Tracking the spatial and temporal links between late Cenozoic extension and magmatism in the Death Valley region (California, USA) Open Access

Reconstructing the magmatic history in the Basin and Range province (western United States and northwestern Mexico; henceforth, “Basin and Range”) from 50 Ma to 12 Ma reveals two distinct space-time trends that highlight the region's complex geologic evolution. The first trend shows a southwestward progression of magmatism across Nevada, while the second trend shows a northwestward progression through western Arizona into southern Nevada along the Colorado River extensional corridor (Fig. 1; Christiansen et al., 1992; Humphreys, 1995; Gans and Bohrson, 1998; Konstantinou et al., 2012; Best et al., 2013; cf. Glazner, 2022). These magmatic trends nearly converge at the narrowest part of the Basin and Range, where significant extension has occurred, particularly in the Death Valley region, which has undergone some of the greatest amounts of extension in the entire Basin and Range (e.g., Stewart, 1983; Wernicke et al., 1988; Wernicke, 1992; McQuarrie and Wernicke, 2005).

The relationship between Cenozoic magmatism and extension across the Basin and Range is complex. In the northern Basin and Range of Nevada, magmatism and extension appear to occur independently (Best and Christiansen, 1991; Colgan et al., 2006). Conversely, in the Colorado River extensional corridor of the southern Basin and Range, magmatism precedes large-magnitude extension (Fig. 1; Glazner and Bartley, 1984; Gans et al., 1989; Armstrong and Ward, 1991; Gans and Bohrson, 1998). This disparity raises intriguing questions about the primary drivers for magmatism in actively extending tectonic settings. Potential drivers in the Basin and Range include mantle plume processes (e.g., Best and Brimhall, 1974; Parsons et al., 1998; Wang et al., 2002), decompression melting resulting from continental extension (e.g., Gans et al., 1989; Christiansen et al., 1992; Wernicke, 1992; McQuarrie and Oskin, 2010; Putirka and Platt, 2012), northward migration of the Mendocino triple junction (e.g., Dickinson and Snyder, 1979; Glazner and Supplee, 1982), slab rollback (e.g., Humphreys, 1995, 2009; Feldstein and Lange, 1999), and variations of these processes.

The Death Valley region is characterized by multiple stages of deformation since ca. 15 Ma (e.g., Wernicke et al., 1988; Wright et al., 1991; Wernicke, 1992; Holm and Dokka, 1993; Davis et al., 1993; Topping, 1993; Miller, 1999; Miller and Friedman, 1999; McQuarrie and Wernicke, 2005; Andrew and Walker, 2009; Fridrich and Thompson, 2011; Norton, 2011; Andrew et al., 2015; Bidgoli et al., 2015; Sizemore et al., 2019; Fleming et al., 2021, 2022) and provides a significant record of late Cenozoic magmatism, featuring well-preserved volcanic deposits and exhumed plutons revealed through tectonic and erosional denudation (e.g., Drewes, 1963; McDowell, 1974; Asmerom et al., 1990; Holm and Wernicke, 1990; McKenna and Hodges, 1990; Wright et al., 1991; Davis et al., 1993; Calzia and Rämö, 2005; Luckow et al., 2005; Miller and Pavlis, 2005; Fridrich and Thompson, 2011; Calzia et al., 2016). The volcanic record provides snapshots of magmatic activity during discrete eruptive events, many of which reflect shorter-lived processes (e.g., Lipman, 1984; Bachmann et al., 2007). These snapshots may obscure the long-term evolution of regional magmatism, complicating our understanding of the temporal links between extension and magmatism. In contrast, the plutonic record offers a more integrated view over extended periods (e.g., Hildreth, 1981; Matzel et al., 2006; Miller et al., 2011; Paterson et al., 2011), enhancing our understanding of the driving forces behind magmatism in extensional settings.

Despite the rich geological record, the relationship between intrusive magmatism and tectonics in the Death Valley region remains poorly understood, primarily due to the lack of modern high-precision crystallization ages for the region's plutons. In this study, we present new chemical abrasion–isotope dilution–thermal ionization mass spectrometry (CA-ID-TIMS) ²⁰⁶Pb/²³⁸U zircon ages and zircon trace element geochemical data for eight Miocene plutons in Death Valley (Fig. 2). This study aims to determine the timing and duration of plutonism in relation to regional tectonic processes. Specifically, we will: (1) assess whether intrusive magmatism occurs episodically or continuously; (2) identify spatiotemporal trends associated with intrusive magmatism; and (3) develop an integrated framework that connects plutonism, volcanism, and tectonism for the late Cenozoic in Death Valley.

Late Miocene extension in the Death Valley region has resulted from 42 km to 105 km of displacement across the region. This displacement was partitioned into a complex network of strike-slip and normal fault systems, which in its latest stage formed a pull-apart basin (e.g., Burchfiel and Stewart, 1966; Stewart, 1983; Wernicke et al., 1988; Serpa et al., 1988; Snow and Wernicke, 1989, 2000; Serpa and Pavlis, 1996; Norton, 2011). The Death Valley region consists of several crustal blocks of significant size, including the (1) Black Mountains, the tectonically exhumed footwall of the Amargosa-Black Mountains detachment; (2) Greenwater Range, a block dextrally offset from the Black Mountains; and (3) Panamint Range, a large hanging-wall block (Fig. 2).

The eastern part of the Death Valley region includes the Kingston Range–Halloran Hills detachment fault (Fig. 2; Burchfiel et al., 1983; Davis et al., 1993; Fowler and Calzia, 1999). The extensional breakaway of the Death Valley region occurs in the Kingston Range, indicated by exposed shallow footwall rocks that show minor amounts of displacement (Davis et al., 1993; Fowler and Calzia, 1999). In the following section, we provide a brief description of the Black Mountains, Greenwater Range, Kingston Range, and Panamint Range, along with the late Miocene plutons exposed in these ranges.

The Black Mountains are a highly exhumed, southeast-tilted crustal block that exposes a structural discontinuity between (1) a footwall of metamorphosed Proterozoic rocks (e.g., Holm and Wernicke, 1990; Miller and Friedman, 1999; Miller and Pavlis, 2005) and mid-crustal Miocene intrusions (>10–20 km emplacement depth; Holm et al., 1992; Pavlis et al., 2018; Fleming et al., 2022) and (2) a hanging wall of Miocene–Pliocene volcanic and sedimentary rocks lying on highly deformed but unmetamorphosed Proterozoic strata with their underlying crystalline basement. This structural discontinuity is marked by the Amargosa–Black Mountains detachment fault (e.g., Holm and Wernicke, 1990; Miller, 1991, 2003; Wright et al., 1991; Holm et al., 1992; Holm and Dokka, 1993; Topping, 1993; Mancktelow and Pavlis, 1994; Miller and Pavlis, 2005). The Black Mountains preserve a complex sequence of repeated metamorphic and high-temperature deformation events since the Cretaceous (e.g., Holm et al., 1992; Miller and Friedman, 1999; Miller, 2003; Miller and Pavlis, 2005; Pavlis et al., 2018; Lima et al., 2018; Fleming et al., 2022).

The Black Mountains are known for three doubly plunging antiformal structures, locally termed as turtlebacks, which are exposed in the footwall rocks of the Amargosa–Black Mountains detachment fault (e.g., Curry, 1938, 1954; Wright et al., 1974, 1991; Holm and Wernicke, 1990; Holm et al., 1992; Holm and Dokka, 1993; Topping, 1993; Mancktelow and Pavlis, 1994). The three turtleback structures—Badwater, Copper Canyon, and Mormon Point (from north to south)—are composed of quartz-feldspar gneiss, pelitic schist, and calcite-dolomite marble rocks that are ductilely sheared, faulted, and folded (e.g., Curry, 1938, 1954; Drewes, 1959; Otton, 1977; Miller, 1991, 1992; Pavlis, 1996; Miller and Friedman, 1999; Çemen et al., 2005; Miller and Pavlis, 2005). Miocene plutons form sill-like intrusions along the structural top of each turtleback and surround the Copper Canyon and Mormon Point turtlebacks (Fig. 2). The basal parts of these sills are mylonitic and are interpreted to represent the exhumed brittle-ductile transition (e.g., Holm et al., 1994b, 1994a; Pavlis, 1996; Miller and Pavlis, 2005). Late Miocene–aged dikes crosscut the Miocene intrusions and turtleback structures (e.g., Wright et al., 1991; Holm et al., 1992; Miller and Friedman, 1999; Dee et al., 2004).

Thermochronology studies of the Black Mountains indicate that high exhumation rates occurred between 10 Ma and 8 Ma (McKenna, 1990; Meurer, 1992; Holm and Dokka, 1993; Bidgoli et al., 2015; Sizemore et al., 2019). This rapid extension resulted in the unroofing of multiple mid-crustal plutons in the western Black Mountains, including the Willow Spring Diorite, Smith Mountain Granite, and quartz monzonite of Gold Valley. Additionally, another diorite intrusion, known as the diorite of Furnace Site, was previously classified as part of the northern Willow Spring Diorite but is distinguishable by its finer-grained texture.

The Greenwater Range is dextrally offset by ~15 km along the Grandview fault from the Black Mountains crustal block (Topping, 1993). This range is composed primarily of late Miocene to Pliocene volcanic rocks (Fig. 2). Plutonic rocks are exposed in several erosional windows beneath a nonconformable contact (Holm et al., 1992) or possibly locally a detachment fault (Calzia and Rämö, 2005). The plutonic rocks range in composition from quartz monzonite to granite (Haefner, 1972; Almashoor, 1983; Holm et al., 1992; Holm, 1995; Petronis et al., 2002; Rämö et al., 2011) and include the granite of Deadman Pass and the Shoshone Granite.

The Kingston Range exposes Proterozoic rocks similar to those in the southern Black Mountains but at lower metamorphic grades (Calzia et al., 2000). These rocks have been minimally exhumed by the Kingston Range–Halloran Hills detachment, which partially reactivates or is localized along a section of Jurassic and Cretaceous thrust faults exposed southeastward in the Mesquite and Clark Mountains (Fig. 2; Burchfiel et al., 1987; Snow and Wernicke, 1989; Walker et al., 1995). The detachment offsets Proterozoic to Paleozoic sedimentary rocks, 14–13 Ma volcanic rocks, and 13–8 Ma supradetachment basin rocks that dip 30° eastward (Davis et al., 1993; Fowler and Calzia, 1999). Following the initiation of ~6 km of WSW-directed displacement along the Kingston Range–Halloran Hills detachment, the Kingston Range Granite was emplaced (Fig. 2; Calzia, 1990; Fowler and Calzia, 1999; Calzia and Rämö, 2000), effectively pinning a segment of the detachment fault. Consequently, a system of strike-slip faults developed to the north and south along with a normal fault to the west, surrounding the granite-pinned segment of the detachment to accommodate continued extension (Davis et al., 1993).

The western boundary of our study area is defined by the Panamint Range (Fig. 2), a structural block situated between two west-dipping faults: the Emigrant detachment fault and Eastern Panamint fault system. The Emigrant detachment fault, which bounds the western side of the Panamint Range, has experienced two phases of deformation: (1) moderate down-to-the-west extension at 15–10 Ma, followed by (2) rapid unroofing with top-to-the-northwest displacement starting at ca. 4 Ma (Andrew and Walker, 2009; Bidgoli et al., 2015). The Eastern Panamint fault system has been active since after 10 Ma, dipping westward beneath the Panamint Range. It is interpreted as a hanging-wall splay of the Amargosa–Black Mountains detachment fault with top-to-the-northwest displacement (McKenna and Hodges, 1990). The hanging wall exposes Paleoproterozoic crystalline basement, Mesoproterozoic to Triassic sedimentary rocks, Mesozoic plutonic and volcanic rocks, and Miocene plutonic and volcanic rocks (e.g., Wasserburg et al., 1959; Lanphere et al., 1964; Stern et al., 1966; McDowell, 1974). Miocene plutonism includes the Little Chief Granite, which intrudes the central Panamint Range (Hunt and Mabey, 1966; McDowell, 1974, 1978; Topping, 1993).

In this study, we focus on eight plutons exposed in the Death Valley region, including several plutons exposed in the footwall of the main extensional fault system, the Amargosa–Black Mountain detachment, and a pluton in the hanging wall (Table 1). The Willow Spring Diorite, Smith Mountain Granite, quartz monzonite of Gold Valley, and diorite of Furnace Site are exposed in the Black Mountains; the granite of Deadman Pass in the Greenwater Range; the Little Chief Granite in the Panamint Range; the Kingston Range Granite in the Kingston Range; and the granite of Rabbit Holes Spring, just west of the Kingston Range (Fig. 2). These plutons all intrude Paleoproterozoic to Neoproterozoic rocks (e.g., Hewett, 1940; Otton, 1977; McDowell, 1978; Holm and Wernicke, 1990; Miller, 1992; Pavlis, 1996; Mahon et al., 2014). Notably, all the felsic granitoids in Death Valley have rapakivi textures, potassium feldspar megacrysts mantled by plagioclase feldspar. A summary of all Miocene intrusions in Death Valley, including those outside the scope of this study, is provided in Table 2.

The Kingston Range Granite is an elliptical rapakivi granite (~150 km²) exposed in the Kingston Range (Fig. 2; Hewett, 1956; Calzia, 1990; Calzia et al., 2000; Calzia and Rämö, 2005). This granitic porphyry intrudes unmetamorphosed Mesoproterozoic to Neoproterozoic Pahrump Group strata (Hewett, 1940; Mahon et al., 2014). The Kingston Range Granite has three phases (from oldest to youngest): (1) northeastern feldspar porphyry, (2) southwestern quartz porphyry, and (3) aplite (Calzia and Rämö, 2005). The Kingston Range Granite contains widespread miarolitic cavities, consistent with shallow emplacement (<4 km; Calzia, 1990; Davis et al., 1993). Mafic xenoliths and rhyolite porphyry dikes are commonly found in both the feldspar porphyry and quartz porphyry phases. The only geochronology for the Kingston Range Granite is a 12.4 Ma hornblende ⁴⁰Ar/³⁹Ar age (Calzia, 1990).

The granite of Rabbit Holes Spring is a small rapakivi intrusion (~8 km²) located west of the Kingston Range (Fig. 2). There are two phases of the granite: (1) gray, medium-grained, monazite-bearing biotite monzogranite and (2) tan, medium- to coarse-grained monzogranite. The tan monzonite phase is typically equigranular but locally porphyritic and commonly contains pegmatite dikes and blocks of gray monzonite (Calzia et al., 2003). The granite of Rabbit Holes Spring was previously interpreted as Cretaceous despite K-Ar biotite ages of 11.0 Ma and 11.6 Ma from the gray and tan granite phases, respectively (Calzia, 1990; Calzia et al., 2003).

The Little Chief Granite is a shallow (<4 km emplacement depth) granite exposed over an area of ~28 km² that intrudes Paleoproterozoic gneiss and Meso- to Neoproterozoic metasedimentary rocks of the Panamint Range (Fig. 2; McDowell, 1978). The Little Chief Granite possesses hypabyssal and rapakivi textures, miarolitic cavities, and mafic inclusions (McDowell, 1974). The granite has two phases: a northern phase, which intrudes the southern phase (McDowell, 1974). The southern phase contains abundant accessory monazite, commonly coarse enough to be visible in hand sample. Prior geochronology for the Little Chief Granite includes: (1) a K-Ar feldspar age of 12.5 ± 1.3 Ma (Stern et al., 1966) from a boulder located at the mouth of the Hanaupah Canyon; (2) a plagioclase-biotite-hornblende Rb-Sr isochron of a similar boulder from Hanaupah Canyon with an age of 10.6 ± 0.2 Ma (Hodges et al., 1990); (3) an ⁴⁰Ar/³⁹Ar biotite age of 11.21 ± 0.08 Ma (McKenna, 1990); and (4) zircon fission-track ages of 11.2 ± 0.5 Ma and 9.8 ± 0.5 Ma for the southern and northern facies, respectively (Topping, 1993).

The Willow Spring Diorite is the largest Miocene intrusion in Death Valley, exposed over ~300 km² in the Black Mountains (Fig. 2; Asmerom et al., 1990). This intrusion is a hornblende gabbro to quartz diorite and appears to form sheets that locally intermingle (Asmerom et al., 1990; Holm, 1995). The Willow Spring Diorite intrudes 1.7 Ga crystalline rocks, Mesoproterozoic metadolomite, and Neoproterozoic Noonday Dolomite (Wright et al., 1974, 1991; Otton, 1977; Holm et al., 1992; Miller and Friedman, 1999; Miller, 2003). The mafic phases of the intrusion are likely sourced from mantle-dominated magmas based on low ⁸⁷Sr/⁸⁶Sr and high εNd. Isotopic variations on the tens-of-meters scale within the Willow Spring Diorite suggest that magma mixing occurred at or close to the depth of emplacement (Asmerom et al., 1990). Prior geochronology includes three U-Pb bulk zircon TIMS ages using older analytical methods (no chemical abrasion): (1) 10.93 ± 0.75 Ma (Wright et al., 1991), (2) 11.6 ± 0.2 Ma, and (3) 11.5 ± 0.1 Ma (Asmerom et al., 1990). Hornblende and biotite ⁴⁰Ar/³⁹Ar ages indicate that the Willow Spring Diorite cooled to ~500 °C by ca. 10 Ma and then to ~350 °C by ca. 8 Ma (McKenna, 1990; Holm and Dokka, 1993).

The Willow Spring Diorite was emplaced as a sill-like body above the turtlebacks (Fig. 2), which represent the floor of the Willow Spring Diorite pluton, in a mid-crustal shear zone (Holm et al., 1992; Pavlis, 1996; Miller and Pavlis, 2005). Ductile fabrics are generally limited to the lowermost part of the pluton (Pavlis, 1996), and structurally higher levels are capped by the curved fault systems of the Death Valley turtlebacks (e.g., Holm and Wernicke, 1990; Miller, 1991, 1992; Pavlis, 1996). Brittle deformation that overprints ductile foliation is observed only in the northern part of the shear zone at the base of the diorite sill (Holm and Wernicke, 1990).

The Smith Mountain Granite is a biotite rapakivi granite exposed over an area of ~40 km² in the southern Black Mountains (Fig. 2). The Smith Mountain Granite intrudes the ductile shear zone of the Mormon Point turtleback (Miller et al., 2004), which includes 1.7 Ga metamorphic basement rocks and the upper portion of the Willow Spring Diorite. The Smith Mountain Granite and Willow Spring Diorite show magma mingling textures, suggesting that the two plutons are comagmatic (Meurer, 1992; Holm et al., 1994a, 1994b; Miller et al., 2004). But in contrast to the diorite, the Smith Mountain Granite lacks ductile deformation structures. Prior ⁴⁰Ar/³⁹Ar geochronology for the Smith Mountain Granite yielded ages ranging from 8.69 ± 0.31 Ma (hornblende; Holm et al., 1992) to 10.85 ± 0.04 Ma (orthoclase; Holm et al., 1994a). Two TIMS U-Pb zircon analyses, using older analytical methods (no chemical abrasion), yielded a crystallization age of 10.45 ± 0.22 Ma (Miller et al., 2004).

The Willow Spring Diorite and Smith Mountain Granite are intruded by a medium-grained quartz monzonite pluton and associated dikes, collectively known as the quartz monzonite of Gold Valley, located east of the Copper Canyon turtleback (Holm, 1995). This quartz monzonite pluton contains minor biotite and amphibole and locally exhibits rapakivi textures (Drewes, 1963). The quartz monzonite of Gold Valley appears as a sheeted, tabular pluton that intrudes into the upper portion of the Willow Spring Diorite as numerous subhorizontal interconnected sills. In some places, the quartz monzonite of Gold Valley and Willow Spring Diorite are interfingering, and in other places they form a faulted contact with each other (Supplemental Material¹). Hornblende ⁴⁰Ar/³⁹Ar dating yields a cooling age of 8.69 ± 0.31 Ma (Holm et al., 1992).

The granite of Deadman Pass is one of several porphyritic and rapakivi quartz monzonite–granite plutons in the Greenwater Range (e.g., Calzia and Rämö, 2005; Calzia et al., 2016). These intrusions were emplaced shallowly (2–2.5 km) and are exposed in erosional windows through Miocene and Pliocene volcanic rocks (Almashoor, 1983). Other granitoids are exposed along the southwest side of the Greenwater Range (Drewes, 1963) and below the Shoshone Volcanics located southeast of the granite of Deadman Pass. The granite of Deadman Pass contains several large xenoliths of gneiss, marble, and quartzite, consistent with intrusion into Neoproterozoic to Mesoproterozoic bedrock (Haefner, 1972).

The diorite of Furnace Site is a fine-grained quartz diorite that was previously mapped as the northern biotite-bearing phase of the Willow Spring Diorite (Otton, 1977). The diorite intruded as a sill-like body below a sheet of quartz monzonite. The diorite is located below the unconformity exposed at Furnace Site. Based on new U-Pb zircon geochronology and zircon trace element data presented herein, we separate this unit out as a distinct diorite pluton younger than the Willow Spring Diorite.

Eight intrusions were sampled across the Death Valley region, with multiple intrusive phases collected from the Kingston Range Granite, Little Chief Granite, Willow Spring Diorite, and Smith Mountain Granite. Samples of the granite of Rabbit Holes Springs were obtained from J. Calzia (personal commun., 2020), a sample of diorite of Furnace Site was provided by R. Thompson and A. Gilmer (personal commun., 2019), and a sample of the quartz porphyry phase of the Kingston Range Granite from T. Bidgoli (personal commun., 2012). Within the largest intrusion, Willow Spring Diorite, rock samples were obtained from different structural levels (Fig. 2). Field photographs can be found in the Supplemental Material.

Zircon separation from rock samples followed standard heavy mineral separation techniques to isolate grains. Selected zircon grains were annealed for 60 h at 900 °C in a muffle furnace, modified after the CA-TIMS method of Mattinson (2005). Annealed grains were mounted in epoxy and polished halfway to expose the interior of the grains. Cathodoluminescence (CL) imaging was conducted using a JEOL 5800 LV scanning electron microscope (SEM) with a K.E. Developments panchromatic CL detector at the U.S. Geological Survey in Denver, Colorado, USA. Based on the internal structures and zoning patterns in the CL imagery (Fig. S1), we targeted specific zircon grains for CA-ID-TIMS U-Pb geochronology and zircon trace element analysis (TIMS-TEA; Schoene et al., 2010). To fully investigate zircon crystallization histories, we selected zircons of varying grain sizes, and where possible, we microsampled grains to isolate tips and cores as informed by CL imagery. Grain fractions that were large enough to break into multiple fragments were separated into two to three pieces and are indicated with a letter at the end of the fraction identifier: T, C, and B correspond to top (tip), center, and bottom (other tip) fragments of a grain, respectively (Table 3; Fig. S1).

Selected grains and grain fragments were loaded into 300 µL fluorinated ethylene propylene (FEP) microcapsules and leached in 50 µL of 29 M HF in a Parr acid digestion vessel for 12 h at 210–215 °C. Grains were subsequently fluxed with 7 M HNO₃ for at least 20 min, rinsed four times with ultrapure water, fluxed again with 6 M HCl for at least 20 min, and finally rinsed four times with ultrapure water. Following chemical abrasion treatment, zircon grains and grain fragments were loaded into individual microcapsules with 50 µL of 29 M HF and ~0.009 g of a mixed EARTHTIME ²⁰⁵Pb-²³³U-²³⁵U tracer (ET535; Condon et al., 2015; McLean et al., 2015). Samples were dissolved in a Parr vessel for ~60 h at 210–215 °C, dried to salts, converted to chlorides by fluxing microcaps with ~50 µL of 6 M HCl overnight either in a Parr vessel at >180 °C or on a hotplate at 100 °C, dried to salts again, and finally redissolved in 50 µL of 3 M HCl for anion-exchange chemistry. Separation and purification of Pb and U from zircon followed a modified procedure from Krogh (1973) using 50 µL HCl-based AG 1-X8 resin. Prior to Pb and U elution, trace elements were eluted dropwise using ~75 µL of 3 M HCl. Pb and U were eluted into the same Teflon beaker with 200 µL of 6 M HCl and 250 µL of ultrapure water, respectively, and were dried down with ~5 µL of ~0.018 M H₃PO₄.

Pb and U isotopic data were collected on an Isotopx Phoenix thermal ionization mass spectrometer at The University of Kansas Isotope Geochemistry Laboratories (KU IGL). Both Pb and U were loaded onto the same degassed zone-refined rhenium filament using a mixed solution of colloidal silica gel and dilute H₃PO₄ (Gerstenberger and Haase, 1997). Pb was analyzed as Pb⁺ via peak-hopping on a Daly ion counting system, while U was analyzed as UO₂⁺ statically on Faraday collectors. Isotopic fractionation corrections for Pb and U are based on repeated analysis of certified Pb reference materials (NBS981 and NBS982) and the estimated ²³³U/²³⁵U in the ET535 tracer solution, respectively (Condon et al., 2015). Data reduction and uncertainty propagation were performed using Tripoli and ET_Redux software packages (Bowring et al., 2011) with algorithms detailed in McLean et al. (2011).

Zircon Th/U ratios are calculated from CA-ID-TIMS U-Pb analysis. The U content was determined by isotope dilution, and the Th content was calculated from the radiogenic ²⁰⁸Pb measurement and ²³⁰Th-corrected ²⁰⁶Pb/²³⁸U date of the zircons, assuming concordance between the U-Pb and Th-Pb systems. Initial ²³⁰Th disequilibrium corrections were applied to the zircon ²⁰⁶Pb/²³⁸U ages assuming that the Th/U of the melt is the same as the whole-rock Th/U (Table S2; Kirkland et al., 2015; Gilmer and Thompson, 2024). For plutons for which whole-rock data do not exist (e.g., Little Chief Granite, granite of Rabbit Holes Spring, and Kingston Range Granite), we assume 3.5 for the whole-rock Th/U ratio, as that is the mean value for granitoids (Kirkland et al., 2015). All age and isotope ratio uncertainties are reported as ±2σ (95% confidence intervals; Table 3).

Zircon trace element analysis (TEA) was performed using the Element2 inductively coupled plasma–mass spectrometer (ICP-MS) with an Elemental Scientific Instruments Apex-Omega HF desolvating nebulizer at KU IGL. Our TEA sample preparation and analysis follow a modified TIMS-TEA procedure from Schoene et al. (2010). Trace element aliquots collected during U-Pb chemistry separation were dried to salts, dissolved with ~7 M HNO₃ to remove organics, dried down again, and finally redissolved in 1 mL of a mixed 0.45 M HNO₃ and 0.02 M HF solution that is doped with ~1 ppb Ir. Ir is used as an internal standard to monitor potential plasma fluctuations and varying nebulizer aspiration rates between samples (Schoene et al., 2010).

Instrument and desolvating nebulizer settings, such as gas flow, radio-frequency power, and torch positions, were optimized for high sensitivity (≥1.2 × 10⁷ cps/ppb ²³⁸U) and minimal oxide formation (UO⁺/U⁺ < 0.4%) based on a 1 ppb multielement calibration solution. The same solution was also used for mass and analog-counting-factor calibrations, which were performed at the start of each ICP session. Analyses were conducted in low mass resolution (m/Δm = 400) mode. The flow rate of the Teflon nebulizer is 125 µL/min, while the uptake time was set to 40–45 s. A small aliquot of each sample was diluted by 100× and analyzed for Zr, Ti, La, Dy, Yb, and Hf for a rapid “dip-test” run. Based on the rapid dip tests, all samples were diluted so that samples did not exceed 40 ppb Zr to minimize large Zr intensities while still being able to measure lower abundance trace elements (e.g., light rare earth elements [REEs]).

When collecting TIMS-TEA analyses, four calibration solutions bracketed 15 unknown samples, and a blank sample was analyzed prior to each standard or sample. Our calibration solutions contained a range of concentrations (4.5–150 ppb Zr, 1–500 ppt REEs). Elements analyzed include Y, Zr, Hf, and REEs. Between samples, a series of rinses using a mixed 0.45 M HNO₃ and 0.02 M HF solution was performed. Procedural blanks were prepared and measured alongside samples.

Data reduction of TIMS-TEA analyses, including corrections for procedural and acid blanks, as well as concentration calculations were performed using MATLAB software. To convert from solution concentrations to zircon concentrations, it is assumed that all measured trace elements substitute for Zr⁺⁴ so that the sum Σ = Zr + Hf + Sc + Y + Nb + Ta + REEs = 497,646 ppm, which is the stoichiometric concentration of the Zr site in zircon (Hoskin and Schaltegger, 2003). Given that the uncertainty in the normalization factor is a systematic error, this error was not propagated because these uncertainties would cancel out for elemental ratios. No significant laboratory trace element contamination is observed from analyzed total procedural blanks. Results are summarized in Table S1 with uncertainties reported as 2σ (95% confidence intervals).

Thirteen samples from eight Miocene intrusions in the Death Valley region (Table 1) were dated by CA-ID-TIMS U-Pb zircon geochronology. These Miocene zircon grains show typical oscillatory and sector zoning patterns. CL images are provided in Figure S1. Isotopic ratios and CA-ID-TIMS U-Pb dates are presented in Table 3 and illustrated in Figure 3. We interpret each rock sample to represent a single magmatic phase of its respective pluton.

Collectively, the ~140 CA-ID-TIMS U-Pb zircon dates define a protracted magmatic history extending from 13.17 Ma to ca. 7.88 Ma (Fig. 3). Ten out of the thirteen CA-ID-TIMS samples exhibit ²⁰⁶Pb/²³⁸U age dispersion that is greater than analytical uncertainties, while three samples have a single resolvable population of zircon crystallization ages (Table 4). For samples with resolvable age dispersion, we use the Bayesian algorithm in the software Chron.jl (Keller et al., 2018) to estimate the maximum (t_max) and minimum (t_min) of the zircon age distribution, which can be interpreted as estimates of the time at which the magmatic system reaches zircon saturation and the solidus, respectively (Table 4; Fig. S2). Given that zircon age distributions are inherently asymmetric (Keller et al., 2018), crystallization time scales (Δt) are the mode of the kernel density estimation distribution of t_max − t_min. These estimates rely on the assumption that zircon crystallization is continuous from the start of zircon saturation until the solidus is reached. Using this approach ensures that the crystallization age range is not artificially expanded by analytical uncertainties (Keller et al., 2018). The largest age dispersion in a single sample is observed in the lower Willow Spring Diorite (786 ^+0.18/_−0.06 k.y.). Protracted crystallization histories (≥550 k.y.) were also documented in individual zircon grains where multiple fractions from the same grain are dated (sample DVCC18-2: fractions XL10B, XL10C, and XL10T; Table 3).

Some Death Valley samples exhibit distinct zircon U-Pb age distributions. For example, the Kingston Range Granite feldspar porphyry phase shows at least two distinct zircon age populations (Fig. 3). Inheritance is not common in these zircon grains, and apart from the granite of Deadman Pass, no sample contains zircon dated at millions or tens of millions of years older than the main age population. Although we targeted zircon grains that did not show obvious xenocrystic cores based on the SEM-CL imaging (Fig. S1), a few zircons dated in the granite of Deadman Pass have Proterozoic ages (1.1–1.6 Ga) and are thus excluded from the data.

A subset of the dated zircon grains was also analyzed for trace elements by TIMS-TEA. The advantage of TIMS-TEA over in-situ techniques is that the trace element data are measured on the same sample volume dated by CA-ID-TIMS, allowing us to track chemical trends through time (Schoene et al., 2010, 2012; Wotzlaw et al., 2013; Samperton et al., 2015). Zircon trace element data are presented in Figures 4–6 as scatter plots of trace element ratios, and the full dataset is provided in Table S1.

U-Pb zircon ages are plotted against different trace element ratios that track magma fractionation (Figs. 4–6). For example, zircon Th/U and Zr/Hf ratios increase and decrease, respectively, with progressive zircon fractionation in an evolving melt and can be used to help differentiate magma sources. This is also true for Lu/Hf ratios, although the heavy REEs (e.g., Lu) are compatible in clinopyroxene, amphibole, and garnet, so if any of these phases crystallized prior to zircon saturation, then Lu would decouple from Hf (Dickinson and Hess, 1982; Watson and Harrison, 1983; Boehnke et al., 2013; Bea et al., 2018). Other useful chemical discriminators include the Eu anomaly (Eu/Eu*), which is sensitive to plagioclase fractionation and oxidation state (Dilles et al., 2015), and Yb/Dy ratios, which are sensitive to amphibole and/or titanite fractionation.

Zircon Th/U ratios are also useful for fingerprinting the magma composition and physical conditions of crystal growth (e.g., temperature). If zircon is in equilibrium with its parent melt, higher-temperature melts are characterized by decreasing zircon Th/U ratios with increasing temperature while fractionated melts are characterized by decreasing Th/U ratios with decreasing temperatures (Kirkland et al., 2015). A compilation of felsic magma zircon Th/U ratios from across the western United States shows that the average Th/U ratio of an igneous zircon is 0.5, though magmatic zircon from granitic to dioritic compositions can have a mean Th/U ratio of 0.75–1.03 (e.g., Bindeman et al., 2006; Kirkland et al., 2015). All zircons in this study exhibit Th/U ratios >0.5. Several of the pluton samples are highly enriched in Th, with zircon Th/U ratios >1.5 in the southern phase of the Little Chief Granite and diorite of Furnace Site (Fig. 4).

The most noticeable temporal zircon trace element trends occur in the Willow Spring Diorite between 11.26 Ma and 10.08 Ma, where Zr/Hf and Eu/Eu* ratios steadily increase and (Yb/Dy)_N ratios, where the subscript N denotes values normalized to chondrite (following Sun and McDonough, 1989), decrease (Fig. 5). In general, the lower Willow Spring Diorite zircons (pink circles in Figs. 4–6) have lower Th/U, Zr/Hf, Lu/Hf, and Eu/Eu* ratios with moderately higher (Yb/Dy)_N ratios compared to the middle Willow Spring Diorite zircons (purple circles). The northern diorite zircons (diorite of Furnace Site, dark blue circles) are distinct from the Willow Spring Diorite with higher Th/U ratios and show some of the largest ranges in trace element ratio values (e.g., Th/U, Zr/Hf, Lu/Hf, and Eu/Eu*; see Figs. 4–6). The Smith Mountain Granite (light and dark green triangles) and quartz monzonite of Gold Valley (yellow squares) zircons are chemically similar to the lower and middle Willow Spring Diorite zircons. The granite of Deadman Pass zircons (red squares) also exhibit low Th/U ratios but high Zr/Hf ratios with a narrow range of Lu/Hf and Yb/Dy ratios.

During zircon crystallization, Hf is preferentially removed from the melt and therefore serves as a useful proxy for magmatic differentiation (Linnen and Keppler, 2002). Zircon trace element ratios are plotted against Hf content in Figure 6 and are separated based on rock composition, with dioritic intrusions in the left column and felsic intrusions in the right column. With exception of granite of Deadman Pass, all other plutons show decreasing (Yb/Dy)_N ratios with decreasing Hf content, indicating co-crystallization of amphibole and zircon. A positive trend in Eu/Eu* ratios with decreasing Hf content is observed in the Smith Mountain Granite and quartz monzonite of Gold Valley zircons, which records decreasing plagioclase fractionation during zircon crystallization and magmatic differentiation. In the dioritic intrusions, high-Hf (>9500 ppm) zircons have larger Eu/Eu* anomalies while low-Hf (<9500 ppm) zircons have smaller Eu/Eu* anomalies. This could indicate either that plagioclase fractionation was more prevalent in the early and more primitive melts or that there was a change in redox conditions to more oxidized magmas, which would suppress the magnitude of Eu anomalies (Tang et al., 2018). The zircons in the felsic intrusions show a clear plagioclase fractionation trend with magmatic differentiation, which is not visible in the zircons from mafic intrusions, indicating that a change to more oxidizing conditions is a more likely explanation.

Our high-precision U-Pb zircon dataset represents the first modern single-grain CA-ID-TIMS U-Pb zircon dataset for late Cenozoic intrusions in the Death Valley region, offering crucial insights into the region's plutonic history. This dataset allows us to divide intrusive magmatism into three distinct phases: (P1) early rapakivi granite magmatism, represented by the Kingston Range Granite, granite of Rabbit Holes Spring, and Little Chief Granite; (P2) the construction of the Black Mountains intrusive complex, which includes the Willow Spring Diorite, Smith Mountain Granite, and the quartz monzonite of Gold Valley; and (P3) late-stage plutonism in the Black Mountains and Greenwater Range, represented by the diorite of Furnace Site and the granite of Deadman Pass.

The first phase (P1) of Miocene plutonism in the Death Valley region began with the construction of the Kingston Range Granite at 13.166 ^+0.136/_−0.028 Ma, based on our measured estimate of the earliest preserved evidence of zircon crystallization (calculations here and throughout) use the methods of Keller et al., 2018). Zircon ages indicate that the feldspar porphyry phase was constructed slowly over 742 k.y. in two distinct pulses (Fig. 3). In contrast, the quartz porphyry phase of the Kingston Range Granite crystallized more rapidly within 281 k.y. (Table 4). This rapid emplacement aligns with field observations indicating that the quartz porphyry phases was intruded as a sill-like body beneath the feldspar porphyry phase (Calzia and Rämö, 2005). Following emplacement, the Kingston Range Granite cooled rapidly, as shown by overlapping zircon U-Pb ages with published ⁴⁰Ar/³⁹Ar hornblende ages (12.36 ± 0.18 Ma; Calzia, 1990). Rapid cooling supports earlier interpretations that the granite was emplaced in the shallow crust (<4 km depth), supported by the widespread presence of miarolitic cavities (Davis et al., 1993; Calzia and Rämö, 2005). Additionally, crosscutting relationships between the Kingston Range–Halloran Hills detachment and the Kingston Range Granite further suggest that this pluton was emplaced within an active detachment fault system (Davis et al., 1993).

The Kingston Range Granite exhibits multiple age populations and abundant disequilibrium textures, such as rapakivi feldspars and mafic xenoliths, indicating the role of mantle-derived melts and magma mixing processes in the generation of Kingston Range Granite magmas (Hibbard, 1981; Cantagrel et al., 1984; Calzia and Rämö, 2005). The zircon Th/U ratios between the two porphyry phases overlap, suggesting that they crystallized from the same or similar melt compositions.

The granite of Rabbit Holes Spring, located ~1 km west of the Kingston Range Granite across alluvial valley fill, began crystallizing at the end of Kingston Range Granite magmatism. The construction of the granite of Rabbit Holes Spring occurred between 12.148 ^+0.091/_−0.038 Ma and 11.830 ^+0.055/_−0.042 Ma (Table 4). Unlike the Kingston Range Granite, granite of Rabbit Holes Spring zircon crystallized as a single, continuous pulse over a period of 320 k.y. Zircon Th/U ratios range from 0.75 to 1.98 and do not show a trend with U-Pb zircon ages (Fig. 4). This suggests either that the granite of Rabbit Holes Spring zircon did not crystallize from a single fractionating melt body or that fractionation occurred on time scales shorter than the precision achievable by CA-ID-TIMS U-Pb ages (e.g., Kirkland et al., 2015).

This was followed by the emplacement of the Little Chief Granite, which began at 11.557 ^+0.093/_−0.070 Ma. This intrusion was constructed relatively rapidly, in ~291 k.y., starting with the southern intrusive phase followed by northern intrusive phase (Table 4). Although both phases exhibit overlapping zircon crystallization ages, field evidence indicates the northern phase intruded into the southern phase (McDowell, 1974). The southern phase is characterized by high zircon Th/U ratios (0.8–2.5; Fig. 4), best explained by the assimilation of enriched continental crust, such as dolomite. McDowell (1974) documented field, petrographic, and geochemical evidence of dolomite assimilation in the Little Chief Granite. The southern phase also contains abundant monazite, commonly coarse enough to be visible in hand samples. It is likely that zircon crystallization predates monazite saturation given that coexisting zircon with monazite would yield low zircon Th/U ratios, similar to metamorphic zircon in monazite-bearing rocks (Vavra et al., 1999; Rubatto et al., 2001; Cesare et al., 2003; Rubatto, 2017).

By comparing emplacement ages with published zircon fission-track (ZFT) ages, we can assess how rapidly the Little Chief Granite was exhumed after it was emplaced. We interpret the youngest estimated time of zircon crystallization (see Fig. 7 and Table 4) as the age at which the granite cooled through the solidus, making it equivalent to the emplacement age. ZFT ages for the northern and southern phases are reported as 9.8 ± 1.0 Ma and 11.2 ± 1.0 Ma, respectively (Topping, 1993). These datasets suggest that the pluton was rapidly cooled following emplacement (McDowell, 1974; Topping, 1993). Groundmass mineral assemblage barometry suggests that crystallization occurred at an estimated depth of ~4 km while emplacement depths are estimated to be <2.8 km (McDowell, 1978).

Phase I intrusive magmatism (P1) was characterized by emplacement of shallow (<4 km) rapakivi granites (Fig. 7; McDowell, 1974; Calzia and Rämö, 2005). Each granite is composed of several mappable intrusive facies with distinct mineralogy, zircon crystallization histories, and chemistry (this study; McDowell, 1974; Calzia and Rämö, 2005).

Phase II intrusive magmatism (P2) is characterized by the synchronous crystallization of multiple mid-crustal plutons in the Black Mountains (Fig. 7). Collectively referred to as the Black Mountains intrusive complex, P2 includes the Willow Spring Diorite, Smith Mountain Granite, and quartz monzonite of Gold Valley. All three intrusions exhibit overlapping zircon crystallization histories (Fig. 3) and display comagmatic relationships, such as magma mingling textures and mutually intrusive contacts (Asmerom et al., 1990; Meurer, 1992; Holm et al., 1994b, 1994a).

The Willow Spring Diorite is the most voluminous late Cenozoic pluton in the Death Valley region. Zircon crystallization began by 11.260 ^+0.115/_−0.024 Ma and continued for at least 1.17 m.y. (Table 4). The oldest and most mafic unit sampled, the lower Willow Spring Diorite, crystallized over 786 k.y. In contrast, the middle Willow Spring Diorite crystallized more rapidly, within 250 k.y. The youngest unit, the upper Willow Spring Diorite, began zircon crystallization at or before 10.086 ^+0.046/_−0.015 Ma. This sample shows a clear single resolvable zircon age population with a mean square of weighted deviates (MSWD) of 0.6 (Table 4). Our U-Pb ages are younger than the previously published age of 11.6 Ma (Asmerom et al., 1990). One possible explanation for this discrepancy is that their multigrain age may include zircon with Mesozoic or Proterozoic inheritance.

Each Willow Spring Diorite sample represents a distinct intrusive phase, as supported by textural evidence, whole-rock chemistry (Figs. S3–S5; Meurer, 1992; Gilmer and Thompson, 2024), and zircon trace element chemistry (Figs. 4–6). This interpretation is consistent with earlier petrologic studies that documented multiple distinct phases of the Willow Spring Diorite (Otton, 1977; Meurer, 1992). The lower and middle Willow Spring Diorite zircon samples form linear trends in trace element ratios consistent with amphibole (clinopyroxene + biotite + accessory phases) and plagioclase fractionation (Figs. 5–6). This suggests that the lower and middle Willow Spring Diorite formed from in situ differentiation of the same magma source. Establishing the relationship between the upper Willow Spring Diorite and other units is more challenging due to the limited TIMS-TEA analyses available for the upper Willow Spring Diorite.

Willow Spring Diorite samples were collected from different structural levels of the east-tilted footwall of the Black Mountains (inset in Fig. 7). By comparing the structural level of each Willow Spring Diorite sample with its emplacement age, or zircon solidus age, we can elucidate the emplacement history of the Willow Spring Diorite. Despite our small sample size, we observe a trend of younger emplacement ages at structurally higher levels. The sample we collected from the lowest structural level has the oldest emplacement age. Emplacement ages become progressively younger at structurally higher levels, suggesting that phases of the Willow Spring Diorite represent a sequence of vertically stacked sills. The mylonitized lowermost part of the Willow Spring Diorite was not sampled.

The felsic plutons of the Black Mountains intrusive complex, the Smith Mountain Granite and quartz monzonite of Gold Valley, began crystallizing zircon at 10.81 ^+0.20/_−0.21 Ma and 10.82 ^+0.16/_−0.10 Ma, respectively, coeval with the middle and upper Willow Spring Diorite. The Smith Mountain Granite crystallized zircon for 580 k.y. while the quartz monzonite of Gold Valley crystallized zircon for 640 k.y., indicating that the felsic plutons formed over similar time scales. Previous U-Pb zircon datasets suggested that Willow Spring Diorite was significantly older than the Smith Mountain Granite (Asmerom et al., 1990; Miller et al., 2004). However, our new U-Pb dataset demonstrates that all intrusions of the Black Mountains intrusive complex crystallized simultaneously, with compositionally distinct magmas crystallizing at the same time. There is field evidence of magma intermingling between the Willow Spring Diorite and quartz monzonite of Gold Valley (see Supplemental Material; Meurer, 1992).

Zircon trace element ratios and fractionation trends in the Smith Mountain Granite and quartz monzonite of Gold Valley overlap with each other and with those of several Willow Spring Diorite samples (Figs. 4–6). These trends are consistent with a comagmatic relationship, which can be further investigated by examining published whole-rock data (Figs. S3–S5; Gilmer and Thompson, 2024). If the Black Mountains intrusive complex is part of a single magmatic system, then the felsic plutons would be fractionated melts, while the Willow Spring Diorite would represent the cumulate material. Chemical tracers for plagioclase (e.g., Eu/Eu*) and amphibole (e.g., La/Yb and Dy/Yb) crystallization do not show linear trends between the dioritic and felsic intrusions, indicating that the latter cannot be derived solely from the differentiation of Willow Spring Diorite magmas (Fig. S5). This conclusion aligns with Meurer's (1992) findings and is further supported by the absence of cumulate textures in the Willow Spring Diorite. However, this does not negate the possibility that the Smith Mountain Granite and quartz monzonite of Gold Valley are petrogenetically related, given that the latter may represent a northern continuation of the Smith Mountain Granite, a hypothesis originally proposed by Holm et al. (1992) now supported by our geochronologic and new geochemical data.

The Black Mountains intrusive complex was emplaced into the middle crust. Crystallization pressures for the Willow Spring Diorite, determined using Al-in-hornblende barometry on a leucocratic phase, range from 260 MPa to 280 MPa. This indicates a minimum emplacement depth of ~10 km, based on a bulk crustal density of 2700 kg/m³ (Meurer, 1992; Holm et al., 1992). Emplacement depths increase to the northwest, suggesting exhumation along an initially steeply dipping fault (Meurer, 1992; Miller and Pavlis, 2005). Al-in-hornblende geobarometry from the Smith Mountain Granite also indicates crystallization at a depth of at least 10 km (Holm et al., 1992). Recent garnet thermobarometry of the turtlebacks indicates temperatures of 660–685 °C and pressures of 500–600 MPa during the Miocene, suggesting that the Willow Spring Diorite was emplaced at even greater depths, closer to 15–20 km (Pavlis et al., 2018; Fleming et al., 2022). Although there are no published geobarometry data for the quartz monzonite of Gold Valley, overlapping zircon crystallization ages (Fig. 3) and mutually intrusive contacts with both the Willow Spring Diorite and Smith Mountain Granite imply that the quartz monzonite crystallized at a similar structural level.

A younger phase of intrusive magmatism, phase III (P3), commenced with the emplacement of the granite of Deadman Pass in the Greenwater Range, where zircon crystallization initiated at 8.176 ^+0.111/_−0.019 Ma and continued for ~309 k.y. (Fig. 7; Table 4). While this granite is chemically similar to other Greenwater Range intrusions (Figs. S3–S5; Calzia and Rämö, 2005; Gilmer and Thompson, 2024), isotopic differences suggest distinct magma sources for the granite of Deadman Pass (Calzia et al., 2016). Additionally, zircon chemistry indicates that the granite of Deadman Pass crystallized from a more evolved magma than other felsic plutons in the Black Mountains intrusive complex (Figs. 4–6).

Concurrently, zircon crystallization began at 7.902 ^+0.057/_−0.015 Ma in the diorite of Furnace Site (Fig. 3; Table 4). Although previously mapped as the Willow Spring Diorite (Otton, 1977; Holm and Wernicke, 1990), the diorite of Furnace Site has younger zircon crystallization ages and higher zircon Th/U ratios (>1.5), indicating it is distinct from the Willow Spring Diorite magmatic system. The diorite of Furnace Site zircon grains are compositionally diverse (e.g., Zr/Hf, Yb/Dy, and Eu/Eu* ratios and Hf content; Figs. 5–6), which means either that these zircons crystallized in discrete melt bodies that were mixed prior to emplacement or that magma differentiation occurred on time scales significantly shorter than those of zircon growth.

The volcanic and plutonic record of Death Valley over the past 16 m.y. offers crucial insights into the development of this tectonically complex region. In this section, we integrate new geochronology from Miocene plutonic rocks, existing volcanic records, and documented tectonic events to examine the spatiotemporal relationship between extension and magmatism. Our compilation in Figure 8 defines six different volcanic episodes (V1–V6; Calzia and Rämö, 2000; Fridrich and Thompson, 2011), three phases of intrusive magmatism (P1–P3; Fig. 7), and four major tectonic stages (T1–T4; Wernicke et al., 1988; Wright et al., 1991; Wernicke, 1992; Holm and Dokka, 1993; Davis et al., 1993; Topping, 1993; Miller and Friedman, 1999; McQuarrie and Wernicke, 2005; Andrew and Walker, 2009; Fridrich and Thompson, 2011; Norton, 2011; Bidgoli et al., 2015; Sizemore et al., 2019; Fleming et al., 2021).

Detachment faulting in the Death Valley region initiated ca. 16–15 Ma, characterized by large-magnitude WSW-directed extension along the Kingston Range–Halloran Hills–Spring Mountains detachment (T1; Burchfiel et al., 1983; Davis et al., 1993; Fowler et al., 1995; Friedmann et al., 1996; Andrew and Walker, 2009). This period also saw WNW-directed extension along several other detachment faults in the Bare Mountain, Bullfrog Hills, and Funeral Mountains (e.g., Hamilton, 1988; Hodges et al., 1989; Hoisch et al., 1997; Fridrich et al., 1999; Fridrich and Thompson, 2011; Lutz et al., 2021). Slip along the Kingston Range–Halloran Hills detachment is bracketed between 13.4 Ma, based on a sill intruded during faulting (K-Ar hornblende; Davis et al., 1993), and 12.4 Ma, which is the published age for the Kingston Range Granite (Calzia, 1990; Fowler and Calzia, 1999). However, new U-Pb zircon ages for the Kingston Range Granite provide a revised minimum of 13.2 Ma for the timing of extension along this detachment.

Magmatism in the region (V1) initiated at ca. 15 Ma with the eruption of several calc-alkaline stratovolcano complexes and associated small intrusive domes in the Wingate Wash–Owlshead Mountains (Fig. 8; e.g., Burchfiel and Davis, 1988; Wagner, 1988; Calzia, 1990; McKenna and Hodges, 1990; Calzia and Rämö, 2000; Niemi et al., 2001; Luckow et al., 2005). Waning volcanism coincided with emplacement of shallow rapakivi granites (P1), which includes the Kingston Range Granite, granite of Rabbit Holes Spring, Little Chief Granite (McDowell, 1974; Calzia, 1990), and the 13.2 ± 0.2 Ma Devil Peak intrusion (Table 2; Guest et al., 2007). Volcanism across the Death Valley region ceased between 12 Ma and 11 Ma (Luckow et al., 2005; Fridrich and Thompson, 2011), with only a few distal tuffs at this time.

The second phase of plutonic activity (P2) is associated with the Black Mountains intrusive complex at ca. 11.3 Ma. This phase lasted for at least 1.2 m.y. and is associated with WNW-directed extension (T2). Although a full tectonic reconstruction is beyond the scope of this project, it is clear that plutonic activity shifted westward, potentially with a minor northward component (e.g., Stewart, 1983; Wernicke et al., 1987; Topping, 1993; Serpa and Pavlis, 1996; Snow and Lux, 1999; Andrew, 2002, 2019; Miller and Prave, 2002; Miller and Pavlis, 2005; Fleming et al., 2021), between ca. 13.2 Ma and 10 Ma (Fig. 7). During this period, extension is accommodated along a series of conjugate strike-slip faults, including NW-striking dextral faults such as the Southern Death Valley fault (e.g., Pavlis and Trullenque, 2021) and the sinistral SW-striking Garlock fault system (Fig. 2; e.g., Serpa and Pavlis, 1996; Andrew et al., 2015).

Major volcanism resumed by 11 Ma (V2) with bimodal compositions but transitioned to felsic-dominated compositions by 10 Ma with eruption of the ca. 11 Ma Sheephead Andesite, ca. 10.5 Ma Rhodes Tuff, and 10.3 Ma Trail Canyon volcanics (Fig. 8; Troxel and Wright, 1987; Wright et al., 1991; McKenna and Hodges, 1990; Calzia and Rämö, 2000; Fridrich and Thompson, 2011). Felsic-dominated volcanism coincided with peak plutonic activity, marked by the emplacement of the Black Mountains intrusive complex (P2). Given the proximity and overlapping ages of the felsic intrusions within the Black Mountains intrusive complex and nearby volcanic units, such as the 10.5 Ma Rhodes Tuff (Fridrich and Thompson, 2011), it is plausible that these felsic intrusions are the plutonic underpinning to some of these volcanic units.

There is a noticeable gap in zircon crystallization ages between 10 Ma and 8.2 Ma (Fig. 3), which may be attributed to several factors. One possibility is sampling bias, given that no rock samples were collected from a >90 km² area in the Black Mountains and large swaths of the Greenwater Range (Fig. 2), potentially leaving intrusions from this period unidentified or undated. Alternatively, younger sedimentary or volcanic rocks may have covered these intrusions, or these intrusions could have been eroded or removed tectonically. This gap might also imply a magmatic lull; however, the eruption of the 8.5–7.5 Ma Shoshone Volcanics in the Greenwater Range (V3; Wright et al., 1991) contradicts this idea.

The most likely explanation for the observed gap is a transition from sill to dike emplacement during this period. Several generations of latite and silicic dikes cut the turtleback shear zone and the intrusive complex, with some exhibiting ductile deformation (Drewes, 1963; Otton, 1977; Wright et al., 1991; Holm et al., 1992; Miller and Friedman, 1999; Dee et al., 2004). Limited geochronology indicates that diking occurred ca. 9.5–7.9 Ma (Holm et al., 1992; Miller and Friedman, 2003). This transition coincides with the solidification of the intrusive complex by 10 Ma (Fig. 3), followed by higher rates of exhumation and cooling (Holm et al., 1992; Holm and Dokka, 1993), reaching ~180 °C by 7.8 Ma (Topping, 1993).

Pluton emplacement resumed by ca. 8.2 Ma (P3), marked primarily by shallow granite intrusions in the Greenwater Range and a diorite intrusion in the northern Black Mountains. Intrusive magmatism appears to have shifted northward within the Black Mountains and Greenwater Range (Fig. 7). The northward age progression in the Greenwater Range is supported by older zircon ages in the Shoshone Granite (ca. 10 Ma laser ablation ICP-MS ²⁰⁶Pb/²³⁸U; Rämö et al., 2011), which is located south of the granite of Deadman Pass. This reinitiation of sill-dominated emplacement coincides with the start of rapid extension (Holm et al., 1992; Holm and Dokka, 1993) and NW-directed dextral transtension (Bidgoli et al., 2015; Sizemore et al., 2019; Fleming et al., 2021). Similar to the Black Mountains, exhumation of the Greenwater Range was rapid enough to allow for erosion of the exhumed footwall prior to burial by the ca. 8.5 Ma Shoshone Volcanics (Wright et al., 1991).

Differences in the emplacement depths between the Willow Spring Diorite and the youngest granites in the region reveal significant tectonic changes over a short period of time. The 11.3 Ma Willow Spring Diorite and 10.8 Ma Smith Mountain Granite are the deepest intrusions, with emplacement depths exceeding 10 km and reaching as deep as 20 km (Pavlis et al., 2018; Fleming et al., 2022). The youngest granite, the 7.8 Ma granite of Deadman Pass, has hypabyssal textures (Haefner, 1972) and is interpreted to have been emplaced shallowly. This implies that there was a profound change in the emplacement depth in a relatively small region of footwall rocks between 10 Ma and 8.2 Ma, which requires rapid tectonic exhumation during this time interval (McKenna and Hodges, 1990; Holm and Dokka, 1993; Topping, 1993; Bidgoli et al., 2015; Sizemore et al., 2019).

Large-magnitude dextral transtension (T3) commenced ca. 8–7 Ma, leading to a period of rapid extension that continued until ca. 4.5 Ma (Topping, 1993; Bidgoli et al., 2015; Sizemore et al., 2019). Transtension may have initiated as early as 10 Ma (e.g., Renik and Christie-Blick, 2013). However, dextral slip observed on NW-striking faults between 10 Ma and 8 Ma could also indicate continued west-directed extension within a conjugate strike-slip system. This system includes the dextral NW-striking Las Vegas Valley shear zone, located east of the Spring Mountains (Fig. 2; Çakir et al., 1998) and the sinistral SW-striking Garlock fault (Andrew et al., 2015), whose combined initial geometries and sense of shear suggests continued east-west extension. Dextral transtension contributed to the exhumation of the Black Mountains through the formation of pull-apart basins located between the Furnace Creek and Southern Death Valley–Sheephead fault zones to the north and south, respectively (Fig. 2; e.g., Troxel and Wright, 1987; Serpa and Pavlis, 1996; Miller and Prave, 2002).

No plutonic activity has been documented in T3; however, basaltic volcanism (V4) continued in the Greenwater Range with the eruption of the ca. 7–5 Ma Greenwater Volcanics (Wright et al., 1991; Calzia et al., 2016). This was followed by ~1 m.y. period of flood basalts (V5), including the ca. 5 Ma Funeral Creek basalts and ca. 6-4 Ma Nova Formation basalts in the northern Panamint Mountains and Darwin Plateau (e.g., McAllister, 1970, 1973; Troxel and Wright, 1987; Hodges et al., 1989; Coleman 788 and Walker, 1990; Snyder and Hodges, 2000). Small-volume basaltic eruptions (V6) continue into the present with the eruption of the 1.5 Ma Shoreline Butte basalts and <1 Ma cinder cones (de Voogd et al., 1986; Wright et al., 1991; Fridrich and Thompson, 2011) and minor rhyolitic eruptions of the 3.6 Ma to ca. 32 ka Mesquite Spring tuff (e.g., Snow and White, 1990; Knott et al., 1999, 2018; Fridrich and Thompson, 2011).

Although no mid-crustal plutons have been identified after P3, young basaltic eruptions and Consortium for Continental Reflection Profiling (COCORP) deep seismic reflection profiles from central Death Valley underscore the existence of a mid-crustal magmatic system below modern-day southern Death Valley (de Voogd et al., 1986; Serpa et al., 1988). This localization of ongoing magmatism may relate to ongoing extensional deformation in this region, which is supported by Holocene slip along the Amargosa–Black Mountains detachment fault (Knott et al., 1999, p. 199; Hayman et al., 2003). However, the main locus of extension shifted west of the study area to the Panamint Valley and Owens Vally fault systems by ca. 4.5 Ma (Knott et al., 1999, 2005; Hayman et al., 2003; Stockli et al., 2003; Andrew and Walker, 2009; Frankel et al., 2016).

Rapakivi textures, characterized by potassium feldspar megacrysts rimmed by plagioclase feldspar, are prevalent in all the Miocene felsic plutons in the Death Valley region (this study; McDowell, 1978; Meurer, 1992; Calzia and Rämö, 2005; Calzia et al., 2016). Rapakivi granites typically crystallize from hot, fractionated, magmas (650–950 °C; e.g., Rämö and Haapala, 1995). The atypical progression from alkali feldspar to plagioclase crystallization can be attributed to two main mechanisms (Vernon, 2016): (1) mixing of mafic and felsic magmas (e.g., Hibbard, 1981; Bussy, 1990; Wark and Stimac, 1992; Andersson and Eklund, 1994) and (2) adiabatic decompression (e.g., Nekvasil, 1991; Eklund and Shebanov, 1999). These processes are likely interconnected, given that hot mantle input can sustain high temperatures during adiabatic decompression. This section examines how these mechanisms contribute to the prevalence of rapakivi granites in Death Valley.

Magma mixing can lead to two outcomes: (1) hot mafic magma may partially dissolve potassium feldspar megacrysts and crystallize new plagioclase rims or (2) the injection of new felsic magma can form plagioclase rims around the potassium feldspar (e.g., Hibbard, 1981; Bussy, 1990; Wark and Stimac, 1992; Andersson and Eklund, 1994). This process may occur in both deep and shallow plutonic rocks in Death Valley. In the Black Mountains intrusive complex, magma mingling textures and isotopic heterogeneities on a tens-of-meters scale are indicative of magma mixing (Asmerom et al., 1990; Meurer, 1992; Holm et al., 1994b; Miller et al., 2004). Additionally, shallow rapakivi granites (P1 and P3), containing mafic xenoliths and inclusions, show geochemical and isotopic signatures that suggest partial melting of the lower crust with juvenile mantle contributions (Calzia and Rämö, 2005).

During adiabatic decompression, potassium feldspar megacrysts crystallize in felsic magmas at mid-crustal depths and are transported in a crystal mush to shallower levels (e.g., Annen et al., 2015; Sparks et al., 2019). Decompression alters the solidus and shifts the two-feldspar stability boundary, favoring plagioclase crystallization (e.g., Nekvasil, 1991; Eklund and Shebanov, 1999). This model aligns well with the shallow (<4 km) Miocene plutons in the Kingston, Greenwater, and Panamint Ranges (Haefner, 1972; McDowell, 1974; Calzia, 1990; Calzia and Rämö, 2005). In contrast, felsic plutons in the Black Mountains were emplaced in the middle crust (>10 km; Holm et al., 1992; Pavlis et al., 2018; Fleming et al., 2022), so for the decompression rapakivi texture mechanism to apply, these plutons would have required rapid tectonic exhumation during late-stage crystallization.

Rapakivi textures are relatively rare in Tertiary granitoids, primarily documented in Archean and Proterozoic granites (e.g., Rämö and Haapala, 1995). Aside from the rapakivi granitoids in Death Valley, there are only two other documented Cenozoic intrusions with these textures—the 48.1 Ma Golden Horn batholith in Washington (Stull, 1978; Eddy et al., 2016) and the 17.4–15.3 Ma Spirit Mountain batholith of the Colorado River extensional corridor in Nevada (Walker et al., 2007). Although magma mixing is a common magmatic process in the formation of felsic plutons across various geologic settings, rapakivi textures are rare. The distinctiveness of the young rapakivi granites in Death Valley and the Spirit Mountain batholith suggests that specific factors, such as localized large-magnitude extension, may have played a significant role in their formation.

We argue that rapid extension of an overthickened crust (≥30 km; e.g., Holm and Wernicke, 1990; Pavlis et al., 2018) is an important factor for generating rapakivi textures. This process likely led to adiabatic decompression, magmatic underplating, and anatexis of the lower crust (Fountain, 1989; Calzia and Rämö, 2005), driving the formation of these textures. Notably, there are striking similarities between the extensional tectonics of Death Valley and that of Proterozoic rapakivi granites in Finland and Greenland (e.g., Haapala and Rämö, 1990, 1992; Brown et al., 1992, 2003; Rämö and Haapala, 1995; Ahl et al., 1999; Haapala et al., 2005), suggesting a link between extensional tectonics and rapakivi textures. While we cannot rule out the magma mixing model, the decompression model fits well within the tectonic framework of a highly extended region (Fig. 8). Further petrologic and geochemical studies are necessary to validate this hypothesis.

Late Cenozoic magmatism in Death Valley is a continuation of the northward-migrating magmatic trend from the Colorado River extensional corridor (Fig. 1; Christiansen et al., 1992; Faulds et al., 2001). These two regions are separated by the ~65 km WSW-striking left-lateral Lake Mead fault zone, which became active after ca. 12.7 Ma. In the Colorado River extensional corridor, the principal extension direction is ENE, whereas in Death Valley, WSW-directed extension is observed along the Kingston Range–Halloran Hills detachment. Subsequently, the principal extension direction shifts to WNW, which is documented in the Amargosa-Black Mountains, Funeral Mountains, and Bare Mountain detachment faults. Changes in the extension polarity along strike are common in extensional rifts (e.g., Lister et al., 1986; Péron-Pinvidic et al., 2017; Balázs et al., 2018).

Similar to those in Death Valley, the plutons of the Colorado River extensional corridor have a narrower age range (17–13 Ma) compared to the volcanic rocks (Faulds et al., 2001). Most of plutons in the extensional corridor were emplaced within 200–500 k.y. of local extension and transitioned from sill emplacement to north-south dike swarms within <400 k.y. after pluton solidification in the northern region (e.g., Spirit Mountain batholith; Miller et al., 2011). Extension typically occurred ~700 k.y. after the northward-migrating pulse of magmatism reached a given area (Gans and Bohrson, 1998).

The northward migration of the Arizona magmatic trend is tectonically linked with the migration of the Mendocino triple junction (e.g., Crough and Thompson, 1977; Best and Hamblin, 1978; Dickinson and Snyder, 1979; Atwater and Stock, 1998; Zandt and Humphreys, 2008). This contrasts with the southward-migrating Nevada magmatic trend, which is associated with slab rollback (e.g., Coney, 1978; Lipman, 1980; Best and Christiansen, 1991). As the northward-migrating Mendocino transform boundary—interpreted as the driver of the northward Arizona magmatic trend—passed across the Death Valley region, slab rollback would have ceased. This aligns with the notable similarities observed between the Death Valley and Colorado River extensional corridors. This shift in tectonic activity highlights the complex interplay between magmatism and tectonic forces in the Death Valley region.

U-Pb zircon geochronology identifies three distinct pulses of Late Cenozoic plutonism in the Death Valley region, beginning with the emplacement of shallow rapakivi granites at ca. 13.2 Ma and lasting for 1.85 m.y. Plutonic activity peaked at 10.5 Ma with the simultaneous crystallization of three mid-crustal plutons associated with the Black Mountains intrusive complex—Willow Spring Diorite, Smith Mountain Granite, and quartz monzonite of Gold Valley. Whole-rock geochemical tracers (e.g., Eu/Eu*, La/Yb, Dy/Yb) indicate that the Smith Mountain Granite and quartz monzonite of Gold Valley are comagmatic but not derived from differentiation of the Willow Spring Diorite. Emplacement ages indicate that the Willow Spring Diorite formed as a sequence of incrementally emplaced, vertically stacked sills.

Zircon crystallization ages reveal continuous intrusive magmatism between 13.2 Ma and 10 Ma, followed by a gap in zircon crystallization that coincides with rapid exhumation of the Black Mountains and a transition from sill to dike emplacement. Late-stage plutonism at 8.2 Ma featured shallow, compositionally diverse intrusions, including the diorite of Furnace Site, which is distinct from the Willow Spring Diorite in U-Pb ages, texture, and zircon trace element ratios (e.g., Th/U, Lu/Hf). Most Death Valley plutons formed incrementally over time scales of 10⁵–10⁶ yr, although a few were constructed instantaneously within the age resolution of CA-ID-TIMS dating. A significant change in emplacement depth between the Willow Spring Diorite and the granite of Deadman Pass reflects rapid tectonic exhumation between 10 Ma and 8.2 Ma, supported by overlapping or closely aligned zircon crystallization ages and published cooling ages.

Late Cenozoic plutonism fits within the broader volcanic history of the region. The spatial and temporal distribution of plutonism and tectonic events from ca. 15 Ma to 8 Ma indicates that plutonism occurred synchronously with major extensional events. The important link between extension and plutonism is further emphasized by the prevalence of rapakivi textures in Miocene felsic plutons of Death Valley, an uncommon feature in Cenozoic plutons globally. These textures are primarily attributed of adiabatic decompression during rapid, large-magnitude tectonic extension. Rapid extension of an overthickened crust (>30 km) is likely a key factor in the generation of rapakivi granitoids in Death Valley.

This project was supported by grants from the Geological Society of America and University of Kansas (Lawrence, Kansas, USA) Department of Geology to C. Chan and by the U.S. Geological Survey's National Cooperative Geologic Mapping Program. We thank the National Park Service for allowing us to collect rock samples. We are very appreciative to Ren Thompson and Amy Gilmer for sharing their Death Valley expertise, providing insightful discussions, giving expert assistance in the field, and sharing a diorite of Furnace Site sample. We also thank James Calzia for the granite of Rabbit Holes Spring samples and Tandis Bidgoli for the Kingston Range Granite feldspar porphyry sample. We would like to acknowledge Nicholas Hayman, an anonymous reviewer, and Associate Editor Terry Pavlis for their thoughtful reviews, and Christopher Spencer for editorial handling. Part of this work was performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory (Livermore, California, USA) under contract DE-AC52-07NA27344, release number LLNL-JRNL-862151.