca. 10 Ma was assigned to the NW-trending Tornillo graben in Big Bend National Park based on the presence of late Miocene fossils in basin-fill sediments (Stevens and Stevens, 1990; Dickerson and Muehlberger, 1994). Although the detrital sanidine ⁴⁰Ar/³⁹Ar date provides a maximum depositional age for the half-graben sediments, it is unclear if the Indio fault accommodated slip prior to 11 Ma, as indicated by thermochronologic data. It is possible that extensional exhumation initiated on the Red Mountain fault and was transferred to the Indio fault at ca. 11 Ma. Alternatively, extension along the Indio fault could have initiated prior to 11 Ma, and basin fill only began to accumulate after a certain threshold of extension had been achieved. Due to the close spatial association between the two faults, and the observation that the Red Mountain fault is not associated with any basin-fill deposits, we propose that slip on the Indio fault initiated prior to 11 Ma (Fig. 15D). This is also supported by the cross-section analysis, which shows that total slip along the Red Mountain fault is less than total slip along the Indio fault.

Previous studies in the Indio Mountains suggested either EW extension from the orientation of calcite veins (Rohrbaugh, 2001) or NE-SW extension from fault kinematic data from the Indio fault (Rodriguez Gonzalez, 2019). Fault kinematic analysis incorporates data from Rodriguez Gonzalez (2019) and documents a dominant NE-SW extension direction ranging from 048–228° to 066–246° along structures with a wide range in orientation (Figs. 11 and 12). In the Santo Domingo basin of northern New Mexico, fault kinematic analysis suggests an overall EW-extension direction, which was rotated clockwise in the later stages of rifting to NW-SE. In the Espanola basin of northern New Mexico, Caine et al. (2017) also document EW extension, although they note that many faults likely reactivated older structures formed during the Laramide orogeny. A separate kinematic study within five Rio Grande rift flank uplifts in southern New Mexico also concluded that many of them record EW extension (Rodriguez Gonzalez, 2019). Our preliminary data set suggests that the southern, NW-trending segment of the rift may record a different extension direction than the central and northern segments of the rift. Future research aimed at collecting fault kinematic data from multiple locations and constraining the timing of fault movement is necessary to more fully document extension in the southern Rio Grande rift.

Eocene rocks preserved in the Indio Mountains, western Texas, offer important insight into the regionally important transition from Laramide contraction to Rio Grande rift extension. Detailed mapping in the southern Indio Mountains highlights a highly irregular angular unconformity that overlies faulted and folded Cretaceous rocks deformed during the Laramide orogeny. Overlying the unconformity is a package of Middle to Late Eocene clastic, volcanic, and volcaniclastic rocks that fill an EW-trending paleovalley cut into Cretaceous rocks. Oldest paleovalleydeposits include chaotic megabreccia units, and sedimentological analysis suggests deposit was from the west. ⁴⁰Ar/³⁹Ar dating constrains filling of the paleovalley from 38.1 Ma to 36.6 Ma and suggests that while Laramide deformation had probably ceased by 38 Ma, relict Laramide topography continued funneling detritus into an EW-trending paleovalley toward the Tornillo basin through the Middle Eocene. These features, including Laramide structures, the angular unconformity, and the paleovalley, were dissected by NW-trending normal faults during development of the Rio Grande rift. Apatite (U-Th)/He thermochronology documents fault-related exhumation had initiated by 27 Ma, and detrital sanidine ⁴⁰Ar/³⁹Ar dating from half-graben basin fill suggests a maximum depositional age of 11.87 Ma. Fault kinematic analysis indicates predominantly NE-SW–directed extension during this timeframe. This rich geologic record serves as a regionally important test site for understanding the complex tectonic, sedimentological, and erosional processes that occurred during transition from the Laramide orogeny to the Rio Grande rift and is a unique southern counterpart to the more thoroughly documented transition to the north.

This research was funded by the U.S. Geological Survey National Cooperative Geologic Mapping Program through EDMAP award G20AC00150 to J.W. Ricketts. A. Conley received support from the West Texas Geological Society. We thank Margo Odlum, Tim Lawton, and Associate Editor Wentao Cao for comments that helped strengthen the manuscript.