The Badwater turtleback, Copper Canyon turtleback, and Mormon Point turtleback are three anomalously smooth, ∼2-km-high basement structures in the Black Mountains of Death Valley, California. Their structural evolution is linked to the Cenozoic tectonic history of the region. To explore their evolution, we apply (U-Th)/He, Ar/Ar, and U-Pb analyses, with multi-domain diffusion modeling to 10 samples from the Badwater turtleback. The cooling history of the Badwater turtleback is used as a proxy for its exhumation history as it uplifted from warmer depths.

We find slow (<2 °C/m.y.) cooling from ca. 32 to 6 Ma, followed by rapid (120–140 °C/m.y.) cooling from ca. 6 to 4.5 Ma, and finally moderate (30–120 °C/m.y.) cooling occurred from ca. 4.5 Ma until the present. When these data are added to previously published cooling paths of the Copper Canyon turtleback and Mormon Point turtleback, a northwest cooling pattern is broadly evident, consistent with a top-to-NW removal of the hanging wall along a detachment fault. We propose a six-phase tectonic history. Post-orogenic collapse and erosion dominated from ca. 32 to 16 Ma. At 16–14 Ma, a detachment fault formed with a breakaway south and east of the Black Mountains, with normal faults in the hanging wall. Moderate extension continued from 14 to 8 Ma causing exhumation of the turtlebacks through the brittle-ductile transition. Dextral transtension at 7–6 Ma produced a pull-apart basin across the Black Mountains with rapid extension. The locus of deformation transferred to the Panamint and Owens Valley fault systems from 4.5 to 3.5 Ma, slowing extension in the Black Mountains until present.


Since the discovery of normal faults in the Basin and Range Province of North America (Gilbert, 1890), the Death Valley region has been used as a natural laboratory to study extensional tectonics. Many important advances in the Basin and Range Cenozoic extensional tectonics were accomplished based on field research in the Death Valley region, as is reviewed in Miller and Pavlis (2005) and Fridrich and Thompson (2011). Death Valley is an ideal location for studying large-magnitude crustal extension due to its hyperarid, barren landscape, and recent nature of its deformation (Sonder and Jones, 1999).

Cenozoic extension in Death Valley has exhumed amphibolite-grade gneissic basement rocks from 10 to 20 km at depth (Curry, 1938, 1954; Wright et al., 1974; Holm and Wernicke, 1990; Holm et al., 1992). These dome-shaped exposures on the steep western edge of the Black Mountains were named “turtlebacks” by Curry (1938; Fig. 1). From north to south, they are the Badwater turtleback (BWT), Copper Canyon turtleback (CCT), and Mormon Point turtleback (MPT; Fig. 2) and they comprise what is commonly known as the Black Mountains Metamorphic Core Complex. Unlike the surrounding mountain ranges, nearly all of the overlying Paleozoic and Mesozoic rocks have been removed from the Black Mountains by faulting (e.g., Stewart et al., 1983; Holm et al., 1992) or by erosion (e.g., Wright et al., 1974; Çemen et al., 1985, 1999), exposing the turtlebacks at the surface.

The main purpose of this study is to provide insight into long-standing uncertainty regarding the timing and rates of deformation in the northern Black Mountains using a new thermochronometric data set for the BWT. The data set was created based on our analysis of samples collected along a transect (Fig. 3) across the BWT, which is the least sampled of the turtlebacks due to difficulty of navigating the terrain. We report 10 (U-Th)/He cooling ages and four sets of muscovite, biotite, and K-feldspar Ar-/Ar cooling ages, including multi-domain diffusion (MDD) models, to create a continuous T-t history for the BWT. In addition, we report U-Pb crystallization ages for three of our samples. We integrate our results with previously published thermochronometric, geochronologic, geobarometric, paleomagnetic, and structural mapping data across Death Valley and present a new structural model of the Cenozoic evolution of the Black Mountains area and associated exhumation of the Death Valley turtlebacks.


Cenozoic extension in the Basin and Range Province is often linked to post-Laramide gravity-driven collapse of the thickened crust (e.g., Wernicke et al., 1987; Jones et al., 1998), asthenospheric flow related to the subducting Farallon plate (e.g., Dickinson and Suczek, 1979), dextral transtension associated with northwest Pacific Plate motion (e.g., Atwater, 1970; Livaccari, 1979), and/or lithospheric delamination of the over-thickened Sierra-Nevada batholith to the west (Ducea and Saleeby, 1996, 1998; Bidgoli et al., 2015). Comparing the timing and spatial patterns of these hypothesized tectonic events to that of local faulting, volcanism, and basin formation is a common method of testing their validity.

Some estimates of Cenozoic extension in Death Valley (Fig. 1) suggest ∼80 km of top-to-NW extension within a triangular region between the Garlock Fault to the south, Death Valley–Furnace Creek (DVFC) fault zone to the north and northeast, and the Death Valley–Black Mountains (DVBM) fault zone to the west (Stewart et al., 1983; Wernicke et al., 1987; Snow and Lux, 1999). Death Valley is in this triangular region of large-magnitude extension with turtlebacks, which only exist in highly extended, recently deformed terranes (Çemen et al., 2005).

The turtlebacks are doubly-plunging antiformal structures composed of calcite-dolomite marble rocks complexly faulted and folded with ca. 1.7 Ga age Mojave Province (e.g., Asmerom et al., 1990; Lima et al., 2018; this study) mylonitic quartz-feldspar gneiss, pelitic schist, and ca. 55 Ma quartz-feldspar pegmatites (Miller and Friedman, 1999). Within the turtlebacks, late Mesozoic to early Cenozoic crustal shortening followed by Cenozoic extension is responsible for a complex geometry of thrust faults crosscut by normal faults (Miller, 2003; Miller and Pavlis, 2005). Tectonite fabrics in the turtlebacks indicate top-to-NW shear (∼N63W at BWT; Miller, 1999a, 1999b), or possibly a more westerly direction followed by lateral clockwise rotation of the footwall (Holm et al., 1993; Serpa and Pavlis, 1996). In the deepest structural levels of each turtleback, mylonitic fabrics suggest early top-to-E and top-to-SE shear, inviting the possibility that the turtlebacks are reactivated early Cenozoic basement thrust faults (Miller, 2003; Miller and Pavlis, 2005). Moreover, older-over-younger fault contacts are observed at all three turtlebacks (Otton, 1976; Miller, 2003), interpreted as Sevier-Laramide Orogeny thrust faults (Miller and Friedman, 1999). Miocene (9.5–6.0 Ma) felsic to mafic dikes exist across the turtlebacks (Holm et al., 1992; Miller and Friedman, 2003).

The BWT contains the northernmost extent of Mojave Province basement rocks in the Black Mountains (Condie, 1992). In normal fault contact to the west, north, and northeast of the BWT (Figs. 2 and 3) are 14 Ma and younger sedimentary and volcanic rocks (Çemen and Wright, 1988; Wright et al., 1999; Miller and Pavlis, 2005) of the Furnace Creek Basin. To the south and southeast, also in normal fault contact, are 11.7–5 Ma (Asmerom et al., 1990; Wright et al., 1991; Miller et al., 2004; this study) plutons named the Willow Spring diorite (WSD), and Smith Mountain granite (SMG; Figs. 2 and 3), which are described in Holm et al. (1994a, 1994b). These plutons make up much of the western-central Black Mountains and mostly surround the CCT and MPT. Al-in-hornblende geobarometry of the WSD in the central Black Mountains indicates a 9.5–12.5 km crystallization depth (Holm et al., 1992). Asmerom et al. (1990) suggest that the WSD is the only intermediate to mafic synrift intrusion exposed at a metamorphic core complex.

Holm et al. (1992) and Holm and Dokka (1993) report mica, feldspar, and hornblende Ar/Ar ages, along with the first appearance of clasts in the hanging wall of sedimentary basins, from the central and northern Black Mountains, indicating cooling from ∼500 to 300 °C at 13–6.8 Ma, and the CCT and MPT, reaching near-surface depths at 4.5–6 Ma. Building off of these results, Bidgoli et al. (2015) report zircon (U-Th)/He ages and thermal modeling in the central and southern Black Mountains indicating cooling through ∼200 °C at 9–3 Ma, and a major inflection of the cooling path at 6 Ma, which they attribute to changes in plate boundary kinematics and lithospheric delamination of the central Sierra Nevada Batholith (Ducea and Saleeby, 1996, 1998). In a different approach, Ferrill et al. (2012) present apatite and zircon fission track (aFT, zFT, respectively) ages in the Black, Funeral, Bare, and Yucca Mountains and Bullfrog Hills, and project them to a transect line parallel to the direction of extension, which suggests a westward migration of cooling of 10–11 mm/yr.

Based on the extensive thermochronometric data set and many other structural, paleomagnetic, sedimentological, and petrological studies, at least two evolutionary models to explain the evolution of the Black Mountains Metamorphic Core Complex exist. In the rolling-hinge model (e.g., Stewart et al., 1983; Hamilton, 1988; Snow and Wernicke, 2000), ∼80 km of horizontal NW displacement of a relatively cohesive hanging wall along a detachment fault unroofed the Black Mountains, followed by progressive isostatic rebound of the footwall, with the turtlebacks as sections of the exposed undulating detachment surface that accommodated this motion (Fig. 4). In the pull-apart model (e.g., Wright and Troxel, 1984; Serpa and Pavlis, 1996; Wright et al., 1999; Miller and Prave, 2002), the Black Mountains were exhumed along many separate, deeply rooted normal faults that formed in response to the transtensional stress between the Sheephead and Furnace Creek fault zones to the south and north, respectively (Fig. 4). In this model, the turtlebacks are exposures of these normal faults, which have been cross-cut by more recent range-front faults (Miller, 1991, 1999a, 1999b). As we point out in our conclusions, we consider both of these models to be conceptual end-members in the evolution of the Death Valley region, and our proposed tectonic evolution includes both of these processes.


Our data set includes zircon and apatite (U-Th)/He ages along with biotite, muscovite, and K-feldspar Ar/Ar ages from some of the same samples, collected along the transect shown in Figure 3. In addition, we report the crystallization ages of the WSD, SMG, and the gneissic rocks to test the Mojave Province correlation of the turtleback rocks based on U-Pb dating.

Thermochronometry is the use of radiometric dating to calculate the elapsed time since a mineral last passed through a specific range of temperatures known as closure temperatures. This age is therefore often referred to as a cooling age. In highly exhumed rocks, the cooling age usually reflects the time since the rock cooled from the higher ambient temperatures of the deeper crust. Thus, the cooling ages most likely reflect the exhumation history of the samples. The exception is those that were reheated by magmatic activity in the cooler shallow crust, a scenario which we discuss later.

For interpretation of the age results, we used closure temperatures of 325–375 °C for Ar in muscovite (mAr; McDougall and Harrison, 1988; Harrison et al., 2009), 275–325 °C for Ar in biotite (bAr; McDougall and Harrison, 1988; Harrison et al., 1985), 150–330 °C for Ar in K-feldspar (kAr; Foland, 1994; Lovera et al., 1991, 1997; McDougall and Harrison, 1999), 160–200 °C for He in zircon (zHe; Reiners, 2005), and 55–80 °C for He in apatite (aHe; Farley, 2002).

Sample Collection and Mineral Separation

We chose our sampling route to be an east-west transect (Fig. 3) with maximum exposure of gneiss, which contains the requisite minerals for thermochronometry, and based on the feasibility of navigating the terrain on foot. The samples consisted of eight gneisses (S-1278, S-1154, S-971, S-795, S-585, S-448, S-193, and S-62; number corresponds to the sample elevation in meters), and two Miocene igneous samples from the Willow Spring diorite (S-WSD; elevation 226 m) and Smith Mountain granite (S-SMG; elevation 1375 m). Samples were collected from outcrops at ∼200 m elevation intervals (Table 1).

We separated zircon, apatite, and K-feldspar from the crushed samples using standard mechanical, magnetic, and heavy liquid (i.e., Bromoform and methylene iodide) techniques targeting clean, euhedral grains with no inclusions or other defects. We handpicked clean, euhedral muscovite and biotite grains from the coarsely crushed sample.

Isotopic Analysis

We conducted U-Pb zircon geochronology on the two igneous samples (S-WSD and S-SMG) and one gneissic sample (S-795) at the University of California at Los Angeles Secondary Ion Mass Spectrometry (SIMS) laboratory, with zircon AS3 as a standard (Paces and Miller, 1993), using an ∼8–12 nA mass-filtered 16O beam focused to spots between ∼20 and 35 μm. We conducted (U-Th)/He zircon and apatite thermochronometry on 10 samples at the Geo- and Thermochronometry Laboratory at the University of Texas at Austin. All analyses were carried out on single grains. All ages were calculated using standard α-ejection corrections using morphometric analyses (Reiners, 2005). We conducted 40Ar/39Ar K-feldspar, muscovite, and/or biotite thermochronometry on five samples (S-1278, S-971, S-448, S-193, and S-62) at the New Mexico Geochronology Research Laboratory. See Data Repository text1 for additional details on sample collection and analytical procedures.


U-Pb Results

Zircon U-Pb analysis from the BWT gneissic sample S-795 has age populations (Table 2) of 1740 ± 20 Ma (207Pb/206Pb), 1462 ± 54 Ma (207Pb/206Pb), and 79.6 ± 5.0 Ma (206Pb/238U), which agree with previously published crystallization or secondary growth age for the Mojave Province (Wasserburg et al., 1959; Dewitt et al., 1984). The ca. 80 Ma age agrees with the Upper Cretaceous phase of crystallization detected through zircon depth-profiling of the gneissic rocks from the MPT (Lima et al., 2018). The ca. 1462 Ma age possibly reflects post-crystallization secondary growth. We report an age of 7.79 ± 1.22 Ma (206Pb/238U) from 10 zircon grains in S-SMG, which is younger than the 10.44 ± 0.22 Ma age by Miller et al. (2004). We also report an age of 10.0 ± 1.0 Ma (206Pb/238U) from two zircon grains in S-WSD, which is younger than the 11.7 ± 0.2 Ma age by Asmerom et al. (1990).

(U-Th)/He Results

Apatite (U-Th)/He (aHe) mean ages from both the gneissic and plutonic samples range from 1.68 to 4.11 Ma, and zircon (U-Th)/He (zHe) mean ages range from 4.52 to 7.16 Ma (Table 3). The ages are generally within error of one another.

Ar/Ar Results

Age spectra for micas are variably complex with biotite yielding slightly to moderately climbing spectra and muscovite with either flat spectra or initially climbing ages that reach an intermediate maximum age before falling and then rising for the higher temperature heating steps (Data Repository figures, see footnote 1). In order to represent a time related to the bulk argon closure temperature, integrated ages are assigned as the preferred ages for the micas. The bAr ages from S-62, S-971, and S-1278 are 7.266 ± 0.015, 7.939 ± 0.013, and 9.581 ± 0.014 Ma, respectively, whereas mAr ages from the same samples are 8.19 ± 0.05, 24.84 ± 0.04, and 32.23 ± 0.03 Ma, respectively.

Integrated KAr ages from S-193, S-448, S-971, and S-1278 are 9.72 ± 0.14, 8.46 ± 0.05, 8.97 ± 0.02, and 27.05 ± 0.17 Ma, respectively. The K-feldspar age spectra generally have initially old ages that step-down to a minimum value before monotonically climbing for much of the spectra (see Data Repository figures). The data quality is variable and samples S-193 and S-1278 appear to have some plagioclase, as shown by low K/Ca and low argon yields. The plagioclase and its likely associated excess argon contribute to the spectra complexity and also likely causes erroneously high apparent ages in parts for the age spectra; therefore, we consider the data from S-448 and S-971 to be most reliable.

K-feldspar Multi-domain Diffusion Results

The argon step-heating results for the four feldspar samples S-193, S-448, S-971, and S-1278 are investigated for thermal history information using the multi-domain diffusion (MDD) model of Lovera et al. (1989, 1997) and Harrison et al. (2005) and a closure temperature of 150–330 °C (Foland, 1994; Lovera et al., 1991, 1997; McDougall and Harrison, 1999). A 46 kcal/mol activation energy is assigned to all the K-feldspar models; however, the measured Arrhenius data for S-971 and S-1278 are compatible with values of 36.8 and 32 kcal/mol, respectively, and are included in the MDD modeling as well. The MDD results (using 46 kcal/mol) for all four samples show slow cooling between at least 10–7 Ma, followed by a significant increase in cooling rate beginning at 7–6 Ma and continuing at a rapid rate until 5–3 Ma, depending on the sample. Modeling parameters and outputs are given in the Data Repository Text and figures. The model fit for S-193 was especially poor with respect to matching measured and modeled age spectra and we emphasize that we did not heavily consider these data for our interpretations.


Cooling History of the Badwater Turtleback

Muscovite, biotite and K-feldspar from the highest-elevation sample S-1278 (Figs. 5 and 6) in the BWT yield discordant ages. It appears that the muscovite high closure temperature domains (∼350–400 °C; Harrison et al., 2009) began retaining argon between ca. 30 and 40 Ma that was followed by slow cooling (<5 °C/Ma) such that lower closure temperature domains in the muscovite and the biotite closed to argon loss around 9 Ma. Based on the most-direct T-t path between ages, cooling was then more rapid at ∼30–40 °C/m.y. from 9 Ma through our lowest aHe closure temperature (70 ± 20 °C; Farley, 2000) at ca. 3 Ma. Sample S-971 has a similar cooling path, except with a slightly higher maximum cooling rate at ∼45–60 °C/m.y., and all respective cooling ages are slightly younger, indicating this sample was structurally lower throughout exhumation (Fig. 6). The cooling path for S-193 (with a bAr age from S-62) has a similar cooling path as the structurally higher samples with a mAr age of ca. 8 Ma rather than ca. 32 Ma. These combined samples have a higher cooling rate at ∼120 °C/m.y.; however, this rate is biased in the positive direction based on the 131 m elevation difference between the S-193 and S-62. Based on a comparison of cooling paths, S-62 and S-193 were likely below the mAr closure depth (∼11–13 km using 30 °C geothermal gradient) until rapid exhumation began at ca. 8 Ma, whereas the higher two samples were above the mAr closure depth ∼20–30 m.y. earlier.

The cause of the age gradients in S-971 and S-1278 muscovite samples is likely related to their complex thermal history where they have resided near their bulk closure temperature for an extended time period. We note that the integrated age of ca. 24 Ma for S-971 is in close agreement with the ca. 24 Ma old mAr aged sample from Holm and Dokka (1993) at the same elevation.

MDD modeling results (Figs. DR5 and DR6, see footnote 1) show the continuous T-t cooling paths of S-971 and S-448 from 100 to 350 °C and from 150 to 350 °C, respectively. For S-971, the model indicates nearly isothermal conditions at ∼325 °C between 9 and ca. 6.5 Ma with the onset of rapid cooling (130–140 °C/m.y.) beginning at 6.5 Ma and persisting until at least ca. 5 Ma and perhaps younger. S-448 shows a very similar result with slow cooling at ∼1.5 °C/m.y. between ca. 10 and 6 Ma followed by rapid cooling at ∼125 °C/m.y. until ca. 4–5 Ma. Both MDD results suggest that the onset of rapid cooling of the BWT at 6–6.5 Ma was 2–4 m.y. later than the “most likely cooling paths” for the BWT of Holm and Dokka (1993).

The close proximity of the WSD and SMG to our samples requires consideration of an episodic loss event rather than simple protracted cooling to explain the age distributions. A hypothetical reheating event at ca. 6 Ma would produce the same model fits as the exhumation pattern shown in the unconstrained cooling only MDD model (see output from “reheat” model in Figure DR8, see footnote 1). However, the occurrence of some gneissic bAr, KAr, and mAr cooling ages much older than the plutons suggest that not all of the gneisses were significantly reheated. Furthermore, the large age gaps between and within the Ar and He ages (Fig. 5) suggest that any episodic heating occurred when the gneisses were at depth (8–10 km or more), as reheating at shallower depths would have reset all the thermochronometers to the same age. We note that S-1278 is laterally close to the SMG in the modern but has bAr and mAr ages much older than the crystallization age of the SMG.

Comparing Cooling Histories between the Turtlebacks

Cooling envelopes for the southern BWT, CCT, and MPT are shown in Fig. 7. The data used to create the BWT envelope are given in Figure DR9 (footnote 1), and the data used to create the envelopes for the CCT and MPT are from Holm and Dokka (1993). We also plotted the U-Pb ages of samples of the WSD and SMG from Holm et al. (1992) and this study (Fig. 7).

The >300 °C cooling history is given by the hAr, mAr, and bAr cooling ages and indicates that the BWT was slowly cooling from ca. 32 Ma until at least 10 Ma, while the CCT and MPT began relatively rapid cooling at 14 and 10 Ma, respectively. This implies that either the BWT originated at shallower depths than the CCT and MPT, exhumed earlier than the CCT and MPT, or that the CCT and MPT (but not the BWT) were reheated to >500 °C by the intrusion of the WSD and SMG, resetting all Ar/Ar ages, as suggested by Miller and Pavlis (2005). We think the latter interpretation is most likely, based on proximity of the plutons to the CCT and MPT, the reconstruction of the plutons atop of the CCT and MPT (Serpa and Pavlis, 1996), and the presence K-feldspar and sillimanite-bearing schists in the CCT and MPT (Holm and Dokka, 1993; Miller and Pavlis, 2005). There is no field or geochemical evidence indicating that these conditions, with the exception of proximity, exist at the BWT. Furthermore, Miller and Pavlis (2005) note that the mylonitic fabrics are recrystallized at the CCT and MPT, but not at the BWT.

Apatite fission track (Dokka et al., 1986; aFT) and zircon fission track (Zeitler et al., 1982; zFT) ages along with the presence of gneissic clasts in the conglomeratic units in the adjacent Furnace Creek basin (Çemen and Wright, 1988; Wright et al., 1999) indicate that the MPT began rapidly cooling from the brittle-ductile transition zone (∼300 °C) between 6 and 10 Ma and reached the surface by ca. 5 Ma. Our MDD modeling results along with aHe ages indicate that the BWT began rapid cooling at 6–7 Ma and reached the surface between 3 Ma and the present. This is in general agreement with the finding by Hayman (2006) that ∼3 km of sedimentary cover has been removed from the Black Mountains crystalline core since the late Pliocene (2.8–3.6 Ma). The sedimentary cover has been deposited in the adjacent Furnace Creek Basin to the north and the Death Valley Basin to the west (Çemen et al., 1985; Çemen and Wright, 1988; Wright et al., 1999; Fig. 1).

The cooling envelopes broadly imply a progressive northwest migration of cooling of the turtlebacks through the ductile crust (Fig. 7). As mentioned earlier, the earlier ductile cooling history of the turtlebacks is likely obscured by the >500 °C magmatic reheating of the CCT and MPT from the intrusion of the WSD and SMG at ca. 12–5 Ma.

Regional Tectonic Drivers of the Cooling Histories

The cooling path of the BWT has at least three distinct phases with two inflection points (Fig. 8). First, slow cooling (<5 °C/m.y.) from ca. 32 to 8 Ma was likely caused by some combination of erosion and faulting during pre– to early–Basin and Range extension (Reheis and Sawyer, 1997; Snow and Wernicke, 2000; Bennett et al., 2003). In the Furnace Creek basin (FCB) to the north, (Fig. 1), the Artist Drive formation marks the earliest sedimentary rock deposition at 14 Ma, coeval with the onset of Basin and Range extension in the Black and Funeral Mountains (Çemen et al., 1985; Çemen and Wright, 1988).

The inflection point at 7–6 Ma in our MDD modeling results (Figures DR4–DR7) marks the transition from slow to rapid cooling at the BWT. Regional onset of dextral transtension possibly occurred around this time with initiation of Sheephead and Furnace Creek strike-slip faults and the onset of pull-apart basin formation (e.g., Stockli et al., 2003; Bidgoli et al., 2015); however, we cannot rule out estimates of initiation slightly prior to this at ca. 10 Ma as noted by Renik and Christie-Blick (2013). The initiation of dextral transtension as part of the Basin and Range extension is likely due to a change in Pacific plate vectors from NW to coast-parallel with western North America (Atwater, 1970; Atwater and Stock, 1998) at ca. 10–8 Ma. The timing of strike-slip to oblique-slip on individual faults from this change appears to propagate westward from at least ca. 10–8 Ma to the present (e.g., Burchfiel et al., 1987; Holm et al., 1992; Hoisch and Simpson 1993; Niemi et al., 2001; Stockli et al., 2003). This propagation appears to have reached the BWT at ∼7–6 Ma based on our results.

Subsidence and deposition intensified in the FCB at 6 Ma, recorded by the relatively thick Furnace Creek (6–5 Ma) and Funeral (5–3 Ma) Formations (Wright et al., 1999). The Death Valley basin (DVB) initiated at ca. 6.5 Ma (Holm et al., 1994a). This period of rapid deposition correlates with the most rapid cooling of the BWT, with an exhumation rate of 4–8 mm/yr assuming a standard Basin and Range 30 °C/km geothermal gradient (Lachenbruch, 1979; Fig. 7). SMG clasts in the Furnace Creek Formation conglomerates with a depositional age of 6–4 Ma also indicate that some portion of the SMG was on the surface by 6 Ma (Wright et al., 1999).

No results exist to calculate a contemporary cooling rate; however, the aHe age cluster along with the MDD model T-t slope broadly implies an exhumation rate between 1 and 4 mm/yr (30–120 °C/m.y. cooling) from ca. 4.5 Ma to the present (Fig. 7). This may be the result of transfer of extension westward into the Panamint Valley (Andrew and Walker, 2009) and later the Owens Valley (Stockli et al., 2003) fault zones. Modeling of a sample transect in the central Black Mountains by Bidgoli et al. (2015) showed that a third inflection point may exist between 3–4 Ma, which they partially attributed to lithospheric delamination and associated uplift of the Sierra-Nevada Batholith. Extension continues presently along the range-front fault on the western edge of the BWT at 1.0 ± 0.2 mm/yr (Frankel et al., 2016), in the Panamint Valley at 2.1–2.7 mm/yr (Hoffman, 2009), and in other places along the Walker Lane Belt at 0.1–0.7 mm/yr (e.g., Berry, 1997; Kirby et al., 2006; Ganev et al., 2010).

Structural Implications of the Cooling Histories

After integrating our thermochronometric results with previously published thermochronometric, geochronologic, geobarometric, paleomagnetic, and structural mapping results across the Black Mountains, we propose a new structural model to explain the Cenozoic extension and exhumation of the turtlebacks with six distinct phases (Fig. 8). The first phase involved crustal thickening and lower-crustal asthenospheric flow related to Sevier-Laramide orogenesis and collapse between 140 and 35 Ma (Livaccari, 1991). This is evidenced by the presence of late Mesozoic to early Cenozoic thrust faults throughout the Black Mountains (Otton, 1976; Holm et al., 1992; Miller, 2003), ca. 61–55 Ma quartz-feldspar pegmatites within each of the turtlebacks (Miller and Friedman, 1999; R. Friedman, 2000, personal commun., cited in Miller and Pavlis, 2005), amphibolite-grade mineral assemblages associated with 700 ± 100 MPa pressure (∼20 km at depth; Whitney et al. 1993), and ca. 66 Ma magmatic overgrowth on zircon grains from the MPT (Lima et al., 2018). Our two BWT basement gneiss U-Pb samples yield ages of ca. 83 and 79 Ma and may also capture magmatic zircon growth related to channelized asthenospheric flow (Lima et al., 2018) or extension from syn- and post-orogenic collapse (Livaccari, 1991).

In the second phase, from 35 to 16 Ma (Fig. 8), the basement rocks within the Black Mountains were exhumed slowly by erosion or collapse of the Sevier Hinterland (Livaccari, 1979). This is evidenced by the long duration between mAr ages of ca. 32, 25, and 8 Ma at the highest, middle, and lowest elevations of the BWT, respectively, and the slow cooling indicated in the K-feldspar MDD models from the sample at the mid-elevation BWT.

At ca. 16–14 Ma (Fig. 8), early Basin and Range detachment faults formed in the southeastern Black Mountains with a shallow northwest or west dip, marking the beginning of the third phase. Listric normal faults formed in the hanging wall and dip-slip immediately began extending the hanging wall causing an isostatic uplift in the footwall. It is possible that the major detachment fault was a reactivated thrust fault of the Sevier Orogeny as evidenced by older mylonitic fabrics in the turtleback shear zones (Miller and Pavlis, 2005).

From ca. 14 to 8 Ma (Fig. 8) in the fourth phase, Basin and Range extension continued and was accommodated by both the detachment fault at depth and the listric normal faults in the hanging wall of the detachment. Isostatic rebound of the detachment footwall during this time exhumed the basement rocks through the brittle-ductile transition zone as evidenced by bAr ages (Holm and Dokka, 1993). Mullion-style ductile folding of the turtleback footwalls (Wright et al., 1974; Norton, 2011) occurred during this period as extension occurred while the turtlebacks were in the mid-crust (Mancktelow and Pavlis, 1994; Dee et al., 2004). A thinning crust and associated asthenospheric upwelling eventually generated intermediate to mafic voluminous intrusions, the WSD and SMG, between 14 and 12 Ma along the ductile shear zone at the base of the detachment, which probably continued to grow and be reheated until at least ca. 10 Ma (Asmerom et al., 1990; Wright et al., 1991; Holm et al., 1993; this study) as extension along the shear zone movement progressed. Conductive heat from these intrusions reset the >300 °C thermochronometers at the CCT and MPT, resulting in 14 and 9 Ma hAr ages, respectively. Internal late-stage crystallization in the SMG would continue until possibly as late as ca. 6 Ma (see individual U-Pb ages in Table DR5, see footnote 1).

A significant kinematic change in plate motions marks the fifth phase (Fig. 8) producing transtension at ca. 7–6 Ma, possibly as early as 10 Ma, until 4.5–3.5 Ma, when a pull-apart basin developed in the Black Mountains as extension intensified (Burchfiel and Stewart, 1966; Norton, 2011). This is also suggested based on the modern rhombohedral shape of the bounding Death Valley–Furnace Creek, Death Valley–Black Mountains, Southern Death Valley, and Sheephead oblique- to strike-slip fault zones (Burchfiel and Stewart, 1966; Fig. 1). Maximum exhumation rates during this time are evidenced by the clustering of zHe and aHe ages at the BWT (this study) and of zHe ages in Sheep Canyon in the central Black Mountains (Bidgoli et al., 2015), steepened MDD modeling path at the BWT, the clustering of zFT and aFT ages at the CCT and MPT (Holm and Dokka, 1993), and the first appearance of SMG clasts in the hanging walls of sedimentary basins at 6–4 Ma (Çemen et al., 1985; Çemen and Wright, 1988; Holm and Dokka, 1993). High crustal attenuation and topographic relief caused rapid deposition in the Furnace Creek and Death Valley basins (Çemen et al., 1999; Wright et al., 1999). As the original detachment surface exhumed to the brittle crust, a new detachment fault formed at depth. The turtlebacks, plutons, and remaining sedimentary and volcanic rocks began extending via a dense network of high-angle normal faults. Oblique slip was occurring on the Death Valley–Furnace Creek and Death Valley–Black Mountains fault zones as well as along the faults between, causing lateral clockwise rotation of the crustal hanging wall blocks between them (Holm et al., 1993; Wright et al., 1991; Serpa and Pavlis, 1996; Miller, 1999a, 1999b). Gravity modeling by Norton (2011) shows thick sediments indicative of a pull-apart basin in the Death Valley basin, and to a lesser-extent in the central Black Mountains west of the Greenwater Range.

The sixth and final phase of extension began between 4.5 and 3.5 Ma (Fig. 8) and continues today. Exhumation in the Black Mountains was slower during this period, as evidenced by the aHe ages in relation to the zHe ages and MDD path, and thermal modeling of zHe ages by Bidgoli et al. (2015). We attribute this slowing of exhumation to a transfer of extension westward into the Panamint (Andrew and Walker, 2009) and Owens Valley fault systems (Stockli et al., 2003). Since the late-Pliocene (ca. 4–3 Ma), the BWT was exhumed from ∼3 km depth to the surface (Hayman, 2006). Oblique slip on the Black Mountains fault zone continues today at a horizontal slip rate of ∼0.8–1.0 mm/yr and along the Panamint fault zone at ∼2 mm/yr (Frankel et al., 2016), with the modern Death Valley topography starting at 2–3 Ma (Knott et al., 2005).


Our new transect-based thermochronometric and geochronologic data from the Badwater turtleback provide constraints on the exhumation of the gneissic rocks of the northern Black Mountains and the adjacent intrusions. Our results demonstrate the importance of placing thermochronometric data in the context of diverse data sets to generate a cohesive 3-dimensional testable structural model, as we did for the Death Valley region (Fig. 1), potentially one of the most highly extended terranes in the world. Our data confirmed the presence of gneissic rocks from the Mojave Province in the Badwater turtleback (Table 2). Moreover, mean U-Pb ages for the Willow Spring diorite and Smith Mountain granite samples indicate younger crystallization in the northern Black Mountains when compared to the same plutons in the central Black Mountains (Figs. 2 and 3).

Apatite and zircon (U-Th)/He ages from 10 samples (Table 3) suggest 30–120 °C/m.y. cooling of the BWT through the apatite and zircon closure temperatures. Comparison of the upper- and middle-elevation sample ages (Fig. 5) at the Badwater turtleback suggests slow cooling of <5 °C/m.y., while the low-elevation sample ages suggest rapid ∼30–120 °C/m.y. cooling (Fig. 6). MDD results (Figs. DR4–DR7) suggest slow cooling of <2 °C/m.y. from ca. 10 to 6 Ma, followed by rapid cooling of 120–140 °C/m.y. from ca. 6 to 4.5 Ma, and slower cooling of 30–120 °C/m.y. from 4.5 Ma until the present. Our data, when combined with previously published data for the Copper Canyon turtleback and Mormon Point turtleback suggests that these turtlebacks exhumed progressively northward (Fig. 7).

Our proposed model is composed of six phases (Fig. 8), beginning with ca. 140–40 Ma Sevier-Laramide orogenesis of a thickened crust. Post-orogenic collapse and erosion dominated from ca. 32 to 16 Ma causing slow exhumation of the Badwater turtleback. At 16–14 Ma, a detachment fault formed with a breakaway south west of the Black Mountains, with listric normal faults in the hanging wall. Basin and Range extension continued from 14 to 8 Ma causing exhumation of the turtlebacks through the brittle-ductile transition. Dextral transtension associated with a change in Pacific plate motion began at 7–6 Ma, producing a pull-apart basin across the Black Mountains and rapid extension rates within. The locus of deformation transferred to the Panamint and Owens Valley fault systems between 4.5 and 3.5 Ma, slowing extension in the Black Mountains.


Fieldwork was assisted by Daphne Douglas and Jessie Roughgarden on the rugged Badwater turtleback. Harold Stowell, Kim Genareau, Johnny Goodwin, Rob Holler, and Rudra Chatterjee provided geochemistry advice and laboratory training and support. Sarah Carson gave invaluable modeling and graphics design support. We benefited tremendously from many discussions and insightful comments of Darrel Cowan, Marli Miller, and Brandon Lutz. Marli Miller and Ian Norton also provided insightful reviews for this manuscript that helped with the overall quality. We thank the National Park Service for permitting this research in Death Valley National Park. Funding was provided by the University of Alabama School of Arts and Sciences, Department of Geological Sciences, Johnson Fund; Geological Sciences Advisory Board Hooks-Graduate Student fellowship grant, and the Graduate Student Research Funds by the University of Alabama Graduate School.

1GSA Data Repository Item 2019136, sample selection and methods, Tables DR1–DR9, and Figures DR1–DR9, is available at http://www.geosociety.org/datarepository/2019, or on request from editing@geosociety.org.
Gold Open Access: This paper is published under the terms of the CC-BY-NC license.