Coexisting Discrete Bodies of Rhyolite and Punctuated Volcanism Characterize 1

Home , Lava Creek Tuff

RESEARCH ARTICLE Coexisting Discrete Bodies of Rhyolite and Punctuated 10.1029/2019GC008321 Volcanism Characterize Yellowstone's Post‐Lava Key Points: • Zircons from Yellowstone's Upper Creek Tuff Caldera Evolution ‐ Basin Member rhyolites yield U Pb Christy B. Till1 , Jorge A. Vazquez2 , Mark E. Stelten2 , Hannah I. Shamloo1 , dates defining crystallization 1,3 populations at ~750–550 and and Jamie S. Shaffer ~350–250 ka 1 2 • Discrete bodies of magma School of Earth and Space Exploration, Arizona State University, Tempe, AZ, USA, U.S. Geological Survey, Menlo Park, characterized the subvolcanic CA, USA, 3Now at Arizona State Geological Sciences, Now at New Mexico State University, Las Cruces, NM, USA system during the Upper Basin Member period and during storage of the Lava Creek Tuff ‐ 206 238 • Abstract Ion microprobe Pb/ U geochronology and trace element geochemistry of the unpolished The geochemical and isotopic ‐ evolution of Yellowstone's rims and sectioned interiors of zircons from Yellowstone caldera's oldest post caldera lavas provide post‐caldera rhyolites suggests a insight into the magmatic system during the prelude and aftermath of the caldera‐forming Lava Creek shift in the magmatic supereruption. The post‐caldera lavas compose the Upper Basin Member of the Plateau Rhyolite and fall assimilation/recharge ratio into two groups based on zircon crystallization age: early lavas with zircon ages between ~750 and 550 ka ‐ Supporting Information: and late lavas with zircon ages between ~350 and 250 ka. Zircons from the early erupted East Biscuit Basin • Supporting Information S1 flow yield U‐Pb dates and trace element compositions, which when considered with the Pb isotopic • Table S1 compositions of their coexisting feldspars and pyroxenes, point to an isotopically distinct parental melt • Table S2 • Table S3 present during crystallization of the Lava Creek magma but untapped by the supereruption. Distinct zircon crystallization ages and Pb‐isotope compositions of major minerals between the early and late Upper Basin Member groups suggest contrasting sources in the magma reservoir. As proxies for melt evolution, Correspondence to: the zircons indicate that Yellowstone's post‐caldera rhyolites became more evolved between mid‐ to C. B. Till, late‐Pleistocene time, during the same interval that melting of hydrothermally altered wall rock and [email protected] recharge by new silicic magmas changed in their relative roles. The results from this study indicate that discrete and ephemeral bodies of silicic magma, at times within a mush dominated reservoir and including Citation: during the prelude to the Lava Creek eruption, have characterized Yellowstone's subvolcanic reservoir. Till, C. B., Vazquez, J. A., Stelten, M. E., Shamloo, H. I., & Shaffer, J. S. (2019). Coexisting discrete bodies of rhyolite and punctuated volcanism characterize 1. Introduction Yellowstone's post‐Lava Creek Tuff caldera evolution. Geochemistry, The volcanic history of Yellowstone caldera after the circa 630‐ka eruption (Matthews et al., 2015) of the Geophysics, Geosystems, 20. https://doi. org/10.1029/2019GC008321 Lava Creek Tuff (LCT) includes episodes of rhyolitic eruptions with intervals of volcanic repose on the order of 104–105 years (Christiansen et al., 2007). At least 23 eruptions of mostly effusive rhyolites have occurred Received 12 MAR 2019 during three apparent intervals (Figure 1) since the Lava Creek eruption, with individual flow unit volumes Accepted 28 JUN 2019 of 2–150 km3 for a cumulative volume of >500 km3 (Christiansen, 2001). Together, these post‐caldera rhyo- Accepted article online 11 JUL 2019 lites track the magmatic evolution of the Yellowstone magmatic system (e.g., Befus & Gardner, 2016; Bindeman et al., 2008; Hildreth et al., 1984, 1991; Girard & Stix, 2009, 2010; Pritchard & Larson, 2012; Stelten et al., 2013, 2015, 2017). A rich body of work on the physical and geochemical nature of silicic magmatic systems (see reviews by Bachmann & Huber, 2016; Cashman & Sparks, 2013; Cashman & Giordano, 2014; Lipman & Bachmann, 2015; de Silva & Gregg, 2014) has revealed that silicic magma reservoirs at caldera volcanoes are variable mixtures of melt, fluid, and crystals that are formed and assembled over multiple levels in the crust, and these phases may evolve over time through complex interactions with wall rocks and additions of new magma. A shared conclusion is that silicic magmas spend the majority of their lifetimes as near‐solidus high‐crystallinity bodies within the upper crust, where the high percentage of crystals relative to silicate liquid (referred to as a “crystal mush”) limits their mobility and in turn their eruptibility (e.g., Bachmann & Bergantz, 2004; Cooper & Kent, 2014; Gualda et al., 2012; Hildreth, 2004; Huber et al., 2009; Miller & Wark, 2008). Outstanding questions remain, including (1) What is the architecture of subvolcanic reservoirs including the spatial relation and interconnectivity of crystal‐rich and crystal‐poor magma domains? (2) What are the timescales for storage and geochemical evolution of silicic magma bodies associated with the large caldera‐forming, as well as smaller effusive eruptions? and (3) What are the relative ‐ ©2019. American Geophysical Union. roles of recharge, crystallization, and assimilation during the long term evolution of a caldera volcano, All Rights Reserved. and how are these factors reflected in the crystal and melt record of magmatic‐volcanic evolution? This

TILL ET AL. 1 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

study addresses these questions at Yellowstone by combining zircon U‐Pb geochronology and trace element geochemistry, the Pb‐isotope composition of pyroxenes, feldspars, and glasses, as well as the petrologic constraints from past studies.

2. Geologic Setting 2.1. Temporal Evolution The Yellowstone Plateau volcanic field has generated three caldera‐ forming eruptions over the past 2.1 Ma (Christiansen, 2001), the most recent of which produced the >1,000‐km3 LCT at 631.3 ± 4.3 ka (2σ; Matthews et al., 2015). The LCT ignimbrite around Yellowstone caldera is composed of two members (LCT‐A and LCT‐B) that represent two pyroclastic phases indistinguishable in age but separated by a cooling break (Christiansen, 2001). A recent study concluded that LCT may have been preceded by two additional tuff units that erupted shortly before members A and B, although the volumes of these new units are unclear (Wilson et al., 2018). Following eruption of the LCT, post‐caldera rhyolites were Figure 1. Summary of eruption ages for the second and third caldera cycles erupted over three apparent intervals (~630–450; ~260; and ~170–75 ka) on the Yellowstone Plateau. Stratigraphic order is from Christiansen et al. (2007). Colored boxes represent eruption age with 2σ error. Eruption ages of mostly effusive intracaldera volcanism (Christiansen, 2001; 40 39 ‐ are based on Ar/ Ar geochronology (Christiansen et al., 2007; Matthews Christiansen et al., 2007). Together with some extra caldera rhyolites, et al., 2015; Stelten et al., 2015; Stelten et al., 2018), with the exception of the the post‐caldera lavas from these intervals compose the formal Plateau North Biscuit Basin and East Biscuit Basin flows, which are based on the Rhyolite, with the rhyolites from the early and middle episodes compos- ‐ youngest U Pb crystallization age from zircon rims (this study). Yellow ing the Upper Basin Member (UBM) and the rhyolites of the youngest epi- boxes denote Central Plateau Member (CPM) units. Early and late divisions of Upper Basin Member (UBM) units are denoted by light and dark green sode composing the Central Plateau Member (CPM; Christiansen & colors, respectively. The pre‐caldera Lewis Canyon‐Mount Jackson units are Blank, 1972; Figure 1). The early UBM rhyolites are exposed in the center denoted in teal and the second caldera Island Park units in black. The and eastern margins of the caldera (Figure 2) and yield 40Ar/39Ar dates caldera‐forming units are denoted in red. Unit abbreviations are as follows: between 527 ± 28 and 489 ± 20 ka (Figure 1; 2σ uncertainties, data of MBB (Middle Biscuit Basin), NBB: (North Biscuit Basin), EBB (East Gansecki et al., 1996, recalculated to the Fish Canyon sanidine monitor Biscuit Basin), LCT (Lava Creek Tuff), MFT (Mesa Falls Tuff). SCL (Scaup δ18 Lake), and SBB (South Biscuit Basin). at 28.17 Ma). These early UBM rhyolites are characterized by O values as low as ~1‰ (Bindeman et al., 2008; Bindeman & Valley, 2001; Hildreth et al., 1984; Pritchard & Larson, 2012). Recently, Wilson et al. (2018) pro- posed a revised caldera margin based on newly recognized LCT exposures, which may mean that early UBM rhyolites near Sour Creek dome erupted from vents just outside of Yellowstone caldera. After a circa 220‐kyr hiatus (Figure 1), the late UBM rhyolites made up of the South Biscuit Basin (SBB) flow (255 ± 22 ka; Bindeman et al., 2008) and Scaup Lake (SCL) flow (262 ± 26 ka; Christiansen et al., 2007) were erupted near the Mallard Lake resurgent dome (Figure S1 in the supporting information). Although the two intervals of UBM volcanism erupted rhyolites with similar plagioclase‐rich mineralogy (Christiansen & Blank, 1972), the early and late groups of UBM rhyolites differ in their radiogenic (e.g., Pritchard & Larson, 2012; Stelten et al., 2013, 2015) and oxygen isotopic (Bindeman & Valley, 2001; Hildreth et al., 1984) compositions. This study focuses on the early and late UBM flows erupted in the center portion of Yellowstone caldera. 2.2. Divisions of the Biscuit Basin Rhyolite The Biscuit Basin flow was originally defined as a single geologic unit and placed in Yellowstone's volcanic stratigraphy by Christiansen and Blank (1972). Largely concealed by younger lavas, Hildreth et al. (1984) estimated a minimum eruptive volume of 2.5 km3 (Figure 2). The unit's pervasive perlitic texture was interpreted to reflect emplacement into a caldera lake (Christiansen, 2001; Hildreth et al., 1984). Christiansen

(2001) recognized the Biscuit Basin rhyolite as one of the least silicic (~72 wt.% SiO2) at Yellowstone, with higher CaO and ΣFeO, and suggested that the unit might be several different flows despite its limited geographic extent. Bindeman & Valley (2001) subdivided the Biscuit Basin flow into the North Biscuit Basin (NBB), Middle Biscuit Basin (MBB), and SBB flows based on their documentation of differing δ18O values, with the subdivi- sions corresponding to low hills and outcrops separated from one another by alluvium and sinter deposits. The units appear equally glassy but somewhat differ in mineralogy with SBB having larger and more

TILL ET AL. 2 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

abundant phenocrysts (17–20 vol.%) relative to MBB (10–12 vol.%) and NBB (12–15 vol.%; Bindeman & Valley, 2001). The 40Ar/39Ar date of 527 ± 28 ka from Gansecki et al. (1996) is for sanidines from an outcrop across from Midway Geyser and thus is the eruption age for the MBB ﬂow. Incremental heating of sanidines from SBB yielded an 40Ar/39Ar date of 255 ± 11 ka (1σ). Bindeman et al. (2008) later revised this date to 261 ± 17 ka (1σ; Watts et al., 2012), supporting the subdivision of the original unit. An additional subdivision termed the East Biscuit Basin (EBB) ﬂow was made by Girard and Stix (2009), who recognized an area of outcrops

composed of low‐SiO2 (~70 wt.%) rhyolite lacking quartz and sanidine but with a population of sieve‐textured plagioclase. Watts et al. (2012) reported an 40Ar/39Ar date of 155 ± 9 ka (1σ) for a sample of the NBB ﬂow, which is too young for the sample to be part of the UBM rhyolites. Instead, this NBB sample is likely to represent a CPM unit, possibly the adjacent Elephant Back ﬂow whose reported K‐Ar date is 153 ± 4 ka (2σ; Christiansen et al., 2007).

Figure 2. Schematic geologic map of third caldera cycle rhyolite lavas at 3. Methods Yellowstone caldera. Bold dotted line is the caldera margin due to the 3.1. Lava Samples and Mineral Characterization Lava Creek Tuff eruption (Christiansen, 2001). Pre‐caldera lavas of the Mt. Jackson and Lewis Canyon Rhyolites are shown with the horizontal line Samples for this study were taken from the basal or upper portions of pattern. Upper Basin Member rhyolites are shown with hatched pattern, vitrophyric outcrops, with the exception of NBB whose glassy flow inter- fl fl with SCL ow highlighted in blue and the Biscuit Basin ows highlighted in ior was sampled. Sample details, localities, and bulk rock analyses are pre- orange. The Central Plateau Member rhyolites are shown in gray. Extra‐ caldera rhyolite flows near Norris are shown with a vertical line pattern. sented in the supporting information (Table S1 and Figures 2 and S1). The Midway Geyser is denoted by small black star. Map is modified from major, minor, and trace element composition of each sample was deter- Vazquez et al. (2009) and Troch et al. (2017) and originally based on the mined using X‐ray fluorescence and inductively coupled plasma mass geologic mapping of Christiansen (2001). MBB = Middle Biscuit Basin; spectrometry at Washington State University's GeoAnalytical Lab follow- NBB = North Biscuit Basin; EBB = East Biscuit Basin; SBB = South Biscuit ing the methods described by Johnson et al. (1999) and Kelly (2018). Basin; SCL = Scaup Lake. Zircons with lengths of 100–300 μm were density separated from washed, crushed, and sieved rock samples using standard heavy liquid techniques and a Frantz magnetic separator. The separated zircons were bathed in dilute hydrofluoric acid for ~5 min to remove adhering glass. Individual zircon crystals were handpicked from the separates using a binocular microscope and were embedded together with presectioned and polished fragments of the Temora‐2 zircon age standard into malleable indium metal hosted in a 2.54‐cm diameter aluminum mount. The selected crystals were oriented so that their {001} faces were parallel to the mount surface as shown in Matthews et al. (2015). To analyze the interiors of the zircons, subsets of crystals from each sample were mounted along with Temora‐2 zircons in epoxy and sectioned to reveal the cores of crystals. All analyzed zircons were documented before SHRIMP analyses via cathodoluminescence (CL) and secondary electron imaging with a scanning electron microscope at Stanford University. CL images of crystal interiors were used to identify crystal zoning and guide targeting (Figure S2). Feldspar and clinopyroxene crystals were hand‐picked from coarsely crushed samples using a binocular microscope and were subsequently mounted in epoxy, sectioned, and polished. In addition to these grain mounts, polished thin sections were prepared for in situ microprobe analyses. The major element composition of the feldspar and clinopyroxene was analyzed with a JEOL 8900 electron microprobe at the U.S. Geological Survey in Menlo Park, California, and a JXA‐8530F Electron Hyperprobe at Arizona State University, Tempe, Arizona. All microprobe analyses were conducted with an accelerating voltage of 15 keV, a beam current of 20–25 nA, and beam diameter 1–5 μm. On peak count times were 10–30 s depending on spectrometer configuration, with the exception of Na for which the count times were 10–15 s.

3.2. Ion Microprobe Methods (Zircon and Pyroxene Analyses) Simultaneous U‐Pb dating and measurement of trace elements for unpolished surfaces on euhedral zircon crystals and their interiors, as well as analyses of trace elements in sectioned pyroxenes, were performed by secondary ion mass spectrometry (SIMS) using the Stanford‐U.S. Geological Survey SHRIMP‐RG ion

TILL ET AL. 3 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

microprobe at Stanford University. Analyses of unpolished rims and sectioned interiors followed the analytical protocol and standardization used by Matthews et al. (2015), with calibration to the Temora‐2 standard (418.6 Ma; Mattinson, 2010). U‐Pb isotopic measurements were conducted using a primary beam of nega-

tively charged O2 ions with ~8‐nA intensity for the zircon analyses and ~13‐nA intensity for the pyroxene analyses. The primary beam was focused using Köhler illumination to produce an elliptical sampling pit, approximately 25 μm in diameter and ~4 μm deep. Targets were selected on the sectioned zircon crystals to avoid melt and mineral inclusions visible in reﬂected light and to sample representative domains, that is, rims, intermediate zones, and cores, apparent in CL images (Figure S2). Targeting of the pyroxene crystals was guided by backscattered‐electron images. Calculated 206Pb/238U dates for the zircons have been derived from calibrated U‐Pb ratios and lower inter- cepts on a Tera‐Wasserburg concordia that is modiﬁed to account for initial 238U–230Th and 235U–231Pa disequilibrium (Wendt & Carl, 1985; Sakata, 2018) assuming zircon‐melt Th/U and Pa/U partitioning of 0.2 (Schmitt, 2007) and 2.9 (Sakata et al., 2017), respectively, and a parental melt in uranium‐series secular equilibrium. Assumed common Pb compositions are from 207Pb/206Pb ratios for respective host rocks or associated feldspars reported by Doe et al. (1982), Vazquez et al. (2009), or Pritchard and Larson (2012). Trace element concentrations except Ti are calibrated to 91500 zircon standard using values reported by Wiedenbeck et al. (2004); Ti concentrations have been calibrated to SL13 zircon (6.1 ppm; Hiess et al., 2008). The U‐Pb results, calculated dates, and additional information are provided in Table S2. Trace element concentrations for the pyroxenes were calculated from secondary ion yields normalized to measured 30Si+ and calibrated to NIST 610 glass using the concentration values reported by Hervig et al. (2006). Results for the pyroxenes are provided in Table S3.

3.3. LA‐ICPMS Pb‐Isotope Methods Feldspar and pyroxene were analyzed by laser ablation multiple‐collector inductively coupled plasma mass spectrometry (LA‐MC‐ICPMS) using a Nu Plasma MC‐ICPMS coupled with a 213‐nm laser ablation system housed in the Interdisciplinary Center for Plasma Mass Spectrometry at the University of California, Davis. Operating conditions included 50% laser energy, 100‐μm spot size, 20‐Hz pulse rate, and 25‐s integration time. All isotopes were measured on Faraday detectors including 204Pb and 202Hg. Analyses of feldspars and pyroxenes were interspersed with measurements of the NIST 610 glass standard. NIST 612 and StHs6/80‐G glasses were used as secondary standards. Mass bias was calculated using the measured Pb isotope ratios of NIST 610, assuming a value of 2.1694 for 208Pb/206Pb (Baker et al., 2004). See Stelten et al. (2013) for additional analytical details.

4. Results 4.1. Petrography and Mineral Compositions Modal abundances of the studied units are presented in Table 1. The early UBM lavas contain 40–60% feldspar phenocrysts. Sanidine occurs as subhedral to euhedral phenocrysts in all of the ﬂows except EBB, with the late UBM ﬂows containing the highest abundances. Truncated zoning, sieve textures, and embayed crystal margins are apparent in cross‐polarized microscopy as well as CL images (e.g., Till et al., 2015). Plagioclase occurs in each of the lavas as euhedral to subhedral phenocrysts or as a component of crystal aggregates or glomerocrysts. Except for the EBB, each lava contains euhedral to subhedral quartz pheno-

crysts, commonly with embayments. Plagioclase ranges in composition between ~Ab60 and Ab75, and sanidine ranges between ~Or40 and Or65 (Figure 3). Sanidines from the late UBM lavas are most diverse, with an overall greater range in composition than for the early UBM sanidines (Figure 3). The late UBM sanidines also have a larger range, as well as a higher limit, of Ba concentrations relative to the early UBM lavas or the LCT (Figure 3). Fe‐rich clinopyroxenes (Fe/(Fe+Mg) = ~0.7) in the late UBM lavas are often zoned, with cores that often have exsolution lamellae, a mantle of relatively high Mg/Fe and low heavy rare‐earth elements (HREE) and Rb, and ﬁnally a rim with somewhat lower Mg/Fe and higher HREE and Rb. Exsolution lamellae in their clinopyroxene cores require slow cooling to near‐solidus temperatures of less than 825 °C (Huebner, 1980; Lindsley, 1983; Figure 4). Zircons from the rhyolites are generally euhedral and resemble zircons described from these and other Yellowstone rhyolites (e.g., Matthews et al., 2015; Stelten et al., 2015). Many of the UBM zircons have

TILL ET AL. 4 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Table 1 Petrography and Modal Abundances of Phases in Studied Upper Basin Member Lavas Lava ﬂow Bulk phenocrysts (vol.%) Phenocrysts present and modal abundance (vol.%) Groundmass texture

East Biscuit Basin 10–15 plag (60%), pyx (35%), Fe‐Ti ox (5%) Perlitic North Biscuit Basin 10–15 san (50%), qtz (30%), plag (15%), pyx (1%), Fe‐Ti ox (4%) Spherulitic Middle Biscuit Basin 10–15 san (40%), qtz (30%), plag (25%), pyx (2%), Fe‐Ti ox (3%) Spherulitic South Biscuit Basin 12–15 san (35%), plag (30%), qtz (25%), pyx (7%), Fe‐Ti ox (3%) Perlitic Scaup Lake 12–15 san (40%), plag (30%), qtz (20%), pyx (7%), Fe‐Ti ox (3%) Perlitic

interiors marked by rounded boundaries and truncated zoning (Figure S2). Inclusions of glass, apatite, and oxides are typical in the zircons. The zircons have trace element patterns typical of magmatic crystallization with relative enrichment in HREE and pronounced europium (Eu/Eu*) and cerium anomalies. Zircons from the early UBM rhyolites are distinctive in that they have lower mean concentrations of Hf, Th, and U, small Eu anomalies, and higher mean Ti concentrations relative to the late UBM (Figure 5). In general, the EBB zircons yield the smallest europium anomalies and lowest Ce and Hf. Zircons from the late UBM rhyolites yield Hf concentrations that are higher than those for zircons from the early UBM rhyolites but similar to those from the CPM lavas and LCT (Figure 5). Similar trace element concentrations characterize the rims and interiors of the late UBM zircons, although Hf concentrations are somewhat lower for SBB zircons. The rare‐earth element patterns of the UBM zircons, as tracked by the ratios of middle to heavy rare‐earth elements (Gd/Yb) and light‐ to heavy rare‐earth elements (Nd/Yb), deﬁne coherent covariations that are followed by Th/U, Hf, and Ti concentrations (Figures 5 and 6).

4.2. U‐Pb Geochronology of Zircon Crystal Faces and Interiors for the Biscuit Basin and SCL Flows 206Pb/238U dates for the Biscuit Basin flow zircons primarily fall within the interval between approximately 250–700 ka, with the interiors of a minority of crystals yielding early Pleistocene‐late Pliocene dates (Tables 2 and S2). Zircons from the late UBM flows yield the youngest dates, whereas those from the early UBM flows yield the oldest (Figure 7). For nearly all of the samples, a range of 206Pb/238U dates are associated with the unpolished rims and sectioned interiors of the zircon crystals, although the unpolished rims tend to yield the youngest dates for a given flow. Application of the mixing model of Sambridge and Compston (1994) that deconvolves the 206Pb/238U dates into discrete Gaussian age populations yields multiple populations for the unpolished rims and sectioned interiors. In general, the unpolished rims on zircons yield populations

Figure 3. Biscuit Basin and Scaup Lake feldspar compositions. (a) Compositions of potassium feldspars from UBM and LCT. Unit abbreviations are as in Figure 1. Data for crystals from LCT members A and B are from Matthews et al. (2015) and Shamloo and Till (2019). Inset shows the projected region of the feldspar ternary. SCL and SBB have similar compositions and exhibit a wider range of orthoclase (Or) content relative to LCT and other UBM feldspars. (b) Feldspar or component versus Ba concentration illustrating distinctive compositional populations. The EBB ﬂow lacks sanidine and thus is not represented on this ﬁgure.

TILL ET AL. 5 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

of dates differing by tens of thousands of years, whereas zircon interiors tend to yield a greater number of apparent populations and a larger span of dates (Table 2 and Figure 7). In the cases where an independent 40Ar/39Ar date is associated with the rhyolite flow, the apparent eruption age is indistinguishable from the youngest population of rim dates (Figure 7 and Table 2). 4.3. Pb Isotopic Composition of Feldspars, Pyroxenes, and Glasses From the EBB, SBB, and SCL Flows Sanidines and clinopyroxenes from the late UBM SCL and SBB flows yield Pb‐isotope compositions with similar ranges of 208Pb/206Pb (2.165–2.185) and 207Pb/206Pb (0.882–0.890; Table S3), which overlap the Pb isotope compositions reported for CPM rhyolites (Figure 8). Plagioclase, clinopyroxenes, and glasses from the early UBM EBB flow are distinct, with lower 208Pb/206Pb and 207Pb/206Pb ratios than for the late UBM and CPM units (Figure 8). These results for the EBB flow are consistent with previously published Pb‐isotope data for other early UBM samples (e.g., Pritchard Figure 4. Trace element geochemistry of Scaup Lake and South Biscuit & Larson, 2012; Watts et al., 2012). Basin clinopyroxenes. Concentrations of Yb versus Rb for core, middle, and rim domains of sectioned clinopyroxene phenocrysts analyzed by SHRIMP‐RG ion microprobe. Other units from this study were not analyzed. 5. Discussion Error bars represent 2σ analytical uncertainties. Abbreviations are as in 5.1. Crystallization Ages Recorded by Early and Late UBM Zircons Figure 1. The 206Pb/238U dates represent crystallization ages, including the unpolished rims, because of the slow diffusion of Pb‐Th‐U in zircon even at magmatic temperatures (Cherniak & Watson, 2003). Individual 206Pb/238U dates might be ~2 kyr older than calculated if 226Ra/230Th is close to zero as anticipated for zircon but unconfirmed by experiments or measurements (Schmitt & Vazquez, 2017) and an additional ~3 kyr older (Matthews et al., 2015) if parental melts

Figure 5. Petrochronology of zircons erupted during the third caldera cycle. Flow unit abbreviations are as in Figure 1. Data for unpolished rims or sectioned interiors of single UBM zircon crystals are denoted by crosses or circles, respectively. Crystallization ages are from disequilibrium‐corrected 206Pb/238U dates (Table S2). Data for zircons from LCT, CPM, and the pre–caldera Lewis Canyon and Mt. Jackson Rhyolites are those reported by Matthews et al. (2015), Stelten et al. (2013, 2015), and Troch et al. (2017), respectively.

TILL ET AL. 6 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

had an initial 20% excess of 230Th relative to 238U as measured for some CPM rhyolites (e.g., Stelten et al., 2015). The similarities between the 40Ar/39Ar dates and the young populations of the unpolished rims from SBB and SCL (Table S2) suggest that the youngest dates from rim analyses provide an effective measure of near‐eruption crystallization. Unpolished rims on zircons from other Yellowstone rhyolites also yield populations of dates that are indistinguishable, within analytical uncertainty, of their eruption ages from 40Ar/39Ar dating. For example, SIMS analyses of unpolished rims on LCT zircons yield 206Pb/238U dates (Matthews et al., 2015) that are indistinguishable from their ~630 ka 40Ar/39Ar eruption age (Mark et al., 2017; Matthews et al., 2015), and unpolished rims on CPM zircons yield 230Th/238U dates like those based on 40Ar/39Ar dating of coexisting sanidines (Stelten et al., 2015). Accordingly, the youngest populations of 206Pb/238U dates from the unpolished zircon rims are fl ‐ Figure 6. Rare‐earth element (REE) ratio variation for third caldera cycle interpreted to re ect near eruption crystallization and in turn serve as zircons. Rare‐earth element ratio variation reflecting relative amounts of an effective measure of eruption age (e.g., Vazquez & Lidzbarski, 2012). major phase and Th‐REE‐rich accessory mineral fractionation. Figure Where there are multiple populations of dates from the unpolished rims, symbols, colors, and data sources are as used in Figure 5. the crystals with older rims are interpreted to be antecrysts that were reentrained from the predominantly solidified portions of their magma chamber and/or represent inclusions that were liberated from major mineral hosts near the time of eruption (e.g., Reid & Vazquez, 2017; Tierney et al., 2019). Dates from the interiors of the zircons provide evidence for recycling of some of Yellowstone's earliest magmatic material. The small number of zircons yielding early Pleistocene ages are likely to represent xenocrysts recycled from plutonic or volcanic material related to the first and second caldera cycles responsible for the Huckleberry Ridge and Mesa Falls eruptions, respectively. The interiors of several EBB and NBB zircons yield dates of ~3 Ma, which suggests recycling of intrusive or volcanic material related to the earliest volcanism on the Yellowstone Plateau. Bindeman and Valley (2001) measured a similar distribution of Mesa Falls‐ aged and older dates for sectioned zircons from various early UBM flows and also interpreted them as reflecting recycling of older intrusive or volcanic material. Late Pliocene dates for a minority of the sectioned zircons are noteworthy because the oldest recognized product of volcanism on the Yellowstone Plateau, Snake River Butte, has an eruption age of 2.14 Ma (Rivera et al., 2017), while the youngest volcanism of the older Heise volcanic center, located west of the Yellowstone Plateau, is ~3.9 Ma (Ellis et al., 2017). These late Pliocene interiors are interpreted to reflect recycling of early Yellowstone Plateau intrusions that do not appear to have had volcanic equivalents or whose volcanic equivalents were obliterated by the younger Quaternary caldera‐forming supereruptions. The apparent ~630‐ka date from the unpolished rims of EBB zircons, as well as its presence in the center of Yellowstone caldera, suggests that this rhyolite flow erupted immediately after the caldera‐forming Lava Creek eruption. Unfortunately, attempts to date the EBB eruption using 40Ar/39Ar geochronology of feldspars have been thwarted by a lack of sanidine, as well as a plagioclase population that is riddled with melt inclusions and excess argon (Watts et al., 2012; and this study). The dates for the interiors of the EBB zircons are mostly tens to approximately a hundred thousand years older than their rims, which is a feature also observed for the many CPM rhyolites that erupted later in the caldera cycle (Stelten et al., 2015) and suggests that EBB zircons contain domains representing recycling of antecrysts (Figure S1). The presence of antecrystic interiors in CPM zircons has been interpreted to reflect scavenging of zircons from a long‐lived reservoir of magma mush, a term which we use to refer to a magma crystallized to the point of being barely mobile and/or necessitating melt extraction through compaction (e.g., Bachmann & Huber, 2019; Hildreth, 2004; Miller et al., 2011). The majority of analyzed interiors from EBB zircons yield dates that form a population centered on ~750 ka, which is a minority population of crystallization ages observed in the interiors of LCT zircons and has been interpreted to be antecrystic (cf. Matthews et al., 2015). A similarity between the dates for EBB and LCT unpolished rims and interiors suggests that these magmas were coeval at the time of their zircon rim crystallization and that they scavenged antecrysts of similar age. Scavenging of antecrystic zircons (e.g., Bindeman & Valley, 2001; Rivera et al., 2014; Stelten et al., 2015; Watts et al., 2012) and sanidines (e.g., Gansecki et al., 1996; Matthews et al., 2015; Rivera et al., 2014, 2016; Stelten et al., 2015, 2018) appears to be a

TILL ET AL. 7 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Table 2 206Pb/238U Age Populations for Both the Unpolished Rims and Sectioned Interiors of UBM Zircon Crystals 206Pb/238U age (± 2σ) Rhyolite ﬂow Rims versus Interiors populations and proportions 40Ar/39Ar eruption age (± 2σ)

East Biscuit Basin None Unpolished rims (n = 27) 635 ± 14 ka (0.87) 744 ± 18 ka (0.13) Interiors (n = 28) 649 ± 42 ka (0.10) 753 ± 15 ka (0.69) 812 ± 34 ka (0.21) North Biscuit Basin None Unpolished rims (n = 16) 580 ± 40 ka (0.41) 653 ± 35 ka (0.59) Interiors (n = 25) 654 ± 9 ka (0.92) 1,530 ± 240 ka (0.04) 2,970 ± 610 (0.04) Middle Biscuit Basin 527 ± 28 ka1 Unpolished rims (n = 22) 546 ± 23 ka (0.25) 607 ± 13 ka (0.75) Interiors (n = 34) 621 ± 14 ka (0.67) 875 ± 19 ka (0.18) 1,143 ± 34 ka (0.12) 2,219 ± 95 ka (0.03) South Biscuit Basin 255 ± 22 kab Unpolished rims (n = 22) 257 ± 9 ka (0.33) 300 ± 11 ka (0.67) Interiors (n=14) 271 ± 25 ka (0.20) 388 ± 26 ka (0.63) 618 ± 51 ka (0.16) Scaup Lake 262 ± 26 kac Unpolished rims (n = 26) 244 ± 9 ka (0.76) 355 ± 23 ka (0.24) Interiors (n = 23) 248 ± 10 ka (0.65) 341 ± 22 ka (0.35) Note. Discrete Gaussian age populations were determined using the mixing model of Sambridge and Compston (1994). 40Ar/39Ar dates for MBB and SLF have been recalculated with a calibration to the Fish Canyon sanidine ﬂuence monitor with an age of 28.172 Ma (Rivera et al., 2011) and decay constant from Min et al. (2000). UBM = Upper Basin Member. a40Ar/39Ar date from Gansecki et al. (1996). b40Ar/39Ar date from Bindeman et al. (2008) c40Ar/39Ar date from Christiansen et al. (2007).

common occurrence for Yellowstone rhyolites, although to varying degrees. Identification of antecrystic sanidine, incompletely degassed of their radiogenic Ar, suggests that scavenging of these crystals occurred shortly before eruption (Andersen et al., 2017; Gansecki et al., 1996; Rivera et al., 2017). The youngest apparent population of the NBB unpolished rims yields an age of 580 ± 40 ka, which suggests an eruption within a few tens of kyr of the MBB eruption. The youngest apparent population of 206Pb/238U dates from the unpolished MBB zircon rims (546 ± 23 ka) is consistent with their eruption age (527 ± 28 ka) based on 40Ar/39Ar dating of MBB sanidines (Tables 2 and S2). The MBB flow appears to have an eruption age that is ~30 kyr older than most of the other early UBM rhyolites, which yield eruption ages of ~495 ka (Figure 1) based on 40Ar/39Ar geochronology (recalculated from Gansecki et al., 1996). The youngest mode of 206Pb/238U dates from the rims of SCL and SBB zircons is within uncertainty of their associated ~260‐ka eruption ages based on 40Ar/39Ar dating of sanidines (Tables 2 and S2), indicating zircon saturation near the time of eruption. Accordingly, the zircons with rims that are significantly older than the eruption age suggest the near‐eruption liberation of antecrysts that had been sequestered from the zircon‐ saturated rhyolite, for example, within solidified reservoir margins or included in larger minerals (e.g., Reid & Vazquez, 2017; Storm et al., 2011). These two rhyolites hold a unique place in the eruptive timeline of Yellowstone because no other post‐caldera rhyolites have been identified with eruption ages of ~260 ka. Their eruption ended an apparent ~220‐kyr hiatus in volcanism at Yellowstone (Figure 1), with

TILL ET AL. 8 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

geospeedometry using trace elements in SCL sanidines suggesting thermal rejuvenation of the subcaldera magmatic system within years of eruption (Till et al., 2015). The radiogenic Pb, Nd, and Hf isotopic compositions of the SBB and SCL rhyolites, including their feldspars and pyroxenes, are similar to each other (Figure 8) and resemble the compositions of CPM rather than early UBM rhyolites. This secular change in radiogenic isotope compositions points to hybridization of the shallow Yellowstone crust via addition of new silicic magmas (Stelten et al., 2017).

5.2. Zircon Trace Element Compositions of Early UBM Rhyolites The trace element concentrations of the early UBM zircons, when tied to their crystallization ages, provide information about the melt evolution of the subvolcanic reservoir during and shortly after the Lava Creek supereruption. The covariations of trace elements in UBM zircons generally follow those defined by LCT and CPM zircons, for example, Eu/Eu* versus Yb, U, or Hf, which at face value are consistent with crystallization from melts related by fractional crystallization (e.g., Claiborne et al., 2006, Claiborne, Miller, & Wooden, 2010). Covariation of Yb/Gd and Nd/Yb, which reflect the slopes of REE patterns for the zircons, is consistent with melts related by fractionation of major phases and accessory minerals in rhyolites. Yb/Gd in zircons is usually positively correlated with indices of differentiation such as Hf (Andersen et al., 2019; Barth & Wooden, 2010; Claiborne, Miller, & Wooden, 2010) and reflects the fractionation of pyroxenes or accessory minerals such as titanite that sequester middle REE (e.g., Andersen et al., 2019; Figure 6). Titanite is absent in Yellowstone rhyolites, but clinopyroxene and orthopyroxene hosting inclusions of apatite, chevkinite, and zircon are common (Stelten et al., 2015; Vazquez et al., 2009, 2014). Nd/Yb typically tracks melts related by fractionation of light REE‐rich accessory minerals such as chevkinite and allanite (Reid et al., 2010). Chevkinite occurs as microphenocrysts in CPM (Vazquez et al., 2014) and UBM (Bindeman et al., 2008) rhyolites, as well as the later erupted portions of LCT while allanite occurs in the early‐erupted portions of LCT (Hildreth, 1981). Accordingly, the covariation in Nd/Yb and Yb/Gd of these rhyolites can be explained primarily by relative fractionation of pyroxenes and chevkinite ± allanite. However, UBM rhyolites define an array with higher Nd/Yb than LCT zircons and a subset of CPM zircons (Figure 6), suggesting a lesser role for fractionation of chevkinite as well as distinctive parental melts (e.g., Andersen et al., 2019). These characteristics could reflect standard crystal‐melt separation and/or residual mineralogy in a partially melted source. Based on their trace element compositions, EBB zircons appear to reflect a least evolved end‐member melt with regard to recording fractionation and crystallization temperature (Figure 5). Within zircons crystallized from a suite of melts related along the same liquid line of descent, smaller europium anomalies and lower Hf concentrations reflect crystallization from less evolved parent melts during magmatic evolution (Claiborne et al., 2006, Claiborne, Miller, & Wooden, 2010). The trace element characteristics of the EBB zircons suggest crystallization from less fractionated silicic magma relative to the LCT and the late UBM rhyolites. This is

also consistent with EBB's distinctly lower (~70 wt.%) SiO2 and high Mg, Ca, and Ba, (Girard & Stix, 2009) relative to other Yellowstone rhyolites. The trace element difference between the EBB and LCT zircons is noteworthy because both yield similar 206Pb/238U dates, which together suggests the presence of different but coeval silicic melts with distinct trace element compositions and lineages. δ18O values for the interiors of EBB zircons yielding near‐eruption ages are like the values for LCT zircons (Watts et al., 2012), suggesting some isotopic similarities between the coeval melts. However, EBB plagioclase and glass have lower δ18O values than for the same LCT phases (Watts et al., 2012), suggesting remelting or assimilation of shallow wall rocks (Troch, Ellis, Harris, et al., 2018) after the caldera eruption.

Using the Ti‐in‐zircon geothermometer (Ferry & Watson, 2007) and assuming a TiO2 activity of 0.8 based on EBB compositions reported by Girard and Stix (2009) and the saturation model of Hayden and Watson (2007), the Ti concentrations of EBB zircons yield crystallization temperatures of ~700–800 °C, with most in the 750–800 °C range. The temperature of zircon saturation for the bulk EBB composition and Zr concentration (Girard & Stix, 2009) is ~825 °C (approximately ±15 °C 1σ uncertainty from the model) using the Boehnke et al. (2013) formulation that revises and has superseded the original Watson and Harrison (1983) model for zircon saturation. Harrison et al. (2007) demonstrated that zircon saturation temperatures calculated from bulk compositions could be misleading without accounting for Zr evolution during the crystallization of major phases. However, the general correspondence between calculated zircon saturation temperatures and the Ti geothermometer for the EBB zircons suggests that

TILL ET AL. 9 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Figure 7. Relative probability and rank order plots of 206Pb/238U zircon date distributions. Plotted 206Pb/238U dates are corrected for initial uranium‐series disequilibrium and are given with 1σ error bars. Orange curves are for unpolished zircon crystal rims (surfaces), and green curves are for sectioned interiors. Dashed gray line for the NBB ﬂow is a combined rim and interior probability curve. Vertical gray bar with dashed black line delimits reported 40Ar/39Ar eruption age with 2σ uncertainty. 40Ar/39Ar dates are from Gansecki et al. (1996), Christiansen et al. (2007), and Bindeman et al. (2008).

TILL ET AL. 10 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Figure 8. Pb‐isotope composition of feldspars and pyroxenes in Yellowstone rhyolites. (a) Plot of 207Pb/206Pb versus 208Pb/206Pb for SBB, SCL, and EBB pyroxenes, feldspar, and glass relative to other Yellowstone rhyolites. Abbreviations and units are as in Figure 1 with the addition of Huckleberry Ridge Tuff Members B and C, LCT Members A and B, second caldera cycle Island Park domes, and the pre‐Yellowstone Heise volcanic center. Data sources indicated in the legend. (b) Inset showing Pb‐isotope ratios from LA‐MC‐ICPMS analyses and illustrating the general overlap of ratios between cores and rims from single rhyolite units and the distinctiveness of EBB minerals and glasses.

they are best explained by growth from their host rhyolite or a compositional equivalent that was saturated in zircon. Zircons from the MBB and NBB ﬂows also yield crystallization temperatures in the 750–800 °C range. These temperatures deﬁne a more restricted range than the approximately 700– 900 °C interval calculated for LCT zircons by Matthews et al. (2015), as well as zircons from the older Huckleberry Ridge Tuff (Rivera et al., 2014) and Mesa Falls Tuff (Rivera et al., 2016), suggesting that the early UBM rhyolites capture a more limited interval of magmatic evolution. It is noteworthy that interiors with >1,000 ppm U concentrations and large Eu anomalies, which are common in LCT zircons (Matthews et al., 2015) and noted in pre‐LCT rhyolite lavas (Troch, Ellis, Schmitt, et al., 2018), appear to be less common in the UBM as well as CPM rhyolites (Figure 5). These U‐rich domains in Yellowstone zircons are likely to represent near‐solidus crystallization at vapor‐saturated conditions representing >80% crystallization of rhyolite (Troch, Ellis, Schmitt, et al., 2018), likely from cooler domains within the magma system where crystallinity is elevated and melt compositions are highly evolved (Matthews et al., 2015; Rivera et al., 2016; Troch, Ellis, Schmitt, et al., 2018). A relative paucity of equivalent domains within the zircons from the UBM and CPM rhyolites suggests storage away from, less recycling, or depletion of the near‐solidus margins of the reservoir that had bounded the LCT magma.

TILL ET AL. 11 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

5.3. Trace Elements in Zircons From the ~260 ka Late UBM Flows The trace element compositions and 206Pb/238U dates of the SCL and SBB zircons are consistent with the geochemistry and similar circa 260 ka 40Ar/39Ar eruption ages for these two late UBM ﬂows. Some difference between the two rhyolites appears to be recorded by the trace elements of their zircons. Despite similar concentrations for the interiors of zircons from both rhyolites, the rims of SCL zircons have higher Hf and Ti concentrations than SBB zircons, which may be taken to indicate crystallization in somewhat more evolved yet hotter melt compositions near the time of eruption. This relation is reversed from the negative Hf‐Ti relation that characterizes most silicic magmas (e.g., Claiborne, Miller, & Wooden, 2010, Claiborne, Miller, Flanagan, et al., 2010). A possible explanation for their different Ti concentrations is that the rhyolitic

magmas had different TiO2 activities, which would result in different partitioning coefficients for Ti even if both magmas had the same temperature. Nevertheless, different Hf concentrations are likely to reflect differences in parental melt composition (Claiborne, Miller, & Wooden, 2010). Clinopyroxenes (Figure 4) and sanidines (Till et al., 2015) from the late UBM flows have zoning suggesting rejuvenation followed by renewed crystallization near the time of eruption. This rejuvenation might be reflected in the somewhat greater range of Ti concentrations for late UBM zircon rims yielding near‐eruption dates (Figure 5). The rims of these late UBM zircons appear to have higher δ18O than their cores (Bindeman et al., 2008), suggesting that renewed crystallization occurred from a melt with a smaller component of recycled low δ18O material.

5.4. Discrete Bodies of Rhyolite During the Prelude to the Lava Creek Supereruption The large number of trace element analyses provide insight into the time‐compositional evolution and architecture of the Yellowstone subvolcanic reservoir leading up to the LCT eruption and during its post‐caldera volcanism. The differences in crystallization ages between rims and interiors of LCT zircons based on high‐ spatial resolution sampling (Matthews et al., 2015) and whole crystal sampling (Wotzlaw et al., 2015) suggest that a relatively limited interval of crystallization (e.g., 40–60 kyr) is recorded by the voluminous caldera‐ forming rhyolite, with little recycling of older material, after being sourced from the melting of middle and shallow crustal lithologies (Wotzlaw et al., 2015). Similarly, short intervals of zircon crystallization appear to be a feature of the older caldera‐forming Mesa Falls and Huckleberry Ridge eruptions on the Yellowstone Plateau (Rivera et al., 2014, 2016; Wotzlaw et al., 2015). In contrast, CPM rhyolites have spreads or “tails” of zircon crystallization ages indicating scavenging of crystals from a long‐lived (tens to hundreds of kyr) reservoir of mush, which together with eruption‐aged U‐Th dates for both the rims of zircons and major phases suggest that each final eruptible magma was extracted from this mush in <10 kyr (Stelten et al., 2015). Like the CPM rhyolites, a tail of dates is apparent from the interiors of early and late UBM zircons (Figure 7 and Table 2), suggesting that the host melts have similarly recycled material from mush or wall rocks of the subvolcanic reservoir. At face value, the difference in the range of zircon ages between the caldera‐forming and the post‐caldera rhyolites (i.e., less than ~60 ka vs. up to ~300 ka, respectively) could be taken to reflect contrasting rates of magma production. However, geospeedometry suggests that the SCL magma came about by remelting/rejuvenation of a near‐solidus source within decades of eruption (Till et al., 2015), with its tail of old zircon ages representing recycling of antecrysts. Similarly, fast rates for “mush‐to‐ eruption” have been inferred at other calderas where tails of zircon crystallization ages are observed (e.g., Allan et al., 2013, 2017). Considering the documented overlap between crystallization ages for zircon rims and their sanidine 40Ar/ 39Ar eruption ages for other Yellowstone rhyolites (e.g., Matthews et al., 2015; Stelten et al., 2015) and the occurrence of the EBB flow within the caldera margins (Girard & Stix, 2009), the 206Pb/238U dates for the EBB zircons suggest that the EBB flow was erupted ~30 kyr or less after the LCT eruption. An alternative is that EBB zircons are wholly (i.e., rims and interiors) an antecrystic cargo recycled by a new and mobile rhyolitic melt that erupted before any new zircon rims could grow (e.g., Klemetti & Clynne, 2014; Tierney et al., 2016). The EBB zircon interiors yielding dates of ~750 ka and older (Figure 7 and Table 2) are likely to be antecrystic, like those in the LCT (Matthews et al., 2015). The trace element compositions indicating that EBB zircons grew from a relatively unevolved and hot rhyolite (e.g., smallest Eu anomaly, low Hf, and high Ti) are consistent with the relatively unevolved and hot composition of EBB melt (e.g., smallest Eu anomaly and highest Zr) based on groundmass and inclusion glass analyses (cf. Girard & Stix, 2009) and point to an affinity between EBB zircons and their host. Moreover, apparent partition coefficients, including for Th/U and Eu/Eu*, derived from average EBB groundmass glass (Girard & Stix, 2009) and

TILL ET AL. 12 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

average EBB zircon rims (Table S2), effectively match the coefficients calculated in other studies using eruption‐aged zircon and host glass (e.g., Reid & Vazquez, 2017; Stelten et al., 2015), suggesting the majority of rims on EBB zircons grew from the melt represented by their host glass. Despite 206Pb/238U dates indicating crystallization at same time as the zircons from the LCT magma, the rims and interiors of EBB zircons have limited compositional overlap with LCT zircons (Figures 5 and 6), suggesting that the melts responsible for the EBB and LCT zircons were coeval but distinct in chemical composition. The compositions of the EBB zircons suggest crystallization from a melt less evolved and hotter than responsible for LCT zircons, which is consistent with the major, minor, and trace element compositions of the bulk rocks. The low‐silica rhyolite bulk composition of the EBB flow suggests that it represents a magma little modified by fractionation after generation via crustal melting (e.g., Jean et al., 2018). A question then arises as to the spatial relation of the melts responsible for the coeval rims on EBB and LCT zircons: Does the EBB melt represent the least evolved portion of a zoned magma body (e.g., Hildreth, 1981) left behind after the majority was excavated in the caldera‐forming eruption or does it reflect a distinct “subchamber” (e.g., Gualda & Ghiorso, 2013; Wotzlaw et al., 2014, 2015) that evolved separately? The zircon trace element concentrations of the two units reveal some key differences in composition, as illustrated in the plot of Nd/Yb versus Gd/Yb, where LCT exhibits a distinct slope relative to EBB. As discussed earlier, this distinction in slope represents different liquid lines of descent reflecting fractionation of pyroxenes and chevkinite ± allanite, with EBB and other UBM rhyolites forming a relatively coherent group that differs from LCT and CPM rhyolites (Figure 6). The coexistence of the EBB and LCT magmas as discrete bodies that did not experience mixing prior to eruption is supported by the Pb‐isotope composition of EBB clinopyroxenes and plagioclase relative to LCT Member A (Figure 8). This difference is supported by the different Sr and Nd isotopic compositions for EBB and LCT whole rocks and glasses (Pritchard & Larson, 2012). Differences in their radiogenic and stable isotope compositions likely reflect differences in the relative roles of crustal melting, fractional crystallization, magma recharge, and shallow wall rock cannibalization, all of which have been inferred for the Yellowstone system (e.g., Bindeman et al., 2008; Hildreth et al., 1991; Stelten et al., 2017; Wotzlaw et al., 2015). The low δ18O values for EBB relative to LCT suggest that remelting or assimilation of hydrothermally altered wall rocks played a larger role in the evolution of the EBB rhyolite (Troch, Ellis, Harris, et al., 2018). Together this evidence suggests that Yellowstone's magma reservoir was characterized by the presence of independent melt bodies during evolution of the Lava Creek rhyolite. Beyond the EBB versus LCT distinction, there is isotopic evidence that LCT Members A and B represent different bodies of magma sequentially tapped from subchambers during the Lava Creek eruption. Until recently, the compositional and mineralogical variations of LCT have been interpreted to reflect a magma chamber vertically zoned in composition and temperature (Christiansen, 2001; Hildreth, 1981). However, zircons from LCT Members A and B yield identical 206Pb/238U dates (Matthews et al., 2015; Wotzlaw et al., 2015) but have distinctly different O and Hf isotope compositions suggesting that they grew from silicic melts that were coeval yet physically separated in subchambers before being tapped for eruption (Wotzlaw et al., 2015). LCT Member A represents about the same volume, ~500 km3, as Member B (Christiansen, 2001), but is more evolved based on trace elements and is correlative with ash fallout that was dispersed over a more restricted area than Member B (Izett, 1981). The results from past studies of the LCT (e.g., Matthews et al., 2015; Wilson et al., 2018; Wotzlaw et al., 2015) suggest that two to four geochemically distinct magma bodies existed during the prelude to the Lava Creek eruption and that there was thermal rejuvenation within decades of eruption (Shamloo & Till, 2019). Our results from the EBB zircons identify an additional contemporaneous melt body albeit one that was not tapped by the caldera‐forming eruption and was instead erupted later to form the EBB flow. Discrete bodies of magma within subvolcanic reservoirs based on glass compositions or crystallization ages have been inferred for caldera‐forming and related eruptions for other volcanic systems such as Long Valley (e.g., Gualda & Ghiorso, 2013), Taupo (e.g., Rubin et al., 2016), and older Yellowstone‐related rhyolites (e.g., Wotzlaw et al., 2014). Distinct compositions of glasses, often separated by concentration gaps, composing the vitric population of distal fallout (e.g., Bégué et al., 2014; Cooper et al., 2012; Gatti et al., 2014; Pearce et al., 2014) or the melt inclusion population of early‐erupted crystals (e.g., Myers et al., 2016; Swallow et al., 2018) suggest the tapping of independent bodies of silicic magma during the course of at least some caldera‐ forming eruptions. Compositional zoning of ignimbrites may reflect a single body of magma that was zoned in composition (e.g., Hildreth, 1981) and temperature or the tapping of compositionally and spatially

TILL ET AL. 13 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

discrete magma bodies (e.g., Gualda & Ghiorso, 2013). In the Yellowstone system, a diversity of melt inclusions observed in the basal fallout of the Huckleberry Ridge Tuff has been interpreted to reﬂect multiple discrete “cupolas” of the main magma body (Myers et al., 2016; Swallow et al., 2018). The tuff of Kilgore, which erupted at 4.5 Ma from the pre‐Yellowstone Heise system located on the Snake River Plain, contains zircons with indistinguishable high‐precision 206Pb/238U crystallization ages but different oxygen isotope compositions, indicating the merging of coeval and independent bodies of rhyolite in the subvolcanic system within millennia of eruption (Wotzlaw et al., 2014). Elsewhere, coeval yet discrete bodies of magmas have been inferred for caldera‐ and post‐caldera rhyolites erupted over a geologically brief period of time based on matching populations of crystallization ages for their zircons (e.g., Barker et al., 2014; Charlier et al., 2003).

5.5. Depths of Storage for the LCT and EBB Magmas The bulk compositions of the EBB and LCT rhyolites can be used to quantify their crystallization depths and provide insight into the architecture of the subvolcanic system around the time of the Lava Creek eruption. The location of the quartz + feldspar cotectics in the haplogranite projection for rhyolitic systems has long been known to provide constraints on the pressure of origin for rhyolitic magmas (e.g., Luth et al., 1964; Tuttle & Bowen, 1958). Updates in experimental petrology during the last 20 years have better constrained

the effect of Ca and dissolved H2O on the location of the ternary minimum and cotectics (Blundy & Cashman, 2001; Gualda & Ghiorso, 2015; Wilke et al., 2017). We calculated pressures for members A and

B of LCT using both literature whole rock and glass compositions and a range of H2O contents consistent with measurements of LCT melt inclusions of 2–4 wt.% H2O (Gansecki, 1998). When used with the Rhyolite‐MELTS geobarometer, the LCT‐A bulk composition returns pressures of 260–300 MPa, whereas the LCT‐A glasses suggest pressures of ~180 MPa. Differences in these estimates suggest a two‐stage storage history (i.e., crystallization at ~300 MPa, followed by shallower glass re‐equilibration at ~180 MPa), as has been inferred for Taupo volcanic zone rhyolites (e.g., Gualda et al., 2018) or the presence of a disequilibrium mineral assemblage such that the bulk composition yields inaccurate estimates (Ganne et al., 2018). The LCT‐B bulk compositions suggest pressures of ~150 MPa, whereas the glasses suggest pressures of ~110 MPa. In contrast, the Wilke et al. (2017) “DERP” haplogranite geobarometer yields pressures of ~230 MPa for the LCT‐B bulk composition, consistent with the observation that the Wilke et al. (2017) geobarometer consistently returns higher pressures than Rhyolite‐MELTS calculations due to the nature of the model calibration (Gualda et al., 2019). Haplogranite geobarometry fails to provide an estimate of absolute pressure for

the EBB composition because the rhyolite is quartz undersaturated. However, the low SiO2 content (~70 wt.%) of the EBB rhyolite and its approximate location on the haplogranite ternary suggests a relatively high‐pressure origin (>500 MPa) for this magma (Gualda & Ghiorso, 2013). The accuracy of these pressures depends on the quality of the reported geochemical analyses and degree to which the major minerals represent equilibrium assemblages. Despite some of these uncertainties, the relative differences between EBB, LCT‐A, and LCT‐B are likely robust and indicate crystallization pressures where EBB>LCT‐A>LCT‐B.

5.6. Distinct Sources for the Early and Late UBM Rhyolites Although high‐resolution petrochronology may be precluded by the ±104‐year uncertainties associated with single 206Pb/238Uor238U‐230Th dates from SIMS analyses (Kent & Cooper, 2017), the approximately half ‐ million‐ year interval of crystallization between the LCT and CPM (Figure 7) is sufﬁcient in duration to resolve long‐term evolution of the magmatic system based on the crystallization ages and compositions of zircons. Coupled with crystallization age, the EBB and early UBM zircons show similar trace element characteristics, consistent with their less evolved compositions relative to the LCT and CPM rhyolites (Figure 5). The increase in Hf concentrations and Eu anomalies in zircons from early‐ to late‐UBM rhyolites, as well as variations in the ratio of REEs (Figure 6), is consistent with a change to more evolved magma compositions reﬂecting crystallization of major and accessory phases (e.g., Claiborne, Miller, & Wooden, 2010). However, a direct relation between the early and late UBM rhyolites by fractionation is ruled out by contrasting radiogenic isotope compositions for feldspars (this study; Pritchard & Larson, 2012; Watts et al., 2012), pyroxenes (Figure 8), glasses (Vazquez et al., 2009), and whole rocks (Hildreth et al., 1991). Moreover, the populations of crystallization ages for late UBM zircons are wholly younger than those from the early UBM rhyolites, that is, little carryover of older crystals, suggesting a different source and origin or effective dissolution of older zircons. The late UBM rhyolites have radiogenic isotope compositions with a larger mantle component than for the early UBM rhyolites (Vazquez et al., 2009; Pritchard & Larson, 2012; Stelten et al., 2015, 2017). Hence,

TILL ET AL. 14 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

the late UBM rhyolites represent distinct melts that may reflect hybridization of intrusive material left from the LCT or early UBM episodes via recharge ± assimilation (e.g., Hildreth et al., 1984, 1991; Pritchard & Larson, 2012) or intrusion of entirely new silicic magma from a mid‐crustal source (e.g., Stelten et al., 2015, 2017). The oxygen isotope compositions of the late UBM rhyolites indicate that assimilation of altered wall rocks remained important (Bindeman et al., 2008; Hildreth et al., 1984; Troch, Ellis, Harris, et al., 2018). 5.7. Late UBM Rhyolites: Distinct Cousins of CPM rhyolites The late UBM rhyolites have closer whole rock Nd, Hf, and Pb isotopic affinities to the younger CPM rhyolites (Stelten et al., 2015, 2017; Vazquez & Reid, 2002; Vazquez et al., 2009; Figure 8). The changes in these radiogenic isotopes together with the oxygen‐isotope compositions of the individual rhyolitic magmas at Yellowstone have been interpreted to be set by the relative roles of silicic recharge from the middle to lower crust and shallow crustal melting and assimilation (e.g., Bindeman et al., 2008; Bindeman & Valley, 2001; Hildreth et al., 1991; Stelten et al., 2015, 2017; Vazquez et al., 2009; Wotzlaw et al., 2015). Based on radiogenic isotopes, Yellowstone silicic magmas are derived from crustal melting and reflect a crust/mantle pro- portion of 0.3–0.4, where unity is entirely crustal (e.g., Jean et al., 2018; Stelten et al., 2017). It is worth restating that the episode of late UBM volcanism occurred ~220 ka after eruption of the early UBM rhyolites and ~90 ka before the episode of CPM volcanism and thus represents a distinct period of volcanism marking a petrologic shift. Temporally distinct volcanism between the first and second caldera cycles similarly marked a petrologic shift on the Yellowstone Plateau, suggesting generation of isolated magma batches during periods of low magma flux between caldera‐forming magma chambers (Rivera et al., 2018). Early UBM magmas have the lowest low δ18O compositions, suggesting the largest role for melting of shallow hydrothermally altered rock during the third cycle (Bindeman & Valley, 2001; Bindeman et al., 2008). Shifts to higher δ 18O values over the course of the post‐caldera volcanism appear to mirror the secular evolution toward radiogenic isotope compositions that are less crustal and more like those observed for Yellowstone basalts (Hildreth et al., 1984, 1991; Stelten et al., 2017). Addition of ~5% juvenile rhyolite, presumably generated at deeper crustal levels, to late UBM rhyolite is needed to form the closer‐to‐mantle Pb‐Hf‐Nd isotopic compositions of the oldest CPM rhyolites and an additional ≤5% juvenile rhyolite to form the youngest CPM rhyolites (Stelten et al., 2017). A role for recharge by silicic derivatives of mantle‐derived basalts is indicated by CPM zircons with near‐mantle Hf‐isotope compositions (Stelten et al., 2015). A general petrogenetic relation between the early and late UBM and CPM magmas is suggested by similar trace element compositions including covariation of REE ratios (e.g., Gd/Yb vs. Nd/Yb; Figure 6). A minority of CPM zircons have cores that yield 238U‐230Th dates extending back to the age of late UBM volcanism, which together with their similar radiogenic isotopic compositions suggests a petrogenetic connection and the potential for a long‐lived (i.e., 150–250 ka) and shallow reservoir of crystal‐rich mush that was hybridized via recharge (Loewen & Bindeman, 2015; Stelten et al., 2015, 2017). Zircons from the pre‐caldera Mount Haynes and Lewis Canyon rhyolites fall along the same REE trend as the UBM and CPM zircons (Figure 6), but their major phases are isotopically different (Figure 8). In contrast, LCT zircons define a distinct trend in terms of REE ratios, suggesting that they crystallized from melt that evolved along a different compositional path than for the UBM, CPM, and at least two of the pre‐caldera rhyolites. Based on the trace elements in zircons and the whole rock/glass radiogenic isotope characteristics of the rhyolites related to Yellowstone caldera, the early UBM, late UBM, and CPM eruptive periods appear to represent distinct chapters of rhyolite production and evolution, perhaps reflecting waxing and waning of magma input and reservoir temperature (e.g., Rivera et al., 2018). 5.8. At What Stage Is the Yellowstone Caldera Cycle? Multiple studies have addressed the question of where the magmatic system at Yellowstone is headed next (e.g., Christiansen et al., 2007; Girard & Stix, 2012), with one interpretation being that the magmatic system has been “dying” since the late UBM eruptions (e.g., Watts et al., 2012) thus posing a low probability for eruption or, alternatively, has started the transition to a new caldera cycle (e.g., Spell et al., 2004). The end of a Yellowstone hot spot caldera cycle is marked by the eruption of tholeiitic basalts though the caldera floor, signifying cooling and complete solidification of the silicic reservoir (Christiansen, 2001; Jean et al., 2018). However, the volcanic expression of the transition to a new caldera cycle is unclear. About two million years of volcanic hiatus marked the transition between the end of the Heise caldera complex and the start of volcanism during the first caldera cycle of the Yellowstone Plateau, but this hiatus appears to be

TILL ET AL. 15 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

anomalously long relative to older caldera complexes of the Snake River Plain and might reflect the hot spot encountering a more refractory part of the lower crust (Anders et al., 2009). In large part, uncertainty exists about whether the magmatic system has cooled, heated, or remained the same since the LCT eruption because there has been discordance between temperatures derived from mineral geothermometers and experimental petrology. For example, oxygen isotope measurements of CPM clinopyroxenes, quartz, zircon, magnetite, as well as the MELTS algorithm yield apparent temperatures ranging between 810 and 1150 °C (Loewen et al., 2015; Loewen & Bindeman, 2016), which suggests the erupted rhyolites and their crystals reflect near‐liquidus melts unassociated with a reservoir of near‐ solidus mush (Loewen & Bindeman, 2016). In contrast, crystallization experiments using actual CPM samples coupled with geobarometry using quartz‐hosted melt inclusions indicate major phase crystallization below ~825°, with crystallization of the observed sanidine‐clinopyroxene‐quartz phenocryst assemblage at 750 ± 25 °C, likely associated with a cool crystal‐rich mush (Befus & Gardner, 2016). Zircon geothermometry based on Ti concentrations and melt geothermometry based on glass Zr concentrations are generally consistent with the lower experimentally derived temperatures, possibly punctuated by periods of heating. A straightforward application of the Boehnke et al. (2013) zircon saturation model is determination of a melt‐quench temperature from zircon‐bearing glass, with eruptive cooling responsible for quenching the zircon‐saturated melt. Groundmass glasses from CPM and late UBM rhyolites have Zr concentrations that inversely correlate with their ~260‐ to 75‐ka eruption ages, suggesting increasing near‐ eruption temperatures from ~725 to 775 °C (Loewen & Bindeman, 2015). The ranges of Zr concentrations preserved by quartz‐hosted melt inclusions from ~170‐ to 75‐ka CPM rhyolites are overlapping (Befus & Gardner, 2016), suggesting that these rhyolites generally record the same ~730–780 °C interval of crystallization. However, the Zr concentrations of CPM groundmass glasses are higher than for most of their associated quartz‐hosted melt inclusions (Befus & Gardner, 2016), suggesting that magmatic heating is a near‐ eruption phenomenon. In contrast, decreasing Ti concentrations in the rims of CPM zircons and quartz suggest a lowering of near‐eruption temperatures over time, from ~775 to 725 °C (Stelten et al., 2015). Based on the change to less crustal radiogenic isotope compositions beginning at ~260 ka, there appears to have been an increased role for inputs of juvenile rhyolite (Stelten et al., 2017). The antecrystic portions of CPM zircons suggest that the proto‐CPM reservoir began to be assembled immediately after eruption of the late UBM rhyolites (Stelten et al., 2015). By the onset of CPM volcanism at ~170 ka, the subvolcanic reservoir was likely to have been characterized by a voluminous crystal‐rich mush from which eruptible melts were generated to produce the CPM rhyolites (Stelten et al., 2015; Loewen & Bindeman, 2015). Resurgence of the caldera's center accompanied the onset of CPM volcanism, likely to have reflected a rise of silicic magma to shallow levels (Christiansen, 2001). Additional recharge of the system by silicic magma continued to change isotopic compositions to more mantle‐like ratios between 170 and 75 ka (Stelten et al., 2015, 2017). Given the uncertain geothermometry, it is unclear whether there has been long‐term cooling or heating of the subvolcanic system during this post‐caldera evolution. It appears that punctuated magmatism and volcanism is a characteristic, with repose durations of approximately 35–95 kyr between eruptions of the first caldera cycle (Rivera et al., 2017) and approximately 90–220 kyr between the early‐late UBM and CPM eruptions (Figure 1). Accordingly, the ~75 kyr since eruption of the last CPM rhyolite may represent an analogous hiatus. Indeed, the crust beneath Yellowstone caldera remains energetic. Gas flux at Yellowstone suggests continued intrusion of mafic magmas beneath the level of the silicic magma reservoir at rates comparable to Kilauea volcano (Lowenstern & Hurwitz, 2008). Geophysical studies have concluded that the present‐day upper crust beneath Yellowstone is characterized by a mushy silicic reservoir with 15 –30% melt (e.g., Chu et al., 2010; Farrell et al., 2014). However, the reservoir may contain ~200–600 km3 of melt (Farrell et al., 2014), although the compartmentalization and potential eruptibility of this melt remains uncertain (Lowenstern et al., 2017).

6. Conclusions Coeval U‐Pb dates for compositionally distinct zircons from the EBB ﬂow indicate synchronous crystallization of different melts during the prelude to the Lava Creek supereruption and point to the presence of multiple subchambers during storage of the Lava Creek magma. Untapped by the Lava Creek eruption, the EBB zircons and their melt erupted into the center of the caldera. The age‐geochemical characteristics of zircons

TILL ET AL. 16 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

from early and late UBM lavas are consistent with the long‐term geochemical evolution evidenced by radiogenic isotopes. The Pb‐isotope compositions of feldspars and pyroxenes from the early and late UBM lavas are consistent with distinct sources for these rhyolites. Together with the observations from mineral geothermometry and past petrologic studies, these results point to a geochemical evolution in which assimilation of altered wall rocks was most prominent during the time of early UBM magmatism and input of juvenile rhyolites was most apparent during the time of late UBM and CPM magmatism. Uncertainties in the temperature‐time evolution of the post‐caldera magmas leave inconclusive the questions as to whether the Yellowstone magmatic system has been heating or cooling to a moribund state or is in a period of repose similar to those separating the different episodes of post‐caldera volcanism.

Acknowledgments References Samples were collected during field seasons in 2012 and 2015 under Allan, A. S., Barker, S. J., Millet, M. A., Morgan, D. J., Rooyakkers, S. M., Schipper, C. I., & Wilson, C. J. (2017). A cascade of magmatic ‐ – Yellowstone National Park research events during the assembly and eruption of a super sized magma body. Contributions to Mineralogy and Petrology, 172(7), 1 34. ‐ permits YELL‐2012‐SCI‐5950 and Allan, A. S., Morgan, D. J., Wilson, C. J., & Millet, M. A. (2013). From mush to eruption in centuries: Assembly of the super sized Oruanui – ‐ ‐ ‐ YELL‐2015‐SCI‐6078. We also magma body. Contributions to Mineralogy and Petrology, 166(1), 143 164. https://doi.org/10.1007/s00410 013 0869 2 acknowledge Geosciences Australia for Anders, M. H., Saltzman, J., & Hemming, S. R. (2009). Neogene tephra correlations in eastern Idaho and Wyoming: Implications for ‐ ‐ providing the Temora‐2 standard. The Yellowstone hotspot related volcanism and tectonic activity Yellowstone hotspot related volcanism and tectonic activity. GSA Bulletin, ‐ – full data sets used for the results in this 121(5 6), 837 856. https://doi.org/10.1130/B26300.1 paper are provided in the supporting Andersen, N. L., Jicha, B. R., Singer, B. S., & Hildreth, W. (2017). Incremental heating of Bishop Tuff sanidine reveals preeruptive radio- – information. We would like to thank genic Ar and rapid remobilization from cold storage. Proceedings of the National Academy of Sciences, 114(47), 12407 12412. https://doi. Matt Coble and Marsha Lidzbarski for org/10.1073/pnas.1709581114 their assistance in the SHRIMP‐RG lab, Andersen, N. L., Singer, B. S., & Coble, M. A. (2019). Repeated rhyolite eruption from heterogeneous hot zones embedded within a cool, – Kari Cooper for discussions about the shallow magma reservoir. Journal of Geophysical Research: Solid Earth, 124, 2582 2600. https://doi.org/10.1029/2018JB016418 ‐ interpretations of our data, Justin Bachmann, O., & Bergantz, G. W. (2004). On the origin of crystal poor rhyolites: Extracted from batholithic crystal mushes. Journal of – Glessner for aiding with the LA‐ICPMS Petrology, 45(8), 1565 1582. https://doi.org/10.1093/petrology/egh019 – Pb analyses, and Leslie Hayden and Bachmann, O., & Huber, C. (2016). Silicic magma reservoirs in the Earth's crust. American Mineralogist, 101(11), 2377 2404. https://doi. ‐ ‐ Bob Jones for electron microprobe org/10.2138/am 2016 5675 imaging and analyses. Many thanks to Bachmann, O., & Huber, C. (2019). The inner workings of crustal distillation columns: The physical mechanisms and rates controlling – the journal‐solicited reviewers Olivier phase separation in silicic magma reservoirs. Journal of Petrology, 60(1), 3 18. https://doi.org/10.1093/petrology/egy103 207 –204 Bachmann, Jonathan Miller, and Katy Baker, J., Peate, D., Waight, T., & Meyzen, C. (2004). Pb isotopic analysis of standards and samples using a Pb Pb double spike and ‐ ‐ ‐ ‐ – Chamberlain and USGS‐solicited thallium to correct for mass bias with a double focusing MC ICP MS. Chemical Geology, 211(3 4), 275 303. https://doi.org/10.1016/j. reviewer Amy Gilmer for their chemgeo.2004.06.030 ‐ thoughtful and constructive reviews of Barker, S. J., Wilson, C. J. N., Smith, E. G., Charlier, B. L. A., Wooden, J. L., Hiess, J., & Ireland, T. R. (2014). Post supereruption magmatic fl – the manuscript. This work was sup- reconstruction of Taupo volcano (New Zealand), as re ected in zircon ages and trace elements. Journal of Petrology, 55(8), 1511 1533. ported by a USGS Mendenhall https://doi.org/10.1093/petrology/egu032 Postdoctoral Fellowship and start‐up Barth, A. P., & Wooden, J. L. (2010). Coupled elemental and isotopic analyses of polygenetic zircons from granitic rocks by ion microprobe, ‐ – support from Arizona State University with implications for melt evolution and the sources of granitic magmas. Chemical Geology, 277(1 2), 149 159. https://doi.org/10.1016/j. to C. B. Till. Any use of trade, firm, or chemgeo.2010.07.017 product names is for descriptive pur- Befus, K. S., & Gardner, J. E. (2016). Magma storage and evolution of the most recent effusive and explosive eruptions from Yellowstone – ‐ ‐ ‐ poses only and does not imply endor- Caldera. Contributions to Mineralogy and Petrology, 171(4), 1 19. https://doi.org/10.1007/s00410 016 1244 x sement by the U.S. Government. Bégué, F., Deering, C. D., Gravley, D. M., Kennedy, B. M., Chambefort, I., Gualda, G. A. R., & Bachmann, O. (2014). Extraction, storage and eruption of multiple isolated magma batches in the paired Mamaku and Ohakuri eruption, Taupo volcanic zone, New Zealand. Journal of Petrology, 55(8), 1653–1684. https://doi.org/10.1093/petrology/egu038 Bindeman, I. N., Fu, B., Kita, N. T., & Valley, J. W. (2008). Origin and evolution of silicic magmatism at Yellowstone based on ion microprobe analysis of isotopically zoned zircons. Journal of Petrology, 49(1), 163–193. https://doi.org/10.1093/petrology/egm075 Bindeman, I. N., & Valley, J. W. (2001). Low‐δ18O rhyolites from Yellowstone: Magmatic evolution based on analyses of zircons and individual phenocrysts. Journal of Petrology, 42(8), 1491–1517. https://doi.org/10.1093/petrology/42.8.1491 Blundy, J., & Cashman, K. (2001). Ascent‐driven crystallisation of dacite magmas at Mount St Helens, 1980–1986. Contributions to Mineralogy and Petrology, 140(6), 631–650. https://doi.org/10.1007/s004100000219 Boehnke, P., Watson, E. B., Trail, D., Harrison, T. M., & Schmitt, A. K. (2013). Zircon saturation re‐revisited. Chemical Geology, 351(C), 324–334. https://doi.org/10.1016/j.chemgeo.2013.05.028 Cashman, K. V., & Giordano, G. (2014). Calderas and magma reservoirs. Journal of Volcanology and Geothermal Research, 288(C), 28–45. https://doi.org/10.1016/j.jvolgeores.2014.09.007 Cashman, K. V., & Sparks, R. S. J. (2013). How volcanoes work: A 25 year perspective. GSA Bulletin, 125(5‐6), 664–690. https://doi.org/ 10.1130/B30720.1 Charlier, B. L. A., Peate, D. W., Wilson, C. J. N., Lowenstern, J. B., Storey, M., & Brown, S. J. A. (2003). Crystallisation ages in coeval silicic magma bodies: 238U–230Th disequilibrium evidence from the Rotoiti and Earthquake Flat eruption deposits, Taupo Volcanic Zone, New Zealand. Earth and Planetary Science Letters, 206(3‐4), 441–457. https://doi.org/10.1016/S0012‐821X(02)01109‐3 Cherniak, D. J., & Watson, E. B. (2003). Diffusion in zircon. Reviews in Mineralogy and Geochemistry, 53(1), 113–143. https://doi.org/ 10.2113/0530113 Christiansen, R. L. (2001). The Quaternary and Pliocene Yellowstone plateau volcanic field of Wyoming, Idaho, and Montana, U.S. Geological Survey Professional Paper 729‐G, 145 p. Christiansen, R. L., & Blank, H. R. (1972). Volcanic stratigraphy of the Quaternary rhyolite plateau in Yellowstone National Park, U.S. Geological Survey Professional Paper, 729‐B,1–28. Christiansen, R. L., Lowenstern, J. B., Smith, R. B., Heasler, H., Morgan, L. A., Nathenson, M., et al. (2007). Preliminary assessment of volcanic and hydrothermal hazards in Yellowstone National Park and vicinity, U.S. Geological Survey Open File Report, 1071, 94. Chu, R., Helmberger, D. V., Sun, D., Jackson, J. M., & Zhu, L. (2010). Mushy magma beneath Yellowstone. Geophysical Research Letters, 37, L01306. https://doi.org/10.1029/2009GL041656

TILL ET AL. 17 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Claiborne, L. L., Miller, C. F., Flanagan, D. M., Clynne, M. A., & Wooden, J. L. (2010). Zircon reveals protracted magma storage and recycling beneath Mount St. Helens. Geology, 38(11), 1011–1014. https://doi.org/10.1130/G31285.1 Claiborne, L. L., Miller, C. F., Walker, B. A., Wooden, J. L., Mazdab, F. K., & Bea, F. (2006). Tracking magmatic processes through Zr/Hf ratios in rocks and Hf and Ti zoning in zircons: An example from the Spirit Mountain batholith, Nevada. Mineralogical Magazine, 70(5), 517–543. https://doi.org/10.1180/0026461067050348 Claiborne, L. L., Miller, C. F., & Wooden, J. L. (2010). Trace element composition of igneous zircon: A thermal and compositional record of the accumulation and evolution of a large silicic batholith, Spirit Mountain, Nevada. Contributions to Mineralogy and Petrology, 160(4), 511–531. https://doi.org/10.1007/s00410‐010‐0491‐5 Cooper, G. F., Wilson, C. J. N., Millet, M.‐A., Baker, J. A., & Smith, E. G. C. (2012). Systematic tapping of independent magma chambers during the 1 Ma kidnappers supereruption. Earth and Planetary Science Letters, 313‐314,23–33. https://doi.org/10.1016/j. epsl.2011.11.006 Cooper, K. M., & Kent, A. J. R. (2014). Rapid remobilization of magmatic crystals kept in cold storage. Nature, 506(7489), 480–483. https:// doi.org/10.1038/nature12991 de Silva, S. L., & Gregg, P. M. (2014). Thermomechanical feedbacks in magmatic systems: Implications for growth, longevity, and evolution of large caldera‐forming magma reservoirs and their supereruptions. Journal of Volcanology and Geothermal Research, 282(C), 77–91. https://doi.org/10.1016/j.jvolgeores.2014.06.001 Doe, B. R., Leeman, W. P., Christiansen, R. L., & Hedge, C. E. (1982). Lead and strontium isotopes and related trace elements as genetic tracers in the Upper Cenozoic rhyolite‐basalt association of the Yellowstone Plateau Volcanic Field. Journal of Geophysical Research, 87(B6), 4785–4806. Ellis, B. S., Szymanowski, D., Wotzlaw, J. F., Schmitt, A. K., Bindeman, I. N., Troch, J., et al. (2017). Post‐caldera volcanism at the Heise volcanic field: Implications for petrogenetic models. Journal of Petrology, 58(1), 115–136. https://doi.org/10.1093/petrology/egx007 Farrell, J., Smith, R. B., Husen, S., & Diehl, T. (2014). Tomography from 26 years of seismicity revealing that the spatial extent of the Yellowstone crustal magma reservoir extends well beyond the Yellowstone caldera. Geophysical Research Letters, 41, 3068–3073. https://doi.org/10.1002/2014GL059588 Ferry, J. M., & Watson, E. B. (2007). New thermodynamic models and revised calibrations for the Ti‐in‐zircon and Zr‐in ‐rutile thermo- meters. Contributions to Mineralogy and Petrology, 154(4), 429–437. https://doi.org/10.1007/s00410‐007‐0201‐0 Ganne, J., Bachmann, O., & Feng, X. (2018). Deep into magma plumbing systems: Interrogating the crystal cargo of volcanic deposits. Geology, 46(5), 415–418. https://doi.org/10.1130/G39857.1 Gansecki, C., Mahood, G., & McWilliams, M. O. (1996). 40Ar/39Ar geochronology of rhyolites erupted following collapse of the Yellowstone caldera, Yellowstone Plateau volcanic field: Implications for crustal contamination. Earth and Planetary Science Letters, 142(1‐2), 91–107. https://doi.org/10.1016/0012‐821X(96)00088‐X Gansecki, CA (1998). 40Ar/39Ar geochronology and pre‐eruptive geochemistry of the Yellowstone Plateau volcanic field rhyolites. Doctoral dissertation, Stanford University, 212 p. Gatti, E., Villa, I. M., Achyuthan, H., Gibbard, P. L., & Oppenheimer, C. (2014). Geochemical variability in distal and proximal glass from the Youngest Toba Tuff eruption. Bulletin of Volcanology, 76(9), 859. https://doi.org/10.1007/s00445‐014‐0859‐x Girard, G., & Stix, J. (2009). Magma recharge and crystal mush rejuvenation associated with early post‐collapse upper basin member rhyolites, Yellowstone Caldera, Wyoming. Journal of Petrology, 50(11), 2095–2125. https://doi.org/10.1093/petrology/egp070 Girard, G., & Stix, J. (2010). Rapid extraction of discrete magma batches from a large differentiating magma chamber: The Central Plateau Member rhyolites, Yellowstone Caldera, Wyoming. Contributions to Mineralogy and Petrology, 160(3), 441–465. https://doi.org/10.1007/ s00410‐009‐0487‐1 Girard, G., & Stix, J. (2012). Future volcanism at Yellowstone caldera: Insights from geochemistry of young volcanic units and monitoring of volcanic unrest. GSA Today,4–10. https://doi.org/10.1130/GSATG143A.1 Gualda, G. A., Bégué, F., Pamukcu, A. S., & Ghiorso, M. S. (2019). Rhyolite‐MELTS vs DERP—Newer does not make it better: A comment on ‘The Effect of Anorthite Content and Water on Quartz–Feldspar Cotectic Compositions in the Rhyolitic System and Implications for Geobarometry’ by Wilke et al. (2017); Journal of Petrology, 58, 789–818. Journal of Petrology, 60(4), 855–864. Gualda, G. A. R., & Ghiorso, M. S. (2013). The Bishop Tuff giant magma body: An alternative to the Standard Model. Contributions to Mineralogy and Petrology, 166(3), 755–775. https://doi.org/10.1007/s00410‐013‐0901‐6 Gualda, G. A. R., & Ghiorso, M. S. (2015). MELTS_Excel: A Microsoft Excel‐based MELTS interface for research and teaching of magma properties and evolution. Geochemistry, Geophysics, Geosystems, 16, 315–324. https://doi.org/10.1002/2014GC005545 Gualda, G. A. R., Ghiorso, M. S., Lemons, R. V., & Carley, T. L. (2012). Rhyolite‐MELTS: A modified calibration of MELTS optimized for silica‐rich, fluid‐bearing magmatic systems. Journal of Petrology, 53(5), 875–890. https://doi.org/10.1093/petrology/egr080 Gualda, G. A. R., Gravley, D. M., Connor, M., Hollmann, B., Pamukcu, A. S., Bégué, F., et al. (2018). Climbing the crustal ladder: Magma storage‐depth evolution during a volcanic flare‐up. Science Advances, 4(10), eaap7567. Harrison, T. M., Watson, E. B., & Aikman, A. B. (2007). Temperature spectra of zircon crystallization in plutonic rocks. Geology, 35(7), 635–634. https://doi.org/10.1130/G23505A.1 Hayden, L. A., & Watson, E. B. (2007). Rutile saturation in hydrous siliceous melts and its bearing on Ti‐thermometry of quartz and zircon. Earth and Planetary Science Letters, 258(3‐4), 561–568. https://doi.org/10.1016/j.epsl.2007.04.020 Hervig, R. L., Mazdab, F. K., Williams, P., Guan, Y., Huss, G. R., & Leshin, L. A. (2006). Useful ion yields for Cameca IMS 3f and 6f SIMS: Limits on quantitative analysis. Chemical Geology, 227(1‐2), 83–99. https://doi.org/10.1016/j.chemgeo.2005.09.008 Hiess, J., Nutman, A. P., Bennett, V. C., and Holden, P. (2008). Ti‐in‐zircon thermometry applied tocontrasting Archean metamorphic and igneous systems: Chemical Geology, 247, 323–338. https://doi.org/10.1016/j.chemgeo.2007.10.012 Hildreth, W. (1981). Gradients in silicic magma chambers: Implications for lithospheric magmatism. Journal of Geophysical Research, 86(B11), 10153–10192. https://doi.org/10.1029/JB086iB11p10153 Hildreth, W. (2004). Volcanological perspectives on Long Valley, Mammoth Mountain, and Mono Craters: Several contiguous but discrete systems. Journal of Volcanology and Geothermal Research, 136(3‐4), 169–198. https://doi.org/10.1016/j.jvolgeores.2004.05.019 Hildreth, W., Christiansen, R. L., & O'Neil, J. R. (1984). Catastrophic isotopic modification of rhyolitic magma at times of caldera subsidence, Yellowstone Plateau volcanic field. Journal of Geophysical Research, 89(B10), 8339–8369. https://doi.org/10.1029/JB089iB10p08339 Hildreth, W., Halliday, A. N., & Christiansen, R. L. (1991). Isotopic and chemical evidence concerning the genesis and contamination of basaltic and rhyolitic magma beneath the Yellowstone Plateau volcanic field. Journal of Petrology, 32(1), 63– 138. https://doi.org/ 10.1093/petrology/32.1.63 Huber, C., Bachmann, O., & Manga, M. (2009). Homogenization processes in silicic magma chambers by stirring and mushification (latent heat buffering). Earth and Planetary Science Letters, 283(1‐4), 38–47. https://doi.org/10.1016/j.epsl.2009.03.029

TILL ET AL. 18 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Huebner, J. S. (1980). Pyroxene phase equilibria at low pressure. Reviews in Mineralogy and Geochemistry, 7(1), 213–288. Izett, G. A. (1981). Volcanic ash beds: Recorders of Upper Cenozoic silicic pyroclastic volcanism in the western United States. Journal of Geophysical Research, 86(B11), 10200–10222. https://doi.org/10.1029/JB086iB11p10200 Jean, M. M., Christiansen, E. H., Champion, D. E., Vetter, S. K., Phillips, W. M., Schuth, S., & Shervais, J. W. (2018). Caldera‐life cycles of the Yellowstone hotspot track: Death and rebirth of the Heise caldera. Journal of Petrology, 59(8), 1643–1670. https://doi.org/10.1093/ petrology/egy074 Johnson, D. M., Hooper, P. R., Conrey, R. M., & Laboratory, G. (1999). XRF analysis of rocks and minerals for major and trace elements on a single low dilution Li‐tetraborate fused bead. Advances in X‐Ray Analysis, 41, 843–867. Kelly, D.S. (2018). Analysis of geological materials by low dilution fusion at the Peter Hooper GeoAnalytical Lab (Washington State University). Washington State University, Pullman, WA, USA. Kent, A. J. R., & Cooper, K. M. (2017). How well do zircons record the thermal evolution of magmatic systems? Geology, 46(2), 111–114. https://doi.org/10.1130/G39690.1 Klemetti, E. W., & Clynne, M. A. (2014). Localized rejuvenation of a crystal mush recorded in zircon temporal and compositional variation at the Lassen Volcanic Center, Northern California. PLoS ONE, 9(12), e113157. https://doi.org/10.1371/journal.pone.0113157 Lindsley, D. H. (1983). Pyroxene thermometry. American Mineralogist, 68(5–6), 477–493. Lipman, P. W., & Bachmann, O. (2015). Ignimbrites to batholiths: Integrating perspectives from geological, geophysical, and geochrono- logical data. Geosphere, 11(3), 705–743. https://doi.org/10.1130/GES01091.1 Loewen, M. W., & Bindeman, I. N. (2015). Oxygen isotope and trace element evidence for three‐stage petrogenesis of the youngest episode (260–79 ka) of Yellowstone rhyolitic volcanism. Contributions to Mineralogy and Petrology, 170(4), 1–25. https://doi.org/10.1007/s00410‐ 015‐1189‐5 Loewen, M. W., & Bindeman, I. N. (2016). Oxygen isotope thermometry reveals high magmatic temperatures and short residence times in Yellowstone and other hot‐dry rhyolites compared to cold‐wet systems. American Mineralogist, 101(5), 1222–1227. https://doi.org/ 10.2138/am‐2016‐5591 Lowenstern, J. B., & Hurwitz, S. (2008). Monitoring a supervolcano in repose: Heat and volatile flux at the Yellowstone Caldera. Elements, 4(1), 35–40. https://doi.org/10.2113/GSELEMENTS.4.1.35 Lowenstern, J. B., Sisson, T. W., & Hurwitz, S. (2017). Probing magma reservoirs to improve volcano forecasts. Eos,1–9. Luth, W. C., Jahns, R. H., & Tuttle, O. F. (1964). The granite system at pressures of 4 to 10 Kilobars. Journal of Geophysical Research, 69(4), 759–773. Mark, D. F., Renne, P. R., Dymock, R. C., Smith, V. C., Simon, J. I., Morgan, L. E., et al. (2017). High‐precision 40Ar/39Ar dating of Pleistocene tuffs and temporal anchoring of the Matuyama‐Brunhes boundary. Quaternary Geochronology, 39,1–23. https://doi.org/ 10.1016/j.quageo.2017.01.002 Matthews, N. E., Vazquez, J. A., & Calvert, A. T. (2015). Age of the Lava Creek supereruption and magma chamber assembly at Yellowstone based on 40Ar/39Ar and U‐Pb dating of sanidine and zircon crystals. Geochemistry, Geophysics, Geosystems, 16, 2508–2528. https://doi. org/10.1002/2015GC005881 Mattinson, J. M. (2010). Analysis of the relative decay constants of 235U and 238U by multi‐step CA‐TIMS measurements of closed‐system natural zircon samples. Chemical Geology, 275(3‐4), 186–198. https://doi.org/10.1016/j.chemgeo.2010.05.007 Miller, C. F., Furbish, D. J., Walker, B. A., Claiborne, L. L., Koteas, G. C., Bleick, H. A., & Miller, J. S. (2011). Growth of plutons by incremental emplacement of sheets in crystal‐rich host: Evidence from Miocene intrusions of the Colorado River region, Nevada, USA. Tectonophysics, 500(1‐4), 65–77. https://doi.org/10.1016/j.tecto.2009.07.011 Miller, C. F., & Wark, D. A. (2008). Supervolcanoes and their explosive supereruptions. Elements, 4(1), 11–15. https://doi.org/10.2113/ GSELEMENTS.4.1.11 Min, K., Mundil, R., Renne, P. R., & Ludwig, K. R. (2000). A test for systematic errors in 40Ar/39Ar geochronology though comparison with U/Pb analysis of a 1.1 Ga rhyolite. Geochimicaet Cosmochimica Acta, 64,73–98. Myers, M. L., Wallace, P. J., Wilson, C. J. N., Morter, B. K., & Swallow, E. J. (2016). Prolonged ascent and episodic venting of discrete magma batches at the onset of the Huckleberry Ridge supereruption, Yellowstone. Earth and Planetary Science Letters, 451(C), 285–297. https://doi.org/10.1016/j.epsl.2016.07.023 Pearce, N. J. G., Westgate, J. A., Gatti, E., Pattan, J. N., Parthiban, G., & Achyuthan, H. (2014). Individual glass shard trace element analyses confirm that all known Toba tephra reported from India is the c. 75‐ka Youngest Toba eruption. Journal of Quaternary Science, 29(8), 729–734. https://doi.org/10.1002/jqs.2741 Pritchard, C. J., & Larson, P. B. (2012). Genesis of the post‐caldera eastern Upper Basin Member rhyolites, Yellowstone, WY: From volcanic stratigraphy, geochemistry, and radiogenic isotope modeling. Contributions to Mineralogy and Petrology, 164(2), 205–228. https://doi. org/10.1007/s00410‐012‐0733‐9 Reid, M. R., & Vazquez, J. A. (2017). Fitful and protracted magma assembly leading to a giant eruption, Youngest Toba Tuff, Indonesia. Geochemistry, Geophysics, Geosystems, 18, 156–177. https://doi.org/10.1002/2016GC006641 Reid, M. R., Vazquez, J. A., & Schmitt, A. K. (2010). Zircon‐scale insights into the history of a Supervolcano, Bishop Tuff, Long Valley, California, with implications for the Ti‐in‐zircon geothermometer. Contributions to Mineralogy and Petrology, 161 (2), 293–311. https:// doi.org/10.1007/s00410‐010‐0532‐0 Rivera, T. A., Darata, R., Lippert, P. C., Jicha, B. R., & Schmitz, M. D. (2017). The duration of a Yellowstone super‐eruption cycle and implications for the age of the Olduvai subchron. Earth and Planetary Science Letters, 479, 377–386. https://doi.org/10.1016/j. epsl.2017.08.027 Rivera, T. A., Furlong, R., Vincent, J., Gardiner, S., Jicha, B. R., Schmitz, M. D., & Lippert, P. C. (2018). Volcanism at 1.45 Ma within the Yellowstone Volcanic Field, United States. Journal of Volcanology and Geothermal Research, 357, 224–238. https://doi.org/10.1016/j. jvolgeores.2018.04.030 Rivera, T. A., Schmitz, M. D., Crowley, J. L., & Storey, M. (2014). Rapid magma evolution constrained by zircon petrochronology and 40Ar/ 39Ar sanidine ages for the Huckleberry Ridge Tuff, Yellowstone, USA. Geology, 42(8), 643–646. https://doi.org/10.1130/G35808.1 Rivera, T. A., Schmitz, M. D., Jicha, B. R., & Crowley, J. L. (2016). Zircon petrochronology and 40Ar/39Ar sanidine dates for the Mesa Falls Tuff: Crystal‐scale records of magmatic evolution and the short lifespan of a large Yellowstone magma chamber. Journal of Petrology, 198–199, egw053–28. https://doi.org/10.1093/petrology/egw053 Rivera, T. A., Storey, M., Zeeden, C., Hilgen, F. J., & Kuiper, K. (2011). A refined astronomically calibrated 40Ar/39Ar age for Fish Canyon sanidine. Earth and Planetary Science Letters, 311(3–4), 420–426. Rubin, A., Cooper, K. M., Leever, M., Wimpenny, J., Deering, C., Rooney, T., et al. (2016). Changes in magma storage conditions following caldera collapse at Okataina Volcanic Center, New Zealand. Contributions to Mineralogy and Petrology, 171(1), 1–4.

TILL ET AL. 19 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Sakata, S. (2018). A practical method for calculating the U‐Pb age of Quaternary zircon: Correction for common Pb and initial disequilibria. Geochemical Journal, 52(3), 281–286. https://doi.org/10.2343/geochemj.2.0508 Sakata, S., Hirakawa, S., Iwano, H., Danhara, T., Guillong, M., & Hirata, T. (2017). A new approach for constraining the magnitude of initial disequilibrium in Quaternary zircons by coupled uranium and thorium decay series dating. Quaternary Geochronology, 38(C), 1–12. https://doi.org/10.1016/j.quageo.2016.11.002 Sambridge, M. S., & Compston, W. (1994). Mixture modeling of multi‐component data sets with application to ion‐probe zircon ages. Earth and Planetary Science Letters, 128(3‐4), 373–390. https://doi.org/10.1016/0012‐821X(94)90157‐0 Schmitt, A. K. (2007). Ion microprobe analysis of (231Pa)/(235U) and an appraisal of protactinium partitioning in igneous zircon. American Mineralogist, 92(4), 691–694. https://doi.org/10.2138/am.2007.2449 Schmitt, A. K., & Vazquez, J. A. (2017). Secondary ionization mass spectrometry analysis in petrochronology. Reviews in Mineralogy and Geochemistry, 83(1), 199–230. https://doi.org/10.2138/rmg.2017.83.7 Shamloo, H. I., & Till, C. B. (2019). Decadal transition from quiescence to supereruption: Petrologic investigation of the Lava Creek Tuff, Yellowstone Caldera, WY. Contributions to Mineralogy and Petrology, 174(4), 32. https://doi.org/10.1007/s00410‐019‐1570‐x Spell, T. L., Smith, E. I., Nastanski, N. M., & Bennett, K. (2004). Establishment and Evolution of a new silicic magma system north of Yellowstone Caldera: Geochronology, geochemistry and petrographic relationships of extracaldera basalts and rhyolites in the Norris‐ Mammoth Corridor. AGU Fall Meeting Abstracts. Stelten, M. E., Champion, D. E., & Kuntz, M. A. (2018). The timing and origin of pre and post‐caldera volcanism associated with the Mesa Falls Tuff, Yellowstone Plateau volcanic ﬁeld. Journal of Volcanology and Geothermal Research, 350,47–60. https://doi.org/10.1016/j. jvolgeores.2017.12.002 Stelten, M. E., Cooper, K. M., Vazquez, J. A., Calvert, A. T., & Glessner, J. J. G. (2015). Mechanisms and timescales of generating eruptible rhyolitic magmas at Yellowstone caldera from zircon and sanidine geochronology and geochemistry. Journal of Petrology, 56(8), 1607–1642. https://doi.org/10.1093/petrology/egv047 Stelten, M. E., Cooper, K. M., Vazquez, J. A., Reid, M. R., Barfod, G. H., Wimpenny, J., & Yin, Q.‐Z. (2013). Magma mixing and the generation of isotopically juvenile silicic magma at Yellowstone caldera inferred from coupling 238U–230Th ages with trace elements and Hf and O isotopes in zircon and Pb isotopes in sanidine. Contributions to Mineralogy and Petrology, 166(2), 587–613. https://doi.org/ 10.1007/s00410‐013‐0893‐2 Stelten, M. E., Cooper, K. M., Wimpenny, J. B., Vazquez, J. A., & Yin, Q.‐Z. (2017). The role of mantle‐derived magmas in the isotopic evolution of Yellowstone's magmatic system. Geochemistry, Geophysics, Geosystems, 18, 1350–1365. https://doi.org/10.1002/ 2016GC006664 Storm, S., Shane, P., Schmitt, A. K., & Lindsay, J. M. (2011). Contrasting punctuated zircon growth in two syn‐erupted rhyolite magmas from Tarawera volcano: Insights to crystal diversity in magmatic systems. Earth and Planetary Science Letters, 301(3‐4), 511–520. https://doi.org/10.1016/j.epsl.2010.11.034 Swallow, E. J., Wilson, C. J., Myers, M. L., Wallace, P. J., Collins, K. S., & Smith, E. G. (2018). Evacuation of multiple magma bodies and the onset of caldera collapse in a supereruption, captured in glass and mineral compositions. Contributions to Mineralogy and Petrology, 173(4), 33. https://doi.org/10.1007/s00410‐018‐1459‐0 Tierney, C. R., Reid, M. R., Vazquez, J. A., & Chesner, C. A. (2019). Diverse late‐stage crystallization and storage conditions in melt domains from the Youngest Toba Tuff revealed by age and compositional heterogeneity in the last increment of accessory phase growth. Contributions to Mineralogy and Petrology, 174(4), 31. Tierney, C. R., Schmitt, A. K., Lovera, O. M., & de Silva, S. L. (2016). Voluminous plutonism during volcanic quiescence revealed by thermochemical modeling of zircon. Geology, 44(8), 683–686. https://doi.org/10.1130/G37968.1 Till, C. B., Vazquez, J. A., & Boyce, J. W. (2015). Months between rejuvenation and volcanic eruption at Yellowstone caldera, Wyoming. Geology, 43(8), 695–698. https://doi.org/10.1130/G36862 Troch, J., Ellis, B. S., Harris, C., Ulmer, P., & Bachmann, O. (2018). The effect of prior hydrothermal alteration on the melting behavior during rhyolite formation in Yellowstone, and its importance in the generation of low–δ18O magmas. Earth and Planetary Science Letters, 481, 338–349. https://doi.org/10.1016/j.epsl.2017.10.039 Troch, J., Ellis, B. S., Mark, D. F., Bindeman, I. N., Kent, A. J., Guillong, M., & Bachmann, O. (2017). Rhyolite generation prior to a Yellowstone supereruption: Insights from the Island Park–Mount Jackson rhyolite series. Journal of Petrology, 58(1), 29–52. Troch, J., Ellis, B. S., Schmitt, A. K., Bouvier, A. S., & Bachmann, O. (2018). The dark side of zircon: Textural, age, oxygen isotopic and trace element evidence of ﬂuid saturation in the subvolcanic reservoir of the Island Park‐Mount Jackson Rhyolite, Yellowstone (USA). Contributions to Mineralogy and Petrology, 173(7), 54. https://doi.org/10.1007/s00410‐018‐1481‐2 Tuttle, O. F., & Bowen, N. L. (1958). Origin of granite in the light of experimental studies in the system NaAlSiO8-KALSi3O8-SiO2-H2O. Geological Society of America Memoir, 74. https://doi.org/10.1130/MEM74 Vazquez, J. A., Kyriazis, S. F., Reid, M. R., Sehler, R. C., & Ramos, F. C. (2009). Thermochemical evolution of young rhyolites at Yellowstone: Evidence for a cooling but periodically replenished postcaldera magma reservoir. Journal of Volcanology and Geothermal Research, 188(1‐3), 186–196. https://doi.org/10.1016/j.jvolgeores.2008.11.030 Vazquez, J. A., & Lidzbarski, M. I. (2012). High‐resolution tephrochronology of the Wilson Creek Formation (Mono Lake, California) and Laschamp event using 238U‐230Th SIMS dating of accessory mineral rims. Earth and Planetary Science Letters, 357,54–67. Vazquez, J. A., & Reid, M. R. (2002). Time scales of magma storage and differentiation of voluminous high‐silica rhyolites at Yellowstone caldera, Wyoming. Contributions to Mineralogy and Petrology, 144(3), 274–285. https://doi.org/10.1007/s00410‐002‐0400‐7 Vazquez, J. A., Velasco, N. O., Schmitt, A. K., Bleick, H. A., & Stelten, M. E. (2014). 238U–230Th dating of chevkinite in high-silica rhyolites from La Primavera and Yellowstone calderas. Chemical Geology, 390, 109–118. Watson, E. B., & Harrison, T. M. (1983). Zircon saturation revisited: Temperature and composition effects in a variety of crustal magma types. Earth and Planetary Science Letters, 64(2), 295–304. https://doi.org/10.1016/0012‐821X(83)90211‐X Watts, K. E., Bindeman, I. N., & Schmitt, A. K. (2012). Crystal scale anatomy of a dying supervolcano: An isotope and geochronology study of individual phenocrysts from voluminous rhyolites of the Yellowstone caldera. Contributions to Mineralogy and Petrology, 164(1), 45–67. https://doi.org/10.1007/s00410‐012‐0724‐x Wendt, I., & Carl, C. (1985). U/Pb dating of discordant 0.1 Ma old secondary U minerals. Earth and Planetary Science Letters, 73(2‐4), 278–284. https://doi.org/10.1016/0012‐821X(85)90076‐7 Wiedenbeck, M., Hanchar, J. M., Peck, W. H., Sylvester, P., Valley, J., Whitehouse, M., et al. (2004). Further characterization of the 91500 zircon crystal. Geostandards and Geoanalytical Research, 28(1), 9–39. https://doi.org/10.1111/j.1751‐908X.2004.tb01041.x Wilke, S., Holtz, F., Neave, D. A., & Almeev, R. (2017). The effect of anorthite content and water on quartz–feldspar cotectic compositions in the rhyolitic system and implications for geobarometry. Journal of Petrology, 58(4), 789–818. https://doi.org/10.1093/petrology/egx034

TILL ET AL. 20 Geochemistry, Geophysics, Geosystems 10.1029/2019GC008321

Wilson, C. J., Stelten, M. E., & Lowenstern, J. B. (2018). Contrasting perspectives on the Lava Creek Tuff eruption, Yellowstone, from new U–Pb and 40Ar/39Ar age determinations. Bulletin of Volcanology, 80(6), 53. https://doi.org/10.1007/s00445‐018‐1229‐x Wotzlaw, J.‐F., Bindeman, I. N., Stern, R. A., D'Abzac, F.‐X., & Schaltegger, U. (2015). Rapid heterogeneous assembly of multiple magma reservoirs prior to Yellowstone supereruptions. Scientiﬁc Reports, 5(1), 1–10. https://doi.org/10.1038/srep14026 Wotzlaw, J.‐F., Bindeman, I. N., Watts, K. E., Schmitt, A. K., Caricchi, L., & Schaltegger, U. (2014). Linking rapid magma reservoir assembly and eruption trigger mechanisms at evolved Yellowstone‐type supervolcanoes. Geology, 42(9), 807–810. https://doi.org/10.1130/ G35979.1

TILL ET AL. 21