The Jim Sage volcanic suite (JSVS) exposed in the Jim Sage and Cotterel Mountains of southern Idaho (USA) consists of two volcanic members composed of ∼240 km3 of Miocene rhyolite lavas separated by an interval of lacustrine sediments. It is capped by rheomorphic ignimbrite and as much as 100 m of basaltic lava flows probably derived from the central Snake River Plain (SRP) province to the north. The occurrence of volcanic vents in the JSVS links the lava flows to their local eruptive centers, while the adjacent Albion–Raft River–Grouse Creek metamorphic core complex exposes ∼3000 km2 of once deep-seated rocks that offer constraints on the composition of the potential crustal sources of these rhyolites. U-Pb zircon ages from the rhyolite lavas of the JSVS range from 9.5 to 8.2 Ma. The Miocene basalt of the Cotterel Mountains has an 87Sr/86Sri composition of 0.7066–0.7075 and ɛNd(i) = –3.7, and the rhyolite lavas of the JSVS have 87Sr/86Sri = 0.7114–0.7135 and εNd(i) values that range from –6.7 to –7.1. Zircon from the rhyolites of the JSVS range in δ18Ozr (Vienna standard mean ocean water, VSMOW) from –0.5‰ to 5.7‰ and have εHf(i) values ranging from –0.8 to –6.8. Based on geochronology, whole-rock major elements, trace elements, isotopes (Sr and Nd), and in situ zircon O and Hf isotopic compositions, we infer that the JSVS is genetically related to the central SRP province. The eruption of the low-δ18O rhyolites of the JSVS, outside of the main topographic extent of the SRP province (without the large calderas inferred for the SRP rhyolites) implies that there might be an alternative mechanism for the formation of the low-δ18O signature other than the proposed assimilation of hydrothermally altered caldera blocks. One suggestion is that the north to south propagation of SRP-type low-δ18O rhyolitic melt along the Albion fault led to off-axis magmatism. Another possibility is that there was prior and widespread (across a region wider than the SRP) hydrothermal alteration of the crust related to its earlier magmatic and faulting history.
The eruption of SRP-type lavas in the hanging wall of an evolving metamorphic core complex helps us outline the role of the SRP magmatic province in the extensional evolution of the northeastern Basin and Range. The lavas of the JSVS imply the addition of basalt, related to the SRP hotspot, to the crust beneath the Raft River Basin that provided a heat source for crustal melting and weakening of the deep crust; this led to a vertical component of crustal flow and doming during extension, after the eruption of the 9.5–8.2 Ma JSVS rhyolites. This younger than 8.2 Ma component of vertical motion during faulting of the Miocene stratified sequence of the Raft River Basin and the rotation of the Albion fault to shallower angles collectively resulted in the subhorizontal detachment structure imaged seismically beneath the Raft River Basin.
Large silicic igneous complexes are commonly associated with continental flood basalt provinces (e.g., Paraná-Etendeka and North Atlantic igneous provinces; Bryan et al., 2002; Bryan, 2007), yet their origin is one of the fundamental debates in igneous petrology and geodynamics. This debate is long standing due to the complex nature of how large silicic magmatic systems are generated and magma stored for large periods of time in heterogeneous continental crust. Understanding the origin and longevity of large silicic complexes is of great importance because this magmatic activity is often associated with mineral deposits and because silicic caldera-forming supereruptions pose major safety and economic risks to society.
In the western United States, the Snake River Plain–Yellowstone (SRP-Yellowstone) vol canic province is a prominent feature, defined by low-relief topography extending ∼750 km from northwestern Nevada to Yellowstone National Park (Wyoming, Montana, and Idaho; Fig. 1 inset; e.g., Pierce and Morgan, 1992, 2009; Coble and Mahood, 2012). It is one of the most productive young silicic igneous provinces on Earth, active from ca. 16.5 Ma to the present day, generating large volumes (tens of thousands of cubic kilometers) of mostly basaltic and rhyolitic magmas (e.g., Christiansen, 2001). Petrologic investigations of silicic, intensely welded outflow ignimbrites (e.g., Branney et al., 2008) surrounding the SRP have demonstrated that the SRP-Yellowstone province is capable of producing large volumes of rhyolitic ignimbrite (>1000 km3) that can travel over great distances (e.g., Andrews and Branney, 2011; Ellis et al., 2011). Individual silicic eruptive centers in the central portion of the SRP are largely inferred because they are mostly buried beneath younger basalt flows, making them inaccessible for direct study. Ongoing borehole drilling studies within the SRP are trying to address some of these problems by sampling the volcanic section directly (e.g., Shervais et al., 2006; Project HOTSPOT, http://www.usu.edu/geo/shervais/Shervais-USU-Geology/Project_Hotspot.html).
Fewer studies have focused on the smaller (<300 km3) eruptive centers adjacent to the SRP province that are not associated with caldera-forming eruptions but are more accessible to study. One of the best examples of such an eruptive center is the Jim Sage volcanic suite (JSVS; Fig. 1), developed within the evolving Miocene Raft River Basin (Konstantinou et al., 2012). Volcanic vents in the JSVS link the lava flows to their eruptive centers, and the adjacent Albion–Raft River–Grouse Creek metamorphic core complex exposes ∼3000 km2 of once deep-seated rocks that offer constraints on the composition of the basement upon which the central SRP was built. This study focuses on the petrogenesis of the volcanic rocks of the JSVS using zircon U-Pb geochronology, whole-rock major element, trace element, and isotopic compositions (Sr and Nd), and in situ zircon O and Hf isotopic compositions.
REGIONAL GEOLOGIC SETTING
Cenozoic Extensional and Magmatic History of the Basin and Range Province
Following the end of regional folding and thrust faulting during the Sevier and Laramide orogenies in the Cretaceous through early Cenozoic (e.g., Burchfiel et al., 1992; DeCelles et al., 1995), Paleocene–Oligocene volcanic rocks erupted across the western U.S., forming a regional pattern of magmatism that becomes younger in a southward direction across the northern Basin and Range (e.g., Christiansen and Lipman, 1972; Stewart et al., 1977; Best and Christiansen, 1991; Christiansen and Yeats, 1992; Fig. 1). Miocene extension in the Basin and Range province has been documented by crosscutting relations, the stratigraphy and age of synextensional basin fill deposits, and low-temperature thermochronology of fault blocks. These data indicate that faulting began after ca. 21 Ma and accelerated at 17–16 Ma in the central part of the northern Basin and Range (Miller et al., 1999; Stockli, 1999; Colgan et al., 2006a, 2006b, 2010; Colgan and Henry, 2009). Extensional faulting is younger at the edges of the province; rapid slip on faults began ca. 15–10 Ma in northwestern Utah and southernmost Idaho (Wells et al., 2000; Egger et al., 2003; Konstantinou et al., 2012) and in northwestern Nevada (Colgan et al., 2006b, 2008; Egger et al., 2010).
While magmatism in the Miocene SRP-Yellow stone province is largely synchronous with Miocene to present-day extension in the Basin and Range province, few studies have tried to explain how Basin and Range extension is accommodated in and across the SRP-Yellowstone province (e.g., Rodgers et al., 1990, 2002; Anders and Sleep, 1992; Parsons et al., 1998; Fig. 1). In addition, there have been very few studies focused on the relation of magmatic addition from the SRP-Yellowstone province to the deep and shallow part of the crust in the region of the Basin and Range province and/or its effects on the extensional evolution of the Basin and Range (e.g., Rodgers and McCurry, 2009). The ∼240 km3 rhyolitic deposits of the JSVS are intercalated with deposits of the Raft River Basin, which formed due to normal faulting associated with the Miocene exhumation of the Albion–Raft River–Grouse Creek metamorphic core complex (Konstantinou et al., 2012). Thus, the JSVS provides both timing constraints and insight as to the possible genetic relationships between magmatism and extensional faulting in the Basin and Range province.
Magmatism in the Snake River Plain
The SRP-Yellowstone is a bimodal province characterized by tholeiitic basalts and hot and dry rhyolitic magmas (e.g., Branney et al., 2008; Christiansen and McCurry, 2008; Fig. 1 inset) that have isotopic compositions that require a significant mantle component (e.g., Hildreth et al., 1991; Nash et al., 2006; McCurry and Rodgers, 2009). The province exhibits an early phase of silicic magmatism that marks a northeast time-transgressive trend from northwest Nevada to Yellowstone, inferred to reflect the passage of the North American plate over a stationary hotspot (see discussions in Morgan et al., 1984; Hooper et al., 2007; Graham et al., 2009). Initial silicic magmatism in the well-exposed High Rock caldera complex, the McDermitt complex, and other dispersed eruptive centers coincided with the early phases of Columbia flood basalt volcanism ca. 16.5 Ma (Coble and Mahood, 2012).
One of the hallmarks of the SRP-Yellowstone province is the abundance of low-δ18O rhyolite magmas, first noted in Yellowstone by Friedman et al. (1974). An estimated >10,000 km3 of low-δ18O rhyolites erupted from the SRP-Yellowstone in the past 15 m.y. (e.g., Hildreth et al., 1984; Bindeman and Valley, 2001; Boroughs et al., 2005; Bindeman et al., 2007; Cathey et al., 2007, 2011). The most widely supported explanation for low-δ18O rhyolites [δ18Owr (whole rock) ∼1‰–4‰] requires a significant component of oxygen to be ultimately derived from negative δ18O meteoric water (–10‰ to –18‰), and therefore requires assimilation (e.g., Taylor, 1986) or remelting (e.g., Bindeman and Valley, 2001) of low-δ18O (2‰ to <–3‰) hydrothermally altered rock. The occurrence of low-δ18O rhyolite in Yellow stone has been explained by the remelting of older collapsed hydrothermally altered caldera blocks (e.g., Taylor, 1986, Bindeman and Valley , 2001). This model was then adopted for the central SRP, where large-scale remelting of buried caldera blocks resulted in the formation of low-δ18O rhyolites (e.g., Bindeman et al., 2007; Watts et al., 2011). In the central SRP, the model implies that the hydrothermal alteration of the upper crust was synchronous with SRP magmatism and was localized above the large magma chambers within the topographic expression of the province. It also predicts that the ignimbrites and early postcaldera lavas associated with the first collapse should show no effect, and there may be a decrease in the δ18O signature of the ignimbrites following multiple cycles of caldera collapse; as reported from the Heise eruptive center of the central SRP (Watts et al., 2011). Later studies from the Picabo eruptive center suggested that that hydrothermal preconditioning during Basin and Range extension, followed by caldera collapse, may have resulted in the generation of overabundant low-δ18O rhyolites (e.g., Drew et al., 2013). An alternative view suggests that the low-δ18O rhyolites in the central SRP-Yellowstone rhyolites (Owyhee-Humboldt, Bruneau-Jarbidge, Twin Falls, and Picabo eruptive centers; Pierce and Morgan, 1992) were produced by melting of crustal rocks that underwent hydrothermal alteration before SRP-Yellowstone magmatism (e.g., Boroughs et al., 2005; Leeman et al., 2008), such as altered Idaho batholith rocks (Criss and Taylor 1983; Criss et al., 1984). This view implies that the low-δ18O magmas reflect a signature of the crustal source that underwent hydrothermal alteration before the Miocene, and thus the alteration may not be limited to just the SRP caldera complexes.
Our study of the JSVS (Figs. 2 and 3) offers insights to the origin of the low-δ18O rhyolites as it represents an eruptive center outside the topographic expression of the SRP province and one that is directly linked to its volcanic vents. In particular, determining the O isotope composition of the rhyolite lavas of the JSVS helps to constrain the possible age and origin of upper crustal hydrothermal alteration. In addition, the adjacent Albion–Raft River–Grouse Creek metamorphic core complex exposes ∼3000 km2 of once deep-seated rocks that may have had similarities in terms of age and composition to the premagmatic crust that underlies the SRP province (Fig. 1). The O-Hf isotope composition of zircon and Sr-Nd whole-rock isotopic compositions from proximal igneous and metamorphic basement rocks exposed in the metamorphic core complex have been well studied (Fig. 1; Strickland et al., 2011b; Konstantinou et al., 2013). Recent hydrogen isotope studies of ductilely deformed rocks in the Raft River Mountains show evidence for rock-meteoric fluid interaction near and along what was once the brittle-ductile transition zone in the crust. These have been interpreted to indicate circulation of meteoric water synchronous with extensional deformation and coeval with development of the mylonites within the metamorphic core complex (Gottardi et al., 2011).
Geologic Mapping and Data Compilation
The JSVS crops out in the Jim Sage and Cotterel Mountains on the eastern side of the Albion Mountains, which form the northern extent of the Albion–Raft River–Grouse Creek metamorphic core complex (Figs. 1 and 2). The JSVS overlies lacustrine sedimentary rocks and is composed of two members of rhyolitic lavas separated by a sedimentary interval (Figs. 2 and 3). It is capped by ignimbrites and basaltic flows (basalt of Cotterel Mountains [Williams et al., 1982]; Fig. 2) that likely represent outflows from the central SRP. Volcanic rocks of the JSVS and associated sedimentary rocks were mapped and described as part of the Salt Lake Formation by Compton (1972, 1975), Williams et al. (1982), Covington (1983), Pierce et al. (1983), Wells (2009), and us (Konstantinou et al., 2012). Borehole, well log, and geophysical (seismic reflection, gravity, and magnetic) data initially collected for the Raft River geothermal project during the early to mid-1970s (Mabey and Wilson, 1973; Williams et al., 1974, 1982; Covington, 1983) are available for the greater Raft River Basin area, and were compiled for the purpose of this study (Konstantinou et al., 2012; Konstantinou, 2013). The Miocene Salt Lake Formation (and the JSVS) accumulated in a topographic depression or major half-graben system that formed as a response to slip along the Albion normal fault system, which bounds the basin to the west (Figs. 1 and 2) (Konstantinou et al., 2012). The onset of extensional fault slip is recorded by the age of ash-fall tuffs deposited near the base of the Miocene section ca. 14 Ma prior to the eruptions of the JSVS rhyolites (Konstantinou et al., 2012; Konstantinou, 2013; Fig. 2; part of lower lacustrine sediments).
Geologic mapping and stratigraphic investigations of the Jim Sage and Cotterel Mountains were carried out between 2009 and 2012, with attention to the basal contacts of the rhyolite flows (Figs. 2–5). These data were used to modify existing maps by Williams et al. (1974) and Pierce et al. (1983) to produce the map in Figure 2 and to better define the stratigraphy of the JSVS (Fig. 3).
In Situ Zircon Sensitive High-Resolution Ion-Microprobe–Reverse Geometry U-Pb Geochronology
In situ zircon U-Pb analyses (n = 137) from five samples from the JSVS were carried out using the sensitive high-resolution ion-microprobe–reverse geometry (SHRIMP-RG) at the Stanford-U.S. Geological Survey Micro Analysis Center (SUMAC). Two samples were collected from the lower member of the JSVS (samples JS11-B, JS1), two samples from the upper member (samples C-RC and C1B), and one from a subvolcanic intrusion (sample C1; Table 1; Figs. 2 and 3; Supplemental Table 11). Zircon was separated by crushing and various separation methods (Gemini table, heavy liquids, Frantz magnetic separator). Zircon was hand-selected for final purity, mounted in a 25-mm-diameter epoxy disc, ground and polished to a 1 µm finish, and imaged in cathodoluminescence (Fig. 6) using a JEOL 5600 scanning electron microscope. U, Th, and Pb isotopes were measured on the SHRIMP-RG (using the procedure described in Strickland et al., 2011b), adjusting the count times for each isotope and the number of five scans (peak-hopping cycles from mass 156 through 270) to six to improve counting statistics for 206Pb and 207Pb, required to obtain Miocene ages with acceptable analytical precision (∼1%–2% for individual analyses, 1σ).
The majority (n = 121) of zircons analyzed have concordant U-Pb ages, had low common Pb, and were interpreted as magmatic zircon that crystallized just prior to eruption of the volcanic rocks. The data were reduced using SQUID 1 (Ludwig, 2003) and concordia diagrams were constructed using Isoplot (Ludwig, 2003). Calculated ages for zircon are standardized relative to R33 (419 Ma; Black et al., 2004). The common Pb composition was based on the value reported by Stacey and Kramers (1975). The calculated concordia intercept ages (Fig. 7) and weighted average 207Pb-corrected 206Pb/238U model ages are reported with 2σ errors (Table 1; Supplemental Table 1 [see footnote 1]). The weighted average ages of the five samples are preferred as the inferred crystallization ages and are all within error of the concordia intercept ages. Zircon concentration data for U and Th were standardized against the well-characterized, homogeneous zircon standard MAD-green (Barth and Wooden, 2010).
Whole-Rock Major and Trace Element Geochemistry
We selected 37 samples for whole-rock major and trace element analyses (Table 1; Fig. 8; Supplemental Table 22), 15 samples from the lower member of the JSVS, 11 from the upper member, and 6 samples of ignimbrite and lava, the stratigraphic positions of which with respect to the JSVS rhyolite flows are not certain due to lack of continuous exposure (outflows). In addition, four samples of the basalt of Cotterel were analyzed. The samples were analyzed using X-ray fluorescence at Washington State University (Pullman, Washington, USA; 13 samples) and Macalester College (St. Paul, Minnesota, USA; 24 samples); 13 samples were dissolved and analyzed for trace and rare earth elements using inductively coupled plasma–mass spectrometry (ICP-MS) at Washington State University (Fig. 9). Analyses of 12 additional samples of Precambrian basement rocks are also included in Figure 8 (compiled from Egger et al., 2003; Strickland et al., 2011b; Konstantinou et al., 2013).
Electron-Microprobe Phenocryst Geochemistry
Three samples from the lower (JS-3, JS-7, JS-10) and three from the upper member (C-RA, C-RD, C-RE) of the JSVS were selected for electron-microprobe analyses of phenocryst chemistry assemblages (Fig. 10; Supplemental Table 33) using the JEOL JXA-733A electron probe at Stanford University (Stanford, California). Pyroxene and feldspar phenocrysts were selected and analyzed for the abundance of major element oxide compositions. The oxide abundances were standardized using well characterized in-house mineral standards and the results were normalized to six oxygen atoms for pyroxene and eight for feldspar (Supplemental Table 3 [see footnote 3]).
We selected 15 samples from the JSVS for whole-rock Sr and a subset of 7 samples for Nd analyses performed at the Stanford ICP-MS-TIMS (thermal ionization mass spectrometry) facility (Table 1; Fig. 11; Supplemental Table 44). For more details about the procedures used for sample dissolution and mass spectrometry, see Appendix 1.
The O isotope compositions of individual zircon domains, dated by SHRIMP-RG (n = 70), from three samples (JS1, JS11-B, C1B) were determined in situ using an IMS-1280 at the University of Wisconsin WiscSIMS (Wisconsin Secondary Ion Mass Spectrometer) laboratory following the procedures described in Kita et al. (2009) and Valley and Kita (2009), and δ18O (Vienna standard mean ocean water, VSMOW) values were calculated with precision reported at 2 standard deviation (SD) level (see Appendix 2 for details of the method; Fig. 12A; Supplemental Table 55). A 133Cs+ beam was focused to ∼10 μm spots and secondary ions of 18O and 16O were counted with dual Faraday cup detectors. Groups of ∼10 sample analyses were bracketed by a total of at least 8 analyses of KIM-5 zircon standard (Valley 2003), which are used to correct for instrument bias. The average value of bias was –0.31‰ and ranged from –0.10‰ to –0.54‰. The reproducibility of these groups of standard data, which varied from 0.19 to 0.52, permil ‰ is considered the best indication of analytical precision, and is reported at 2 SD. For the purpose of this study, we have used the δ18Ozr (zr—zircon) values to calculate δ18Owr using the equation rearranged from Lackey et al. (2008): δ18Owr ≈ δ18Ozr + 0.0612 (wt% SiO2) – 2.5.
Hf isotopic compositions were determined in situ on the same zircons that were dated and analyzed for O isotope compositions, from five samples of the JSVS (n = 56). The analyses were performed at Washington State Uni versity’s LA-ICP-MS facility (LA—laser ablation), following the procedure of C.M. Fisher (2013, personal commun.); for detailed methods, see Appendix 3. Individual zircon crystals from two lower member samples (samples JS11-B, JS-1), two upper member samples (C1B, C-RC), and a subvolcanic intrusion sample (sample C1) were analyzed (Fig. 12B; Supplemental Table 66).
Stratigraphy of the Jim Sage Volcanic Suite
Distinguishing between high-temperature rheomorphic ignimbrites and rhyolite lavas that erupted from the SRP is challenging because the intense welding of these ignimbrites gives them a lava-like appearance (Branney et al., 2008). Differentiating between individual ignimbrite units is equally difficult due to their similar appearance and geochemistry (e.g., Hildreth and Mahood, 1985). Many workers have used high-precision geochronology, trace element geochemistry, and phenocryst geochemistry to correlate individual ignimbrite units and estimate the volumes of individual eruptions (e.g., Hildreth and Mahood, 1985; Bonnichsen et al., 2008; Ellis et al., 2012). In this study, we used the physical characteristics of the volcanic deposits (thickness, aspect ratios, terminations of the flows and basal contacts; outlined by Branney et al., 2008) to differentiate between lavas and ignimbrites in the JSVS.
The rhyolite flows of the JSVS are exposed along two north-south elongate antiformal structures that make up the rugged Jim Sage and Cotterel Mountains (Fig. 2). The JSVS is composed of two members separated by a thin sequence of sediments (Fig. 3). A dark weathering, moderately to densely welded ignimbrite is exposed in the Raft River Basin, ∼10–11 km southwest of the Jim Sage Mountains, but these exposures are separated by alluvial cover from the lava flows exposed at the Jim Sage Mountains (Fig. 3). The base of the lower member overlies fine-grained lacustrine deposits of the Salt Lake Formation (Konstantinou et al., 2012; Fig. 3). Underlying sedimentary rocks locally preserve evidence of soft-sediment deformation and high-temperature fusion at the contact with the lower rhyolite lava flow (Fig. 3). The two members have nearly constant thickness (370–440 m for each member) along their strike length, except at their northern and southern exposures, where they taper rapidly and have ∼30–50-m-thick terminal lobes. The two members are exposed over areas of ∼260 km2 and have aspect ratios between 1:40–1:50, and calculated volumes, based on cross sections and map patterns, ranging from 110 to 130 km3 (Fig. 2).
The lower member is best exposed in the canyons along the eastern side of the Jim Sage Mountains (Fig. 2). The base of the member consists of a 5–30-m-thick rhyolitic autobreccia, with vitreous clasts and a matrix of strongly altered glass (Figs. 3, 4A, and 4G). The rest of the lower member is made up of a thick (350–400 m) sequence of massive or layered, glassy rhyolite lava with occasional columnar jointing and devitrified interiors (Fig. 3). These rocks have aphanitic or glassy matrix and 20%–25% phenocrysts of euhedral plagioclase (andesine) > anhedral augite and pigeonite > Fe-Ti oxides > quartz > sanidine >> zircon, and are similar to the phenocryst compositions of rhyolite erupted from the SRP province (e.g., Bonnichsen and Citron, 1982; Honjo et al., 1992; Andrews et al., 2008; Ellis and Wolff, 2012; Fig. 4A).
The lower and upper members are separated by a relatively thin (0–100 m) sedimentary breccia containing clasts derived from the underlying lower member (Fig. 3). The breccia is topped by a ∼50–100-m-thick lacustrine sedimentary unit composed of thickly bedded calc-arenite and tuff, with thin beds of volcanic sandstone composed of glass shards (Fig. 3).
The upper member is best exposed in the Cotterel Mountains, and has a locally auto brecciated base with occasional peperites and clasts of the older rhyolite lavas (Figs. 3 and 4E). The peperites provide evidence that the lavas flowed in water or wet sediment. The majority of the upper member is made up of ∼400 m of intercalated massive and banded lava with thick sequences characterized by columnar joints (Fig. 3). These lavas are similar in appearance to the lower member unit, with aphanitic or glassy matrix and 20%–25% phenocrysts composed of euhedral plagioclase (andesine) > anhedral augite and pigeonite > Fe-Ti oxides > quartz > sanidine >> zircon (Figs. 4B, 4E, 4F). Thus, the basal contacts of both the lower and upper member are characterized by autobreccias, typical of silicic lavas, rather than a basal surge deposit characteristic of a welded ignimbrite (Branney et al., 2008; Figs. 4A, 4B, 4E–4G). The thicknesses (lack of exposures with thickness of <30 m), aspect ratios, and abrupt lobate terminations are all consistent with the physical attributes of silicic lava flows of the SRP (Branney et al., 2008; Fig. 2). In addition, based on seismic reflection data and borehole data, it was demonstrated (Konstantinou, 2013) that these rhyolites are not widely present in the Raft River Basin, buried beneath younger sediments. Instead, the eastern limit of the rhyolite flows is their approximate present-day exposure in the Jim Sage and Cotterel Mountains. Had these rhyolite flows been ignimbrites, one would expect they would have spread out more thinly to fill the east-west extent of the floor of the Raft River Basin, but the rhyolites were not sampled in the deep boreholes of the Raft River Basin that penetrated the lower plate of the metamorphic core complex (Williams et al., 1974, 1982; Covington, 1983).
Small subvolcanic intrusions (vents) and lava domes are also present. These are identified by their silicified interiors, and the fact that they dome the strata they were intruded into (Figs. 2 and 5). Based on a positive magnetic anomaly that covers the area of exposed rhyolite domes (Fig. 2) and on seismic reflection profiles that image what appear to be intrusions (Mabey and Wilson, 1973; Williams et al., 1974, 1982; Covington, 1983; Konstantinou, 2013), an elliptical (∼4 by 2 km) cryptodome probably underlies the eastern portion of the Jim Sage Mountains (Fig. 2). We infer the exposed domes and the cryptodome to be part of the vent system for the lavas of the JSVS.
Above the upper rhyolite member of the JSVS is a thin (<100 m), densely welded ignimbrite with an interior zone that displays intense rheomorphic flow and transposed and/or deformed shards (Fig. 4C). A less welded distal phase of this ignimbrite preserves small collapsed pumice lapilli (Figs. 4C, 4D). This ignimbrite covers most of the northern Cotterel Mountains, forming a relatively smooth dip slope on its northwestern side (Fig. 2). The ignimbrite is also exposed in the hills near the town of Albion (Fig. 2). Above this ignimbrite is a capping sequence of basaltic flows exposed in the northern Cotterel Mountains (basalt of Cotterel Mountains of Williams et al., 1982), which is as much as ∼100 m thick, and was dated as 9.2 ± 1.5 Ma (1σ; K-Ar; Armstrong et al., 1975). Both the ignimbrite and the basalt of Cotterel are inferred to be erupted from the central SRP.
Lacustrine sediments conformably overlie the upper rhyolite flow, are poorly exposed along the flanks of the Cotterel Mountains, and are known from borehole localities in the eastern part of the Raft River Basin (Fig. 4). These deposits consist mostly of thick-bedded reworked ash-fall tuffs and marls and contain clasts of rhyolite from the JSVS.
SHRIMP-RG U-Pb Geochronology of Zircon
Sample JS11-B, from the lower member of the JSVS, is a two-pyroxene rhyolite lava containing euhedral zircon (Fig. 6A). The weighted mean of 23 207Pb-corrected 206Pb/238U ages yields an age of 9.53 ± 0.10 Ma (2σ) with a mean square of weighted deviates (MSWD) of 2.3 (Fig. 7A). Zircon analyses (n = 14) from a second sample, collected from a higher stratigraphic position of the lower member of the JSVS (sample JS1), yielded a weighted mean age of 9.34 ± 0.12 Ma (2σ; MSWD = 0.9; Fig. 7B). Three zircon grains resulted in slightly older U-Pb ages (ca. 10–11 Ma) and are interpreted as antecrystic (e.g., Bindeman et al., 2001). Sample C1, collected from the subvolcanic intrusion in the southern Cotterel Mountains, yields a weighted mean age of 9.19 ± 0.14 Ma (2σ; MSWD = 2.7; n = 26). The high MSWD may reflect two zircon populations with an irresolvable age difference (Fig. 7C).
Two samples from the upper member of the JSVS were also collected and dated. Sample C-RC is a two-pyroxene rhyolite lava that yielded a weighted mean age of 8.43 ± 0.12 Ma (2σ; MSWD = 1.5; n = 26; Fig. 7D). Four zircon grains resulted in slightly older ages (ca. 9.5 and 11 Ma) and are interpreted to be antecrystic, similar to observations from the Yellowstone caldera (Vazquez and Reid, 2002). A second, very similar sample collected from a higher stratigraphic position (C1B) yielded a weighted mean age of 8.19 ± 0.14 Ma (2σ; MSWD = 1.4; n = 14; Fig. 7E).
A small number of zircon cores (n = 9) from all five Miocene samples were identified and dated; these ages are plotted on the same concordia diagram (Fig. 7F). U-Pb isotope ratios from individual zircon define two discordia lines with poorly constrained upper intercepts ca. 2550 and 1745 Ma (±140–150 m.y.) and Mesozoic lower intercepts (ca. 75 and 95 Ma ± 270 m.y.; Fig. 7F). These data points are too few for any conclusive interpretation, but there are widespread exposures of Archean basement in the region dated as ca. 2550 Ma, and most Eocene–Oligocene plutons of the Albion–Raft River–Grouse Creek metamorphic core complex that have been studied have similar inherited ages (Strickland et al., 2011b; Konstantinou et al., 2011, 2013).
Results from Major and Trace Element Geochemistry
The lower member of the JSVS has a limited range in composition, varying from low-silica alkali rhyolite to low-silica rhyolite (SiO2 = 72%–74%; Fig. 8). The upper member of the JSVS is a low-silica alkali rhyolite with slightly less silica content (SiO2 = 70%–71.5%; Fig. 8) than the lower member. The four mafic samples are basaltic in composition and most of the outflow ignimbrite and lava samples (away from the Jim Sage and Cotterel Mountains) have compositions similar to that the lower member of the JSVS (Fig. 8). All of the JSVS samples are ferroan (high FeOt/MgO ratios), and distinctly different from the evolved calc-alkalic composition of the underlying Archean basement in the region (Fig. 8B). The rare earth element (REE) patterns of the rhyolite samples from the lower and upper members of the JSVS are very similar, and show strong enrichment in light (L) REEs relative to heavy (H) REEs, and large negative Eu anomalies (Fig. 9). The basalt of Cotterel Mountains has a small enrichment in LREEs relative to HREEs and a small positive Eu anomaly (Fig. 9).
Geochemistry of Phenocryst Assemblages
The lower and upper rhyolite members of the JSVS have andesine feldspar phenocrysts that are homogeneous in composition (Fig. 10). The two members of the JSVS have distinctly different pyroxene pair compositions. The lower member is characterized by an augite-pigeonite pyroxene pair, and the upper member of the JSVS has an augite-pigeonite pair that contains more MgO than the lower member (Fig. 10). The pyroxene pairs in the rhyolite lavas are consistent with being in equilibrium at high temperatures (∼850–900 °C; Lindsley, 1983) and low pressures (<3 kbar), and are similar to typical central SRP rhyolites (e.g., Bonnichsen and Citron, 1982; Honjo et al., 1992; Andrews et al., 2008; Ellis et al., 2010). In addition, a single population of augite and pyroxene pair in both flow units is consistent with similar observations of central SRP lavas, whereas ignimbrites from the province often have multiple populations of pyroxene pairs (Cathey and Nash, 2009; Ellis and Wolff, 2012).
Whole-Rock Multicollector-ICP-MS Sr and Nd Isotope Analyses
The whole-rock Sr and Nd isotope results were used to calculate age-corrected 87Sr/86Sri and εNd(i) values (Fig. 11; Supplemental Table 4 [see footnote 4]; Bouvier et al., 2008). The lower member has 87Sr/86Sri = 0.7114–0.7121 and εNd(i) values that range from –6.7 to –7.1; the upper member has 87Sr/86Sri = 0.7130–0.7135 and εNd(i) values of –7.1 (Fig. 11). The basalt of Cotterel Mountains has 87Sr/86Sri composition of 0.7066–0.7075 and εNd(i) = –3.7 (Fig. 11; Supplemental Table 4 [see footnote 4]). The isotopic compositions of the rhyolite lavas of the JSVS and the basalt of Cotterel Mountains are very similar to published data from rhyolites and basalts of the central SRP (Fig. 11; Nash et al., 2006; Christiansen and McCurry, 2008).
In Situ Zircon Secondary Ion Mass Spectrometer O Isotope Analyses
The O isotope compositions of individual analytical spots from Miocene JSVS zircons from the lower member range in δ18Ozr (VSMOW) from –0.5‰ to 4.6‰, and from the upper member range from 0.5‰ to 4.9‰ (Fig. 12A; Supplemental Table 5 [see footnote 5]). However, most of the zircon analyses cluster around values of δ18Ozr (VSMOW) ranging from 0‰ to 2.5‰ (43 of 58 analyses), with only one analysis having a δ18Ozr value >5‰. The whole-rock oxygen isotope compositions calculated from δ18Ozr range from δ18Owr ≈ 6.7‰ to 1.4‰ (but mostly cluster around 2‰–4.5‰) and are similar to values reported from the 12–6 Ma rhyolites of the central SRP (see compilation in Watts et al., 2010; Fig. 12A). A few large (∼200 μm in diameter) zircon grains were analyzed with multiple secondary ion mass spectrometer spots, and the results show isotopic zoning within individual grains (Figs. 6A, 6B). In this limited number of examples, we observed a decrease in δ18Ozr from core (δ18Ozr ∼ 4‰) to rim (δ18Ozr ∼ 2.2‰), similar to observations from Yellowstone zircon reported in Bindeman et al. (2008).
In Situ Zircon LA-ICP-MS Hf Isotope Analyses
The Hf isotope results and the calculated εHf(i) values from the upper and the lower members of the JSVS are summarized in Figure 12B (Supplemental Table 6 [see footnote 6]). The Hf isotope analyses yielded zircon εHf(i) values ranging from –0.8 to –6.8, but most of the analy ses have a small range from ∼–4 to –6.5 and the values obtained from each individual sample are homogeneous (Fig. 12B; Supplemental Table 6 [see footnote 6]). The lower member of the JSVS has zircon εHf(i) values that range from –0.8 to –6.8, and the upper member has a range of values from –1.6 to –6.6 (Fig. 12B). The total range of values within each unit is larger than the 2σ error on individual zircon analysis (±1.5 εHf units). Zircons with εHf(t) values from –0.8 to –2.1 are suggested to be antecrystic because they have Hf isotope compositions that are more radiogenic than the majority of zircon analyses, which cluster around –4 and –6.5 εHf(t) values (Fig. 12B).
Possible Crustal and Mantle Sources for Geochemical Modeling
Based on stratigraphic considerations alone, the Albion–Raft River–Grouse Creek metamorphic core complex represents a relative vertical uplift of a minimum of 10 km, from the top of the Paleozoic section to the Archean (Konstantinou et al., 2012). Thus the metamorphic core complex exposes extensive areas of once deep-seated rocks, similar to those that may underlie the central SRP and the JSVS. The Sr-Nd-Hf-O isotope compositions of several igneous and metamorphic rocks of different ages now exposed in the core complex are well characterized (Strickland et al., 2011b; Konstantinou et al., 2013). Here we summarize the stratigraphy and the crustal structure of rock units currently exposed in the Albion–Raft River–Grouse Creek metamorphic core complex. Because of their regional nature, these same rock units might be expected to be at depth beneath the SRP and southern Idaho. The highest stratigraphic and/or structural levels compose a thick (6–8 km) sequence of Paleozoic carbonates now partly exposed in the Black Pine Mountains (Fig. 1; Compton et al., 1977) and in mountain ranges surrounding the Albion–Raft River–Grouse Creek metamorphic core complex. It is unlikely that this upper part of the crust would be involved as a major source of contamination for Miocene magmas, because it is not thermodynamically favorable to melt and assimilate carbonate rocks in igneous systems. A thick sequence (∼5 km) of Neoproterozoic quartzites and schists is beneath this carbonate sequence and constitutes the base of the passive margin succession of western North America. Even though these rocks have been tectonically attenuated (to ∼1 km) in the Albion–Raft River–Grouse Creek metamorphic core complex (Fig. 1), they are probably thicker to the north (e.g., Compton et al., 1977; Compton 1972, 1975), and may have underlain extensive areas of southern Idaho. The schist units have compositions favorable for having undergone partial melting during Miocene magmatism. The structurally deepest and oldest rocks exposed today in the metamorphic core complex are Archean crystalline rocks of the Green Creek Complex, which constitutes depositional basement for the Neoproterozoic and Phanerozoic cover sequence. The Green Creek Complex is composed primarily of augen orthogneiss with lesser amphibolite, diorite, tonalite, and metasedimentary rocks (e.g., Compton, 1972, 1975), all of which have favorable compositions to undergo melting during Miocene magmatism. Rocks of similar Archean age and isotopic composition have been incorporated as xenoliths in basaltic magma of the SRP (Leeman et al., 1985). Finally the Albion–Raft River–Grouse Creek metamorphic core complex exposes a wide variety of Eocene–Oligocene calc-alkaline plutons that also have compositions favorable to undergo melting during Miocene magmatism. Similar Eocene plutons are exposed in the Challis volcanic field (e.g., Criss and Taylor , 1983; Criss et al., 1984), and the extensive magmatic roots of these Cenozoic plutons may underlie large areas of southern Idaho, including the region of the SRP province.
For the purpose of modeling the petrogenesis of the Miocene magmas (see following discussions), we made some simplifying assumptions about possible crustal reservoirs, assuming average isotopic compositions of the lower crust, upper crust, and a mantle source for the magmatic system (following the reasoning in Konstantinou et al., 2013) and summarized in the following. However, the structure and composition of the deep crust are likely more variable and complex than assumed here and shown in Table 2 and Figure 13.
Silicic rocks like the Proterozoic schist units, the Archean orthogneiss and the Eocene–Oligo cene plutons that intrude them, are the dominant rock type exposed in the Albion–Raft River–Grouse Creek metamorphic core complex (e.g., Egger et al., 2003; Strickland et al., 2011b; Konstantinou et al., 2013). We infer that similar rocks may underlie the whole region of southern Idaho, and we collectively refer to these potential sources and/or contaminants as upper crust (Fig. 13; Table 2). Specifically, we used the composition of Oligocene plutons to represent the composition of the upper crust reservoir. The bulk Sr-Nd isotopic composition of the upper crust was estimated to be 87Sr/86Sr10 = 0.716, and εNd(10) = –28 based on measured values from the Oligocene plutons (Table 2; Strickland et al., 2011b; Konstantinou et al., 2013; Fig. 13A). The εHf(10) of the potential upper crust reservoir was estimated to be εHf(10) = –32, based on values obtained in zircon from Oligocene magmas in the Albion–Raft River–Grouse Creek metamorphic core complex (Konstantinou et al., 2013; Table 2). The values selected for εHf(10) and εNd(10) do not follow the terrestrial array relationship established by Vervoort et al. (2011), but are based on the measured isotopic ratios of Oligocene plutons. Zircon δ18O values analyzed from the Oligocene plutons and the Archean basement were used to estimate whole-rock magmatic values. These calculated δ18Owr for the Archean and Oligocene rocks have a very narrow range of 6.5‰–7.5‰, with an average value of δ18Owr = 7.1‰ (Konstantinou et al., 2013). However the δ18Owr of the upper crustal potential crustal reservoir was estimated to be either –1.5‰, similar to values estimated for the SRP (e.g., Watts et al., 2011), or ∼2.5‰, similar to values measured from hydrothermally altered rocks of the Idaho batholith (Criss and Taylor, 1983; Criss et al., 1984). This low-δ18Owr value reflects postmagmatic hydrothermal alteration by meteoric waters in the region of southern Idaho. The cause of this alteration discussed in the following.
Eocene plutons in the Albion–Raft River–Grouse Creek metamorphic core complex are thought to have extensive deep roots, which are not exposed, based on the time span of Eocene magmatism and the compositions of these plutons (Konstantinou et al., 2013). In addition, the Archean basement is inferred to include rocks of more mafic compositions at depth, based on the isotopic composition of the Cenozoic igneous rocks (Strickland et al., 2011b; see detailed discussion in Konstantinou et al., 2013). A potential lower crust reservoir could be represented by the roots of Eocene plutons as well as by intermediate-composition Archean crystalline rocks. Based on the interpretation (made in Konstantinou et al., 2013) that the Eocene plutons in the Albion–Raft River–Grouse Creek metamorphic core complex represent reworking of the deep crust, we assume that the isotopic composition of the Eocene plutons now exposed to the surface may reflect the average composition of the potential lower crust reservoir. The Sr, Nd, O, and Hf isotopic composition of this potential lower crust reservoir was estimated to be 87Sr/86Sr10 = 0.7108, εNd(10) = –18, δ18Owr = 7.1‰, and εHf(10) = –22 based on values determined from the Eocene Emigrant Pass pluton (Table 2; Fig. 13; Konstantinou et al., 2013). This deep crustal reservoir has normal δ18Owr ≈ 7.1‰, similar to values calculated by analyzing zircon from the Eocene plutons (Table 2; Fig. 13B; Konstantinou et al., 2013). The normal δ18Owr of the lower crustal reservoir reflects the inability of meteoric fluids to infiltrate beneath the brittle-ductile transition zone (depths of ∼10 km; Person et al., 2007) and shift the oxygen isotope composition of the deep crust to lower and/or negative values.
The geochemical compositions of possible asthenospheric or lithospheric mantle reservoirs that contributed melts to the lower crust during the Cenozoic is poorly constrained, primarily because of the extensive contamination of mantle-derived basalts by lower crustal melts. However, He isotope compositions from Miocene and Pleistocene basalts within the SRP indicate that these magmas may be mixtures between a mantle plume and a lithospheric mantle reservoir (Graham et al., 2009). For this study, we estimated that the unmodified isotopic composition of the Miocene mantle-derived basalts represented the mean composition of three end-member mantle reservoirs, using the range of values summarized in Rollinson (1993). The three mantle reservoirs used were mantle plume and/or prevalent mantle, ancient enriched lithospheric mantle (lower εNd, εHf than plume value), and a fluid-enriched (higher 87Sr/86Sr than plume value) lithospheric mantle (Fig. 13; Table 2; references in Rollinson, 1993).
Petrogenesis of the Jim Sage Volcanic Suite
The similarity of zircon εHf(i) values measured from an exposed subvolcanic intrusion and the lava samples of the JSVS indicates that this (and possibly other) subvolcanic intrusion is part of the volcanic vent system for the eruption of the lavas of the JSVS (Fig. 12B). Thus, the occurrence of lava flows, lava domes, and subvolcanic intrusions within the Raft River Basin implies that the JSVS erupted within the basin, and the rhyolite lavas did not flow very far from their vents (<30 km). The northern part of the JSVS was capped by ignimbrite and basalt flows derived from the central SRP.
The age range of the rocks in the JSVS, together with their whole-rock major, trace element, and isotopic compositions and their zircon oxygen and Hf isotope compositions, are very similar to geochemical data sets collected on the high-temperature rhyolites from the central SRP (e.g., Bonnichsen and Citron, 1982; Honjo et al., 1992; Boroughs et al., 2005; Cathey et al., 2007, 2008; Branney et al., 2008; Christiansen and McCurry, 2008; Figs. 8, 11, and 12), suggesting that the JSVS is clearly a part of the SRP province. The high temperature of the rhyolite lavas (estimated from pyroxene compositions; Fig. 10) is consistent with a large flux of astheno sphere-derived basalt (hotspot) that in turn led to widespread crustal melting, as suggested for the magmas of the SRP province (e.g., Nash et al., 2006; McCurry and Rodgers, 2009).
Isotopic Modeling of the Miocene Magmas of the JSVS and Central SRP
Our extensive isotopic data set from once deep-seated rocks of the Albion–Raft River–Grouse Creek metamorphic core complex allows us to use these rocks as proxies for the composition of the upper and lower crust in modeling measured compositions in Miocene volcanic rocks. To model and evaluate the relative roles and components of mantle versus crustal material in the Miocene magmas of the JSVS and the central SRP province, we used the isotopic constraints provided by the upper and lower crust reservoirs (see discussion of possible crustal and mantle sources for geochemical modeling), and the equations of simultaneous assimilation and fractional crystallization (AFC) derived by DePaolo (1981). We also used the equation for estimating the ratio of mass of assimilated crust versus initial mass of magma from Aitcheson and Forrest (1994). We simultaneously modeled the isotopic data (Sr, Nd, O, and Hf) together with the Sr, Nd, and Hf elemental compositions and tried a range of geologically reasonable values for the independent variables, using A:FC rates (r value in DePaolo, 1981) and bulk partition coefficients (the final values used are reported in Table 2). The final Sr-Nd-O-Hf isotopic compositions of hybrid magmas predicted from our AFC model are compared to measured whole-rock values (87Sr/86Sri-εNd(i); Fig. 13A) and to the calculated δ18Owr and εHf(i) based on measured isotopic values from Miocene zircon (calculated δ18Owr and εHf(i); Fig. 13B).
We used a two-stage assimilation model for magma genesis (e.g., Gans et al., 1989; Grunder, 1993; Watts et al., 2010) that involves the high-temperature interaction of mantle-derived basalts with partial melts from the modeled normal δ18O (≈7.0‰) lower crust reservoir (see discussion of possible crustal and mantle sources; Table 2) and high (r = 0.96) A:FC rates, approaching zone-refining values, shown by the black lines in Figures 13A and 13B. The high A:FC rates (r = 0.96) were chosen to reflect the high flux of basaltic melts that intruded the lower crust during the Miocene. Based on this modeling, we suggest that the basalt of the Cotterel Mountains and other basalts in the SRP may be contaminated with small amounts (∼10%) of lower crust melts (Fig. 13), although the iso topic values for the basalt may also reflect contamination from a lithospheric mantle source.
This contaminated magma was further hybridized by variable amounts of partial melts from the modeled upper crust reservoir (red lines in Figs. 13A, 13B; isotopic values in Table 2), but at lower A:FC rates (r = 0.34). The upper crust assimilant has a negative δ18Owr ≈ –1.5‰, similar to values used in modeling of the SRP province (e.g., Watts et al., 2010, 2011). Our modeling suggests that the rhyolites of the JSVS and the SRP can be explained by the high-temperature interaction of crust-contaminated basalt (similar to the Cotterel basalt) followed by large amounts of assimilation of upper crust rocks. More specifically, our AFC modeling indicates that the Miocene rhyolites of the SRP may represent hybrid magmas composed of ∼60% mantle-derived basalt, and ∼40% crustal melts derived from two modeled crustal reservoirs; a normal δ18O (≈7.1‰) lower crust, and a negative δ18O (≈–1.5‰) hydrothermally altered upper crust reservoir (Fig. 13). The relatively low A:FC rates (r = 0.34) used in modeling the interaction of contaminated basalt with upper crustal rocks indicates high crystal fractionation rates, which were probably enhanced by the residence of rhyolitic magma chambers at shallow crustal levels (6–10 km). Based on regional stratigraphic studies in southern Idaho, the crust at 6–10 km may have been dominated by the presence of carbonate rocks (see discussion of possible crustal and mantle sources). However, depending on the extensional histories of the crust in this region, structurally deeper crystalline basement and overlying Neo protero zoic quartzites and schists may have been at shallower levels or 5–10 km deep (they are at the surface in the Albion–Raft River–Grouse Creek metamorphic core complex). Residence of rhyolitic magma chambers at shallow, relatively cool crustal levels (∼5–10 km) has been documented today by the geophysical imaging of low-velocity zones interpreted to be shallow crystal-rich mushes beneath Yellowstone (e.g., Chu et al., 2010).
The low A:FC rates used to model the rhyolites of the SRP and the JSVS indicate that these low-silica rhyolite may have evolved by ∼80% crystal fractionation (modeled at a partial melt fraction F = 0.2), coupled with assimilation (Fig. 8). Crystal fractionation helps to decrease the volume of the melt in the magmatic system undergoing AFC processes, thus producing low-δ18O values with smaller amounts of crustal component than those predicted by simple mixing models. To illustrate this, we were able to successfully model rhyolite magma of δ18Owr ≈ 2.0‰ using the AFC model and assimilating ∼30% crust with δ18Owr = –1.5‰ by ∼70% of already contaminated (by 10% of crustal melts) basalt with δ18Owr = 5.8‰ (Fig. 13B). If a simple mixing model is used with the same crust-basalt end members (δ18Owr = –1.5‰ and 5.8‰ for crust and contaminated basalt, respectively), and crust to contaminated basalt mixing ratios of 30%:70%, the resulting δ18Owr would be ∼3.6‰, and not ∼2.0‰ predicted by the AFC model.
Low-δ18O Rhyolites of the JSVS
The low-δ18O values of the JSVS rhyolite lavas that were erupted within the Raft River Basin, ∼50 km south of the SRP province, require the assimilation of hydrothermally altered crust at shallow levels as proposed for the SRP rhyolites. However, there are no known long-lived caldera systems within the Raft River Basin, thus the low-δ18O signature of the JSVS cannot be attributed to repeated caldera collapse events that drop hydrothermally altered rocks into the magmatic system (e.g., Bindeman and Valley, 2001; Bindeman et al., 2007; Watts et al., 2011; Drew et al., 2013). We suggest two ways to explain the low-δ18O signature of the JSVS lavas that erupted outside the SRP province. Each of these suggestions (model A and model B) (Fig. 14) has its strengths and limitations.
Model A: Low-δ18O Silicic Melt Propagation Along a Preexisting Normal Fault
The eruption of low-δ18O rhyolites outside the main SRP province may be explained by the shallow-level (<10 km) propagation of hundreds of cubic kilometers of low-δ18O rhyolite magma from the central SRP magma chambers to a magma chamber beneath the JSVS (Fig. 14; model A). The low-δ18O signature of the rhyolite melt is produced within the SRP-Yellowstone province via one of the processes described herein (see discussion of Cenozoic extensional and magmatic history of the Basin and Range province; e.g., cf. Bindeman et al., 2007; Boroughs et al., 2005, 2012; Fig. 14; model A), and via AFC processes (as described in the discussion of isotopic modeling). This model is successful at explaining the similar compositions of all the isotopic systems (Sr-Nd-O-Hf) between the SRP and the JSVS, because the JSVS is essentially connected with the magmatic system of the central SRP at shallow depths (see the discussion of isotopic modeling). The propagation of the rhyolite beneath the Raft River Basin occurs along north-south conduits perhaps controlled by the geometry of the major north-south–trending normal fault (the Albion fault) that bounds the Albion Mountains on its eastern side. The Albion fault was an important Miocene crust-penetrating structure and possibly served as a structural weakness in the upper crust (Konstantinou et al., 2012; Konstantinou, 2013). Based on the extensional history of the Albion–Raft River–Grouse Creek metamorphic core complex (Konstantinou et al., 2012; Konstantinou, 2013), the elastic crust in the region of southern Idaho underwent high extensional strain rates from ca. 14 until after 8 Ma, and during that time major SRP magma chambers may have been just north of the Albion–Raft River–Grouse Creek metamorphic core complex. This magma chambers may have acted as large hetero geneous bodies in the elastic crust having completely different physical properties (e.g., viscosity, stiffness) relative to the extending elastic crust. This less stiff inclusion may have focused the extensional strain in the elastic crust along the northern and southern tips around the magma chamber and may have assisted in the propagation of rhyolitic melt along north-south–trending avenues in the shallow crust. This process may be analogous (at a different scale) to borehole wall breakouts, which form symmetrically at the orientation of least principal horizontal stress, due to the concentration of stress around the borehole.
The major pitfall of this model of explaining the JSVS is achieving the physical propagation of rhyolite melt within the crust, over large distances (∼50 km; Fig. 14; model A). This is problematic due to the high viscosity of silicic melt that would cause them to quench during injection over large distances. Few studies have focused on silicic melt propagation along structures in the upper crust, but an example may be the 222–196 Ma, 18 km linear trend of a silicic dike swarm in the Candelaria mining district in western Nevada (Thomson et al., 1995). Lateral migration of silicic magma along faults in the Aso Caldera, southwestern Japan has also been demonstrated (Miyoshi et al., 2013). Another problematic aspect of this model is that large amounts of hydrothermally altered rocks with negative δ18O are required to produce the large-volume magmas of the SRP (and the JSVS; see the discussion of isotopic modeling). These extremely low δ18O rocks are rare in the geologic record of the region, but we envision that some of these rocks are faulted and buried beneath younger volcanic rocks within the SRP, and thus inaccessible for study.
Model B: Large-Scale Crustal Melting Beneath the Raft River Basin
An alternative explanation for the occurrence of the low-δ18O rhyolite lavas of the JSVS (Fig. 14; model B) might entail melting of hydrothermally altered crust outside the SRP-Yellowstone province and directly beneath the Raft River Basin. We envision that a normal δ18O (≈5.6‰) basaltic plumbing system of the JSVS may have been connected to that of the SRP-Yellowstone province at deeper levels, either within the lower crust or at the base of the crust (Fig. 14; model B). In model B, the low-δ18O rhyolite magmas of the JSVS may have formed by the inter action of a large flux of basalt with hydrothermally altered crust at depths of 6–8 km beneath the Raft River Basin, resulting in large-scale crustal melting but not significant mixing and/or assimi lation of these crustal melts with the basalt (Fig. 14). Melting of the crust on an extensive scale (rather than AFC processes; see the discussion of isotopic modeling) would require a smaller volume (∼500 km3) of low-δ18Owr upper crust, with δ18Owr of ∼2.5‰ to produce the rhyolites of the JSVS.
Because rocks with δ18Owr of 2‰–3‰ are more common in the region (e.g., Criss and Taylor, 1983; Criss et al., 1984; <200 km north of the Raft River Basin), both the δ18Owr ≈ 2.5‰ and the volume of the hydrothermally altered crust are not as problematic as the values required for modeling the rhyolites of the SRP province (δ18Owr ≈ –1.5‰; see the discussion of isotopic modeling).
Because there are no known caldera complexes in the Raft River Basin to explain the hydrothermal alteration of the crust, we propose alternative mechanisms for the formation of ∼500 km3 of low-δ18O crust beneath the Raft River Basin. Measured zircon δ18O values for deep Archean–Oligocene crustal rocks in the Albion–Raft River–Grouse Creek metamorphic core complex indicate that they have normal magmatic δ18Owr values (δ18Owr ≈ 7.1‰; see discussion of possible crustal and mantle sources; Strickland et al., 2011b; Konstantinou et al., 2013). These relationships imply that the alteration was synchronous, or younger than Oligocene, and a plausible mechanism for the hydrothermal alteration of the upper crust may be related to extensional deformation during the formation and localized rise of metamorphic core complexes. Oligocene granite-cored gneiss domes first rose diapirically to the base of the upper crust over a protracted period of time (32–25 Ma; Fig. 14; Strickland et al., 2011a, 2011b; Konstantinou et al., 2012). The regions around and between the granite-core gneiss domes may have been intensely faulted and fractured (see discussion in Konstantinou et al., 2012, and numerical modeling by Rey et al., 2011). These fractures would have brecciated large volumes of the Archean basement and the Proterozoic schist units above and away from the gneiss domes. A large number of closely spaced faults may have potentially provided multiple conduits for meteoric fluids to percolate at depths of 6–10 km. The emplacement of the Oligocene plutons and the rise of the hot crust of the gneiss domes may have provided the necessary heat to exchange oxygen between low-δ18O meteoric fluids and the Archean basement, and Proterozoic metapelites thus may have resulted in the formation of a low-δ18O crustal reservoir in the brecciated and metamorphosed region between and around the gneiss domes.
Oxygen isotope studies of rocks within and around other metamorphic core complexes in the western U.S. indicate that low-δ18O zones are common in the brecciated and chloritized regions around the complexes. For example, in the Valhalla gneiss dome, 1–2-km-thick upper plate chloritic breccias have δ18Owr and δ18Ofeld (feld—feldspar) that range from ∼5‰ to –5‰, even though the mylonite zones δ18Owr and δ18Ofeld values that range from 11‰ to 2‰ (Holk and Taylor, 2007). Another example is the Bitterroot metamorphic core complex, where chloritic breccias above the mylonite zones have δ18Owr and δ18Ofeld that range from ∼6‰ to –4‰, and the mylonite zones have δ18Owr and δ18Ofeld values that range from 12‰ to –1‰ (Kerrich and Hyndman, 1986). In the Albion–Raft River–Grouse Creek metamorphic core complex, retrogression of garnet and staurolite to white mica and chlorite away from the Oligocene gneiss domes is widespread in the Precambrian schist units. This retrogression affects much larger volumes of crust than those that are metamorphosed to sillimanite-grade rocks in the carapaces of the actual gneiss domes. However, there has been no systematic study of the δ18O values of these rocks, but such a study might offer insight into the δ18Owr composition of the crust in this region.
Another way of possibly forming a low-δ18O crustal reservoir is via the percolation of meteoric fluids deep into the crust along the Miocene Albion fault system (Konstantinou et al., 2012), when large volumes of meteoric fluids may have percolated to ∼8 km and exchanged oxygen with the silicic rocks of the Archean basement, the Proterozoic schist units, or Cenozoic granites, thus forming a localized low-δ18O reservoir. Providing support for fault-related hydrothermal circulation systems are recent studies of ductilely deformed rocks in the Raft River detachment that, based on their hydrogen isotopes, show evidence of percolation of meteoric fluids at deep crustal levels that are now exposed to the surface (Gottardi et al., 2011). These studies show that normal fault systems provide the necessary conduits for meteoric fluids to circulate to depths of 10 km in the crust, to the brittle-ductile transition zone.
Model B portrayed in Figure 14 has a major limitation. If the JSVS represents large-scale crustal melting of rocks similar to the Archean basement, the Proterozoic schist units and the Oligocene granites exposed today in the adjacent metamorphic core complex, then the Sr-Nd-Hf isotope compositions of the JSVS should be nearly identical to the compositions of rocks exposed in the complex. However, the rhyolites of the JSVS have much less evolved Sr-Nd-Hf compositions than the compositions represented by the silicic rocks within the Albion–Raft River–Grouse Creek metamorphic core complex (e.g., Strickland et al., 2011b; Konstantinou et al., 2013; Fig. 13). Thus model B may not effectively predict all the measured isotopic compositions of the magmas at the JSVS.
Implications for the Evolution of the Raft River Detachment and Basin and Range Faulting
The eruption of ∼240 km3 of high-temperature rhyolite in the synextensional Raft River Basin has important implications for the evolution of Basin and Range faulting and the exhumation of the adjacent Albion–Raft River–Grouse Creek metamorphic core complex. The data (and preceding discussion) establish a direct link between magmatism in the JSVS and the central SRP-Yellowstone province based on similar timing, geochemistry, and isotopic compositions. The data might also imply that a silicic Miocene pluton with possible mafic roots could directly underlie the Raft River Basin (Fig. 14; model B). This inferred high-level Miocene silicic magmatic system would, as for the SRP, be driven by a high flux of basalt and thermal input to the middle and lower crust of southern Idaho and may have led to extensive crustal melting at shallow levels of the crust beneath the region of the Raft River Basin and the Albion–Raft River–Grouse Creek metamorphic core complex. The structural history of the Raft River Basin (as detailed by Konstantinou et al., 2012; Konstantinou, 2013) indicates that the rotation of the Albion fault to shallow angles and all of the normal faulting and rotation of Miocene basin fill strata occurred after the eruption of the JSVS (after 8.2 Ma; Konstantinou et al., 2012). This young faulting and extension is most easily explained by coeval doming with a component of vertical uplift during horizontal stretching of the crust beneath the Raft River Basin. Horizontal extension and vertical uplift resulted in the present-day subhorizontal geometry of the basal detachment imaged seismically beneath the Raft River Basin (Covington, 1983; Konstantinou et al., 2012; Konstantinou 2013).
The possible existence at depth of a silicic magmatic system implies (as it does for the SRP-Yellowstone) addition of basaltic material at greater depth in the crust and concomitant heating of the middle and lower crust leading to possibly extensive crustal melting beneath the region of the Albion–Raft River–Grouse Creek metamorphic core complex during the Miocene. This crustal melting event may have resulted in increased mobility of the deep crust during and after the eruption of the rhyolite lavas. We therefore interpret the extensional deformation and vertical component of uplift after ca.8.2 Ma to be a direct consequence of greater mobility of the immediately underlying crust.
The JSVS is composed of ∼240 km3 of rhyolite lavas, capped by central SRP–derived rheomorphic ignimbrites and a thin sequence (<100 m) of Miocene basalt flows. The rhyolite lavas of the JSVS erupted from small centers within the Raft River Basin based on their physical attributes (aspect ratios and basal contact) and the occurrence of subvolcanic intrusions and domes. The ignimbrite and basalts are probably outflows from the central SRP. The major and trace element and isotopic compositions of the rhyolite lavas are consistent with having an origin similar to that of the magmas in the central SRP-Yellowstone province. This observation, coupled with the fact that the rhyolites have a low-δ18O signature, suggests that the in situ assimilation of hydrothermally altered caldera blocks proposed to explain the origin of low-δ18O rhyolites in the SRP may not have been the case here. We offer two models that might explain the low-δ18O signature of the lavas in the JSVS, each with their own strengths and limitations. The first model suggests that SRP-type rhyolitic melt propagated southward along the Albion fault to erupt in the center of the Raft River Basin. The second model proposes that the hydrothermal alteration of the crust in the region of southern Idaho may be more widespread than previously thought and was related to the earlier igneous and gneiss dome history or the normal faulting history of the region, and that the eruption of the JSVS was localized above a magma chamber off axis of the SRP. It is possible that a low-δ18O silicic pluton beneath the Raft River Basin could have provided a heat source for crustal melting and weakening that led to extension accompanied by vertical doming and rise of the Raft River Basin after the eruption of the JSVS rhyolites.
We thank Stanford University for the Stanford Graduate Fellowship that supported this research for three years, ExxonMobil for a science grant, and the Leventis Foundation and the Stanford McGee Funds for providing financial support for the analytical part of this project. Support was also provided by the National Science Foundation (NSF) Tectonics Division (grants EAR-0809226 and EAR-0948679 to Miller). Gail Mahood, Matt Coble, and Eric Gottlieb provided helpful discussions and insights about the scientific work presented herein. Mike McCurry, Ilya Bindeman, and Ben Ellis provided valuable reviews that greatly improved the paper. We thank Noriko Kita and Kouki Kitajima for assistance with secondary ion mass spectrometer (SIMS) analysis of oxygen isotope ratios. The University of Wisconsin WiscSIMS is partially supported by NSF grants EAR-0319230, EAR-0744079, and EAR-1053466. The Washington State University Radiogenic Isotope and Geochronology Laboratory is partially supported by NSF grants EAR-0844149, EAR-1019877, and EAR-1119237. We thank Karrie Weaver and Caroline Harris for their help with the whole-rock Sr and Nd isotope analyses.
APPENDIX 1. METHODS FOR WHOLE-ROCK TRACER ISOTOPE ANALYSES
Samples were selected and processed at the Stanford University ICP-MS-TIMS (inductively coupled plasma–mass spectrometry–thermal ionization mass spectrometry) facility. The samples were chipped, and rock chips were hand-selected and pulverized to a very fine powder using an agate shutter box bowl. A small aliquot of each sample (40–60 mg) was dissolved in Savillex Teflon vials using a combination of 29 N HF, followed by a mixture of 3:1 ratio of 10.5 N HCl and 14 N HNO3. An aliquot of the solution was dried and the salt was redissolved in HCl to prepare for column chromatography. Strontium and the rare earth elements (REE) were separated from major elements and other trace elements using an HCl elution on 200–400 mesh AG50-X8 cation exchange resin. The Sr was further isolated from Rb using Eichrom 20–50 μm Sr-spec resin in Teflon microcolumns. Elemental Nd was purified from the REE aliquot of the cation exchange column by column chromatography using Eichrom 50–100 μm LN-spec resin. The chem istry was monitored to ensure adequate separation of Nd from Ce and Sm, the isobaric interferences of which may compromise the quality of the isotopic analyses. The resulting purified Sr and Nd salts were prepared for mass spectrometry analysis by redissolving them in 2% HNO3. Isotope analyses were performed using the Nu-Instruments Plasma HR multicollector ICP-MS.
Each Sr isotope analysis consisted of 50 cycles of 10 s integration time, split into 2 blocks, during which 6 isotopes were measured simultaneously (83Sr, 84Kr, 84Sr, 85Rb, 86Sr, 87Sr, 88Sr) on Faraday cups. Instrumental mass fractionation and 86Kr interference on 86Sr were corrected for by subtracting the 84Kr to obtain the natural 84Sr/88Sr ratio of 0.00675476 while iteratively solving for the exponential fractionation factor and Kr correction on 86Sr, an approach similar to that of Jackson and Hart (2006). Baselines were measured at 0.25 amu from peak center to account for scattering of 40Ar2 that creates a nonlinear baseline across the Sr masses. Interference on 87Sr from 87Rb was small and corrected using 85Rb and the instrumental mass fractionation factor for Sr; 43 analyses of SRM 987 resulted in an average 87Sr/86Sr value of 0.710144 ± 52 (2σ). Sample ratios were multiplied by a normalization factor determined by correcting SRM 987 analyses to a value of 87Sr/86Sr = 0.71025 ± 1 (2σ; Balcaen et al., 2005) and the external error based on the 31 SRM 987 analyses was added to the analytical error for each sample. The total mass of sample analyzed ranged from 20 to 50 ng, and a blank analysis yielded an estimated mass of Sr < 0.01 ng. A secondary standard of BHVO-1 was analyzed 7 times and yielded an average value of 0.703483 ± 14 (2σ) [accepted value is 0.703475 ± 17 (2σ); Weiss et al., 2005] after correction using the normalization factor obtained from SRM 987.
Each Nd isotope analysis consisted of 40 cycles of 10 s integration time, split into 2 blocks, during which 5 isotopes were measured simultaneously (143Nd, 144Nd, 145Nd, 146Nd, and 144Sm), monitored for Ce and Sm interferences. Mass bias was exponentially corrected based on a 146Nd/144Nd ratio of 0.7219. Sample ratios were multiplied by a normalization factor determined from 26 analyses of JNdi-1 resulted in an average 143Nd/144Nd value of 0.512087 ± 22 (2σ). The total mass of sample analyzed ranged from 10 to 30 ng, and a blank analysis yielded an estimated mass of Nd < 0.005 ng. A secondary standard of BHVO-1 was analyzed 3 times and yielded an average value of 143Nd/144Nd 0.512969 ± 20 (2σ). This compares to the value for BHVO-1 of 143Nd/144Nd = 0.512986 ± 18 (2σ) from Weiss et al. (2005).
APPENDIX 2. METHODS FOR OXYGEN ISOTOPE ANALYSES
Oxygen isotope analysis was performed at the University of Wisconsin WiscSIMS (Wisconsin Secondary Ion Mass Spectrometer) laboratory using a CAMECA IMS-1280 ion microprobe following the procedures outlined in Kita et al. (2009) and Valley and Kita (2009). The zircon crystals were mounted with the KIM-5 O isotope standard (Valley, 2003; δ18O = 5.09‰ Vienna standard mean ocean water, VSMOW), and polished prior to O isotope analyses. Extra care was taken to achieve a smooth, flat, low-relief polish. An ∼2 nA primary beam of 133Cs+ was focused to 10-μm-diameter spots on the sample. A normal-incidence electron gun and carbon coat were used for charge compensation. The secondary ion acceleration voltage was set at 10 kV and the O isotopes were collected in 2 Faraday cups simultaneously in multicollection mode. Four consecutive measurements on KIM-5 zircon standard (δ18O = 5.09‰ VSMOW; Valley, 2003) were analyzed before and after groups of 10–20 sample spots. The 2 SD precision for bracketing groups of eight standard analyses is reported as external precision for sample data. The average values of the standard analyses that bracket each set of unknowns were used to correct for instrumental bias. The average precision (reproducibility) of the bracketing standards for this study ranged from ±0.13 to ±0.35 and averaged ±0.26‰ (2 SD). After O isotope analysis, ion microprobe pits were imaged by secondary electron microscopy and pits that were located on obvious cracks or inclusions were excluded from the data set.
APPENDIX 3. METHODS FOR HAFNIUM ISOTOPE ANALYSES
Hf isotopic compositions were determined on zircons from five Miocene samples at Washington State University’s LA-MC-ICP-MS (laser ablation–multicollector–inductively coupled plasma–mass spectrometry) facility, using the procedure from C.M Fisher (2013, personal commun.). The resulting isotopic ratios were evaluated based on the quality and duration of the Hf analysis; 40 zircon Hf isotope analyses yielded 176Hf/177Hf ratios that were corrected for interlaboratory bias using the R33 standards (correction factor = 1.000159), and the corrected ratios, along with measured 176Lu/177Hf and sensitive high-resolution ion microprobe–reverse geometry ages, were used to calculate εHf(i) and εHf(t) values. The results are reported in Supplemental Table 6 (see footnote 6). Isotopic ratios with values ranging from εHf(t) –25 to –44 were excluded from further consideration (n = 8) because they could represent mixed isotopic compositions resulting from ablation and analysis of Miocene magmatic zircon and inherited zircon cores.