Abstract

Despite numerous studies on the structural evolution of metamorphic core complexes, there is still little consensus on the set and sequence of processes that bring deep levels of the crust to the surface during extension. This problem is partially related to the fact that core complexes expose polydeformed rocks, the history of which has been challenging to decipher. New geochronological and structural data combined with existing data provide improved insight into the Cenozoic extensional evolution of the Albion–Raft River–Grouse Creek (ARG) metamorphic core complex. The Cenozoic extensional history of the core complex can be divided into several distinct stages based on the geochronology and structure of igneous and metamorphic rocks in the lower plate of the complex combined with the geochronology and regional geologic context of sedimentary and volcanic rocks flanking the complex. Initial volcanism and plutonism was Eocene age (42–34 Ma), related to a regional southward-younging magmatic event. The development of high-temperature (sillimanite grade) metamorphic fabrics and mineral assemblages in footwall rocks was mostly Oligocene (ca. 32–25 Ma), synchronous with the diapiric rise and intrusion of evolved plutons to mid-crustal depths (∼10–15 km), formed by partial melting and remobilization of the deeper crust. There is no evidence for associated volcanism or basin development at the surface during this time span. The metamorphic and plutonic rocks of the core complex apparently remained at depth for ∼10–12 m.y. until the Middle Miocene (ca. 14 Ma), when they were exhumed by Basin and Range faulting.

Detrital zircon studies of continental basin sediments demonstrate that the synextensional Raft River Basin, bounded by the Albion–Raft River fault system, began to develop along the eastern side of the ARG metamorphic core complex ca. 13.5 Ma, synchronous with footwall cooling and uplift between 13.5 and 7 Ma recorded by apatite fission track ages. The evolution of sediment sources in the Raft River Basin help define three phases of Miocene tectonism. (1) Between 13.5 and 10.5 Ma, rapid slip on the Albion fault, which rooted into a ductile-brittle transition zone represented by the Raft River detachment, exhumed Paleozoic strata that, together with Miocene volcanic rocks, sourced the basin. (2) Between 10.5 and 8.2 Ma, continued slip resulted in a topographic depression filled with volcanic rocks and detritus derived from footwall metamorphic and crystalline rocks as well as prior sources. (3) After 8.2 Ma, the sedimentary basin was cut, rotated, and repeated by a set of younger north-south–striking normal faults that extended the basin in an east-west direction, structurally uplifting the basin sediments to erosion. These younger faults die out to the south and minimally displace the fault system that bounds the metamorphic core of the Raft River Mountains.

The Cenozoic evolution proposed for the ARG metamorphic core complex indicates that the formation of Oligocene granite-cored gneiss domes and their high-temperature metamorphic carapace and overlying detachments, are distinctly older (ca. 10 Ma), and thus unrelated to the younger exhumation by high-angle faulting.

INTRODUCTION

Metamorphic core complexes are structural culminations of exhumed metamorphic and igneous rocks that occur in extensional settings. They were first described in the Basin and Range province of the western U. S. (Fig. 1A), where they are typically intruded by synextensional plutons and bound by detachment faults associated with zones of very high ductile extensional strains (e.g., Coney, 1980; Wernicke, 1981; Armstrong, 1982; Miller and Gans, 1989). Regions flanking metamorphic core complexes typically expose supracrustal rocks, including Cenozoic volcanic and sedimentary sequences that record the final exhumation of the complexes (e.g., Eberly and Stanley, 1978; Dickinson, 1991; Davis et al., 2004; Miller and Gans, 1989; Miller et al., 1999; Colgan and Metcalf, 2006; Colgan and Henry, 2009).

Despite numerous studies on the structural evolution of metamorphic core complexes, there is still little consensus on the set and sequence of lower crustal processes that exhume the deep levels of the crust in an extensional tectonic setting, and how these processes are reflected in the surface geology (Fig. 2). The concept of large-offset low-angle normal faults was first used by Wernicke (1981) to explain the northern Snake Range and other core complex detachment faults. Since then, this model has been widely used to explain the geometry of extensional fault systems and core complexes, both in surface exposure and in seismic reflection profiles. Motion on low-angle normal faults is mechanically difficult to explain based on rock physics experiments and the fact that most seismically active normal faults have steep (>30°) dips (e.g., Jackson and White, 1989; Thatcher and Hill, 1991; Collettini and Sibson, 2001; Collettini and Barchi, 2002). The origin of the high-strain ductile extensional fabrics and mylonites associated with metamorphic core complex detachment faults is also a matter of debate. The low-angle normal fault model explains the mylonitic fabrics as forming in extensional shear zones that penetrate the deep crust and bound a relatively rigid footwall block (e.g., Wernicke, 1981; Wernicke and Axen, 1988; Wells et al., 2000; Fig. 2). An alternative explanation of these fabrics is that they form at the top of zones of crustal flow and diapirism, aided by elevated geotherms and the presence of partial melts (e.g., Rehrig and Reynolds, 1980; Miller et al., 1988; MacCready et al., 1997; Foster et al., 2001; Whitney et al., 2004; Teyssier et al., 2005; Rey et al., 2009) (Fig. 2). In these views, high-strain fabrics and mylonites develop between the flowing part of the crust, i.e., the lower plate, and its brittle cover, the upper plate, and represent the ductile-brittle transition zone of the crust.

In the Albion–Raft River–Grouse Creek (ARG) complex, the age of high-strain fabrics and the depth to which they have developed in the crust are debated. Early studies in the ARG metamorphic core complex by Armstrong (1968a), Compton et al. (1977), Miller (1980), and Todd (1980) related the dominant, gently dipping, and high-strain foliation in the lower plate to attenuation and extension of the crust during intrusion of Oligocene plutons. These earlier conclusions are supported by geochronologic studies (Egger et al., 2003; Strickland et al., 2011a, 2011b). Alternatively, studies by Wells et al. (1998, 2012), Harris et al. (2007), Hoisch et al. (2008), and Wells and Hoisch (2008) used metamorphic pressure-temperature estimates in conjunction with geochronology and thermochronology to argue that many of the high-strain fabrics in the Grouse Creek and Albion Mountains are related to alternating periods of contraction and extension in Cretaceous–Eocene time.

Unraveling the exact timing of events in core complexes is important because the history of these deep crustal rocks forms the basis for our understanding of the processes that brought them to the surface and the nature of the transition from regional shortening to extension in the North American Cordillera. In addition, the evolution and kinematic history of core complexes have a critical effect on our ability to balance structural sections and calculate amounts of extension across the Basin and Range (Fig. 2). Specifically, the two end-member models in Figure 2 represent significantly different horizontal strains (β-factor of 1.33 versus 2.0) and also imply substantially different initial crustal thicknesses. The lack of consensus on how metamorphic core complexes form represents a challenge in terms of our ability to quantitatively evaluate preextensional crustal thicknesses and their implied paleoelevations, which are sensitive to calculated amounts of crustal extension (e.g., DeCelles, 2004; Ernst, 2010). We address these questions by compiling existing data and combining it with new data from the ARG metamorphic core complex, where the timing of Cenozoic events is now documented with greater clarity (Figs. 1 and 3). Specifically, we compare the timing of magmatic and deformational events in the lower plate of the ARG metamorphic core complex to the faulting and depositional history of the flanking Raft River Basin (Fig. 3). Using these data, we demonstrate that the Cenozoic extensional history of ARG metamorphic core complex is both multistage and polygenetic.

BACKGROUND

Regional Cenozoic Tectonic and Magmatic Setting of the ARG

The northern Basin and Range province of the western U.S. (Fig. 1A) is a broad, active continental rift, characterized by northward-striking normal fault blocks formed by east-west extension, a lower crust with subhorizontal seismic reflectivity, and a sharp Moho at a uniform depth of 28–32 km (e.g., Klemperer et al., 1986; Hauser et al., 1987; Catchings and Mooney, 1991; Catchings, 1992; Gashawbeza et al., 2008). These present-day characteristics are the result of a protracted history of both magmatism and deformation, the details of which elusive and controversial. Following the end of regional folding and thrust faulting in the latest Cretaceous–earliest Cenozoic (Sevier and Laramide orogenies; e.g., Armstrong, 1982; Burchfiel et al., 1992; DeCelles, 1994, 2004), Cenozoic volcanic rocks erupted across the northern Basin and Range; they describe a prominent north to south younging pattern (e.g., Armstrong and Ward, 1991; Best and Christiansen, 1991; Christiansen and Yeats, 1992; Gans et al., 1989; Figs. 1A, 1B). The onset of magmatism began in southern Canada, northern Idaho, and Montana ca. 55 Ma and migrated to southern Nevada by ca. 20 Ma (Christiansen and Yeats, 1992; Armstrong and Ward, 1991; Best and Christiansen, 1991; Figs. 1A, 1B). Based on geochemistry, these magmas are generally interpreted as derived from mafic parent magmas that were generated by the melting of hydrated mantle. The cause of this magmatism is thought to be related to the progressive delamination of the shallowly dipping subducted Farallon slab, inferred to have underlain the northern Basin and Range province during the Laramide orogeny (ca. 80–60 Ma) (e.g., Dumitru et al., 1991; Burchfiel et al., 1992; Humphreys, 1995; Humphreys et al., 2003,Christiansen and McCurry, 2008). The southward sweep of magmatism left a landscape that had only limited topographic relief and was covered by flatlying volcanic rocks; this is documented by the widespread unconformity developed at the base of volcanic strata across the northern Basin and Range, and by the unimpeded, channelized east-west flow of Eocene–Oligocene ignimbrites in paleovalleys (e.g., Gans et al., 1989; Henry, 2008; Colgan and Henry, 2009; Van Buer et al., 2009; Long, 2012; Fig. 4). A compilation of caldera locations and ignimbrite flow directions suggests that the topographic divide of this low-relief region was in the central part of the northern Basin and Range during the Eocene (Henry, 2008), but the absolute elevation of this plateau is controversial (e.g., DeCelles, 2004; Mulch et al., 2006; Best et al., 2009; Henry, 2008; Ernst, 2010; Henry et al., 2011).

Basin and Range faulting leading to the present-day topography began between 20 and 15 Ma, with rapid slip on faults in the central part of the northern Basin and Range occurring ca. 17–16 Ma (Miller et al., 1999; Stockli, 1999; Colgan and Metcalf, 2006; Henry, 2008; Colgan and Henry, 2009). The onset of Basin and Range faulting is younger toward the borders of the province; rapid slip on faults began ca. 15–10 Ma in both the northwestern and northeastern parts of the province, including the region in and around the ARG metamorphic core complex (e.g., Wells et al., 2000; Egger et al., 2003, 2010; Colgan and Metcalf, 2006; Colgan et al., 2007; Fosdick and Colgan, 2008).

Geology and Geochronology of Footwall Rocks of the ARG

The oldest rocks exposed in the ARG metamorphic core complex are crystalline Archean orthogneiss, schist, and amphibolite (collectively termed the Green Creek Complex), unconformably overlain by quartzites and metapelites that represent the base of the passive margin succession of the Cordillera (Fig. 3; Armstrong, 1968b; Compton et al., 1977; Stewart, 1980). These metasedimentary rocks are Proterozoic in age, on the basis of association of their high δ13C values (Wells et al., 1998) and detrital zircon signatures (Link and Johnston, 2008). Higher stratigraphic units include the Ordovician Pogonip Group and the Pennsylvanian–Permian Oquirrh Group (included in the Phanerozoic unit in Fig. 3). The Cambrian and parts of the Silurian–Devonian sections of the passive margin succession are missing from the ARG and have been inferred to have been removed by normal faulting in the Late Cretaceous–early Cenozoic (Wells et al., 1998).

Regional studies of Mesozoic deformation in the Sevier belt east of the ARG indicate that shortening was partitioned in time between 145 and ca. 55 Ma along four major thrust systems (e.g., Armstrong, 1963; DeCelles, 1994, 2004; Burtner and Nigrini, 1994; Yonkee and Weil, 2011), and that the motion on these systems migrated eastward through time, taking place mostly in the Mesozoic with much less shortening continuing into the Cenozoic (Wiltschko and Dorr 1983; Heller et al., 1986; DeCelles, 1994, 2004; Burtner and Nigrini; 1994; Yonkee and Weil, 2011; Appendix 1). Geochronologic and thermochronologic studies of the deeper parts of the stratigraphic section exposed in the ARG indicate that this part of the hinterland of the Sevier thrust belt was subject to burial and metamorphism in the Mesozoic (e.g., Hoisch et al., 2002, 2008; Harris et al., 2007; Wells et al., 2012). These data are compatible with the fact that the estimated original thickness of the pre-Cenozoic stratified section across this region is ∼9–10 km (e.g., Compton et al., 1977; Hintze, 1988) and that internal thrust faulting may have thickened this section (Fig. 1A).

Cenozoic igneous rocks are exposed both within and flanking the ARG (Fig. 3) and were intruded or erupted during three distinct events in the Late Eocene (42–34 Ma), the Oligocene (32–25 Ma; Fig. 5A), and the Late Miocene (14–8 Ma; Compton, 1983; Perkins et al., 1995). Late Eocene magmatism in the ARG is represented by the Emigrant Pass plutonic complex, composed of calc-alkaline intermediate to felsic magmas (Fig. 3). The Emigrant Pass plutonic complex (Fig. 3) was emplaced at shallow crustal levels (5–10 km) (Egger et al., 2003) and was coeval with the eruption of felsic volcanic rocks (Compton, 1983; Kistler and Lee, 1989; Nutt and Ludington, 2003). To the southeast of the ARG, andesite flows, dacite, and rhyolite ignimbrites of the 41–39 Ma northeast Nevada volcanic field represent a comparable suite of magmas (Brooks et al., 1995; Palmer and MacDonald, 2002). Based on their geochemistry and tracer isotope compositions, Eocene magmas incorporated variable amounts of ancient continental crust (Armstrong and Hills, 1967; Compton et al., 1977; Wright and Wooden, 1991; Egger et al., 2003). The Emigrant Pass plutonic complex intrudes a tilted Paleozoic stratigraphic section and several normal faults, which were folded during the intrusion of the pluton (Egger et al., 2003). However, regional and local evidence for significant Eocene fault-bound sedimentary basins is generally lacking both in and around the ARG metamorphic core complex and across the greater northern Basin and Range. Exceptions include the northern Nevada Bull Run and Copper Basins, which expose thick sequences of conglomerate and lacustrine sediments that record the onset of faulting as early as ca. 43 Ma (e.g., Axelrod, 1966a, 1966b; Clark et al., 1985; McGrew et al., 2008; Henry et al., 2011). Reconstruction of the Cenozoic unconformity across the broader region of northern Nevada and the ARG suggests that mostly Pennsylvanian, Permian, and Triassic strata were exposed at the surface prior to and during the eruption of Eocene volcanic rocks (Fig. 4B). The great stratigraphic thickness (∼7 km) of the Pennsylvanian–Permian Oquirrh Group in the region around the ARG, however, makes the reconstructed Cenozoic unconformity a less viable method for estimating the depth of erosion in this region, as compared to elsewhere in the northern Basin and Range (e.g., Gans et al., 1989; Long, 2012).

Oligocene plutons of the ARG, collectively termed the Cassia plutonic complex (Strickland et al., 2011b; Figs. 3 and 5A), are geochemically more evolved compared to earlier Eocene magmas. Although Oligocene plutons appear limited in map extent due to dissection by normal faults and burial by basin fill (Fig. 3), geologic cross sections and geochronological and petrologic investigations have been used to infer that the plutons underlie an extensive region of the ARG (Figs. 3, 5A, and 6) (Strickland et al., 2011b). The oldest is the Middle Mountain pluton, dated as 32 Ma (Strickland et al., 2011b) (Figs. 5A and 6), and the youngest is the Red Butte pluton, dated as 25 Ma (Compton et al., 1977; Todd, 1980; Egger et al., 2003) (Figs. 3, 5A, and 7). The composite plutons of the Cassia plutonic complex have more evolved Sr and Nd isotopic signatures relative to the Eocene Emigrant Pass plutonic complex, implying a greater degree of crustal assimilation (Wright and Wooden, 1991; Strickland et al., 2011b). The zircon crystals in the Oligocene plutons have Oligocene magmatic rims and ubiquitous Neoarchean cores, supporting this inference (Fig. 5A; Egger et al., 2003; Strickland et al., 2011b). However, the young zircon overgrowths have light δ18O values (5.40‰ ± 0.6‰), implying that they crystallized in equilibrium with mantle values (Strickland et al., 2011b). Unlike earlier Eocene magmas, there is no evidence that plutons of this age ever erupted; their intrusion spans a period of well-documented volcanic quiescence in the northern Basin and Range (e.g., Armstrong and Ward, 1991; Burton, 1997; du Bray, 2006; Fig. 1B).

Detailed geologic mapping and geochronology of the Oligocene plutons and their contact aureoles (Figs. 6 and 7) (Todd, 1980; Egger et al., 2003; Strickland et al., 2011a, 2011b) provide convincing evidence for extreme thinning of roof rocks during the ascent and crystallization of these plutons, which would be expected if their rise was diapiric. The plutonic rocks are also involved in these extreme strains, especially along their western sides (Todd, 1980; Strickland et al., 2007, 2011a, 2011b) (Figs. 6 and 7). In the Albion and Middle Mountains, the metamorphic mineral assemblages developed in the attenuated roof rocks of these plutons indicate intrusion and final crystallization at depths close to the aluminum silicate triple point, based on the presence of staurolite + kyanite ± sillimanite ± occasional andalusite in country rocks (Figs. 6 and 7; Holdaway, 1971; for an alternative interpretation of the ages of these metamorphic mineral assemblages, see Wells et al., 2012). Asymmetric fabrics and lineations, defined (in part) by the growth of elongate bundles of fibrolitic sillimanite, indicate top-to-the-northwest shear during northwest-southeast stretching at amphibolite facies conditions, leading us to conclude that the rise and emplacement of the Oligocene Cassia plutonic complex was intimately related to the dominant deformational and metamorphic event mapped in its country rocks (Strickland et al., 2011a, 2011b) (Figs. 6 and 7). Earlier thermochronologic studies reported 45–37 Ma 40Ar/39Ar ages from rocks in the western Raft River Mountains and the northern Grouse Creek Mountains (Saltzer and Hodges, 1988; Wells et al., 2000) that were interpreted as recording initial cooling of footwall rocks during Eocene motion along the Middle Mountain shear zone (Saltzer and Hodges, 1988; Wells et al., 2000; Hoisch et al., 2008).

In the Grouse Creek Mountains, the intrusion of the 25 Ma Red Butte plutons was accompanied by amphibolite facies metamorphism and extreme attenuation of section in response to west-northwest–east-southeast to east-west stretching (Fig. 7; Compton et al., 1977; Todd, 1980; Egger et al., 2003). This west-northwest–east-southeast fabric is superimposed on an earlier north-south–trending lineation and flattening fabric (Compton et al., 1977; Todd, 1980) that has not been directly dated, but is interpreted as Cretaceous in age by correlation to a dated north-south fabric developed in Ordovician rocks at higher structural levels (Wells et al., 2008).

Field and thin section relations suggest that kyanite, sillimanite, and andalusite grew during and shortly after the second deformation, depicted by west-northwest–east-southeast– to east-west–trending lineations (Figs. 7A, 7B, 7D, 7E). Stretching and attenuation of section accompanied the development of the Ingham Pass detachment fault under conditions where staurolite and kyanite were stable (Fig. 7F). The relations and mineral assemblages thus suggest that the Oligocene Red Butte pluton crystallized at depths close to the aluminum-silicate triple point, similar to the conditions inferred in the Albion and Middle Mountains region (Figs. 6 and 7; Holdaway, 1971). The 40Ar/39Ar thermochronology of muscovite and biotite from country rocks to the Red Butte pluton indicate that the region slowly cooled below ∼350 °C (biotite closure temperature) by ca. 21 Ma (Sheely, 2002), compatible with the inferred ∼10–15 km depths of the emplacement of the pluton. The northwest-southeast–stretching amphibolite facies fabrics associated with the Ingham Pass detachment and the detachment fault are cut and displaced by large offset Miocene normal faults along the western and eastern margins of the range (Figs. 3 and 7; Compton, 1983; Martinez, 2000; Egger et al., 2003).

In summary, data from the various parts of the Cassia plutonic complex and its wall rocks suggest that the lower crust beneath the ARG metamorphic core complex was at sufficiently high temperatures to undergo partial melting across a protracted time interval spanning at least 7 m.y., from ca. 32 Ma to 25 Ma. Most likely the interval of crustal melting spanned a broader range of time (∼17 m.y.), from 42 to 25 Ma, considering the ages of intrusion of the older Eocene magmas (Egger et al., 2003; Strickland et al., 2011b). During this time span, mobilization of the lower and middle crust resulted in the diapiric rise of the composite Cassia plutonic complex and its gneissic carapace, as indicated by the extreme vertical attenuation of its roof and wall rocks. The diapiric rise and emplacement of the granite-cored gneiss domes occurred during boundary conditions that were extensional, thus allowing strain to focus along the top of the plutonic complex and along normal sense shear zones recorded by the top-to-the west indicators in the strongly foliated and lineated (west-northwest–east-southeast stretching) country rocks. Although strains are very high at deeper structural levels, elevated temperatures fostered amphibolite facies fabrics rather than mylonitic fabrics with grain-size reduction (Figs. 6 and 7). In contrast, both the eastern Raft River Mountains and the western flank of Middle Mountain (Figs. 3 and 6) expose high-strain quartzites with lower temperature (greenschist facies) mylonitic fabrics formed as a consequence of vertical flattening and approximately northwest-southeast to east-west stretching and shear (Compton et al., 1977; Wells et al., 2000; Wells, 2001; Strickland et al., 2011a, 2011b). The 40Ar/39Ar ages of muscovite, coupled with quartz microstructures indicative of greenschist facies conditions (∼350–490 °C; Gottardi et al., 2011), led Wells et al. (2000) to interpret that most of the exhumation of the Raft River Mountains occurred along the east-dipping Raft River detachment in the Middle to Late Miocene (15–7 Ma) (Wells et al., 2000). This interpretation is consistent with fission track thermochronology from the Raft River Mountains indicating rapid cooling and inferred slip on the detachment between 13.5 ± 2.2 and 7.4 ± 2.0 Ma (2σ) (Wells et al., 2000). The eastward-younging ages were interpreted as related to migration of a rolling hinge detachment fault using the model of Buck (1988). Apatite and zircon fission track ages that range from 15.1 ± 2.4 to 12.1 ± 3.4 Ma (2σ; weighted average 13.4 ± 1.0 Ma) in the Grouse Creek and Albion Mountains are also interpreted to record the time of rapid exhumation of the western side of the core complex crystalline rocks to near surface conditions ca. 13.4 Ma (Egger et al., 2003).

METHODS

Geologic Mapping of Synextensional Basin Sediments

Cenozoic sedimentary and volcanic sequences flank the ARG metamorphic core complex on its northern, eastern, and western sides (Fig. 3). This study focuses on the evolution of the Raft River Basin developed along the eastern side of the Albion Mountains and northern flank of the Raft River Mountains (Figs. 3, 8, and 9). Although parts of the Raft River Basin section were previously studied and the deposits were identified as part of the Salt Lake Formation (Compton, 1972, 1975; Williams et al., 1974, 1982; Smith, 1982; Covington, 1983; Pierce et al., 1983; Wells, 2009), the exact age range of the strata and their relationship to the uplift history of the footwall rocks of the ARG were not clear before this study. Geologic mapping of the sedimentary and volcanic sequences on the north side of the Raft River Mountains was combined with existing map data (Compton, 1972, 1975; Williams et al., 1974, 1982; Smith, 1982; Covington, 1983; Pierce et al., 1983; Wells, 2009). Borehole, well log, and geophysical (seismic reflection, gravity, and magnetic) data initially collected for the Raft River geothermal project during the early 1970s through mid-1970s (Mabey and Wilson, 1973; Williams et al., 1974, 1975, 1982; Covington, 1977a, 1977b, 1977c, 1978, 1983; Oriel et al., 1978, 1979a, 1979b) were added and these data were used to produce the maps and cross section shown in Figures 8 and 9, as well as the stratigraphic columns displayed in Figure 10.

Geochronology

Single grain U-Pb ages were determined by LA-ICP-MS (University of Arizona Laserchron facility) on detrital zircons that were separated from five samples of Miocene sedimentary strata and from two samples of sandy limestone from the underlying Pennsylvanian–Permian Oquirrh Group. Three samples of Miocene volcanic rocks and one tuff from the base of the Miocene sedimentary section were selected for zircon U-Pb SHRIMP-RG (sensitive high-resolution ion microprobe, reverse geometry) geochronology, performed at the Stanford-U.S. Geological Survey facility using the methods described in Strickland et al. (2011b) (Tables 1–3). One sample of a reworked tuff was analyzed using both the LA-ICP-MS and SHRIMP-RG techniques to obtain a more precise depositional age.

Zircon was separated by traditional crushing and grinding methods, followed by standard separation methods that utilize the hydraulic, density, and magnetic properties of zircon (Gemini table, heavy liquids, and a Frantz magnetic separator). For detrital zircon geochronology, seven samples (five Miocene and two Pennsylvanian–Permian) were processed following the methods and analytical procedures of Gehrels et al. (2006, 2008) (Appendix 2). The LA-ICP-MS analytical data are reported in Table 1. Uncertainties shown are at the 1σ level and include only measurement errors. Analyses that were >20% discordant (by comparison of the 206Pb/238U and 206Pb/207Pb ages) or >5% reversely discordant were excluded from further analyses. Many detrital zircon grains from the Miocene sediments have large analytical errors in terms of their 206Pb/207Pb ages, both because of their young age and low 207Pb counts, thus yielding either high 207Pb/235U or 206Pb/207Pb errors, and high discordance estimates, and were also excluded from further analyses. Several detrital zircon grains from the Miocene sediments yielded Miocene (younger than 15 Ma) 206Pb/238U ages with acceptable errors (±3 m.y.). These young zircon grains were used to calculate the maximum depositional age of the enclosing sediments (Fig. 10).

The corrected 206Pb/238U (for zircon grains younger than 1000 Ma) and 206Pb/207Pb ages (for zircon grains older than 1000 Ma) were used to construct relative probability diagrams and cumulative probability diagrams (Fig. 11), using Isoplot (Ludwig, 2003). The relative age probability diagrams show each age and its uncertainty as a normal distribution and sum all ages from a sample into a single curve, where the area under the curve equals unity. Cumulative probability plots normalized each curve according to the number of constituent analyses such that each curve contains the same area (Gehrels et al., 2008).

The depositional age for each of the five Miocene samples was determined using the weighted average of the youngest detrital zircon grains from each of the samples (Ludwig, 2003). These data were plotted on weighted average diagrams and are presented in Figure 10. Miocene-age zircon grains were excluded from the suite of zircon used to construct cumulative and relative probability diagrams, so as to best compare the older zircon populations in the samples. A matrix of P-values was calculated, using the K-S (Kolmogorov-Smirnov) statistic (Gehrels et al., 2008), in order to better compare the detrital zircon populations of the six (of seven) samples that contain zircon grains older than Miocene (Appendix 3; results shown in Fig. 11D).

The Miocene detrital zircon suites were examined with cathodoluminescence (CL) imaging after completion of the U-Pb analyses, in order to identify in more detail the textural characteristics of the detrital zircon grains that were dated (Figs. 5D, 5E).

RESULTS

Stratigraphy and Age of the Raft River Basin Deposits

The Salt Lake Formation, deposited in the Raft River Basin, is best exposed north of the Raft River Mountains, east and south of the Jim Sage Mountains, east of the Cotterel Mountains, and locally west of the Black Pine Mountains (Figs. 8 and 9). Parts of the Salt Lake Formation were described by Williams et al. (1982), who noted three tuffaceous members (lower, middle, and upper) and the rhyolite and basalt flows of the Jim Sage volcanic member (Fig. 10E). Here we present evidence that there are two more units (identified as unit 1 and unit 2) below the lower tuffaceous member of Williams et al. (1982), and we describe the full thickness of their lower member (unit 3 of this study), locally >1000 m thick and not 200–500 m, as first described by Williams et al. (1982) (Figs. 10A, 10E; Appendix 4).

The sedimentary rocks of the Salt Lake Formation are cut by several normal faults, which displace and rotate the section (Figs. 8 and 9). Average tilts of strata are ∼30° to the west with dips as steep as 50° and as shallow as 15°. The fault blocks (blocks 4–6) with the largest amount of rotation of strata (∼48°; red symbols in Fig. 10C) are responsible for exposing the basal unconformity and the deepest part of the sedimentary section, in the region north of the Raft River Mountains (Fig. 9). In map view, the traces of the normal faults trend approximately north-south and are truncated to the north by an accommodation structure south of the Jim Sage Mountains (the Narrows structure of Covington, 1983). To the south, the same faults die out in terms of their displacement and locally displace minimally (<500 m) the system of faults that bounds the Raft River Mountains along their northern flank and separates metamorphic rocks of the footwall from late Paleozoic and Cenozoic strata (Figs. 8 and 9). The tilted fault blocks expose the entire stratigraphy of the Cenozoic section (Figs. 9 and 10), but the inferred maximum stratigraphic thickness of the Salt Lake Formation (∼4.5 km) is a composite, put together from partial sections measured in individual fault blocks.

The Salt Lake Formation can be divided into 7 units; units 1–6 are the focus of this study and are described in greater detail in Appendix 4. Units 1–3 are almost continuously exposed along the northern flank of the Raft River Mountains, where they were deposited unconformably on strata of the Pennsylvanian–Permian Oquirrh Group, and attain a maximum thickness of ∼2.3 km (Figs. 9 and 10). Unit 1 is mostly composed of breccias, fanglomerates, and pebble conglomerates deposited on top of the Pennsylvanian–Permian Oquirrh Group. Clast counts from the rocks of unit 1 indicate that this unit records an unroofing sequence, where the oldest rock clasts become progressively more abundant upsection (Fig. 10D; Appendix 4). Metamorphic (schist) clasts, presumably from the deeper parts of the metamorphic core complex, become abundant only in the pebble conglomerates of unit 2. Granitic clasts from either the Archean orthogneiss or the Oligocene Almo pluton were not observed in either unit 1 or 2 (Appendix 4). Unit 3 is composed of a thick sequence (∼1000 m) of lacustrine fine-grained ash-fall tuffs and marls.

Units 4–6 are composed of ∼1 km of volcanic rocks that are best exposed in the Jim Sage (units 4 and 5) and Cotterel Mountains (unit 6) (Figs. 3, 8, and 10; Appendix 4). The volcanic rocks in the Jim Sage Mountains were deposited locally above an angular unconformity of ∼5°–15° (depending on location) developed across older Miocene sedimentary rocks of units 1–3 (Williams et al., 1974, 1982; Pierce et al., 1983; Covington, 1983; this study) and are gently tilted by normal faults, forming an antiformal structure. Units 4–6 are displaced by ∼2–3 km along the Albion fault and are also cut by steep (50°–60°), small-offset normal faults. The stratigraphic section of unit 4 was measured in the southeastern part of the Jim Sage Mountains, whereas units 5 and 6 were measured in the southern Cotterel Mountains (Appendix 4). Unit 7 is poorly exposed along the southern flank of the Cotterel Mountains (upper tuffaceous member of Williams et al., 1982) and along the western flank of the Black Pine Mountains (previously mapped as undifferentiated Cenozoic sediments by Smith, 1982; Wells, 2009). The maximum thickness of unit 7 is ∼1.2 km based on drill core data from the eastern part of the Jim Sage Mountains (Williams et al., 1982; Pierce et al., 1983; Covington, 1983; Figs. 8 and 10E), and the younger strata of unit 7 have been interpreted to be Pliocene in age (Williams et al., 1982). Unit 7 is cut by normal faults and dips gently west ∼10° except where it is exposed on the western flank of the Black Pine Mountains, where it dips ∼30° to the east (Fig. 8). The Miocene–Pliocene Salt Lake Formation is overlain by the early Pleistocene Raft Formation, middle Pleistocene basalts of the Snake River group, and by unconsolidated late Pleistocene alluvium (Fig. 10E; Williams et al., 1982).

Detrital Zircon Geochronology

Detrital zircon geochronology of five Miocene samples is discussed from the oldest to the youngest sample. Sample AKR-09–2, from the stratigraphically lowest tuff, was collected ∼300 m above the basal unconformity of the section (Fig. 10A; Tables 2 and 3), and has abundant reworked glass shards. Zircon crystals from this sample were dated both by LA-ICP-MS (nine zircon grains) and with the SHRIMP-RG (seven zircon grains). A large number (∼30) of zircon grains analyzed by LA-ICP-MS resulted in Miocene 206Pb/238U ages (ca. 18–13 Ma), but had high errors (due to their young age) or high common lead (204Pb), and so were excluded from the calculation of depositional age. The 16 concordant zircon analyses obtained by both geochronology techniques represent a unimodal population with a weighted average 206Pb/238U age of 13.45 ± 0.28 Ma (with a low MSWD [mean square of weighted deviates] of 0.17), interpreted as a depositional age (Fig. 10A).

Sample AKR-09–1A is from the matrix-supported pebble conglomerate deposited at the top of the fanglomerates of unit 1 (Fig. 10A; Table 2). Of the 100 detrital zircon grains analyzed using LA-ICP-MS, 55 were concordant and older than 20 Ma, and were used to plot relative and cumulative probability curves (Figs. 11B, 11C). Zircon grains range in age from Paleoproterozoic through Mesoproterozoic (2000–1000 Ma), with multiple peaks, notably ca. 1800 Ma, 1500 Ma, and 1000 Ma (Fig. 11). The sample has Archean zircons (ca. 3300–2600 Ma) and a small number of zircon grains that are early Paleozoic (ca. 550–450 Ma). One zircon crystal is Cretaceous in age (ca. 120 Ma) and one discordant zircon (not included in the diagram) has an Eocene age (ca. 40 Ma). Nine detrital zircon grains, with low error Miocene ages, yield a weighted average age of 13.25 ± 0.3 Ma (Fig. 10A), interpreted to be close to the depositional age of unit 1. Approximately 35 detrital zircon analyses resulted in ages with either high 206Pb/238U or high 207Pb/235U errors, mostly because these zircon grains are young (Miocene).

Sample JSR-09–6 is a coarse-grained muscovite-rich sandstone, deposited near the top of unit 2 (Fig. 10A; Table 2). Of the 100 detrital zircon grains analyzed, 67 are concordant and older than Miocene, and used to plot relative and cumulative probability curves (Fig. 11). The zircon population in this sample is very similar to AKR-09–1A, with Mesoproterozoic and Paleoproterozoic zircon populations (Figs. 11B, 11C). It has Archean zircon (ca. 2500–3300 Ma) and several zircon grains that are early Paleozoic (500–450 Ma). Only two detrital zircon grains yielded low error, concordant Miocene ages from this sample and were combined with Miocene zircon grains from sample AKR-09–3, which was collected from the same unit, to calculate an approximate depositional age for unit 2 of 10.5 ± 0.4 Ma. Three detrital zircon grains from this sample yield discordant ages and 24 yielded analyses have high 206Pb/238U or high 207Pb/235U errors, mostly because of their young (Miocene) age.

Sample AKRR-09–3 is a medium-grained quartz and feldspar-rich sandstone, deposited near the top of unit 2 (Fig. 10; Table 2). A population of 54 detrital zircon analyses yielded concordant ages, with acceptable errors and ages older than 20 Ma, and were used to plot relative and cumulative probability curves (Fig. 11). The zircon population age range is similar to that of the previous samples, but the abundances of different ages differ in a significant way (Fig. 11). This sample lacks appreciable amounts of Phanerozoic zircon but has moderate populations ranging in age from Paleoproterozoic through Mesoproterozoic (Figs. 11B, 11C). It has more Archean zircon than the last two samples, with a large peak ca. 2600 Ma, similar in age to dated exposures of the Green Creek Complex (Strickland et al., 2011b). One zircon has a discordant Oligocene age (possibly derived from the Cassia plutonic complex) and was not included in Figure 11. Based on the similarity of the stratigraphic position of AKRR-09–3 and JSR-09–6 (Fig. 8), the young detrital zircon grains from these two samples were combined to calculate a weighted average age of 10.5 ± 0.4 Ma, interpreted as the approximate depositional age of unit 2 (Fig. 10A). This sample also yielded 10 discordant zircon analyses and 14 additional analyses with high 206Pb/238U or high 207Pb/235U errors, which have apparent Miocene ages. The CL images of the detrital zircon grains with apparent Miocene ages have similar textures to low-U zircon from Miocene volcanic rocks (cf. group 2 of Fig. 5D with 5C).

Sample JSR-09–5 is a tuffaceous sandstone from the middle of unit 3 (Table 1). Of 100 zircon grains analyzed, only 37 were concordant with acceptable errors and ages older than 20 Ma, and were used to plot relative and cumulative probability curves (Fig. 11). The relative probability diagram of this sample is similar to AKR-09–1A and JSR-09–6 in that it contains zircon populations ranging in age from Paleoproterozoic through Mesoproterozoic (2000–1000 Ma), with large peaks ca. 1600 Ma and 900 Ma (Figs. 11B, 11C). The sample has a strong early Paleozoic age peak (ca. 450 Ma), a few ca. 2450–2300 Ma zircon grains, and one Archean (ca. 3050 Ma) zircon. Two Eocene zircon grains (ca. 35 Ma) are discordant and were not included in the diagram. A population of 15 young detrital zircon grains with low errors was used to calculate a weighted mean of 9.7 ± 0.7 Ma representing the approximate depositional age for the upper part of unit 3 (Fig. 10A). Several zircon grains were excluded from further statistical analysis due to high 206Pb/238U or high 207Pb/235U errors or because they were discordant. Of the 54 excluded zircon grains, most (38) yielded approximate Miocene ages with high 206Pb/238U errors while 16 yielded discordant ages. CL imaging of the zircon grains from this sample indicate that several of the discordant zircon grains have large chaotic cores and zoned rims, very similar to zircon dated from the Oligocene plutons of the core complex (cf. group 3 of Fig. 5E to 5A and 5B) (Strickland et al., 2011b). In addition, most of the zircon grains with high errors and apparent Miocene ages have CL images similar to the low-U zircon from Miocene volcanic rocks (cf. group 2 of Fig. 5E to 5C).

To test for similarity between the detrital zircon populations of sedimentary rocks of the Salt Lake Formation to the older rocks of the Pennsylvanian–Permian Oquirrh Group (Figs. 11B, 11C), detrital zircons were analyzed from two Pennsylvanian–Permian samples. Sample DC-1 is a calcareous sandstone from the Lower Pennsylvanian–Permian Oquirrh Group near Dove Creek Pass (Fig. 11), and has zircon age populations that range from Paleoproterozoic through Mesoproterozoic (2000–1000 Ma), with large peaks at 1750, 1650, and 1500 Ma (Fig. 11B). The sample has an early Paleozoic zircon population (ca. 450 Ma) and several Archean through Paleoproterozoic (ca. 3050–2450 Ma) zircon grains. Sample PER-SS is an Early Permian sample from the northern Black Pine Mountains (northern block of Smith, 1982; Figs. 8 and 11B; Table 2). This sample is a fine-grained, silty limestone with very small (<50 μm) zircon grains. The detrital zircon population of PER-SS is almost identical to that of DC-1 with a large Paleoproterozoic to Mesoproterozoic zircon suite ranging in age from 2000 to 1000 Ma, several early Paleozoic (ca. 450 Ma) zircon grains, and several Archean to Paleoproterozoic (ca. 3050–2450 Ma) zircon grains.

Zircon U-Pb SHRIMP-RG Analyses of the Jim Sage Volcanic Suite

The three Miocene volcanic rocks collected from the Jim Sage and Cotterel Mountains (Fig. 8A; Tables 1 and 3), dated using SHRIMP-RG, provide constraints on the end of the deposition of unit 3 and the onset of magmatism related to the Snake River Plain, in the Raft River Basin. The 207Pb-corrected 206Pb/238U ages of suites of zircon grains were used to calculate the weighted mean age for each sample. The lower rhyolite (unit 4) has an age of 9.46 ± 0.09 Ma (2σ; MSWD = 1.5; n = 23) at its base and 9.33 ± 0.10 Ma (2σ; MSWD = 0.86; n = 16) for the top of the unit (Figs. 5 and 10A); the two reported ages for unit 4 are within error of each other. A sample from the top part of unit 6 results in a weighted average age of 8.21 ± 0.15 Ma (2σ; MSWD = 1.4; n = 14). These three ages are interpreted to represent the eruptive ages of unit 4 and unit 6.

Sources for the Salt Lake Formation

A comparison of the detrital zircon populations from the Miocene Salt Lake Formation to those from late Paleozoic sediments shows a striking similarity (Figs. 11B, 11C). The zircon age peaks and approximate percentage of each age population in the Miocene rocks are very similar to those from late Paleozoic strata (Fig. 11). This observation is compatible with the fairly large P-values obtained by performing the K-S test (P = 0.09–0.51; Fig. 11D; Gehrels et al., 2006). The P-values in the K-S test from detrital zircon suites test the null hypothesis that any two randomly selected detrital zircon populations are not similar. High P-values disprove the null hypothesis, thus the two populations in question can be interpreted to be similar. The high P-values and the apparent similarity in detrital zircon populations between the late Paleozoic rocks and the Miocene strata lead us to infer that late Paleozoic strata were the dominant detritus source for the earliest part of the Miocene Raft River Basin. This inference is compatible with and supported by abundant late Paleozoic clasts in the basal Miocene conglomerates of unit 1, and by the observed relationship that across this region mostly Pennsylvanian–Permian strata were exposed at the Cenozoic erosional surface (Fig. 4B). The other major sources of detrital zircon for the Salt Lake Formation were Miocene ash-fall tuffs, ignimbrites, and volcanic flows that were deposited proximal to the region (Figs. 5D, 5E).

A notable exception to the nearly identical detrital zircon populations of the Miocene and the late Paleozoic samples is the sandstone sample AKR-09–3 from unit 2 (Fig. 10). This sample contains a larger percentage of Neoarchean (ca. 2600 Ma) zircon (Figs. 11B, 11C), indicating that rocks from the Green Creek Complex may have been exposed and eroding at that time (ca. 10.5 Ma). Sample JSR-09–5 also contains several zircon grains having CL images that are characterized by chaotic (metamict) cores with magmatic overgrowths, and that look very similar to zircon dated from the Cassia plutonic complex (Strickland et al., 2011b; cf. group 3 of Fig. 5E to Figs. 5A, 5B). Both of these observations indicate that the deep metamorphic and plutonic rocks of the ARG core complex were at the surface and exposed to erosion by ca. 10.5 Ma, consistent with the fact that Miocene (ca. 9 Ma) volcanic rocks are mapped as unconformably overlying Archean basement rocks in the core of the Albion Mountains and Miocene rhyolites unconformably overlie metamorphic rocks along the west flank of the Albion Mountains (Armstrong et al., 1975; Compton, 1972, 1975; Miller et al., 2008).

The general absence of Middle Eocene age (55–45 Ma) zircon populations in the Miocene sediments of the Raft River Basin indicate that the ca. 52–45 Ma Challis volcanic field located north of the ARG and the Snake River Plain (Fig. 1) (Janecke and Snee 1993; Dostal et a., 2003; Gaschnig et al., 2010) was not a major source of detritus as the Raft River Basin developed. Given the widespread exposure of the Challis volcanic rocks and the fact that zircon crystals from this region were transported as far west as the coast of California (Dumitru et al., 2013), we postulate that a physiographic barrier, possibly a Miocene age Snake River Plain thermal bulge (e.g., Anders and Sleep 1992), or a valley like the present-day Snake River Valley, may have restricted sediment transport from north to south during the development of the Raft River Basin in the Miocene. The absence of Challis-age zircon in the basin is also indirect evidence in favor of the lack of development of Eocene–Oligocene catchments or basins in the region of the ARG. These earlier basins, if they existed, would likely have contained sediments with 52–45 Ma zircon that would in turn have been reworked into the younger Miocene basins.

Slip History of the Albion Fault and Raft River Detachment

The Albion fault, which strikes north-south in map view and dips ∼20°–30° to the east (Figs. 8 and 9), most likely represents the fault that first moved ca. 14 Ma to create the accommodation space that was then filled by Miocene sediments of the Salt Lake Formation. The fault mapped as the Raft River detachment (Figs. 8 and 9) possibly represents an exhumed ductile-brittle transition zone into which the Albion fault partially rooted. The faults mapped within the south part of the Raft River Basin displace rocks as young as ca. 9.5 Ma and have displacements ranging from ∼500 m to ∼2.5 km (Fig. 8; cross-section B–B′). The faults cut and repeat the Miocene sedimentary sequence and appear to die southward, minimally (<500 m) displacing the mapped trace of the Raft River detachment (Figs. 8 and 9). We interpret these crosscutting relationships to indicate that the normal faults mapped in the Raft River Basin are younger than both the Albion fault and the high-strain fabrics of the Raft River detachment.

The east-west trend of the trace of the Raft River detachment separating lower plate rocks beneath the domed detachment from upper plate Miocene fault blocks can be interpreted as a having a major strike-slip component in its early history of slip between 13.5 and 9.5 Ma. An alternative interpretation is that the normal faults that cut and rotate the Miocene basinal sequence sole into a basal detachment beneath the Raft River Valley, and this fault system has been cut and downdropped by a down-to-the north normal fault along the northern (and southern) flanks of the Raft River Mountains, uplifting the Raft River Mountains with respect to the faulted Miocene basin sections and exposing the basal detachment fault, which might have formed the Miocene ductile-brittle transition zone (Figs. 8 and 9). While these two scenarios are not mutually exclusive, we prefer the former scenario, given that the lower plate rocks at the north edge of the Raft River Mountains have north-dipping foliation planes but east-west–trending lineations, thus precluding ductile top-to-the-north stretching, which would produce north-south–oriented lineations. Another piece of evidence precluding much top-to-the-north motion along the northern edge of the Raft River Mountains is the fact that the dips of the Miocene strata do not change toward the Raft River detachment, which might be expected if the northern edge of the Raft River detachment had top-to-the-north motion (Fig. 9). Regardless of the exact kinematic relationships, the entire evolution of the complex resulted in the domal nature of the elongate Raft River Mountains core complex, which appears to wrap around the northern, southern, and eastern edges of the mountain (Fig. 9).

Apatite and zircon fission track ages from the Raft River Mountains (Wells et al., 2000) and the Albion Mountains (Egger et al., 2003) range from 15.1 ± 2.4 to 7.4 ± 2.0 Ma (2σ), and were interpreted to represent cooling via the rapid exhumation of the footwall to near surface conditions ca. 13.4 Ma (Egger et al., 2003; Fig. 12; summary of apatite fission track thermochronology) and to record the progressive unroofing of the Raft River detachment during the migration of a rolling hinge from ca. 13.4 to 7 Ma (Wells et al., 2000; Wells, 2001). The new data from the Raft River Basin allow us to delineate a more detailed history of the Miocene faulting and deposition temporally related to the 13.4–7 Ma low-temperature uplift and cooling history of adjacent footwall rocks.

The oldest Miocene tuff dated (13.45 Ma) records the beginning of deposition of the Salt Lake Formation in the Raft River Basin. Faulting along the Albion fault began ca. 14 Ma, with rapid slip occurring between 13.5 and 10.5 Ma (Figs. 12, 13A, and 13B). During this time, the basin developed its greatest accommodation space and was filled with clastic detritus derived almost exclusively from upper plate Paleozoic rocks (Figs. 11B, 11C, 13A, and 13B).

By between ca. 10.5 and 9.5 Ma, sediment sources to the basin included structurally deeper rocks of the footwall (Archean and Cenozoic intrusive rocks) with continued contribution of debris from the hanging wall (Paleozoic), consistent with a history of rapid exhumation documented by the reverse clast stratigraphy (Fig. 10D) and detrital zircon signatures of basin sediments (Figs. 11A, 11B, and 13C). Rapid slip on the Albion fault was responsible for the rise of the Albion Mountains and final exposure at the surface of deep crystalline basement, as evidenced by unconformities preserved in parts of the Albion Mountains, where ca. 9 Ma volcanic rocks (Armstrong et al., 1975) overlie metamorphic and igneous rocks. The topographic depression formed between 13.5 and 10.5 Ma by motion on the Albion fault was quickly filled by ash-fall tuffs and rhyolite flows. The rotation of the Albion fault and the basin fill deposits (by ∼5°–15°) appears to have occurred during this time span, as indicated by the small angular unconformity developed between unit 3 and unit 4 (Fig. 13C).

After ca. 9.5 Ma, extension was accommodated by a series of intrabasin planar normal faults that cut, rotate, and repeat the Miocene sequence, including the volcanic rocks (Figs. 3, 8, 9, and 13D). Extension after ca. 9.5 Ma was also responsible for the rotation of the intrabasin normal faults to shallower angles. This interpretation is supported by the dips of bedding in units 1–3 which, where mapped, are about the same independent of stratigraphic position, suggesting that most rotation occurred after the deposition of unit 3 (Fig. 10C). This relation is best observed in fault block 4, where an ∼2.5-km-thick section is rotated by the same amount (current dip 48°W; red symbols in Figs. 9 and 10C).

Drill core data indicate that the younger than 8.2 Ma unit 7 (upper tuffaceous member of Williams et al., 1982) was deposited mostly the eastern Raft River Basin (Williams et al., 1974, 1982) and that faulting and sedimentation progressively stepped eastward toward the center of the basin.

The depositional and faulting history of the Raft River Basin indicates that the Albion fault likely initiated at a fairly steep angle (∼60°) and accommodated a minimum of 4 km of vertical component of slip based on the thickness of units 1–3 (2500 m) and the present-day relief from the basin to Cache Peak, Idaho (∼1400 m). We infer that this fault system may have been the same as the plane of initial slip along what is now the Raft River detachment (Fig. 13). We envision that the entire Albion fault system began to move ca. 14 Ma, rotated by 5°–15° between 14 and 9.5 Ma, and likely soled into a shallow ductile-brittle transition zone, which was later exhumed and is now the Raft River detachment. Thus, in our interpretation the mylonite shear zone of the Raft River detachment represents the boundary between a plastic and/or viscous lower plate and its elastic lid (upper plate). Most of the rotation of the Miocene section and its bounding faults occurred after ca. 9.5 Ma; this is based on the fact that all Cenozoic sedimentary units older than 9.5 Ma are involved in the same amount of rotation during faulting. Therefore most of the flexural rotation to shallower angles of the intrabasin fault blocks, the Albion fault and the Raft River detachment, occurred after ca. 9.5 Ma. During this time, rocks beneath the present-day exposures flowed in the ductile regime and underwent stretching (Fig. 13D).

The minimum amount of Miocene extension can be calculated by cross-sectional line-length balancing of the Cenozoic unconformity across the region north of the Raft River Mountains (Fig. 8; cross-section B–B′). The present-day east-west width of the exposure of the unconformity is ∼38 km (Dfinal) and the cross-sectional length of the unconformity, after the restoration of fault slip, is 22 km (Doriginal). The calculated β-factor from these estimates is 1.72 and the minimum amount of Miocene extension is ∼72%.

DISCUSSION

The geochronologic, thermochronologic, and structural data from the ARG discussed herein (summarized in Fig. 12) reveal a complex and protracted Cenozoic extensional history for this complex. The various Cenozoic events affecting this region differ in terms of their timing and the crustal processes they represent, but each event sets the stage for subsequent events. Resolving and understanding this history is crucial to answering many of the puzzling questions about the timing of events in metamorphic core complexes and the nature of detachment faults that bound them.

The most important conclusion of the data presented here is that the extreme attenuation of units at amphibolite (sillimanite) facies in the ARG developed at least 10 m.y. prior to the final exhumation of the metamorphic core complex during Basin and Range faulting. This documented difference in timing makes it clear that this metamorphic core complex and its bounding detachment faults did not develop during a single protracted event, as suggested by many models, but rather, its evolution took place in a series of steps or stages (Figs. 12 and 14).

Stage 0: Preextensional Deformation

Regional studies of Mesozoic deformation in the Sevier belt east of the ARG indicate that shortening was partitioned in time between 145 and ca. 55 Ma, along four major thrust systems (e.g., Armstrong, 1963; DeCelles, 1994; Burtner and Nigrini, 1994; DeCelles, 2004; Yonkee and Weil, 2011). These studies document that the motion on these thrust systems migrated eastward through time mostly in the Mesozoic, with minor shortening continuing into the Cenozoic (Wiltschko and Dorr, 1983; Heller et al., 1986; DeCelles, 1994, 2004; Burtner and Nigrini; 1994; Yonkee and Weil, 2011; Appendix 1). Geochronologic and thermochronologic studies of the deeper parts of the stratigraphic section exposed in the ARG suggest that this part of the hinterland of the Sevier thrust belt was subject to burial and metamorphism in the Mesozoic (e.g., Hoisch et al., 2002, 2008; Harris et al., 2007; Wells et al., 2012) but the crustal thickness at the end of Mesozoic shortening is not well quantified and is somewhat controversial (e.g., Coney and Harms, 1984; Gans et al., 1989; DeCelles, 2004; Colgan and Henry 2009; Ernst, 2010). Our beginning diagram (Fig. 14) shows a crustal thickness of ∼45–50 km, similar to estimates from east-central Nevada based on the palinspastic restoration of Cenozoic extensional structures (Gans et al., 1989).

Stage 1: Precursor Eocene Magmatism, 42–34 Ma

Following Mesozoic–early Cenozoic crustal thickening, calc-alkaline magmatism migrated from southern Canada to southern Nevada between ca. 55 and ca. 20 Ma (Armstrong and Ward, 1991; Best and Christiansen 1991; Christiansen and Yeats, 1992; Gans et al., 1989). The southward sweep of magmatism is attributed to asthenospheric upwelling following the delamination of the shallowly dipping Farallon slab, which led to the intrusion of mafic magmas into the crust (e.g., Armstrong and Ward, 1991; Best and Christiansen, 1991; Christiansen and Yeats, 1992; Humphreys 1995; Humphreys et al., 2003; Fig. 1), resulting in widespread heating and melting of the lower and middle crust, forming large MASH zones (mixing-assimilation-storage-hybridization; Fig. 14).

In the region of the ARG, this event is represented by ca. 42–34 Ma calc-alkaline intermediate to felsic igneous rocks intruded as both shallow-level plutons (the Emigrant Pass plutonic complex) and erupted volcanic sequences (Figs. 12 and 14) (e.g., Compton, 1983; Kistler and Lee, 1989; Brooks et al., 1995; Nutt and Ludington, 2003; Egger et al., 2003). This early magmatism was driven by mantle-derived magmas, which interacted and assimilated significant amounts of crust beneath this region (Fig. 14), resulting in hybrid magmas (Grunder 1993, 1995; Gans et al., 1989; Humphreys et al., 2003; Dostal et al., 2003). Despite the evidence for voluminous magmatic activity (as well as hybridization at depth) and the possibility that significant weakening of the crust might have occurred during this time span, there is limited evidence for large offset high-angle faults and significant basin development in the ARG and surrounding region. Specifically, the regional Cenozoic unconformity map (Fig. 4) suggests that only late Paleozoic strata were exposed at the surface before Cenozoic magmatism began. However, we acknowledge that using the erosional surface presented in Figure 4 as a datum results in a somewhat crude estimate of the amount of exhumation prior to the Cenozoic, especially in the region of the ARG where the Pennsylvanian and Permian strata are as much as 7–8 km thick. Thus multiple high-angle normal faults may have developed across the top of the ARG as long as they only cut and offset Pennsylvanian–Permian strata and did not bound deep basins.

Stage 2: Diapiric Ascent of Pluton-Cored Gneiss Domes, 32–25 Ma

Following the shallow-level emplacement and eruption of magmas between 42 and 34 Ma, heating likely continued at depth, leading to broader melting and ultimately mobilization of the deeper crust. We envision that the maturation of the MASH zone established during stage 1 resulted in the episodic production of crustal melts that became the plutonic cores of Oligocene crustal welts (gneiss domes), which rose diapirically to depths of ∼10–15 km under extensional boundary conditions (Fig. 14). The diapiric rise of these welts to the base of the upper crust was accompanied by extreme thinning and stretching of roof rocks at amphibolite (sillimanite) facies, forming the well-developed foliation and extensional lineation dated by U-Pb monazite and zircon as 32–25 Ma (Egger et al., 2003; Strickland, 2010; Strickland et al., 2011a, 2011b; Fig. 14). This mobile infrastructure was separated from the overlying brittle crust by detachment faults, such as the Ingham Pass fault and the high-temperature portion of the Middle Mountain shear zone (Egger et al., 2003; Strickland et al., 2011a, 2011b). These high-temperature shear zones were diachronously developed, but represent the result of the development of a protracted ductile-brittle transition in the crust (or top of channel flow) (Fig. 14).

There is no evidence for volcanism during the 32–25 Ma time interval of gneiss dome development or for the formation of deep fault-bounded sedimentary basins. It is possible that supracrustal extension was extreme but took place only across the width of the crustal diapirs, thus representing only about ∼22 km of localized extension across the ARG (Fig. 14). Normal faults that may have moved during this time interval were subsequently deformed and uplifted during the rise of the gneiss domes. Flow of more mobile lower to mid-crustal rocks into the regions of plutonism and crustal thinning is likely (e.g., Gans, 1987; McKenzie et al., 2000), thus the region of the ARG metamorphic core complex might have been a site of uplift and erosion rather than deposition. One of the important hallmarks of the geology of the ARG metamorphic core complex, recognized for many years, is that ductile extensional fabrics related to vertical thinning are superimposed on a complexly faulted Paleozoic section (Compton et al., 1977; Miller, 1978, 1983; Wells, 1997; Wells et al., 1998, 2000). This rather unusual superposition of ductile fabrics on earlier brittle structures may provide support for this speculative interpretation.

Stage 3: Basin and Range Faulting, 14 to After 7 Ma

A 10–12 m.y. hiatus separated the diapiric rise of Oligocene plutons and their entrained wall rocks from the onset of Basin and Range faulting. The timing of faulting along the Albion and Raft River fault systems is recorded by the inception of deposition into the Raft River Basin ca. 14 Ma (Figs. 10, 12, 13, and 14), synchronous with the development of Miocene basins south of the Raft River Mountains and along the western flanks of the Grouse Creek Mountains (Compton, 1983; Todd, 1980; Martinez, 2000, 2001; Egger et al., 2003). Specifically, the Grouse Creek Basin developed in the hanging wall of a Miocene fault system that cuts older Oligocene fabrics and the Ingham Pass detachment in the Grouse Creek Mountains (Fig. 7). The high-temperature Middle Mountain shear zone is also cut by the same fault system to the north (Figs. 3 and 6). Rapid uplift of the ARG metamorphic core complex occurred between ca. 14–10.5 Ma, by faulting on both sides of the complex. After ca. 8 Ma, extension was accommodated by systems of intrabasinal faults that cut and offset the 13.5–10.5 Ma sediments and 10–8 Ma volcanic rocks. Rapid slip on both the early parts of the Albion fault as well as along the faults that cut the Salt Lake Formation correspond in age to the low-temperature cooling history of footwall rocks that began at 14 Ma and progressed to after 7.5 Ma (Fig. 12; Wells et al., 2000; Egger et al., 2003).

During Basin and Range faulting, hotspot-generated magmatism progressed eastward along the Snake River Plain–Yellowstone trend (e.g., Draper, 1991; Pierce and Morgan, 1992; Ellis et al., 2010). The overlap in age between the nearest eruptive centers of the Snake River Plain (the 10.5–8.5 Ma Twin Falls and the 10.2–9.2 Ma Picabo centers) and the Miocene extension in the ARG metamorphic core complex (Figs. 10, 13, and 14) has important implications for the evolution of the complex. This magmatism represents a major thermal event that may have resulted in crustal heating and melting, thus decreasing the strength of the crust and assisting younger and deeper crustal flow during Basin and Range extension and uplift of the ARG metamorphic core complex.

CONCLUSIONS

This study provides new structural and geochronologic data regarding the timing of sedimentation, extensional faulting, and exhumation of the Albion–Raft River–Grouse Creek metamorphic core complex. Coupled with existing geochronologic data that bracket the age of lower plate magmatism, metamorphism, and diapiric rise of pluton-cored crustal welts, we are able to distinguish three different stages leading to the development and exhumation of this core complex that involved changing lithospheric-scale processes (Fig. 14).

Deciphering this history has illustrated that the faults that exhume the complex (stage 3) are a distinctly younger (and unrelated) event than the amphibolite (sillimanite) grade fabrics associated with the diapiric rise of the granite-cored gneiss domes in the Oligocene (stage 2), because they are separated by a substantial time gap of 10–12 m.y. This observation is significant because it implies that not all fabrics and mylonites in metamorphic core complexes formed at the same time or via the same process, and thus should not be kinematically linked in the interpretation of metamorphic core complexes. This study highlights that core complexes of the Basin and Range may have polygenetic and multiphase Cenozoic extensional histories that need be accounted for when modeling their histories and determining the total offset and amount of extension they represent (Fig. 2).

This work was made possible through grants awarded to Alexandros Konstantinou and Elizabeth Miller. Field work funds were provided by a Leventis Foundation grant, a Stanford Graduate Fellowship, and a Stanford McGee Fund awarded to Alexandros Konstantinou. Field and analytical work were funded by National Science Foundation Tectonics Division grants EAR-0809226 and EAR-0948679 to Miller. We thank Michael Wells, Mike Williams, and an anonymous reviewer for comments that helped to significantly improve this manuscript.

APPENDIX 1. TIMING OF SHORTENING AND LIST OF REFERENCES

Table A1 summarizes the timing of shortening recorded in the different thrust fault systems of the Mesozoic–early Cenozoic Sevier and Laramide orogenies in the northeastern Utah and Wyoming regions. Also shown is the list of references used in this study to summarize the timing of shortening.

APPENDIX 2. GEOCHRONOLOGY METHODS

A large split of detrital zircon grains is incorporated into a 1 in (2.54 cm) epoxy mount together with fragments of our Sri Lanka standard zircon. The mounts are sanded down to a depth of ∼20 μm, polished, imaged, and cleaned prior to isotopic analysis.

U-Pb geochronology of zircon grains is conducted by laser ablation–multicollector–inductively coupled plasma mass spectrometry (LA-MC-ICPMS) at the Arizona LaserChron Center (Gehrels et al., 2006, 2008). The analyses involve ablation of zircon with a New Wave UP193HE Excimer laser using a spot diameter of 30 μm. The ablated material is carried in helium into the plasma source of a Nu HR ICPMS, which is equipped with a flight tube of sufficient width that U, Th, and Pb isotopes are measured simultaneously. All measurements are made in static mode, using Faraday detectors with 3 × 1011 ohm resistors for 238U, 232Th, 208Pb-206Pb, and discrete dynode ion counters for 204Pb and 202Hg. Ion yields are ∼0.8 mv/ppm. Each analysis consists of 1 15 s integration on peaks with the laser off (for backgrounds), 15 1 s integrations with the laser firing, and a 30 s delay to purge the previous sample and prepare for the next analysis. The ablation pit is ∼15 μm in depth.

For each analysis, the errors in determining 206Pb/238U and 206Pb/204Pb result in a measurement error of ∼1%–2% (at 2σ level) in the 206Pb/238U age. The errors in measurement of 206Pb/207Pb and 206Pb/204Pb also result in ∼1%–2% (at 2σ level) uncertainty in age for grains that are older than 1.0 Ga, but are substantially larger for younger grains due to low intensity of the 207Pb signal. For most analyses, the crossover in precision of 206Pb/238U and 206Pb/207Pb ages occurs ca. 1.0 Ga.

The 204Hg interference with 204Pb is accounted for measurement of 202Hg during laser ablation and subtraction of 204Hg according to the natural 202Hg/204Hg of 4.35. This Hg is correction is not significant for most analyses because our Hg backgrounds are low (generally ∼150 cps at mass 204).

Common Pb correction is accomplished by using the Hg-corrected 204Pb and assuming an initial Pb composition from Stacey and Kramers (1975). Uncertainties of 1.5 for 206Pb/204Pb and 0.3 for 207Pb/204Pb are applied to these compositional values based on the variation in Pb isotopic composition in modern crystal rocks.

Interelement fractionation of Pb/U is generally ∼5%, whereas apparent fractionation of Pb isotopes is generally <0.2%. In-run analysis of fragments of a large zircon crystal (generally every fifth measurement) with known age of 563.5 ± 3.2 Ma (2σ error) is used to correct for this fractionation. The uncertainty resulting from the calibration correction is generally 1%–2% (2σ) for both 206Pb/207Pb and 206Pb/238U ages.

Concentrations of U and Th are calibrated relative to our Sri Lanka zircon, which contains ∼518 ppm of U and 68 ppm Th.

The analytical data are reported in Table 3. Uncertainties shown in the tables are at the 1σ level, and include only measurement errors. Analyses that are >20% discordant (by comparison of 206Pb/238U and 206Pb/207Pb ages) or >5% reverse discordant are not considered further.

The resulting interpreted ages are shown on Pb*/U concordia diagrams and relative age-probability diagrams using the routines in Isoplot (Ludwig, 2003). The age-probability diagrams show each age and its uncertainty (for measurement error only) as a normal distribution, and sum all ages from a sample into a single curve. Composite age probability plots are made from an in-house Excel program that normalizes each curve according to the number of constituent analyses, such that each curve contains the same area, and then stacks the probability curves.

APPENDIX 3. STATISTICAL CALCULATIONS

The K-S statistic is a statistical calculation takes into account the squared differences of the cumulative probability curves between any two samples and tests the hypothesis that two samples are different from each other and are therefore not obtained from the same parent population. A P-value <0.05 confirms the hypothesis and indicates that the two samples are statistically not similar in terms of their detrital zircon population signatures and that the similarities can be explained by a random distribution of detrital zircon age. In contrast, a P-value >0.05 indicates that the samples may be similar in terms of their detrital zircon population (Gehrels et al., 2008) and similarities between samples cannot be explained by random distribution.

APPENDIX 4. DETAILED STRATIGRAPHY OF THE SALT LAKE FORMATION

Unit 1, best exposed in fault blocks 2 and 4, is ∼700 m thick and composed primarily of breccias, fanglomerates, and pebble conglomerates deposited on top of the Pennsylvanian–Permian Oquirrh Group (Figs. 3, 8, and 10). The lower part of unit 1 is composed of clast-supported, poorly sorted breccias and fanglomerates derived primarily from the underlying Paleozoic limestone sequence. The upper part of unit 1 is composed of both clast and matrix-supported pebble conglomerate intercalated with coarse sandstone. Rare imbrications in the pebble conglomerate indicate that the paleocurrent directions were predominantly from the west to the east.

Clast counts of conglomerate indicate that the basal part of unit 1 are composed mostly of 3–15-cm-diameter, subangular clasts of gray-blue carbonates (inferred to be clasts of the Pennsylvanian–Permian Oquirrh Group) with lesser amounts of 2–5-cm-diameter angular, white, massive quartzite clasts, similar to the Ordovician Eureka Quartzite (Fig. 10D). Upsection the quartzite clasts increase in abundance, while small (1–2 cm in diameter) dark gray, mica-bearing carbonate clasts (inferred as clasts of the Fish Haven Dolomite) are also present, together with inferred clasts from the Oquirrh Group (Fig. 10D). The upper parts of unit 1 contain significant amounts of 1–4-cm-diameter subangular clasts of mica-bearing metamorphosed calc-silicate rock, inferred to be clasts of the metamorphosed part of the Ordovician Pogonip Group and the Proterozoic Schist of Mahogany Peaks (Fig. 10D). Thus, the stratigraphy of unit 1 records an unroofing sequence, where the oldest rock clasts become progressively more abundant upsection (Fig. 110D). No granitic clasts were identified (e.g., clasts derived from sources that include the Almo pluton or the Archean basement); therefore footwall igneous rocks were not exposed at the surface during the early phases of motion on the basin-bounding fault (ca. 13.5 Ma; this study).

Unit 1 grades abruptly upward into red to buff colored, coarse, occasionally cross-bedded sandstones of unit 2. The sandstones are composed primarily of moderately angular quartz, lithic fragments, feldspar, and abundant muscovite (Fig. 10), and are interbedded with matrix-supported pebble conglomerates with well-rounded 1–2 cm clasts of carbonate, quartzite, and schist. Collectively, the sandstones and conglomerates of unit 2 are interpreted as fluvial deposits. The abundant detrital muscovite in the sandstone, as well as the quartzite and schist clasts in the pebble conglomerate, indicate mixed sources for unit 2, which is interpreted to have been derived from detritus of Paleozoic carbonates and quartzites as well as metamorphosed Proterozoic strata. Granitic clasts derived from the Almo pluton or from the Green Creek Complex were not observed in unit 2.

Unit 2 grades upward into unit 3, which has a maximum thickness of ∼1200 m and is largely composed of lacustrine deposits (lower tuffaceous member of Williams et al., 1982). This unit is best exposed and described in fault blocks 2, 3, and 4 north of the Raft River Mountains (Fig. 9). A nearly continuous ∼1.2 km thick exposure of unit 3 crops out on the southeast flank of Jim Sage Mountains, previously interpreted as part of the upper tuffaceous member (Williams et al., 1982). However, the sedimentary rocks mapped in this outcrop do not contain rhyolite clasts and are more highly tilted than the overlying rhyolite flows. The sedimentary strata consistently dip ∼32° to the east (blue symbols in Fig. 10C), while the rhyolite flows dip at most ∼18° to the east. Furthermore, the contact of the sedimentary rocks with the rhyolite flows indicates that the volcanic rocks flowed on top of, and disrupted, the sedimentary rocks. These observations indicate that the exposure of sedimentary rocks strata here is in fact older than the rhyolites, and that its lithology is consistent with that of unit 3. In general, the lacustrine deposits contain intercalated beds (to several tens of meters thick) of (unit 3-a) greenish-weathering, well-sorted ash-fall tuffs with abundant glass shards and evidence of soft-sediment deformation; (unit 3-b) off-white, well-sorted and sometimes cross-bedded calcarenite, composed of subrounded grains of quartz and calcite; (unit 3-c) yellow to buff weathering cross-bedded, fluvial calcarenites with quartz and calcite sand (unit 3-d), fine-grained, thinly laminated chalks and marls composed of smectite, and fine calcite with abundant pyrite, occasional gypsum, and bedding surfaces that exhibit abundant mud cracks (Fig. 10). The presence of gypsum and mud cracks indicates the existence of shallow ephemeral playa lakes in the Raft River Basin during that time.

The base of unit 4 is a dark weathering, incipiently to moderately welded ignimbrite, exposed ∼7–8 km west of the Narrows, that contains phenocrysts of plagioclase, clinopyroxene, and Fe-Ti oxides. Stratigraphically above, but never in direct contact with the ignimbrite is a thick sequence (∼450 m) of massive, glassy rhyolitic lava flows with occasional columnar jointing. The phenocrysts of these lavas consist of plagioclase (andesine), augite, and pigeonite (Figs. 10A, 10E).

Unit 5 is a lacustrine sedimentary unit, composed of thickly bedded calcarenite and tuff, with thin beds of volcanic sandstone composed of glass shards and lava flows or breccias containing clasts derived from underlying unit 4 (Figs. 10A, 10E).

The bottom part of unit 6 is in sharp contact with unit 5, and is a massive rhyolite lava flow and locally an autobrecciated ignimbrite with abundant pepperites, which provide evidence that the lavas and the ignimbrite flowed in water or wet sediment. The rest of unit 6 is made up of ∼450 m of intercalated massive, glassy, lava flows, thinly bedded lavas, densely welded ignimbrites, and columnar-jointed rhyolite flows (Figs. 10A, 10E). These rocks contain plagioclase (andesine), augite, pigeonite, and minor quartz phenocrysts. The top of unit 6 is a capping sequence of basaltic flows (basalt of northern Cotterel Mountains of Williams et al., 1982), to ∼100 m thick, exposed in the northern Cotterel Mountains (Figs. 3 and 10) and dated as 9.2 ± 1.5 (1σ) Ma (K-Ar; Armstrong et al., 1975, recalculated as 9.25 ± 1.5 Ma with new decay constants; but the standard used in their study was not reported). The volcanic rocks of units 4 and 6 are coeval with, and have phenocryst assemblages similar to, volcanic rocks of the central Snake River Plain (e.g., Perkins and Nash, 2002; Ellis et al., 2010). We interpret these rhyolites to represent the southern extent of the Miocene magmatic province of the Snake River Plain (Konstantinou, 2011).

Unit 7 is very poorly exposed along the flanks of the Cotterel and Black Pine Mountains (Figs. 3 and 8), and is mostly composed of thickly bedded reworked tuffs that include clasts of rhyolite from units 4–6. Where unit 7 is exposed on the western flank of the Black Pine Mountains, it is composed of intercalated conglomerate containing angular clasts of the Oquirrh Group, and medium- to coarse-grained sandstone (Smith, 1982; Wells, 2009). The sedimentary strata are tilted and dip as much as 30° to the east (Fig. 8). Their tilts toward the range front, as well as the brecciated nature of the contact with the Oquirrh Group, led us to reinterpret the contact as a west-dipping normal fault. Even though the exposures of unit 7 west of the Black Pine Mountains are lithologically similar to the description of unit 1 (this study), their age is likely younger than 9 Ma, because rhyolites similar to the 9.5–8.2 Ma Jim Sage volcanic rocks are described in the nearby Strevell borehole. When these volcanic rocks are projected to the surface, they appear to underlie the conglomerates adjacent to the Black Pine Mountains, thus making the conglomerates equivalent in age to unit 7 of this study (Fig. 8).