Magmatic lulls in the Sierra Nevada captured in zircon from rhyolite of the Mineral King pendant, California

Whereas studies of Mesozoic magmatism in the Sierra Nevada arc (California, USA) are biased toward studies of voluminous plutonic units (e.g., Bateman, 1992; Chen and Moore, 1982; Glazner et al., 2004, Lackey et al., 2008; and many others), Triassic, Jurassic, and Cretaceous volcanic pendants and septa within the Sierra Nevada batholith commonly contain metamorphosed volcanic rocks that either predate or are coeval with the plutonic rocks that bound them (e.g., Kistler and Swanson, 1981; Busby-Spera, 1983; Saleeby et al., 1990; Fiske and Tobisch, 1994; Schweickert and Lahren, 1999; Barth et al., 2012). This includes notable volcanic packages at Erskine Canyon, in the southern Sierra Nevada, in the Oak Creek, Boyden Cave, and Goddard pendants in the Kings Canyon region, and in neighboring areas of the Alabama Hills and Inyo Range; a group of pendants with abundant metavolcanic rocks is found between Yosemite National Park and the Long Valley caldera. Although it is well known that metavolcanic rocks in pendants around Kings Canyon and to the north overlap in age and in some cases may be correlative, the Mineral King pendant in the south-central Sierra Nevada represents a key domain between shallower and deeper levels of the batholith, as well as major intrabatholithic structural breaks (Saleeby et al., 1978; Saleeby and Busby, 1993).

These metavolcanic remnants are a sparse but important record of magmatism in high levels of the Sierran Arc. However, interpretation of the history of these volcanic units is problematic because they are extensively deformed, hydrothermally altered, and contact metamorphosed (Hanson et al., 1993; Sorensen et al., 1998). Zircon contained within such volcanic rocks affords the clearest window into original magmatic conditions because its refractory nature (Harley and Kelly, 2007) makes it unaffected by the myriad effects of secondary alteration. In particular, U-Pb age data and isotopic and trace element information from zircon provide fresh insight into volcanism, as exemplified by studies of zircon from Triassic volcanic rocks in the Saddlebag Lake pendant (Lackey et al., 2008; Dietterich et al., 2009; Barth et al., 2012).

Aggregate plutonic and volcanic ages (Irwin and Wooden, 1999, 2001) show that rates of magmatism across the arc were not constant over its 150 m.y. life span, and prominent lulls in activity have been noted (Paterson, 2012; Barth et al., 2013). Little is known about these periods of quiescence due to their paucity within the geologic record. Any remnants from these periods of low magmatic activity contain information that is lost across much of the arc and begs the question of how magmatism was changing from the Late Triassic arc (Barth et al., 2012, 2013) to the Cretaceous batholith emplacement (e.g., Ducea, 2001; Glazner et al., 2004). Two end-member models could exist: (1) magmatism was very infrequent during these lulls, so mature magmatism systems would be rare, or (2) much of the magmatic record of magmatism has been lost (to erosion or burial) and mature magmatic system that exhibit abundant crystal recycling and rhyolite volcanism are active across the arc during the lulls. By assessing these patterns of magmatism across the history of the western North American margin, a better understanding of tempos of magmatism can accompany tectonic models for the evolution of the margin (Paterson, 2012).

In this study we present new zircon age and compositional data on four metavolcanic units from the Mineral King pendant in the central Sierra Nevada. These four units provide information about volcanic and intrusive activity on the western margin of North America during the Early Jurassic and Early Cretaceous. These units are rhyolite deposits related to arc magmatism, and the zircons from these rhyolites reveal the changing parameters of magmatism from ca. 197 Ma to ca. 134 Ma, suggesting an increase in the influence of continental crust and recycling of magmatic precursors. The zircons also imply that even during periods of low overall magmatic productivity in the western North American margin, voluminous rhyolite volcanism was occurring, with all the hallmarks of long-lived magmatic systems where magma is extracted from a crystal mush, promoting abundant crystal recycling. The data also show how the petrogenesis of altered volcanic rocks can be determined by the combination of single zircon spot ages and trace element analyses combined with bulk zircon oxygen isotopic analyses. This provides the opportunity to examine the important parameters of petrogenesis (fractional crystallization, assimilation, and mixing; e.g., see Claiborne et al., 2010a; Walker et al., 2010; Barth et al., 2012) on suites of silicic, zircon-bearing rocks that have hitherto been considered too altered to be assessed by whole-rock methods.

The Mineral King pendant is located in the central Sierra Nevada in California (Fig. 1) and is one of the many septa and screens representing the wall rock into which the Sierra Nevada batholith intruded. The Mineral King pendant is constructed from metaigneous volcanic (rhyolite to andesite) and intrusive (gabbro and granite) units, breccia sandstones, marbles, and marine sediment (Busby-Spera, 1983; Saleeby and Busby, 1993; Sisson and Moore, 2013). Many of these units were altered by the intrusion of fluids and shearing (Bateman, 1992; D’Errico et al., 2012). These units are bound within the pendant by multiple faults and shear zones (Busby-Spera, 1983). The package of rock was interpreted as a marine sequence from the paleocontinental shelf of western North America (Busby-Spera and Saleeby, 1987; Saleeby and Busby, 1993), possibly representing 70 m.y. of continuous stratigraphy (Busby-Spera, 1983). The Mineral King pendant is bound to the east by the quartz diorite of Empire Mountain (ca. 109 Ma; D’Errico et al., 2012) and the granite of Coyote Pass (97 ± 2 Ma; Busby-Spera and Saleeby, 1987) and to the west by the granodiorite of Castle Creek (98 ± 2 Ma; Busby-Spera and Saleeby, 1987). Emplacement barometry of the eastern contact with the Empire quartz monzodiorite suggests shallow depths of <3 km (D’Errico et al., 2012). Overall, the Mineral King pendant has been broadly correlated with other pendants in the King Series (Saleeby and Busby, 1993).

There are four rhyolite units (R₀, R_1, R₂, R₄; Fig. 1) within the Mineral King pendant that have been interpreted as tuffs and flows from a submarine (or proximal) caldera (Busby-Spera, 1984, 1986). In total, ∼90 km³ of rhyolite tuffs have been documented in the Mineral King pendant, each ∼20–25 km³ in volume, along with 10 km³ of andesite lavas and volcaniclastic units (Busby-Spera, 1984). These units were originally dated by Busby-Spera (1983) using multigrain zircon dissolution analyses combined with stratigraphic interpretation of the pendant. The R₀, R₁ and R₃ rhyolites were thought to be Early Triassic (215 ± 2 Ma). The R₄ (189 ± 2 Ma) and the R₂ were assigned to the Early Jurassic, the sediment in the pendant marking quiescence between eruptions (Busby-Spera, 1983). A volumetrically minor dacite intrusive breccia on the easternmost contact was dated as 240 ± 7 Ma (Busby-Spera, 1983). Structurally, the Empire fault separates the R₂ and R₄ units and the Farewell fault separates the R₀ and R₁ units from the R₂ and R₄ units (Fig. 1). Busby-Spera (1983) interpreted that the volcanism in the Mineral King pendant as the transition from an Early Triassic island arc (R₀, R₀₁, R₃, R₄ units) to an Early Jurassic continental arc (R₂ unit) as a continuous depositional sequence.

The R₀ rhyolite (10JS03) is intruded by the Castle Creek granodiorite to the west, and the contact between the two units is indistinct and interfingering (Fig. 1). This unit is considerably altered to quartz sericite schist; few plagioclase feldspar phenocrysts remain. However, primary bedding of ashy and/or sandy layers are evident (Fig. 2A). Large (>1 cm to tens of centimeters) quartz veins with boudinage textures are present in outcrops near the contact with the Castle Creek. This unit has an estimated modal abundance of 55% quartz, 35%–45% mica (mostly biotite, some muscovite), and 5%–10% plagioclase feldspar.

The R₁ rhyolite (11MK02) is near the southwestern boundary of the pendant. It is bound to the west by a sandstone breccia, calc-silicate sediment, and an undated aplitic intrusion, and to the east by a slate unit that extends along the valley bottom (Fig. 1). As with all samples within the pendant, it has been altered to quartz sericite schist but retains relic textural fabrics of a rhyolite tuff (e.g., ghost-like pumice clasts; Fig. 2B). This unit has a modal abundance similar to that of R₀.

The R₂ rhyolite has no visible contacts with surrounding units and is bound to the east by calc-silicates and to the west by slate (Fig. 1). It is heavily altered, with relict phenocrysts visible on the weathered surfaces but not visible on fresh surfaces (Fig. 2C). There is abundant mica, and from west to east the outcrop grades from a rhyolite tuff to a flow-banded rhyolite. The unit has been altered to quartz sericite schist. Both biotite and muscovite micas are aligned in a clear foliation direction, particularly in the flow-banded portion of the unit. This unit has a modal abundance similar to that of R₀, but with a higher proportion of plagioclase feldspar versus micas.

The R₄ rhyolite is the least visibly altered of all the units, but is still classified as a quartz sericite schist. It is bound to the east by the Empire quartz monzodiorite (109 Ma; D’Errico et al., 2012) and to the west by the R₂ rhyolite, although the direct contact is not observed (Fig. 1). The outcrop displays a platy weathering pattern, and fresh surfaces have visible phenocrysts (Fig. 2D). Alignment of mafic crystals produces a slight fabric. This unit has an estimated modal abundance of 55% quartz, 20% mica (mainly biotite), 20% plagioclase feldspar, and 5% amphibole.

Samples of the four major rhyolite units (R₀, R₁, R₂, R₄) within the Mineral King pendant were collected during the summers of 2010 and 2011. Although all of the units are metamorphosed (likely through heating and fluid involvement to hornfels facies), care was taken to choose samples that were fresh and still showed primary volcanic textures. Approximately 3 kg of sample were crushed and zircons were extracted from all samples using standard mineral separation techniques (see Klemetti et al., 2011) at Pomona College and California State University Bakersfield. Zircon crystals larger than 57 mm were mounted in epoxy and polished at the Stanford University–U.S. Geological Survey Micro Analysis Center (SUMAC laboratory). Prior to SHRIMP-RG (sensitive high-resolution ion microprobe–reverse geometry) analysis, all samples were imaged via cathodoluminescence (CL) (Fig. 3) to assess the zoning and dissolution patterns along with melt and/or mineral inclusions within the grains. Analyses for U-Pb ages (Supplemental Table 1¹) and trace elements (Supplemental Table 2²) in the zircons were performed via SHRIMP-RG using methods from Mazdab and Wooden (2006) and all data were reduced using Isoplot 4 and SQUID2 software (http://www.bgc.org/isoplot_etc/squid.html) (Table 1; Supplemental Table 1 [see ^{footnote 1}]; Ludwig, 2003). We report ²⁰⁶Pb/²³⁸U ages corrected by measured values of ²⁰⁷Pb (Supplemental Table 1 [see ^{footnote 1}]). Conventional concordia (Supplemental Figure³) and weighted mean age diagrams (Fig. 4) were constructed using Isoplot (Ludwig, 2003). Samples of the R33 zircon (419 Ma) standard were analyzed to correct for instrumental drift and reproducibility for U-Pb age determination and samples of the MAD in-house zircon were used for trace element reproducibility (Supplemental Table 2 [see ^{footnote 2}]). All errors on final ages are 1s unless noted otherwise.

Because zircons in the Mineral King rhyolites show simple population traits by CL (i.e., no rim or core CL differences that suggest secondary ion mass spectrometry analysis is warranted; Fig. 3; Lackey et al., 2013), bulk laser fluorination oxygen isotope analysis was used to provide an accurate and precise measure of magmatic δ¹⁸O (King et al., 1997). Oxygen isotopic analyses of bulk zircon separates were performed at the Stable Isotope Laboratory at the University of Oregon. Zircon concentrates were prepared for δ¹⁸O analysis by sequentially cleaning bulk aliquots in HNO₃, HF, and HCl to remove mineral impurities and radiation-damaged domains from the crystals (Lackey et al., 2008). Oxygen was liberated from zircon by laser fluorination at the University of Oregon Stable Isotope Lab using a 35W New Wave CO₂ infrared laser to heat samples in the presence of BrF₅ reagent (Bindeman, 2008). Oxygen was cryogenically cleaned, pumped through hot Hg to capture excess F₂, and converted to CO₂ with a hot carbon rod before analysis. Isotope ratios were measured with a Finnigan MAT 253 mass spectrometer and measured values were corrected to the Gore Mountain garnet standard (UWG-2) analyzed throughout the session using the accepted δ¹⁸O value of 5.80‰ (Valley et al., 1995). The average raw δ¹⁸O of UWG-2 for 2 days of analyses (n = 6) in this study is 5.75‰ ± 0.12‰ (2σ).

We measured major and selected trace elements for three samples by X-ray fluorescence (Supplemental Table 3⁴) at Pomona College. Rock powder and flux were vortex blended in a 1:2 ratio (3.5 g powder to 7.0 g dilithium tetraborate), and fused to glass beads in graphite crucibles heated at 1000 °C for 10 min. Beads were reground and fused a second time, polished on diamond laps, and analyzed. Concentrations were determined according to calibrations based on 55 certified reference materials that span a large range of compositions (see Lackey et al., 2012, for further details regarding X-ray fluorescence analysis at Pomona College). Typical analytical 2σ uncertainties for analyses (in wt%) of major elements, determined conservatively as reproducibility of multiple beads produced from a sample or standard powder, are ±0.15% SiO₂; 0.06%–0.07% Fe₂O₃, Al₂O₃; 0.02%–0.05% Na₂O, K₂O, MgO, CaO; <0.004% TiO₂, MnO, P₂O₅. Trace element uncertainties at ppm ± 2s are 1 ppm La, Pr, Hf; 2 ppm Nd, Cu, Ga, Nb, Pb, Rb, Th, U, Y, Zr; to 3 ppm Cr, Sc, Sr, V; other elements have higher uncertainties: Ce 4 ppm; Zn 6 ppm; Ni 7 ppm; Ba 30 ppm.

CL images of zircons from the Mineral King metavolcanic rhyolite units show that most of the crystals have not undergone much if any alteration since their primary igneous formation (Fig. 3). The zircon samples are typically a few hundred micrometers in length. The crystals are variably cored, some with dark, U-rich cores that exhibit minor dissolution (rounded shapes relative to the euhedral outer zoning) with regular, oscillatory zoning with euhedral crystal shapes (Fig. 3). Some other cores observed have mottled cores with abundant melt inclusions and light colored cores that lack any evidence of zoning. There is no noticeable difference in the shapes and sizes of zircons or in zonation between zircons from any of the rhyolite units.

We performed 51 spot analyses on zircons culled from the 4 Mineral King metavolcanic units. Table 1 summarizes the age information determined from the ²³⁸U-²⁰⁶Pb analyses and Supplemental Table 1 (see ^{footnote 1}) contains the full complement of U-Pb analysis data. In all, three units (R₀, R₁, R₂) yielded weighted mean ages of 136.5 ± 2.7, 134.2 ± 0.7, and 135.5 ± 1.4 Ma, respectively, and one unit (R₄) yielded an age of 195.7 ± 1.4 Ma (1σ errors; Fig. 4; Table 1). We have chosen weighted mean ages for the geochronology in this study as it encapsulates the complex nature of zircon populations in igneous rocks by examining the range of ages, looking for the highest density of ages for the chosen age. However, using the youngest ages derived from zircons that did not exhibit Pb loss as an eruption age does not significantly change the age interpretation of the units (Fig. 4; Table 1). Most samples had at least one grain that showed potential Pb loss with ages younger than the overall population, and these grains were rejected when calculating the weighed mean ages unless noted. Two zircons analyzed on the R₀ rhyolite had younger ages (128.9 ± 1.2 Ma and 129.5 ± 0.9 Ma, 1σ error) that caused a higher MSWD (mean square of weighted deviate; 12) for the weighted mean age (Fig. 4). However, the exclusion of these zircons does not affect the age determination significantly, although it reduces the MSWD. If the youngest 2 grains are excluded, the weighted mean age is 137.6 ± 2.0 Ma (MSWD 5.2) and if an additional 2 oldest ages are excluded, the weighted mean age becomes 135.5 ± 1.1 Ma (MSWD 0.86). We choose to retain all the zircon from this sample in the calculation of the weighted mean age even though it does not produce the lowest MSWD, as all calculated ages are within error. The R₀ rhyolite also yielded a single xenocrystic zircon with an age of 1.63 Ga ± 8.4 Ma (1σ; Table 1).

Trace element analyses (n = 65) from Mineral King zircons were performed (some of which do not have corresponding U-Pb analyses), a compilation of which is in Supplemental Table 2 (see ^{footnote 2}). The Mineral King rhyolite units show the general patterns seen in zircon, with enriched heavy rare earth element (HREE) versus light REE (bulk silicate Earth normalized; Sano et al., 2002), positive Ce anomalies, and modest Eu anomalies (Fig. 5). Two zircon analyses (10JS03–3.1, 10JS04–2.1) show enrichment across all trace elements (Fig. 5; Supplemental Table 2 [see ^{footnote 2}]); this may reflect growth in more fractionated liquid or small inclusions within the zircon that did not affect all trace elements analyzed. Zircons with anomalous values (such as high values of K, Ca, or Fe) were rejected from this compilation because they probably included analyses of mineral or glass inclusions (Supplemental Table 2 [see ^{footnote 2}]).

Zircons from the Mineral King rhyolite units range in Hf concentration from 7525 to 12800 ppm (Fig. 6). Overall, the R₀ and R₁ rhyolites have the widest range in Hf concentrations (8088–12800 ppm), the R₂ rhyolites has the lowest average Hf concentration (7525–10280 ppm), and the R₄ rhyolites is the most restricted (9081–10270 ppm; Fig. 6). These patterns are reflected across most trace element patterns. The R₀ and R₁ rhyolites represent the wider range of values, while the R₂ and R₄ rhyolites are more restricted, although many times the patterns overlap (Fig. 6). The U concentrations vary from 511 to 5127 ppm, and Th concentrations vary from 47 to 2618 ppm (Fig. 6), producing a range of Th/U of 0.114–1.449 (Fig. 7).

Zircons show a wide range of Eu/Eu*, from 0.11 to 0.49 (Figs. 6 and 7). The R₀ and R₁ zircons show the greatest variability, spanning the entire range of Eu/Eu* in Mineral King rhyolite zircon. The R₂ zircons have a smaller range, only reaching Eu/Eu* of ∼0.3, and the R₄ zircons have the most limited range, all between 0.22 and 0.26. The range in Yb/Gd ratios over all the units is 6.96–37.88 (Fig. 7), with similar patterns: the R₀ and R₁ rhyolites encompass the entire range, the R₂ rhyolite ranges from 10 to 20.3, and the R₄ rhyolite ranges from 14.1 to 26.7 (Fig. 7).

Bulk zircon oxygen isotopic compositions of zircons provide an accurate and precise measure of overall magmatic δ¹⁸O in altered volcanic rocks (cf. King et al., 1997). Three of the Mineral King rhyolite units yielded a range of δ¹⁸O (zircon, Zrn) from 5.33‰ to 7.33‰ (Table 1). The oldest unit (R₄, 195.7 ± 1.4 Ma) had the lowest δ¹⁸O(Zrn) (5.33‰ ± 0.09‰), while the youngest units (R₀ and R₁, 136.5 ± 2.7 and 134.2 ± 0.7 Ma, respectively) had the highest δ¹⁸O(Zrn) (6.78‰ ± 0.08‰ and 7.33‰ ± 0.16‰, respectively). Fractionation of oxygen isotopes between zircon and magma, Δ¹⁸O(Zrn), can be calculated according to SiO₂ content of magma (Lackey et al., 2008; see Supplemental Table 3 [see ^{footnote 4}] for whole-rock analyses). Thus, magmatic Δ¹⁸O values for the Mineral King metarhyolite units are 7.33‰ to 9.33‰.

Previous studies (Busby-Spera, 1983; Saleeby and Busby, 1993) that examined the metavolcanic rocks of the Mineral King pendant used bulk zircon dissolution U-Pb geochronology and fossil correlation and/or stratigraphy to determine the relative age relationships of the units (Fig. 1). The conclusion was that the youngest rocks of the pendant were located on the east side in the R₄ rhyolite and the rhyolite units (and interbedded metasediments) became older to the west, with Triassic ages for the R₀ rhyolite (Busby-Spera, 1983). There were some complications to this order, because the intrusive dacite breccia on the far eastern side of the pendant was dated as 217 Ma (Busby-Spera and Saleeby, 1987), but overall, Busby-Spera (1983) interpreted the metavolcanic and metasedimentary rocks of the Mineral King pendant as a nearly continuous record of volcanic and sedimentary deposition from the Triassic to Early Jurassic.

Our findings indicate that the stratigraphy across the Mineral King pendant is much more complex (Fig. 1) than as concluded by Busby-Spera (1983). As the U-Pb analyses of single zircon grains from multiple igneous units across the pendant show (Fig. 4; Table 1), there is not a clear age progression from east to west across the pendant. When compared to ages from Busby-Spera (1983), starting from the east, the R₄ rhyolite is Early Jurassic in age, with a weighted mean age of 195.7 ± 1.4 Ma; this age is slightly older than the previous age of 189 ± 2 Ma. The R₂ rhyolite produced an age of 135.5 ± 1.4 Ma, significantly younger than the inferred Early Jurassic age. The R1 rhyolite produced a similar age of 134.2 ± 0.7 Ma rather than Early Triassic, while the R0 rhyolite yielded an age of 136.5 ± 2.7 Ma, compared to Early Triassic. Multiple samples from Busby-Spera (1983) were discordant and the presence of xenocrystic zircon with ages older than 1.6 Ga in the R₀ rhyolite would support the idea that the bulk zircon dissolutions were contaminated by these much older grains, thus skewing the data toward older, discordant ages.

One important question to resolve when examining the metavolcanic units in the Mineral King pendant is whether they represent discrete events or a dismembered single volcanic package. Based solely on the geochronology, it is clear that the R₄ rhyolite is separated by more than 60 m.y. from the other dated rhyolite units within the pendant. However, the R₂, R₁, and R₀ units are within error of each other in age, so other factors must be considered in order to determine if they are separate rhyolite units (as suggested by Busby-Spera, 1983, 1984) or dismembered remains of a single rhyolite unit. One way to approach this issue is to look at the trace element compositions of the zircons found in the various metavolcanic units. Although the units are altered, the refractory nature of zircon allows for it to record magmatic conditions even if much of the rock host has been altered (Harley and Kelly, 2007). By comparing the compositional populations of zircon within each unit, it can be determined if R₂, R₁, and R₀ are likely to be the same rhyolite unit.

There are clear differences in the trace element compositions of zircons from the R₄ rhyolite and the R₀, R₁, and R₂ rhyolites (Figs. 4–8). The tight clustering of the zircons from the R₄ across multiple trace element parameters (Figs. 6 and 7) suggests that the rhyolite does not contain a large component of recycled zircon, either from the host rocks or from precursory silicic intrusions. This is supported by the lowest δ¹⁸O(Zrn) values in the Mineral King metarhyolites in the R₄ rhyolite (Table 1), precluding abundant assimilation of continental crust that tends to have higher values due to abundant quartz and feldspar (James et al., 1985). The R₄ rhyolite also has the lowest REE abundances of any metarhyolite unit within the Mineral King pendant (Fig. 5).

The R₀, R₁, and R₂ rhyolites are more complex, with a wider range of values across various trace element parameters (Figs. 6–8), suggesting an increased importance of zircon recycling within these youngest rhyolite units. The R₀, R₁, and R₂ rhyolite may share compositional and temporal attributes (Busby-Spera, 1983). However, our new data support the idea that these do not appear to be a single rhyolite unit sheared by the structures within the Mineral King pendant. By using the trace element compositions of zircons from these units, some broad observations can be made. First, the zircon from the R₂ rhyolite is compositionally distinct from zircon from the R₀ and R₁ rhyolites. Although all three of these metarhyolites have ages within error of each other, the R₂ shows distinct zircon trace element compositions (Figs. 6 and 7): lower Hf concentrations along with higher Th/U, lower Yb/Gd, and higher Ce/Ce* at constant Hf concentrations. These trace element parameters suggest that the R₂ rhyolite was hotter (Miller et al., 2003), although the Eu/Eu* (Fig. 6) for the R₂ rhyolite is within the same range as the R₀ and R₁ rhyolites. The R₂ rhyolite also has overall higher REE concentrations versus the R₀ and R₁ rhyolites (Fig. 5).

The R₀ and R₁ rhyolites have a wider range in most trace element parameters relative to the R₂ rhyolite and have minor overlap with R₂ rhyolite compositions. These differences could mean that the R₂ rhyolite is the product of a different eruptive center that was brought into the Mineral King pendant via structural motion along a fault, or that the R₂ rhyolite is a compositionally earlier magma produced by the same eruptive center as the R₀ and R₁ rhyolites. Although the ages of these three units are within error, the presence of zircons with compositions similar to those of the R₂ in the R₀ and R₁ rhyolites suggests that they may have been derived from the R₂ residue. However, it is not possible, based on the evidence presented here, to conclusively differentiate between separate nearby eruptions or temporally separated eruptions from the same eruptive center.

The rhyolites of the Mineral King pendant capture the magmatism within the western margin of North America during a period of relative magmatic quiescence (Barth et al., 2013; Figs. 9A, 9B). The zircon compositions of these rhyolite units allow a glimpse into the magmatic processes at work when these rhyolites formed and how those processes have changed over the 60 m.y. recorded in the pendant. Through the trace element and oxygen isotope composition of these zircons, three general observations can be made across the history of the Mineral King pendant (Fig. 1). (1) The zircon compositional populations change from tightly clustered in the 195.7 Ma R₄ rhyolite to more dispersed in the 134–136 Ma R₀, R₁, and R₂ rhyolites (Fig. 4). (2) The δ¹⁸O(Zrn) (Table 1) increases from the R₄ rhyolite (∼5.33‰) to the R₀, R₁, and R₂ rhyolites (6.78‰–7.33‰). (3) The Th concentrations in zircons increase from <100 ppm in the R₄ rhyolite to >100 ppm in the R₀, R₁, and R₂ rhyolite (Fig. 6). All these parameters suggest an increased role in recycling and/or incorporation of continental material and/or silicic intrusive material in the upper crust across the history of the Mineral King units.

The uniformity of the zircon composition in the R₄ rhyolite suggests that the zircon grains it contains are not derived from diverse sources, such as preexisting crust or previous intrusions. Substantial incorporation of antecrystic zircon can produce wide variation in zircon trace element compositional populations (Klemetti et al., 2011). Combined with the high partition coefficient of most REEs into zircon, especially HREEs (Hanchar and van Westrenen, 2007), the tight clustering of the R₄ zircons across most trace element parameters (Figs. 6 and 7) would require a restricted magmatic source (compositionally and thermally) without much addition of crust or magma mixing. This could be produced by direct melting of the lower crust to produce the rhyolite or fractionation of the rhyolite from a basaltic parent, without significant interaction with the crustal column or a crustal column lacking in continental material of differing composition. In contrast, the diverse zircon compositional populations in the R₀, R₁, and R₂ rhyolites (Figs. 6 and 7) would be expected in a well-established magmatic system with abundant crystal recycling (Walker et al., 2010; Klemetti et al., 2011; Stelten and Cooper, 2012). The presence of xenocrystic zircon (1.63 Ga) in the R₀ rhyolite also supports the idea that continental crust had to have been incorporated into the batches of rhyolite magma.

These patterns in trace element compositions could be the product of fractional crystallization orassimilation of crustal material and/or precursory silicic intrusions (AFC). To determine which process plays the most significant role in the development of the trace element patterns for the Mineral King rhyolites, we constructed simple models of fractional crystallization of a rhyolite, using a low-Hf (high temperature) zircon from the R₁ unit (Table 2) as a starting composition and assimilation of crustal, silicic material into rhyolitic magmas (Fig. 10). Using partition coefficients (Supplemental Table 4⁵) for selected trace elements (U, Yb, Gd, Hf), an estimate of the trace element composition of the rhyolitic liquid in which the zircon crystallized was calculated (Table 2); this was used to model fractional crystallization using a modal assemblage of a typical unaltered rhyolite (55% quartz, 20% biotite, 20% plagioclase feldspar, 5% amphibole, 0.003% zircon) comparable to those found at the Mineral King pendant. Fractionation of this assemblage cannot account for the U/Yb diversity of zircon compositions found in the R₀, R₁, and R₂, because U/Yb is insensitive to fractionation (Fig. 10; Grimes et al., 2007). Instead, addition by assimilation of U-enriched continental crust (∼5.4 ppm) by rhyolitic magma (Table 2; Fig. 10), likely accompanied by some fractional crystallization, best fits the pattern of increasing U/Yb with increasing Hf concentrations (Fig. 10). We chose to use a nearby granite that likely represents the composition of the middle to upper crust in the region (North Mountain; Wenner and Coleman, 2004). Although it is younger than the Mineral King rhyolites, these granites are voluminous and a good mix of melts extracted from the middle to upper crust, so they probably reflect the compositions of granitic melts (Table 2) that developed within the continental crust on the western margin of North American during the Jurassic to Cretaceous. The oldest R₄ (197 Ma) rhyolite required the least input of any crustal material, while the R₀, R₁, and R₂ (134–136 Ma) require the most input of crustal material to explain their compositional diversity (Fig. 10) in U/Yb versus Hf. The compositional spread of all zircons analyzed (Fig. 10; Supplemental Table 2 [see ^{footnote 2}]) can be explained by the wide range of partition coefficients for U, Yb, and Hf in zircon and potential small differences in parental rhyolite batches.

The δ¹⁸O(Zrn) values from the Mineral King rhyolites are considerably higher than the measured δ¹⁸O (whole rock) values of 4.5‰–5.1‰ reported for two Mineral King metarhyolites (D’Errico et al., 2012). Thus, whole-rock δ¹⁸O values of rhyolites in Mineral King have been lowered by exchange of these rocks by surface waters, possibly seawater or meteoric water. High-silica rhyolite and dacite metavolcanic rocks in the Sierra Nevada produced during high magmatic fluxes in the Mesozoic are in the range of 5.0‰–8.0‰ (Lackey et al., 2005, 2008). Thus, the Mineral King values are similar to values produced during high flux periods (Fig. 9C), rather than to values produced during magmatic lulls defined by Barth et al. (2013). Moreover, their δ¹⁸O values do not indicate remelting of hydrothermally altered antecedent volcanic rocks, which would produce low δ¹⁸O magma and possibly intracrystalline zoning such as seen at the Yellowstone caldera (Watts et al., 2012). Oxygen isotopic compositions are lowest [δ¹⁸O(Zrn) of ∼5.3‰] in the R₄ rhyolite, suggesting that this rhyolite was derived from a more mantle-like parent (∼5.3‰; Valley et al., 1998). Significant incorporation of continental crust or recycling of plutonic cumulates (mush) from the middle and upper crust that has not undergone significant hydrothermal alteration with low δ¹⁸O (meteoric) waters should drive the δ¹⁸O(Zrn) in these magmas to higher values, as seen in the R₀ and R₁ rhyolites, where δ¹⁸O(Zrn) predicts magmatic values of ∼8.7‰–9.3‰, higher than most known volcanic units of the same time period (Fig. 9C). These are δ¹⁸O values similar to those observed in the modern central Andes of Chile, where highly thickened continental crust has a great influence on the compositions of the magmas erupted (James et al., 1985; Hildreth and Moorbath, 1988; Feeley and Sharp, 1995).

The Th compositions of zircons can also be sensitive to the presence of continental crustal or fractionated silicic material; the partition coefficient D_Th for zircon is not well constrained, but is likely high (estimates for peraluminous granite are as high as 22; Bea et al., 1994), so changes in the Th composition of crust being introduced into the magmatic system will be readily apparent. Magmas that have resided in and assimilated upper crustal material should have increased Th contents, as the upper crust has almost 10 times higher Th concentrations than the lower crust and ∼1.6 times higher values than the middle crust (Rudnick and Gao, 2003). In addition, fractionating basaltic magmas should become enriched in Th, because it is highly incompatible (D < 0.01 in basalt; Weaver et al., 1986). Thus, any assimilation of previously existing crystal mushes will also increase Th concentrations of melt. The older R₄ rhyolites have relatively low Th contents, all <100 ppm (Fig. 6C). However, the younger R₀, R₁, and R₂ rhyolites have variable but higher Th contents, ranging into the hundreds to thousands of parts per million (Fig. 6C). This would support the incorporation of continental crust or silicic mushes into the magmas.

All of these lines of evidence (trace element models, oxygen isotopic compositions, Th content) suggest that the R₄ rhyolite, formed during the Early Jurassic, was minimally influenced by continental material or previously existing crystals mushes. This is different than the Triassic rhyolite in the Saddlebag Lake pendant studied by Barth et al. (2012), for which extraction from crystal mushes via rejuvenation is invoked for their formation. However, this is also in broad agreement with conclusions from Barth et al. (2013), where magmas produced during the Late Triassic–Early Jurassic magmatic lull show less evidence of crustal interactions. However, the Early Cretaceous R₀, R₁, and R₂ rhyolites reflect extraction from a preexisting crystal mush with contributions from continental crust.

The rhyolites of the Mineral King pendant provide a window into two periods of caldera volcanism along the western margin of North America; they also record information from periods considered as “background level” of magmatic production compared to episodes of voluminous magmatism in the Middle Jurassic and middle to Late Cretaceous (Irwin and Wooden, 2001; Barth et al., 2013; Fig. 10). Therefore, they also provide useful insight into the sources of magmas during these quiescent periods. Recent discussion of changing zircon trace element chemistry in the detrital record suggests that magmatic lulls record different magma compositions that may be linked to lower budgets of slab-derived fluid and minimal crustal involvement, possibly as a result of large-scale extension or transtension in the arc (Barth et al., 2013). Comparison of the Mineral King data to those key ratios of U/Yb, Th/U, and Yb/Gd used by Barth et al. (2013) shows that the Mineral King rhyolite magmas are broadly similar in composition to magmas sampled in the detrital zircon record for the two lulls identified by Barth et al. (2013) in the Late Triassic–Early Jurassic (lull 1) and Late Jurassic–Early Cretaceous (lull 2). However, there are some key differences from the conclusions of Barth et al. (2013). The R₄ rhyolite, which was within lull 1, has much lower Yb/Gd, suggesting that zircon grew from a more HREE-depleted melt than other lull 1 magmas or at higher crystallization temperatures (Figs. 8 and 11). The R₀, R₁, and R₂ rhyolites have much lower Th/U than other lull 2 magmas (Fig. 9E), while Yb/Gd and U/Yb span larger ranges (Figs. 8 and 11). These observations suggest that the R₀, R₁, and R₂ rhyolite magmas were influenced little by slab fluids. These rhyolites also have a higher whole-rock δ¹⁸O [based on δ¹⁸O(Zrn)] than the few other volcanic units sampled within lull 2 (Fig. 9C).

Barth et al. (2013) implied that magma produced during the magmatic lulls had minimal crustal involvement. However, the δ¹⁸O(Zrn) values of the 134–136 Ma rhyolites (Table 1) clearly show a stronger influence of crustal material in their evolution, as do the trace element models (Fig. 10) and Th concentrations (Fig. 6C) found in zircon. The production of rhyolite with δ¹⁸O(Zrn) values of 5.3‰ at 195 Ma (R₄) is consistent with partial melting of lower crust and little involvement of crustal melts; however, the later (134–136 Ma) rhyolites (R₀, R₁, and R₂) in the pendant show higher crustal inputs, albeit not at levels shown for some >10‰ granites in the Cretaceous (Lackey et al., 2008; Fig. 9). Therefore, the notion of magmatic lulls being background magma production periods with less crustal influence does not appear to be supported by the Early Cretaceous rhyolites at the Mineral King pendant. However, relative to the peak magmatism in the Sierra Nevada from the Triassic to the Cretaceous, the 134–136 Ma rhyolites of the Mineral King pendant do not show contamination that is as extensive, likely reflecting the lower rates of magma supply into the arc during these lulls.

The Early Jurassic magmatism (R₄) has much less evidence for the influence of continental crust or magmatic precursors, but by the Early Cretaceous (R₀, R₁, and R₂), it is clear that continental crust and magmatic recycling played a much larger role in magma genesis. This mirrors the tectonic evolution of western North America, where the continued accretion of exotic terranes to the continent caused a change in the nature of the crust (Schweickert and Lahren, 1993; Ernst, 2010; Barth et al., 2012; Holland et al., 2013). Arc magmatism off North America was probably in the form of island arcs that docked with North America during the Jurassic (Irwin and Wooden, 1999, 2001). The R₄ rhyolite is likely the product of one of these more primitive arcs, as the rhyolite shows little influence of continental crust in oxygen or trace element compositions. This might be similar to rhyolite magmas erupted in island arcs such as the modern Kermadec Islands (Graham et al., 2008) or Mariana Islands (Reagan et al., 2008).

However, as the accretion of arc material progressed into the Early Cretaceous, the influence of continental crust was greater. The δ¹⁸O of rhyolite erupted during the Early Cretaceous is much higher than that of the Early Jurassic (Fig. 11C); this is typically interpreted as a signal of continental crust influence (Lackey et al., 2008; Barth et al., 2013). The wider range of trace element compositions in the R₀, R₁, and R₂ rhyolite zircons also reflect an increased component of crustal recycling, as zircons from different pockets of subvolcanic crystallizing magma capture the variations in magma composition as new magmas intrude (Claiborne et al., 2010a; Klemetti et al., 2011) and smaller batches of magma fractionate (Storm et al., 2011). These crystals are subsequently sampled during each eruption of the volcano associated with each eruptive package (Claiborne et al., 2010b; Klemetti et al., 2011; Stelten and Cooper, 2012). This change reflects the thickening of the crust leading toward the intrusion of the Sierra Nevada batholith. It may be this priming of the crust during the millions of years before the large pulse of granitic magmas that allows for such voluminous magmatism in the crust (e.g., Lipman, 2007; Grunder et al., 2008; Walker et al., 2010). Even though these Early Cretaceous rhyolites of the Mineral King pendant may be scarce in the modern surface rock record across the Sierra Nevada province, the eruption of voluminous rhyolites that show the evidence of long-lived magmatism may be one of the key events that allowed for the intrusion of the batholith in the Late Cretaceous.

The rhyolite tuffs and flows found in the Mineral King pendant, dated using U-Pb spot ages of zircon, represent a combination of allochthonous Late Triassic (195.7 ± 1.4 Ma) and Early Cretaceous volcanic deposits (134.2–136.5 Ma). The distribution of ages from the Mineral King pendant suggests a complex origin. The original chronology for the units in the septum from Busby-Spera (1983) had the rhyolite units becoming younger to the west as a continuous depositional sequence and all units being Triassic in age. However, the new zircon ages presented here show that only one unit is Triassic in age and the rest are Early Cretaceous, suggesting that the pendant most likely was brought together through structural means after the emplacement (134.2 Ma) of the youngest rhyolite.

The 196 Ma and 134–136 Ma rhyolite units represent one of the few records of magmatism for two periods recognized as lulls in magmatic activity across the Late Triassic to Early Cretaceous in western North American margin. The evidence from zircon ages and compositions suggests that magma from the Sierra Nevada arc during the Early Cretaceous was interacting with North American crust, with increasing crustal influence from the Early Jurassic to Early Cretaceous, following the evolution of the western margin of the continent from an island to continental arc. This evidence includes xenocrystic zircon with ages older than 1.6 Ga, δ¹⁸O(Zrn) compositions of zircons from the 134–136 Ma rhyolite that are ∼6.7‰–7.3‰ (rhyolite composition of ∼8.7‰–9.3‰), and trace element compositions that are consistent with a highly fractionated composition similar to granite of the Late Cretaceous Sierran Arc. There is also strong evidence in U/Yb and Th compositions of zircon that requires the input of continental crust into the 134–136 Ma rhyolite units relative to the 197 Ma rhyolite within the Mineral King pendant. This evidence of Early Cretaceous magmatism shows that even though the record of volcanism is scant within the Sierra Nevada arc, voluminous caldera magmatism was occurring along the western margin of North America as continental crust developed.

These data also show how examination of spot ages and trace element analyses of zircon combined with bulk zircon oxygen isotopic data can help unravel the petrogenesis of altered volcanic rocks. The signatures of fractional crystallization and assimilation are recorded in the trace element composition of zircon, and the resistance of zircon to alteration (unlike whole-rock or major phases) allows for the estimation of the relative importance of these processes in the formation of volcanic units. Improved constraints on the partitioning of trace elements into zircon at different compositions and conditions will permit further exploitation of these methods to explain the petrogenesis of ancient volcanic rocks.

We thank David Greene, Megan D’Errico, and Staci Loewy for assistance during field work (summer 2010, 2011) and Joe Wooden and Matt Coble (Stanford University) for assistance with the SHRIMP-RG (sensitive high-resolution ion microprobe–reverse geometry) analyses. We also thank Ilya Bindeman (University of Oregon) for assistance with the oxygen isotope analyses and the National Park Service for permitting sampling for this project. Funding for this research was provided by the Keck Geology Consortium and the Denison University Undergraduate Research Fund. The Pomona College XRF Laboratory is partly supported by National Science Foundation Division of Undergraduate Education grant CCLI 0942447 and Pomona College. Partical support comes from NSF grant EAR-0948706 to Lackey. We also thank Jamey Jones, Craig Grimes, and a third anonymous reviewer for their comments on earlier versions of this manuscript.