Journal of Petrology Advance Access published October 29, 2016 · 2019. 10. 25. · ice blocks and the associated ash-flow tuff, including zir-con morphology, trace element chemistry,

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

Tiffany A. Rivera1*, Mark D. Schmitz1, Brian R. Jicha2 and

James L. Crowley1

1Isotope Geology Lab, Department of Geosciences, Boise State University, 1910 University Drive, Boise, ID 83725,

USA; 2Department of Geoscience, University of Wisconsin–Madison, 1215 W. Dayton Street, Madison, WI 53706,

USA

*Corresponding author. Present address: Westminster College, 1840 South 1300 East, Salt Lake City, UT

84105, USA. E-mail: [email protected]

Received July 6, 2015; Accepted August 12, 2016

ABSTRACT

The 1�3 Ma Mesa Falls Tuff (MFT), the second and volumetrically smallest of the Yellowstonecaldera-forming eruptions, was examined using a joint zircon petrochronological and sanidine40Ar/39Ar approach to constrain the thermal and chemical evolution, autocrystic growth, antecrystic

recycling, and eruptive age of the host magma. A total of 451 laser ablation inductively coupled

plasma mass spectrometry in situ spot analyses collected from 323 zircon crystals from five pum-

ice blocks and two welded ash-flow tuff samples provide trace element and Ti-in-zircon thermom-etry data, which are in turn complemented by high-precision 206Pb/238U dates from over 50 of those

grains. Sanidine grains from two of the pumices were analyzed by incremental step-heating or total

fusion 40Ar/39Ar dating techniques performed on single crystals using a multi-collector mass spec-

trometer, yielding an eruption age of 1�300 6 0�001 Ma. Zircon dates range from 1�57 to 1�30 Ma.Rare grains older than 1�37 Ma may contain inherited cores recycled from the Huckleberry RidgeTuff (HRT) or other associated smaller volume, effusive Yellowstone magmas; however, the bulk ofthe Mesa Falls Tuff crystal load cannot be attributed to a long-lived, residual Huckleberry Ridge

Tuff magma body. Zircon compositions define trends of strengthening negative europium anomaly

and increasing incompatible trace element concentrations over �150 �C of cooling. Crystals defin-ing this full compositional spectrum range in age from 1�33 to 1�30 Ma; the dominant mode of 19grains yields a mean crystallization age of 1�303 6 0�002 Ma, within uncertainty of the sanidine40Ar/39Ar age, attesting to the rapidity of magma accumulation, differentiation and crystallization

prior to eruption. A subset of composite grains composed of extremely differentiated core compos-itions overgrown by Mesa Fall Tuff-like rims probably represents earlier solidified sidewall or roof

accumulations later remobilized within the main Mesa Falls Tuff magma. Fractional crystallization

modeling utilizing temperature-dependent zircon–melt partition coefficients is successful in repro-

ducing the trends in incompatible trace element enrichment within zircon grains as a function of

decreasing temperature and increasing europium anomaly. Zircon geochemistry thus provides a

robust proxy for magma evolution, from the time of zircon saturation through differentiation anderuption. Integrated sanidine and zircon dates coupled with the thermochemical trends indicate

that the Mesa Falls Tuff magmatic system differentiated over a period of< 30 kyr, with the bulk of

zircon crystal nucleation and growth occurring within 10 kyr of eruption. These petrochronological

VC The Author 2016. Published by Oxford University Press. All rights reserved. For Permissions, please e-mail: [email protected] 1

J O U R N A L O F

P E T R O L O G Y

Journal of Petrology, 2016, 1–27

doi: 10.1093/petrology/egw053

Original Article

Journal of Petrology Advance Access published October 29, 2016

by guest on Novem

ber 13, 2016http://petrology.oxfordjournals.org/

Dow

nloaded from

http://www.oxfordjournals.org/http://petrology.oxfordjournals.org/

studies of the MFT and HRT clearly illustrate that the long-term volumetric extrusive rate at

Yellowstone (�2 � 10–3 km3 a–1) is punctuated by episodes of much higher magmatic flux (�2 �10–2 to �2 � 10–1 km3 a–1).

Key words: geochemistry; magma chamber; U–Pb dating; rhyolite; radiogenic isotopes; igneouspetrology; differentiation; Ar–Ar dating; Yellowstone

INTRODUCTION

Zircon crystals have been used as proxies for their host

magma composition (Murali et al., 1983; Heaman et al.,

1990; Bea, 1996; Belousova et al., 2002; Hoskin &

Schaltegger, 2003), and as records of the evolution and

differentiation of silicic magma systems (Reid et al.,

2011; Stelten et al., 2013; Wotzlaw et al., 2014). The sat-

uration of Zr and associated crystallization of zircon

over a range of petrologically relevant pressures and

temperatures in most intermediate to felsic silicate

magmas, the incorporation of many normally incom-

patible trace elements into zircon, and the sluggish kin-

etics of diffusive exchange in zircon at magmatic

conditions all make this accessory mineral particularly

suited to the preservation of magmatic history through

intra-crystalline zonation. Prior studies have shown that

an integration of (1) proxies for intra-crystalline zon-

ation (e.g. cathodoluminescence imaging), (2) trace

element concentration and titanium-in-zircon thermom-

etry by in situ laser ablation inductively coupled plasma

mass spectrometry (LA-ICP-MS) or secondary ion mass

spectrometry (SIMS), and (3) high-precision 206Pb/238U

dating by chemical abrasion isotope dilution thermal

ionization mass spectrometry (CA-ID-TIMS)—collect-

ively termed petrochronology—can aid in the recon-

struction of growth histories within populations of

zircon crystals, quantification of the chemical evolution

proceeding from cooling of the host magmas, the tim-

ing and tempo of that chemical evolution, and the iden-

tification of populations of crystals arising from discrete

pulses of magmatism (Miller & Wooden, 2004; Rivera

et al., 2013, 2014; Chelle-Michou et al., 2014; Klemetti &

Clynne, 2014; Wotzlaw et al., 2014, 2015; Chamberlain

et al., 2015; Stelten et al., 2015). These intra-crystalline

records thus provide a unique window into the com-

plexity and longevity of a magma system from the time

of zirconium saturation to eruption.

Large-volume rhyolitic magma bodies that produce

caldera-forming eruptions have been the focus of con-

siderable interest and debate, most of which centers on

the mechanisms and timescales of magma assembly

and storage. Models for the generation of large-volume

rhyolitic magmas vary from fractionation of mafic melts

(e.g. Halliday et al., 1989; Hildreth et al., 1991) to whole-

sale crustal melting and recycling (e.g. Ewart & Stipp,

1968), or some combination of these two processes

(e.g. McCulloch et al., 1994; Charlier et al., 2005;

Pritchard & Larson, 2012; Szymanowski et al., 2015). In

the Yellowstone Volcanic Field the physical nature of

the source from which the magmas are derived remains

unresolved. Isotopic modeling by Hildreth et al. (1991)

suggested ‘hybridization’ of partial melts of Archaean

crust and mafic, mantle-derived magmas, leading to the

production of the large-volume Yellowstone silicic mag-

mas. Oxygen isotope data for caldera-forming and

inter-caldera rhyolite flows reveal isotopically zoned zir-

con crystals (from core to rim) that have been inter-

preted to reflect remelting and cannibalization of

previously crystallized and altered, shallow crustal ma-

terial (Hildreth et al., 1984; Bindeman & Valley, 2001;

Bindeman et al., 2001, 2007; Ellis et al., 2010; Watts

et al., 2011, 2012; Drew et al., 2013; Bindeman &

Simakin, 2014; Wotzlaw et al., 2014, 2015). However,

many researchers have presented crystal-scale evi-

dence of a ‘crystal mush’ or near-solidus subvolcanic

crystal and melt mixture that may be present in a var-

iety of tectonic settings, including other caldera-forming

eruptions along the Yellowstone hotspot track (e.g.

Bachmann et al., 2002; Bachmann & Bergantz, 2004;

Hildreth & Wilson, 2007; Ellis & Wolff, 2012; Eppich

et al., 2012; Bragagni et al., 2014; Cooper & Kent, 2014;

Ellis et al., 2014; Klemetti & Clynne, 2014; Stelten et al.,

2015); this is supported by tomographic models of the

Yellowstone magma chamber (Farrell et al., 2014). In

the ‘crystal mush’ model, small proportions of melt seg-

regate from the crystallizing phases, producing a

crystal-poor rhyolite melt lens, which is subsequently

evacuated by eruption. A key distinction between

melting-dominated and mush-dominated models for

crystal-poor silicic magma genesis is the time scale of

melt generation, the former (crustal melting and caldera

cannibalization) being potentially rapid, dependent

upon mantle heat flux, whereas the latter requires tens

to hundreds of thousands of years for the segregation

of super-eruption sized volumes of rhyolite. Recent

work on the timescales of storage for these segregated

melts prior to eruption within the Yellowstone system,

based on crystal-scale parameters and modeling, gives

a range from months (Till et al., 2015) to tens of thou-

sands of years (Rivera et al., 2014; Wotzlaw et al., 2014;

Matthews et al., 2015; Stelten et al., 2015).

The Yellowstone Volcanic Field is a natural labora-

tory in which to test and model in greater detail the fi-

delity of zircon-derived records of magma evolution, as

well as address basic questions regarding the tempo of

storage and differentiation of silicic magmas. The

Yellowstone caldera system at the northern terminus of

the Snake River Plain hotspot track hosts a number of

candidates for investigation (Fig. 1). The earliest

2 Journal of Petrology, 2016, Vol. 0, No. 0

by guest on Novem


Dow

nloaded from

http://petrology.oxfordjournals.org/

(2�08 Ma; Rivera et al., 2014; Singer et al., 2014) andlargest eruption (�2500 km3; Christiansen, 2001) pro-duced the Island Park Caldera and Huckleberry Ridge

Tuff (HRT). The most recent eruptive phase produced

the Yellowstone Caldera and Lava Creek Tuff (LCT;

0�63 Ma; Matthews et al., 2015), with an eruptive volumeof �1000 km3 (Christiansen, 2001). The smallest of thethree caldera-forming eruptions, with an initial em-

placement volume of �280 km3 (Christiansen, 2001),was the Mesa Falls Tuff (MFT), which generated the

Henry’s Fork Caldera.

Unlike the complex eruptive histories of the HRT and

LCT, which produced multiple cooling units, the MFT

was erupted in a single pulse (Christiansen, 2001). Its

smaller size and simpler eruptive history led us to inves-

tigate the MFT to decipher pre-eruptive processes and

the duration of magmatic evolution prior to eruption. In

this contribution MFT zircon crystals from single pum-

ice blocks and the associated ash-flow tuff, including zir-

con morphology, trace element chemistry, Ti-in-zircon

thermometry, and 206Pb/238U dates are examined. We

explore the geochemical variability within and between

zircon crystals using models of fractional crystallization

calibrated by Ti-in-zircon thermometry, including

temperature-dependent zircon partition coefficients. We

compare the zircon populations between discrete pum-

ice blocks and comminuted ash-flow tuff samples to es-

tablish the degree of homogenization of the MFT

magma at the time of eruption. The relative importance

of intra-crystal variability for CA-ID-TIMS U–Pb ages is

assessed using a quantitative volumetric mixing model

calibrated by in situ geochemical measurements. These

zircon data are then compared with new high-precision40Ar/39Ar sanidine dates for the MFT, obtained using a

new generation, multi-collector noble gas mass spec-

trometer, allowing for a re-examination of the eruption

age, zircon pre-eruption residence time, and tempo of

associated magmatic processes.

GEOLOGICAL BACKGROUND AND PREVIOUSGEOCHRONOLOGY

Volcanism within the Yellowstone Volcanic Field and

associated Eastern Snake River Plain has been studied

extensively, with volcanic activity largely credited to a

stationary hotspot (e.g. Armstrong et al., 1975; Pierce &

Morgan, 1992; Smith & Braile, 1994; Camp, 1995) gener-

ating a large crustal magma reservoir (Farrell et al.,

2014), although other geological phenomena, such as

return flow around a segmented portion of a subducted

slab (James et al., 2011; Fouch, 2012), have been pro-

posed. Caldera-forming eruptions began at �17 Maalong the Nevada–Oregon–Idaho border. As the North

American tectonic plate has drifted over this localized

area of mantle upwelling, the calderas have become

progressively younger to the east, with the youngest

rhyolitic volcanism occurring at c. 70 ka within the

present-day Yellowstone National Park.

Yellowstone National Park boundaryMONTANAWYOMING

MONTANA

IDAHO

KILOMETERS0 20

W112˚00’ W111˚00’ W109˚45’

N45˚00’

N44˚06’

Island Park Caldera

Yellowstone Caldera

Henry’s ForkCaldera

ID

MT

WY

UTCO

AZ NM

B A

SAMPLINGLOCATIONS

Fig. 1. (a) Location map of the calderas of the Yellowstone Plateau and Yellowstone National Park. Mesa Falls Tuff sample locationsare indicated by the stars. Map modified from Christiansen (2001) and Matthews et al. (2015). (b) Location of Yellowstone NationalPark (YNP) within the interior west of the USA. Modified from Matthews et al. (2015).

Journal of Petrology, 2016, Vol. 0, No. 0 3

by guest on Novem


Dow

nloaded from


Caldera-forming eruptions within the Yellowstone

Volcanic Field occurred in three cycles. Prior to and fol-

lowing each caldera-forming eruption, smaller volume

rhyolite flows erupted along the caldera margins and

within the newly formed caldera. Subsequent basaltic

lavas erupted along the margins and as caldera-filling

flows; basalts associated with the youngest caldera-

forming eruption have not yet erupted (Christiansen

et al., 2007). Figure 2 provides a simplified stratigraphy

of the rhyolitic volcanic cycles of the Yellowstone

Volcanic Field, drawn from the extensive review of the

geological history of the three volcanic cycles by

Christiansen (2001). We summarize the geological evo-

lution of the first two cycles below.

The first volcanic cycle and the Huckleberry

Ridge TuffThe Huckleberry Ridge Tuff (HRT) was the earliest and

most voluminous eruption within the Yellowstone

Volcanic Field, and generated the Island Park Caldera.

Although the HRT has been interpreted as a single erup-

tive unit, three members (A, B, and C) have been identi-

fied with various bulk volumes (Christiansen & Blank,

1972; Christiansen, 2001). The eruption ages of the

three units have been debated, including the possible

existence of a temporal hiatus between eruption of

members B and C (e.g. Ellis et al., 2012). Wotzlaw et al.

(2015) examined zircon grains from each of the three

members and concluded that all members erupted at

Rhyolite of Snake River Butte2.15 – 2.16 Ma

Huckleberry Ridge Tuff2.08 Ma

Big Bend Ridge Rhyolites~ 2 Ma

Big Bend Ridge Rhyolites~ 1.3 Ma

Mesa Falls Tuff1.30 Ma

Island Park Rhyolites~ 1.3 Ma

Mount Jackson & Lewis Canyon Rhyolites1.22 to 0.61 Ma

Lava Creek Tuff0.63 Ma

Plateau Rhyolite0.52 to 0.07 Ma

Age VolcanicCycle

Pre-calderaRhyolite

Caldera-formingEruption

Post-calderaRhyolite

Ple

isto

cene

Plio

cene

First

Third

Second

Fig. 2. Simplified volcanic stratigraphy for the rhyolitic eruptions of the three volcanic cycles forming the Yellowstone VolcanicField. Modified from Christiansen (2001). Ages compiled from Obradovich (1992), Gansecki et al. (1996), Bindeman et al. (2001),Lanphere et al. (2002), Christiansen (2001), Watts et al. (2012), Rivera et al. (2014), Singer et al. (2014), Matthews et al. (2015), Rivera& Jicha (2015), Stelten et al. (2015) and Wotzlaw et al. (2015).

SiO2

70 71 72 73 74 75 76 77 78 79

Na 2

O +

K2O

7.0

7.5

8.0

8.5

9.0

9.5

10.0

Huckleberry Ridge Tuff (A & B)

Mesa Falls TuffHuckleberry Ridge Tuff (C)

Snake River Butte flowpost-HRT flowspre-MFT flows

** Mesa Falls Tuff** Huckleberry Ridge Tuff

SiO2

70 71 72 73 74 75 76 77 78 79

TiO

2

0.0

0.1

0.2

0.3

0.4

0.5

206Pb/204Pb

16.5 16.9 17.3 17.7 18.1 18.5

207 P

b/20

4 Pb

15.4

15.5

15.6

15.7

15.8(a) (b) (c)

Fig. 3. (a) Total alkalis and (b) TiO2 vs silica for whole-rock analyses of Huckleberry Ridge Tuff (HRT; members A, B, and C), MesaFalls Tuff (MFT), Snake River Butte Flow, and the Big Bend Ridge Flows separated into post-HRT and pre-MFT units. (c) Whole-rockPb isotopic compositions for the same units as in (a) and (b). **Pb isotopic ratios for feldspars extracted from the MFT and HRT(member B) obtained in this study and by Rivera et al. (2014) are shown for comparison. Whole-rock data compiled from Boyd(1961), Hamilton & Leopold (1962), Hamilton (1965), Witkind (1969), Doe et al. (1982), Hildreth et al. (1984, 1991), Bindeman &Valley (2001) and Christiansen (2001). Major element analyses normalized to 100%.


by guest on Novem


Dow

nloaded from


2�08 Ma, which is consistent with the 40Ar/39Ar sanidinedates for member B of Rivera et al. (2014) and Singer

et al. (2014). Further, Rivera et al. (2014) completed a zir-

con petrochronological investigation by acquiring cath-

odoluminescence images, trace element signatures,

titanium-in-zircon thermometry, and high-precision U–

Pb ages for this member, leading to an interpretation of

rapid magma accumulation and differentiation within a

few thousand years prior to eruption.

All three members of the HRT have phenocryst as-

semblages of quartz, sanidine, and plagioclase, with

minor iron–titanium oxides, clinopyroxene, and acces-

sory phases; variations in abundance and phenocryst

size are present between the three members

(Christiansen, 2001). Geochemically, members A and B

are similar in their major and minor element abun-

dances, whereas member C is offset from these two

members (Fig. 3). The Sr, Nd, and Pb isotopic compos-

itions of member C also differ from those of members A

and B, attributed to a greater contribution of crustal ma-

terial to the magma that produced member C (Hildreth

et al., 1991).

One rhyolite dome, Snake River Butte (SRB), has been

identified as a precursor to the eruption of the HRT. This

flow can be seen along the eastern margin of Big Bend

Ridge, a topographic high that has been interpreted as a

caldera segment produced from the eruption of the HRT

(Christiansen, 2001). Recent petrochronology of the SRB

rhyolite has suggested that the magma producing this

effusive eruption was active some 70–80 kyr prior to the

caldera-forming eruption, and was compositionally dis-

tinct from the magma generating the HRT (Rivera et al.,

2014; Wotzlaw et al., 2015).

The second volcanic cycle and the Mesa Falls TuffThe MFT has been interpreted as a single eruptive unit,

sourced from the Henry’s Fork caldera in the Island

Park, Idaho region (Fig. 1). Previous 40Ar/39Ar single-

crystal sanidine analyses yielded an eruption age of

1�321 6 0�012 Ma (Gansecki et al., 1998), whereas multi-crystal sanidine incremental heating experiments

yielded an age of 1�312 6 0�014 Ma (Lanphere et al.,2002). Bindeman et al. (2008) reported 206Pb/238U SIMS

ages for two MFT zircon crystals—one at 1�45 6 0�03 Mafor a core area and the other consisting of a core and

rim analysis, both yielding an age of 1�49 6 0�05 Ma.Wotzlaw et al. (2015) published a 206Pb/238U CA-ID-TIMS

zircon crystallization age of 1�3004 6 0�0073 Ma for 13grains from the MFT.

Most proximal exposures of the MFT are 30–70 m

thick and have a distinctive pinkish color. The base of

the MFT is characterized by a thick ash and pumice de-

posit; the lower part hosts pumice blocks commonly up

to 30 cm in diameter, although larger pumices can be

found, whereas the pumice size decreases (to �3 cm)further upward in the unit. Quartz and feldspar (sanidine

and plagioclase) phenocrysts up to 1 cm in diameter are

conspicuous within the pumice and host matrix, along

with less abundant pyroxene, iron–titanium oxides, and

accessory phases including zircon and the rare earth

element (REE)-bearing phase chevkinite (Christiansen,

2001). The upper part of the unit consists of a more

densely welded ash-flow sheet with the same pheno-

cryst assemblage.

Similar to the first volcanic cycle that produced the

HRT, the MFT is underlain by several pre-caldera rhyo-

lites. Several rhyolite flows and one tuff form the Big

Bend Ridge Rhyolite sequence that erupted between

the deposition of the HRT and the MFT. Two older flows

(Blue Creek and Headquarters) yield K–Ar ages of c. 1�8Ma, whereas the younger flows (Bishop Mountain Flow,

Tuff of Lyle Spring, Moonshine Mountain Flow, and

Green Canyon Flow) yield K–Ar ages of 1�10 6 0�02 to1�32 6 0�02 Ma (Obradovich, 1992; Christiansen, 2001).

MATERIALS AND METHODS

Sample locations and descriptionsSeven samples were collected from road cuts along

Highway 20 north of Ashton, Idaho. Samples 13MFT-1

to 13MFT-5 were collected at a single locality

(44�10�050’N, 111�25�416’W); each sample represents asingle pumice block, ranging in size from 20 to 30 cm.

Sample 13MFT-6 was collected at the same location,

but consists of the bulk ash-flow tuff with embedded

pumice fragments of 5–6 cm length. Sample 13MFT-7

was collected to the north at a second road cut

(44�07�311’N, 111�26�497’W) and represents the moredensely welded part of the unit. Mineral separation of

zircon and sanidine followed standard magnetic and

density separation techniques, followed by handpicking

at Boise State University (BSU).

U–Pb zircon trace element geochemistry anddating by LA-ICP-MSZircon crystal preparation followed the general proced-

ures of Rivera et al. (2013, 2014) as summarized here.

Cathodoluminescence (CL) imaging was performed

with a Gatan MiniCL detector coupled to a JEOL JSM-

T300 scanning electron microscope. Following internal

morphological characterization, areas were selected for

laser ablation-inductively coupled plasma-mass spec-

trometry (LA-ICP-MS) analysis. These analyses were

conducted in situ using a 213 nm frequency-quintupled

Nd-YAG NewWave laser coupled to a ThermoElectron

X-Series II ICP-MS system, with a 10 Hz at 5 J cm–2

pulsed laser and 30 lm spot size. NIST SRM-610 andSRM-612 glasses served as primary standards for trace

element concentrations and the Ple�sovice zircon was

used for U–Pb calibration. Zircon standards were meas-

ured after every 10 unknowns, whereas glass standards

were measured at the beginning and end of a 150-spot

analysis cycle. As measured 204Pb signals were indistin-

guishable from background, reported 206Pb/238U dates

are not corrected for common Pb. U–Th–Pb isotope

ratios and error propagation for each analysis include


by guest on Novem


Dow

nloaded from


uncertainty contributions from counting statistics, back-

ground subtraction, and standard calibration.

Zircon saturation and Ti-in-zircon thermometryZircon saturation in silicate melts is a systematic func-

tion of temperature and composition, and for felsic, per-

aluminous systems, such as the HRT and MFT, zircon

saturation temperatures may range from 750 to 1020�C,

dependent upon the compositional parameter M [(Na þKþ 2Ca)/(Al � Si) on a molar basis] (Watson & Harrison,1983; Boehnke et al., 2013). Measured whole-rock zirco-

nium contents can be used to calculate minimum zircon

saturation temperatures in systems that contain rela-

tively few inherited grains, or maximum temperatures

for those with significant xenocrystic or antecrystic

populations (Miller et al., 2003). Zircon saturation tem-

peratures for the MFT were calculated using literature

whole-rock data (Hildreth et al., 1984, 1991; Bindeman

& Valley, 2001; Christiansen, 2001) and the equations of

Boehnke et al. (2013). The calculated zircon saturation

temperatures range from 768 to 836�C, with a mean of

799�C. Zircon saturation temperatures for the HRT were

previously reported by Rivera et al. (2014) and range

from 782 to 900�C using the same calibration.

Titanium (Ti) incorporated into the tetrahedral site of

the zircon crystal lattice can serve as a proxy for the

crystallization temperature of that zircon within the host

magma (Watson & Harrison, 2005; Watson et al., 2006;

Ferry & Watson, 2007). Titanium concentrations were

measured in MFT zircon crystals by LA-ICP-MS simul-

taneously with the other trace elements as described

above, and Ti-in-zircon temperatures were calculated

following the equations of Ferry & Watson (2007).

Although the thermometer was developed for rutile-

saturated conditions, Watson et al. (2006) noted that

partitioning of Ti into zircon could serve as a thermom-

eter for rutile-free melts with constraints on the activ-

ities of SiO2 (aSiO2) and TiO2 (aTiO2). The MFT meltswere saturated in quartz, therefore we use an activity of

unity for aSiO2. An examination of aTiO2 for other large-volume silicic systems in the western USA led to the

choice of 0�55 for aTiO2 (Hayden & Watson, 2007; Warket al., 2007; Campbell et al., 2009; Hofmann et al., 2014;

Reid et al., 2011). These values are consistent with those

chosen by Rivera et al. (2014) and by Wotzlaw et al.

(2015), allowing us to directly compare zircon crystal-

lization temperatures for multiple caldera-forming erup-

tions within the Yellowstone Volcanic Field.

We calculate a 610�C uncertainty on the Ti-in-zircon

temperatures based upon the 610% (1r) reproducibilityof an Orapa (Botswana) kimberlite zircon standard

(Rivera et al., 2013), which is appropriate for differential

thermometry within a magma system buffered to near-

constant aTiO2 (Hayden & Watson, 2007; Wark et al.,2007). We acknowledge that the absolute temperature

uncertainty of the Ti-in-zircon thermometer is also

Table 1: Pb isotopic compositions for sanidine from Mesa Falls Tuff and Huckleberry Ridge Tuff

Sample Leach 206Pb/204Pb 207Pb/204Pb 208Pb/204Pb 207Pb/206Pb 208Pb/206Pb

13MFT-1 L4 17�2425 15�5469 38�2989 0�9017 2�221213MFT-1 L5 17�2498 15�5571 38�3323 0�9019 2�222213MFT-2 L4 17�2521 15�5593 38�3373 0�9019 2�222213MFT-2 L5 17�2516 15�5577 38�3320 0�9018 2�222013MFT-3 L4 17�2521 15�5587 38�3366 0�9018 2�222113MFT-3 L5 17�2502 15�5561 38�3276 0�9018 2�221913MFT-4 L4 17�2506 15�5570 38�3306 0�9018 2�222013MFT-4 L5 17�2536 15�5611 38�3450 0�9019 2�222513MFT-5 L4 17�2461 15�5515 38�3118 0�9017 2�221513MFT-5 L5 17�2376 15�5401 38�2743 0�9015 2�220413MFT-6 L4-1 17�2369 15�5399 38�2700 0�9015 2�220313MFT-6 L4-2 17�2328 15�5356 38�2592 0�9015 2�220113MFT-6 L4-3 17�2347 15�5385 38�2682 0�9016 2�220413MFT-6 L5-1 17�2344 15�5370 38�2633 0�9015 2�220213MFT-6 L5-2 17�2363 15�5403 38�2743 0�9016 2�220613MFT-6 L5-3 17�2433 15�5498 38�3061 0�9018 2�221513MFT-7 L5 17�2447 15�5524 38�3149 0�9019 2�221813MFT-7 L4 17�2433 15�5504 38�3087 0�9018 2�2217Average 17�2440 15�5494 38�3051 0�9017 2�2214Standard deviation 0�0071 0�0087 0�0295 0�0001 0�0008HRT-1* L4 16�9796 15�4771 37�9850 0�9115 2�2371HRT-1* L5 16�9817 15�4807 37�9981 0�9116 2�2376HRT-1* L6 16�9813 15�4779 37�9877 0�9115 2�2370HRT-1 L7 17�0082 15�5083 38�0918 0�9118 2�2396HRT-1 L8 17�0190 15�5226 38�1391 0�9121 2�241013HRTB-1 L4 17�0074 15�5135 38�1154 0�9120 2�240413HRTB-1 L5 17�0082 15�5083 38�0918 0�9118 2�2396Average 16�9979 15�4983 38�0584 0�9118 2�2389Standard deviation 0�0164 0�0191 0�0658 0�0002 0�0016

All analyses measured as one run of 600 ratios except for sample 13MFT-5, in which leaches 4 and 5 were analyzed in three install-ments of 200 ratios.*Data from Rivera et al. (2014).


by guest on Novem


Dow

nloaded from


affected by the assumed value of aTiO2. Our chosenvalue for aTiO2 of 0�55 is a likely minimum value forhigh-silica rhyolites, and thus translates into maximum

temperatures. For example, increasing aTiO2 to 0�7would result in a systematic decrease in calculated tem-

peratures of 20–30 �C at the low and high ends, respect-

ively, of the Ti concentration range recorded in MFT

zircons, This systematic offset is less important for the

differential thermometry emphasized in this

contribution.

U–Pb zircon dates obtained by CA-ID-TIMSSelected zircon grains were prepared for chemical abra-

sion thermal ionization mass spectrometry (CA-ID-TIMS)

according to the procedures reported by Davydov et al.

(2010) and Rivera et al. (2013). Single grains were an-

nealed and chemically abraded, spiked with the

EARTHTIME 205Pb–233U–235U isotope dilution tracer solu-

tion (ET535; Condon et al., 2015; McLean et al., 2015), dis-

solved, and Pb and U purified by ion chromatography.

Isotope ratios were measured by TIMS at Boise State

University. All dates and uncertainties were calculated

following the algorithms of Schmitz & Schoene (2007)

and U decay constants of Jaffey et al. (1971). 206Pb/238U

dates and ratios were corrected for initial 230Th disequi-

librium using a Th/U[magma] value of 4�64 6 0�30 (2r),based upon whole-rock Th and U measurements on the

MFT (Hildreth et al., 1984, 1991; Christiansen, 2001), and

an estimate of initial 230Th/238U[magma] activity of 1�074 60�0566 (Vazquez & Reid, 2002). A laboratory Pb blank of0�3 pg was estimated from long-term reproducibility ofblank measurements at BSU (Rivera et al., 2014). Excess

common Pb above the estimated blank levels may be

attributed to the abundant inclusions of silicates, oxides

and/or glass within the MFT zircon grains. Isotope ratios

of coexisting sanidine crystals were used to estimate the

initial Pb isotopic composition. Approximately 200 mg of

handpicked sanidine grains were sequentially acid

leached following the method of Housh & Bowring

(1991). Pb was purified from sequential dilute hydro-

fluoric acid leachates using ion exchange chromatog-

raphy in dilute hydrobromic acid, and isotope ratios

were measured via static Faraday cup TIMS. Data are re-

ported in Table 1.

40Ar/39Ar sanidine age determinationsHandpicked sanidine grains were loaded into an alumi-

num sample disk along with Alder Creek Rhyolite sani-

dine (1�1860 6 0�0016 Ma; Rivera et al., 2013) as theneutron fluence monitor and irradiated for 5 h at the

cadmium-lined in-core (CLICIT) facility at the Oregon

State University TRIGA reactor. Argon isotopic analyses

were conducted at the University of Wisconsin–Madison

using a 60 W CO2 laser and a Nu Instruments Noblesse

multi-collector mass spectrometer, equipped with a

Faraday detector and four ETP ion counting electron

multipliers—two at high mass (IC0 and IC1) and two at

low mass (IC2 and IC3) positions. The Alder Creek

Rhyolite standard and 0�5–1�0 mm sanidine grains frompumice 13MFT-3 were fused in one step at 7 W, whereas

sanidine crystals from pumice 13MFT-2 were incremen-

tally heated. All analyses were performed on single sani-

dine crystals. Analyses of unknowns, blanks, and

neutron fluence monitor minerals were carried out in

identical fashion with a routine involving one peak hop.

Detailed description of the measurement protocol, de-

tector efficiency and mass fractionation corrections, and

data reduction have been given by Jicha et al. (2016).All ages are reported with 2r uncertainties. 40Ar/39Ar

dates are reported relative to the Fish Canyon Tuff sani-

dine monitor age of 28�201 Ma (Kuiper et al., 2008) andinclude the uncertainty in the irradiation parameter J,

when available.

RESULTS40Ar/39Ar sanidine geochronologyPlateaux were achieved for each of 11 incremental heat-

ing experiments of 13MFT-2, with nine of the experi-

ments including more than 85% of the 39Ar (Fig. 4).

Plateau ages range from 1�2985 6 0�0034 to 1�3019 60�0025 Ma, with a weighted mean age of 1�3006 60�0008 Ma [0�06%; MSWD (mean squared weighted de-viation) ¼ 0�63, p (probability) ¼ 0�85, n ¼ 11]. Blank-cor-rected raw data (including standard analyses) are

provided in the Supplementary Data (supplementary

data are available for downloading at http://www.pet

rology.oxfordjournals.org).

Single-crystal fusion analyses of 18 sanidine grains

from pumice 13MFT-3 yield a multi-modal distribution

with dates ranging from 1�2988 6 0�0049 to 1�3480 60�0092 Ma (Fig. 5). The youngest 12 grains yield aweighted mean age of 1�2996 6 0�0011 Ma (0�08%,MSWD ¼ 0�23, p ¼ 1�00), older grains form a populationranging from 1�31 to 1�33 Ma (n ¼ 4), and two additionalgrains yielded dates from 1�34 to 1�35 Ma (Table 2;Supplementary Data). The significance of the older

grains will be discussed in a subsequent section.

Combining the analyses from both incremental heating

and fusion experiments (Fig. 6) yields a weighted mean

age of 1�3001 6 0�0006 Ma (0�04%; n ¼ 24; MSWD ¼0�53, p ¼ 0�98). Propagation of the uncertainty on thedecay constant (Min et al., 2000) results in an external

error of 60�0025 (0�20%) Ma.

Zircon morphology and zoning patternsThere is clear qualitative similarity in external morph-

ology and internal zoning patterns between the zircon

populations of the discrete pumice and ash-flow tuff sam-

ples analyzed in this study (Supplementary Data). All

sample populations are characterized by elongate pris-

matic grains (aspect ratios of 1:5 to 1:10), with a variety of

sizes (Supplementary Data Figs 1–7). Grains of longer as-

pect ratio tend to have strongly planar internal oscillatory

zoning in CL, whereas grains with a smaller aspect ratio

exhibit generally more even luminescence with only


by guest on Novem


Dow

nloaded from

http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://www.petrology.oxfordjournals.orghttp://www.petrology.oxfordjournals.orghttp://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/

subtle oscillatory zoning. All grains share a common

outer mantle of relatively luminescent zircon, more

clearly manifested on elongate prismatic grains as pyr-

amidal tips. There is little evidence for internal zoning

truncations or unconformities within the growth bands of

these luminescent crystals. Mineral and glass inclusions(bleb-like to irregular CL-dark domains) are so ubiquitous

in these grains as to induce a weakly paramagnetic

character to most zircons from the MFT. Accompanying

these predominant crystal populations are a set of gener-

ally more equant crystals with distinctive non-

luminescent (CL-black) cores overgrown by lighter rims,

qualitatively similar to the overgrowth mantles of the for-

mer. The overgrowths span a continuum from apparentconcordance with prismatic dark cores to obvious dis-

cordance with resorbed and truncated dark cores.

1400

1350

1300

1250

1200

1150

Age

(ka)

1400

1350

1300

1250

1200

1150

Age

(ka)

1400

1350

1300

1250

1200

1150

Age

(ka)

1400

1350

1300

1250

1200

1150

Age

(ka)

0.10.0 0.80.60.40.2Cumulative 39Ar Fraction

Plateau age = 1298.9 ± 2.9 kaMSWD = 1.5, prob. = 0.18

incl. 100% of the 39Arn/N = 6/6


incl. 88.9% of the 39Arn/N = 7/11


incl. 68.3% of the 39Arn/N = 6/9
















incl. 79.6% of the 39Arn/N = 7/10



(a) (c)(b)

(d) (e) (f)

(i)(h)(g)

(j) (k)

Fig. 4. (a–k) 40Ar/39Ar plateau diagrams for incremental heating experiments on single sanidine crystals from pumice sample13MFT-2. Box heights are 2r; ages presented with 2r uncertainty including a J-error of 0�147% (J ¼ 0�0012259 6 0�0000006).Plateau steps are in orange; rejected steps are white. n/N is the number of steps that define a plateau out of the total steps.


by guest on Novem


Dow

nloaded from


LA-ICP-MS zircon U–Pb geochronology, traceelement geochemistry, and Ti-in-zirconthermometryIn total, 451 spot analyses were conducted on 323 zir-

con grains from the seven samples. CL images, with

their spot locations, LA-ICP-MS trace element data, cal-

culated Ti-in-zircon temperatures and 206Pb/238U dates

are given in the Supplementary Data. Four distinct

chemical domains were identified based on trace elem-

ent (e.g. U, Nb, Hf, Th, Y) correlations versus Th/Y, Eu

anomaly (Eu/Eu*) and Ti-in-zircon temperatures (Fig. 7).

Eu/Eu* was determined as the quotient of chondrite-

normalized measured Eu divided by the square root of

the product of chondrite-normalized measured Sm and

Gd. CI carbonaceous chondrite normalizing values are

from Sun & McDonough (1989). REE diagrams for each

chemical domain are provided in the Supplementary

Data.Chemical domain 1 (CD1, 16% of analyses; Fig. 8)

consists of analyses with smaller Eu anomalies (values

of Eu/Eu* > 0�25), temperatures from 800 to 875�C, lowTh/Y (�0�06), low incompatible trace element concen-trations and a shallow slope of heavy rare earth elem-

ents (HREE). Chemical domain 2 (CD2, 30%) has

moderate values of Eu/Eu* (0�10–0�30), slightly lowertemperatures (725–850�C), low incompatible trace elem-ent concentrations and a slight steepening of the HREE

as compared with CD1 analyses. Chemical domain 3

(CD3, 43%) exhibits larger Eu anomalies (smaller values

of Eu/Eu*), high Th/Y (0�10–0�18), generally lower tem-peratures (725–775�C), increasing incompatible trace

element concentrations and a progressively steepening

slope of HREE. The fourth group, chemical domain 4

(CD4, 11%), differs from the previous three groups in

exhibiting extreme incompatible trace element concen-trations (e.g. U up to 20̂000 ppm; Nb up to 700 ppm), Eu/

Eu* less than 0�05 with many analyses approachingzero, Th/Y extending from 0�2 to 1�2 and the steepestslope of HREE. Despite these differences, Ti-in-zircon

thermometry yields temperatures of c. 700–775�C, simi-

lar to CD3.

Chemical domains can be linked to morphological

features revealed by CL imaging (Supplementary Data).CL-black cores consistently have CD4 compositions, but

CD4 compositions also occur as CL-gray rims over-

growing the CL-black cores, or rarely as a CL-gray inter-

ior overgrown by either a CD3 or CD4 rim. Most

importantly, the rims of these CD4 areas are character-

ized by CD3 composition. CD1 and CD2 compositions

are found primarily within the elongate prismatic

grains, along edges or interiors, and on CL-bright to CL-gray grains. These compositions never occur with CL-

black cores; their overgrowths are consistently of the

CD3 composition. Further, all three of these compos-

itions can be found within a single grain, where CD1

characterizes the interior CL-brightest region, CD2

occurs as the interior CL-gray region, and CD3 as the

rim overgrowth. Finally, no grains contain a CD3 or CD4

interior with a CD1 or CD2 overgrowth. In summary, the

majority of zircon grains contain interiors of hotter,more primitive CD1 or CD2 compositions, mantled with

cooler, more evolved CD3 compositions, with a minor-

ity of grains containing highly evolved CD4 compos-

itions correlating to CL-black interiors, overgrown by

either CD3 or CD4 compositions.

The low uranium contents of the MFT zircon grains

(92% of analyses contain less than 250 ppm U) result in

low radiogenic Pb yields and imprecise LA-ICP-MS206Pb/238U dates, whose accuracy is relatively sensitive

to small fluctuations in background counts in the Pb

mass spectrum. Nonetheless, applying a simple 207Pb-

Rel

ativ

e pr

obab

ility

43.192.1 1.331.321.311.30

Age (Ma)1.381.371.361.35

Total fusion age = 1299.6 ± 1.1 kaMSWD = 0.23, prob. = 1.00 n = 12/18

Fig. 5. Probability distribution function (continuous black line)of single-crystal total fusion analyses of sanidine extractedfrom pumice sample 13MFT-3. Analyses used in the weightedmean calculation are indicated by circles. Analyses excludedfrom the weighted mean calculation are indicated by squares.Bars on single data points represent 2r uncertainty.

Table 2: 40Ar/39Ar total fusion experiment dating summary forpumice sample 13MFT-3

Age1 (Ma) 2r %40Ar* K/Ca

1�2988 0�0049 93�97 26�31�2989 0�0029 90�98 30�81�2990 0�0022 91�09 19�71�2992 0�0024 95�05 24�91�2994 0�0026 91�79 25�31�2994 0�0026 89�85 24�51�2994 0�0030 89�01 25�91�2996 0�0023 93�77 21�51�3004 0�0019 95�75 19�51�3006 0�0053 90�59 28�51�3007 0�0026 89�29 26�11�3009 0�0047 95�40 25�21�3097 0�0020 88�14 27�81�3136 0�0031 92�73 27�41�3180 0�0046 80�69 27�81�3196 0�0041 71�45 28�11�3243 0�0043 93�91 22�01�3411 0�0029 90�89 26�41�3480 0�0092 48�30 27�61Ages calculated relative to Alder Creek Rhyolite sanidine at 1�186 Ma (Rivera et al., 2014); uncertainties include the error onJ (J ¼ 0�0012072 6 0�0000007).


by guest on Novem


Dow

nloaded from

http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/

1.30

1.25

1.20

1.35

1.40

1.45

Age

(M

a)

13MFT-3; total fusion experiments13MFT-2; incremental heating experiments

Weighted mean 40Ar/39Ar date of the youngest 24 sanidine grains:

1.3002 ± 0.0005 Ma MSWD = 0.59; p = 0.94

95% confidence interval

Weighted mean 238U/206Pb date of the youngest 19 zircon grains:

1.3031 ± 0.0020 Ma MSWD = 0.55; p = 0.94

95% confidence interval

Chemical domain 2Chemical domain 3Chemical domain 1Chemical domain 4

(a) (b)

Fig. 6. Ranked dates for Mesa Falls Tuff sanidine (a) and zircon (b) grains. Bars include 2r uncertainty. Dotted line indicates the prob-ability distribution function; continuous line represents weighted mean of youngest grains. (a) Single crystal Mesa Falls Tuff sanidinegrains analyzed by both incremental heating and total fusion experiments. Youngest 24 grains yield a weighted mean age of 1�3002 60�0025 Ma. (b) Single crystal Mesa Falls Tuff zircon grains analyzed by CA-ID-TIMS. Colors correspond to chemical domains deter-mined by LA-ICP-MS analyses. Bars with multiple colors represent different chemical domains preserved within one grain. (See textfor a full explanation of each chemical domain.) Youngest 19 grains yield a weighted mean age of 1�3031 6 0�0020 Ma.

950

900

850

800

750

700

650

T˚C

(Ti

-in-z

ircon

)

05.00 0.400.300.200.10Eu/Eu*

To ~ 30 ppm Nb to700 ppm

020 161284Nb

To ~ 0.25

Th/Yto 1.2

02.00 0.160.120.080.04Th/Y

Eu/

Eu*

0

0.50

0.40

0.30

0.20

0.10

00010 800600400200U

U to20000 ppm

1 10 1000100Nb

02.00 0.160.120.080.04Th/Y

02.00 0.160.120.080.04Th/Y

U

0

500

400

300

200

100 U to

200

00 p

pm

Th/Y to 1.2

02.00 0.160.120.080.04Th/Y

Th/Y to 1.2

Hf t

o 18

000

ppm

14000

12000

10000

8000

6000

Hf Th/U

2.0

1.5

1.0

0.5

02.00 0.160.120.080.04Th/Y

Th/Y to 1.2

Th/

U to

2.0

HRT CD 1HRT CD 2HRT CD 3

Chemical domain 2 Chemical domain 3Chemical domain 1 Chemical domain 4

Th/Yto 1.2

(a) (c)(b)

(d) (e) (f)

(i)(h)(g)

± 10˚C ± 10˚C ± 10˚C

Fig. 7. Selected incompatible trace element concentrations of Mesa Falls Tuff zircon plotted as bivariate diagrams. (a–c)Temperature (�C) vs Eu/Eu*, Nb, and Th/Y. For comparison, fields for Huckleberry Ridge Tuff chemical domains 1, 2, and 3 areshown (Rivera et al., 2014). Blue text accompanying these fields indicates the extent of values for chemical domain 3. (d–f)Europium anomaly vs U, Nb, and Th/Y for Mesa Falls Tuff zircon chemical domains. (g) Uranium vs Th/Y. (h) Hf vs Th/Y. (i) Th/U vsTh/Y. Green arrows in (b)–(d) and (f)–(i) indicate the extent of the values for Mesa Falls Tuff chemical domain 4. Uncertainty on Ti-in-zircon temperatures is 6 10 �C determined through analytical reproducibility (Rivera et al., 2013).


by guest on Novem


Dow

nloaded from


based signal criterion to filter out biased analyses (–

0�005 < 207Pb/235U< 0�005) yields a median 206Pb/238Udate for 151 spot analyses of 1�316 þ0�066/–0�050 Ma(95% confidence interval), within error of the 40Ar/39Ar

sanidine date. Under the same exclusion criterion, a

similar result is obtained for 56 spots with U> 250 ppm

(1�259 þ0�057/–0�052), of which the majority are CD4 do-mains. The median error of a single LA-ICP-MS spot

analysis is 0�33 Ma, clearly too imprecise to reveal vari-ations on magmatic time scales, but adequate to re-

solve the presence of significantly older, xenocrystic

cores. Thus it is striking that no pre-Pleistocene zircon

domains are apparent in the screened database of 151

spot analyses. Similarly, only a handful of LA-ICP-MS

spot ages are resolvably older than 1�3 Ma, and pos-sibly occupy zircon crystal cores inherited from the ear-

lier HRT cycle of magmatism.

CA-ID-TIMS zircon U–Pb geochronologyFifty-two grains (Figs 6 and 9) spanning the compos-

itional and thermal spectrum were analyzed by CA-ID-

TIMS and yield a multi-modal distribution of 206Pb/238U

dates ranging from 1�568 6 0�011 to 1�278 6 0�068 Ma(Table 3). Two grains older than 1�5 Ma may containcores of zircon as old as the HRT; however, no discrete

HRT grains were identified. Another three grains define

an apparent sub-population with a mean of 1�374 60�012 Ma. The youngest 19 grains define a peak at1�3031 6 0�0020 Ma (MSWD ¼ 0�55, p ¼ 0�94). Usingmixture modeling deconvolution (Sambridge &

Compston, 1994), additional peaks are identified at

1�350 6 0�002, 1�328 6 0�002, 1�322 6 0�002, and 1�318 60�002 Ma (Fig. 10).

The range and modes of dates obtained are apparent

in the morphology and geochemical variability of the

zircon grains. Zircon age distributions within each pum-

ice and ash-flow tuff sample span a similar range of

dates, and all samples yield multi-modal distributions,

consistent with the dominant modes determined

through mixture modeling (Fig. 10). In general, zircon

grains forming the 1�303 Ma peak are elongate, CL-bright, and exhibit planar zoning, whereas the grains

forming the intermediate age peaks are characterized

by the afore-mentioned CL-dark cores. Dated grains of

chemical domains 1, 2, and 3 form the youngest popula-

tions, whereas grains with CD4 compositions dominate

the intermediate modes of 1�318, 1�322, 1�328, and1�350 Ma, and are generally absent from the populationcomprising the youngest and oldest grains (Figs 6 and

10). The oldest populations, present in pumices 13MFT-

2 and 13MFT-3 and the ash-flow tuff sample 13MFT-6,

have CD3 compositions.

DISCUSSION

Distribution of crystal populations betweenpumice and pyroclastic flowThe large pumice clasts in the outflow MFT provide an

opportunity to examine the heterogeneity in the zircon

crystal cargo between discrete packets of MFT magma.

Between-sample heterogeneity is in fact minor—all

pumice and ash-flow tuff samples share similar zircon

morphologies, display the same degree of zircon chem-

ical variability, and illustrate the same intracrystalline

relationships between CD1 and CD2 interiors with CD3

overgrowths. Additionally, all samples contain grains

with evolved CD4 compositions. Figure 8 visualizes

these similarities between the seven samples in histo-

grams of chemical domain frequency.

Minor variations in zircon morphology exist between

pumice samples, particularly in the degree of lumines-

cence and aspect ratios. For example, pumices 13MFT-

2 and 13MFT-3 have a greater abundance of grains with

longer aspect ratios and planar zoning relative to the

other pumice samples, which contain more grains of

shorter aspect, oscillatory-zoned zircon. Despite minor

differences in zircon morphology between pumice sam-

ples, the relative abundance of chemical domains repre-

sented by the zircon population sampled in each

pumice is insignificant, with CD1 forming �15–20% ofanalyses, CD2 forming �25–30% of analyses, CD3forming more than �40% of analyses, and CD4 formingless than �20% (Fig. 8). The exception to this is pumice13MFT-4, which contains nearly equal proportions of

CD3 (29%) and CD4 (23%) analyses. This may be due to

the smaller number of zircon grains analyzed by LA-ICP-

MS (n ¼ 29 with 35 LA-ICP-MS spots) compared withthe larger sample pool for the other pumice samples

(e.g. 13MFT-2 n ¼ 61 with 88 LA-ICP-MS spots).The two ash-flow tuff samples exhibit similar distri-

butions of chemical domains and morphologies to the

pumice samples, irrespective of the increased welding

or abundance of a phenocryst-poor matrix. Zircon

grains tend to be smaller than those present in the pum-

ice samples, and are better characterized by the shorter

grains with oscillatory zoning, although longer, planar

CD1: 16% CD2: 30% CD4: 11%CD3: 43%P

erce

nt o

f ana

lyse

s pe

r sam

ple

60%

0%

50%

40%

30%

20%

10%

Pumice samples13MFT-1 13MFT-2 13MFT-3 13MFT-4 13MFT-5 13MFT-6 13MFT-7

14

30

39

1720

28

47

6

17

25

49

10

14

34

29

23

11

32

53

5

20

3533

13

16

32

38

14

CD

1

CD

3C

D2

CD

4

Ash flow tuff

Fig. 8. Distribution of chemical domains within each MesaFalls Tuff sample. Numbers within bars represent per cent oftotal LA-ICP-MS spots per sample. Bold numbers above barsindicate per cent of total LA-ICP-MS analyses from all samples(n ¼ 451).


by guest on Novem


Dow

nloaded from


MFT2-z171.568

± 0.011 Ma

420

419

MFT2-z31.514

± 0.006 Ma

323

MFT3-z121.379

± 0.008 Ma

528

MFT2-z161.367

± 0.014 Ma

371

373

372

MFT6-z11.369

± 0.022 Ma

280

279

MFT6-z81.360

± 0.024 Ma

308

306

MFT3-z71.350

± 0.002 Ma

489

488

MFT4-z51.347

± 0.006 Ma

459

MFT3-z111.344

± 0.099 Ma

505

MFT3-z91.343

± 0.049 Ma

450

MFT1-z51.340

± 0.014 Ma

360

MFT1-z41.337

± 0.033 Ma

333

MFT3-z81.333

± 0.009 Ma

442

443

MFT4-z61.331

± 0.004 Ma

461

MFT3-z21.328

± 0.011 Ma

465

466

MFT4-z41.329

± 0.003 Ma

457

458

MFT4-z81.328

± 0.018 Ma

464

(a)

CD 2CD 3

CD 1

CD 4

MFT2-z111.323

± 0.009 Ma

327

MFT4-z101.325

± 0.020 Ma

444

MFT2-z61.320

± 0.011 Ma

290

MFT1-z21.322

± 0.011 Ma

327

326

MFT3-z101.326

± 0.003 Ma

513

MFT3-z31.323

± 0.027 Ma

449

MFT1-z91.322

± 0.002 Ma

344

343

MFT1-z81.318

± 0.002 Ma

340

339

MFT2-z121.317

± 0.009 Ma

305

MFT1-z31.316

± 0.012 Ma

335

336

MFT2-z51.311

± 0.008 Ma

321

320

MFT2-z141.316

± 0.006 Ma

317

MFT2-z41.311

± 0.006 Ma

285

284

MFT1-z61.315

± 0.013 Ma

332

MFT1-z11.315

± 0.018 Ma

328

329

MFT4-z31.314

± 0.021 Ma

474

473

(b)

CD 2CD 3

CD 1

CD 4

Fig. 9. CA-ID-TIMS dated Mesa Falls Tuff zircon grains, arranged by decreasing 206Pb/238U date, shown with 30 lm LA-ICP-MS spotlocation coded by chemical domain. (a) Oldest grains dated within this study. (b) Intermediate aged grains. The abundance ofgrains with CD4 compositions and CL-black cores in (a) and (b) should be noted. (c) Youngest zircon grains demonstrate lack of CL-black interiors and CD4 compositions. All dates include 2r analytical error.


by guest on Novem


Dow

nloaded from


zoned grains are also represented. Grains with dark in-

teriors with CD4 compositions are also present, forming

�15% of the analyses. However, relative to the pumices,the ash-flow tuff samples tend to have fewer CD3 ana-

lyses (33% and 38%). We interpret the similarities in zir-

con form and chemical variability as signs of

homogenization of the penultimate MFT magma prior

to eruption. Our detailed imaging and analysis of the

zoning patterns of more than 300 zircon crystals ex-

tracted from the MFT pumices and the ash-flow tuff

support earlier interpretations for a homogeneous

large-volume Mesa Falls Tuff magma (e.g. zircon d18Oresults of Bindeman et al., 2008; Wotzlaw et al., 2015;

quartz d18O results of Hildreth et al., 1984).

Pre-eruptive differentiation of the Mesa Falls Tuffand Huckleberry Ridge Tuff magmasFollowing zircon saturation of a presumably hotter and

relatively less evolved ‘early’ MFT magma, its down-

temperature progressive differentiation may be recon-

structed using the Ti-in-zircon thermometer and indices

of fractionation such as Eu/Eu* and Th/Y directly meas-

ured in zircon. CD1 compositions, initially defined based

on correlated Th/Y and Eu/Eu*, correspond to the high-

est Ti-in-zircon temperatures (800–875�C) and are inter-

preted as the earliest autocrystic zircon crystallization.

These temperatures correspond to the high end of the

range of calculated zircon saturation temperatures from

published whole-rock major element oxide data

(Hildreth et al., 1984, 1991; Bindeman & Valley, 2001;

Christiansen, 2001). Although we note that the highest

Ti-in-zircon temperatures observed in CD1 analyses

exceed the calculated saturation temperature, this may

be due to evolution of magma compositions subse-

quent to early zircon growth (Harrison et al., 2007).

CD2 and CD3 compositions define progressively

lower Ti-in-zircon temperatures, with the lowest CD3

values reaching �720�C. This progressive cooling trendis matched by correlated increases in incompatible

trace elements (ITE) and the magnitude of the Eu anom-

aly in zircon (Fig. 7; Supplementary Data), consistent

with the progressive differentiation of the magmatic

system through crystallization including feldspar 6quartz. CD3 compositions, which give the lowest tem-

peratures, have pronounced Eu anomalies and elevated

Th/Y, and probably represent the final stage of differen-

tiation prior to eruption, an interpretation supported by

the occurrence of CD3 compositions as outer rims and

tips of zircon crystals.

The CD4 domains, with their extreme ITE concentra-

tions and strongly negative Eu anomalies, appear to be

the result of further differentiation beyond CD3 compos-

itions. However, the restriction of CD4 compositions to

the cores of the crystals, and the superimposition of

less evolved CD3 composition rims onto these cores,

does not permit a simple origin for the CD4 domains

from a single monogenetic differentiation series. One

hypothesis is that these CD4 cores represent crystalliza-

tion from highly evolved residual liquids trapped within

crystal cumulates formed during the earlier evolution of

the MFT magma system, which were subsequently

remobilized, disaggregated, and overgrown during resi-

dence in the climactic magma composition.

Alternatively, these cores could represent crystallization

MFT3-z61.309

± 0.012 Ma

474

MFT6-z71.307

± 0.035 Ma

318

319

MFT4-z91.309

± 0.024 Ma

462

MFT6-z21.308

± 0.028 Ma

270

MFT4-z21.306

± 0.020 Ma

453

454

MFT3-z51.305

± 0.015 Ma

461

462

MFT2-z11.304

± 0.008 Ma

311

310

MFT2-z131.304

± 0.004 Ma

302

MFT3-z41.303

± 0.003 Ma

421

422

423

MFT2-z21.290

± 0.014 Ma

330

329

MFT2-z71.299

± 0.010 Ma

318

319

MFT1-z111.290

± 0.025 Ma

397

MFT2-z151.278

± 0.067 Ma

349

MFT6-z41.304

± 0.073 Ma

324

MFT2-z91.291

± 0.022 Ma

370

MFT3-z11.300

± 0.028 Ma

455

454

453

MFT4-z11.297

± 0.023 Ma

442

443

MFT2-z101.302

± 0.027 Ma

401

400

MFT6-z31.302

± 0.079 Ma

286

287

(c)

CD 2CD 3

CD 1

CD 4

Fig. 9. Continued


by guest on Novem


Dow

nloaded from

http://petrology.oxfordjournals.org/lookup/suppl/doi:10.1093/petrology/egw053/-/DC1http://petrology.oxfordjournals.org/

Tab

le3:

CA

-ID

-TIM

SU

–Pb

iso

top

icd

ata

Ra

dio

ge

nic

iso

top

era

tio

sR

ad

iois

oto

pic

da

tes

Sa

mp

le(a

)G

rain

(a)

Sp

ot

(a)

Th

/U(b

)

20

6P

b*

�1

0–1

3

mo

l(c

)

mo

l%

20

6P

b*

(c)

Pb

*/

Pb

c(c

)

Pb

c(p

g)

(c)

20

6P

b/

20

4P

b(d

)

20

8P

b/2

06P

b(e

)

20

7P

b/

23

5U

(e)

%e

rr(f

)C

orr

.2

06P

b/

23

8U

(e)

(Ma

)%

err

(f)

coe

f.2

07P

b/

23

5U

(g)

6 (f)

20

6P

b/2

38U

(g)

6 (f)

MFT-2

z15

349

0�2

61

0�0

011

32�1

0�1

30�1

926�6

0�0

91

0�0

0003

57436

0�0

001982

5�2

30�0

25

0�0

315�9

1�2

778

0�0

669

MF

T-2

z23

29

/33

00�7

37

0�0

11

36

8�9

0�7

20�4

25

7�9

0�2

55

0�0

01

17

78

0�0

00

20

01

1�0

70�0

85

1�1

90�9

31�2

90

00�0

13

8M

FT

-1z1

13

97

0�4

34

0�0

03

05

6�3

0�3

80�2

04

1�3

0�1

51

0�0

00

76

24

50�0

00

20

01

1�9

10�0

73

0�7

81�9

01�2

90

10�0

24

6M

FT

-2z9

37

00�7

46

0� 0

05

45

9�4

0�4

70�3

14

4�4

0�2

58

0�0

01

16

26

90�0

00

20

03

1�6

80�0

37

1�1

83�1

71�2

91

00�0

21

7M

FT

-4z1

44

2/4

43

0�6

74

0�0

06

85

7�7

0�4

30�4

14

2�5

0�2

33

0�0

01

22

15

10�0

00

20

12

1�7

60�0

67

1�2

41�8

71�2

97

20�0

22

8M

FT

-2z7

31

8/3

19

0�8

00

0�0

11

77

8�3

1�1

90�2

78

3�0

0�2

76

0�0

01

28

30

0�0

00

20

15

0�7

50�1

65

1�3

00�3

91�2

98

90�0

09

7M

FT

-3z1

45

3/4

54

/45

50�7

61

0�0

05

63

7�0

0�1

90�8

02

8�0

0�2

63

0�0

01

27

14

10�0

00

20

16

2�1

60�0

87

1�2

91�8

11�2

99

60�0

28

0M

FT

-6z3

28

6/2

87

0�3

79

0�0

01

92

4�0

0�0

90�5

22

3�5

0�1

32

0�0

00

34

50

25

0�0

00

20

19

6�0

60�0

26

0�3

41

7�2

1�3

01

70�0

78

9M

FT

-2z1

04

00

/40

10�6

84

0�0

05

55

6�7

0�4

20�3

54

1�7

0�2

36

0�0

01

00

39

90�0

00

20

20

2�0

60�0

49

1�0

24�0

51�3

01

90�0

26

8M

FT

-3z4

42

1/4

22

/42

30�5

72

0�0

70

37

7�5

1�0

31�7

57

7�3

0�1

98

0�0

01

27

30�0

00

20

22

0�2

20�5

44

1�2

90�0

41�3

03

10�0

02

9M

FT

-6z4

32

40�7

70

0�0

01

63

1�0

0�1

50�3

02

6�1

0�2

66

0�0

01

31

18

56

0�0

00

20

22

5�6

10�0

18

1�3

32

4�6

1�3

03

50�0

73

1M

FT

-2z1

31

0/3

11

0�9

11

0�0

14

38

1�9

1�5

30�2

69

9�8

0�3

13

0�0

01

25

19

0�0

00

20

23

0�5

80�1

81

1�2

70�2

41�3

04

20�0

07

6M

FT

-2z1

33

02

1�0

55

0�0

37

09

0�0

3�1

60�3

41

80�2

0�3

62

0�0

01

29

30�0

00

20

24

0�3

10�4

77

1�3

10�0

41�3

04

50�0

04

0M

FT

-3z5

46

1/4

62

0�6

88

0�0

11

26

2�0

0�5

10�5

84

6�8

0�2

38

0�0

01

24

65

0�0

00

20

25

1�1

20�1

18

1�2

50�8

11�3

05

20�0

14

6M

FT

-4z2

45

3/4

54

0�6

78

0�0

06

26

2�3

0�5

20�3

14

7�8

0�2

34

0�0

01

00

23

50�0

00

20

26

1�5

60�0

56

1�0

22�3

91�3

06

00�0

20

4M

FT

-6z7

31

8/3

19

0�6

68

0�0

02

94

7�9

0�2

90�2

63

4�6

0�2

31

0�0

01

41

40

80�0

00

20

27

2�6

90�0

34

1�4

35�8

21�3

06

60�0

35

1M

FT

-6z2

27

00�6

34

0�0

04

25

4�2

0�3

70�3

03

9�4

0�2

19

0�0

01

12

26

80�0

00

20

30

2�1

00�0

53

1�1

33�0

41�3

08

50�0

27

5M

FT

-4z9

46

20�6

08

0�0

05

55

7�8

0�4

30�3

44

2�7

0�2

10

0�0

01

12

21

20�0

00

20

31

1�8

00�0

52

1�1

32�4

01�3

08

80�0

23

6M

FT

-3z6

47

40�6

63

0�0

13

94

9�8

0�3

11�1

93

4�9

0�2

29

0�0

01

18

24

0�0

00

20

31

0�8

90�2

58

1�2

00�2

81�3

09

10�0

11

7M

FT

-2z5

32

0/3

21

0�7

66

0�0

17

98

1�4

1�4

30�3

49

7�2

0�2

64

0�0

01

15

19

0�0

00

20

34

0�5

80�1

84

1�1

70�2

31�3

11

30�0

07

6M

FT

-2z4

28

4/2

85

0�8

59

0�0

22

68

4�6

1�8

40�3

41

17�2

0�2

95

0�0

01

31

90�0

00

20

35

0�4

60�2

60

1�3

20�1

21�3

11

40�0

06

0M

FT

-4z3

47

3/4

74

0�7

06

0�0

07

65

5�6

0�4

00�5

14

0�2

0�2

44

0�0

01

27

11

70�0

00

20

38

1�5

60�0

72

1�2

91�5

01�3

13

80�0

20

5M

FT

-1z6

33

20�6

70

0�0

12

17

1�6

0�8

00�4

06

3�5

0�2

32

0�0

01

33

18

0�0

00

20

39

0�9

70�2

59

1�3

50�2

51�3

14

60�0

12

7M

FT

-1z1

32

8/3

29

0�9

33

0�0

04

66

4�4

0�6

20�2

15

0�6

0�3

20

0�0

01

28

91

0�0

00

20

41

1�4

00�0

92

1�3

01�1

91�3

15

40�0

18

4M

FT

-2z1

43

17

1�0

36

0�0

14

58

4�8

1�9

50�2

21

18�4

0�3

55

0�0

01

33

60�0

00

20

42

0�4

70�3

84

1�3

50�0

81�3

15

90�0

06

2M

FT

-1z3

33

5/3

36

0�9

64

0�0

07

47

4�4

1�0

00�2

17

0�5

0�3

31

0�0

01

28

38

0�0

00

20

42

0�8

90�1

38

1�3

00�4

91�3

16

00�0

11

6M

FT

-2z1

23

05

0�6

26

0�0

18

17

3�6

0�8

70�5

56

7�4

0�2

16

0�0

01

28

11

0�0

00

20

43

0�6

60�3

30

1�3

00�1

41�3

16

70�0

08

7M

FT

-1z8

33

9/3

40

0�6

06

0�1

14

09

5�8

7�1

30�4

24

27�8

0�2

10

0�0

01

35

10�0

00

20

45

0�1

40�5

58

1�3

70�0

21�3

17

90�0

01

9M

FT

-2z6

29

00�8

29

0�0

13

17

6�1

1�0

50�3

47

5�5

0�2

85

0�0

01

20

49

0�0

00

20

48

0�8

20�1

08

1�2

10�6

01�3

19

80�0

10

8M

FT

-1z2

32

6/3

27

0�7

84

0�0

12

57

4�6

0�9

60�3

57

0�9

0�2

70

0�0

01

32

18

0�0

00

20

50

0�8

60�2

51

1�3

40�2

41�3

21

50�0

11

3M

FT

-1z9

34

3/3

44

0�6

40

0�1

38

59

7�6

12�8

70�2

87

51�7

0�2

21

0�0

01

35

10�0

00

20

51

0�1

20�5

42

1�3

70�0

11�3

22

20�0

01

5

(co

nti

nu

ed

)


by guest on Novem


Dow

nloaded from


Tab

le3.

Co

nti

nu

ed

Ra

dio

ge

nic

iso

top

era

tio

sR

ad

iois

oto

pic

da

tes

Sa

mp

le(a

)G

rain

(a)

Sp

ot

(a)

Th

/U(b

)

20

6P

b*

�1

0–1

3

mo

l(c

)

mo

l%

20

6P

b*

(c)

Pb

*/

Pb

c(c

)

Pb

c(p

g)

(c)

20

6P

b/

20

4P

b(d

)

20

8P

b/2

06P

b(e

)

20

7P

b/

23

5U

(e)

%e

rr(f

)C

orr

.2

06P

b/

23

8U

(e)

(Ma

)%

err

(f)

coe

f.2

07P

b/

23

5U

(g)

6 (f)

20

6P

b/2

38U

(g)

6 (f)

MF

T-3

z34

49

0�7

20

0�0

05

85

3�6

0�3

70�4

13

8�8

0�2

48

0�0

01

29

84

0�0

00

20

52

2�0

20�1

16

1�3

11�1

11�3

22

80�0

26

8M

FT

-2z1

13

27

0�5

85

0�0

17

17

0�3

0�7

30�6

15

9�8

0�2

02

0�0

01

38

14

0�0

00

20

53

0�7

00�2

63

1�4

00�1

91�3

23

00�0

09

3M

FT

-4z1

04

44

0�8

88

0�0

07

36

2�9

0�5

70�3

64

8�6

0�3

05

0�0

01

26

13

70�0

00

20

55

1�4

80�0

62

1�2

81�7

51�3

24

60�0

19

6M

FT

-3z1

05

13

0�6

90

0� 0

82

87

6�7

1�0

22�1

67

4�6

0�2

38

0�0

01

33

20�0

00

20

57

0�2

00�6

01

1�3

50�0

31�3

26

00�0

02

6M

FT

-4z8

46

40�6

65

0�0

08

36

4�1

0�5

70�3

95

0�2

0�2

29

0�0

01

35

11

40�0

00

20

60

1�3

80�0

61

1�3

71�5

61�3

27

80�0

18

3M

FT

-3z2

46

5/4

66

0�9

47

0�0

14

17

4�8

1�0

20�3

97

1�7

0�3

25

0�0

01

33

21

0�0

00

20

61

0�8

40�2

00

1�3

50�2

81�3

28

50�0

11

1M

FT

-4z4

45

7/4

58

1�0

21

0�0

58

89

5�2

6�9

70�2

43

78�8

0�3

50

0�0

01

34

20�0

00

20

62

0�2

00�4

94

1�3

60�0

31�3

29

40�0

02

7M

FT

-4z6

46

10�7

17

0�0

43

49

0�5

3�0

90�3

81

90�6

0�2

47

0�0

01

35

40�0

00

20

65

0�2

80�3

54

1�3

70�0

51�3

31

10�0

03

8M

FT

-3z8

44

2/4

43

0�6

64

0�0

18

54

5�2

0�2

51�9

33

1�8

0�2

29

0�0

01

25

21

0�0

00

20

68

0�7

10�2

52

1�2

70�2

71�3

32

70�0

09

5M

FT

-1z4

33

30�7

22

0�0

03

34

9�9

0�3

20�2

73

6�0

0�2

49

0�0

01

30

16

00�0

00

20

75

2�4

50�0

79

1�3

12�1

01�3

37

30�0

32

8M

FT

-1z5

45

00�6

32

0�0

12

55

1�2

0�3

21�0

13

6�0

0�2

18

0�0

01

38

30

0�0

00

20

79

1�0

10�2

22

1�4

00�4

11�3

40

10�0

13

5M

FT

-3z9

36

00�6

73

0�0

02

84

2�0

0�2

30�3

23

1�1

0�2

32

0�0

01

32

57

50�0

00

20

84

3�6

50�0

45

1�3

47�7

31�3

43

00�0

49

0M

FT

-3z1

15

05

1�1

86

0�0

00

92

5�3

0�1

20�2

32

4�1

0�4

05

0�0

01

67

10

68

0�0

00

20

85

7�3

60�0

30

1�7

01

8�1

1�3

44

10�0

98

9M

FT

-4z5

45

90�9

65

0�0

27

97

8�1

1�2

10�6

68

0�9

0�3

31

0�0

01

40

15

0�0

00

20

90

0�4

70�1

99

1�4

20�2

11�3

46

90�0

06

4M

FT

-3z7

48

8/4

89

0�7

91

0�1

32

09

7�8

14�7

40�2

48

25�8

0�2

72

0�0

01

38

10�0

00

20

95

0�1

30�5

40

1�4

00�0

11�3

50

20�0

01

7M

FT

-6z8

30

6/3

08

0�8

79

0�0

04

76

0�5

0�5

20�2

54

5�7

0�3

02

0�0

01

44

26

10�0

00

21

10

1�7

30�0

42

1�4

63�8

31�3

59

90�0

23

5M

FT

-2z1

63

71

/37

2/3

73

0�6

84

0�0

12

85

7�4

0�4

20�8

14

1�4

0�2

36

0�0

01

44

46

0�0

00

21

21

1�0

10�1

45

1�4

60�6

71�3

67

40�0

13

8M

FT

-6z1

27

9//2

80

0�6

56

0�0

06

16

1�2

0�5

00�3

24

6�5

0�2

26

0�0

01

55

12

20�0

00

21

23

1�5

90�0

64

1�5

81�9

31�3

68

50�0

21

7M

FT

-3z1

25

28

0�5

84

0�0

21

37

8�2

1�1

10�5

08

2�1

0�2

01

0�0

01

41

80�0

00

21

39

0�5

70�3

54

1�4

30�1

11�3

78

90�0

07

8M

FT

-2z3

32

30�6

01

0�0

30

18

6�4

1�9

90�3

91

32�7

0�2

06

0�0

01

59

70�0

00

23

49

0�4

10�2

96

1�6

10�1

21�5

14

30�0

06

3M

FT

-2z1

74

19

/42

00�5

61

0�0

17

55

4�7

0�3

61�2

43

8�7

0�1

92

0�0

01

57

15

0�0

00

24

32

0�7

10�2

84

1�5

90�2

51�5

67

70�0

11

1

(a)

z1,z2

,e

tc.

are

lab

els

for

sin

gle

zirc

on

gra

infr

ag

me

nts

an

ne

ale

da

nd

che

mic

ally

ab

rad

ed

aft

er

Ma

ttin

son

(20

05

).A

na

lyse

sin

bo

ldfo

nt

we

reu

sed

toca

lcu

late

we

igh

ted

me

an

ag

es.

Sp

ots

are

ass

oci

ate

dL

A-I

CP

-MS

an

aly

ses

pe

rfo

rme

do

nth

ose

gra

ins.

(b)

Mo

de

lT

h/U

rati

oca

lcu

late

dfr

om

rad

iog

en

ic2

08P

b/2

06P

bra

tio

an

d2

07P

b/2

35U

ag

e.

(c)

Pb

*a

nd

Pb

cre

pre

sen

tra

dio

ge

nic

an

dco

mm

on

Pb

,re

spe

ctiv

ely

;m

ol%

20

6P

b*

wit

hre

spe

ctto

rad

iog

en

ic,b

lan

ka

nd

init

ialco

mm

on

Pb

.(d

)M

ea

sure

dra

tio

corr

ect

ed

for

spik

ea

nd

fra

ctio

na

tio

no

nly

.F

ract

ion

ati

on

est

ima

ted

at

0�1

56

0�0

3%

pe

ra

.m.u

.fo

rD

aly

an

aly

ses,

ba

sed

on

an

aly

sis

of

NB

S-9

81

an

dN

BS

-98

2.

(e)

Co

rre

cte

dfo

rfr

act

ion

ati

on

,sp

ike

,a

nd

com

mo

nP

b;

up

to0�4

pg

of

com

mo

nP

bw

as

ass

um

ed

tob

ep

roce

du

ral

bla

nk:

20

6P

b/2

04P

b¼

18�0

39

60�6

4%

;2

07P

b/2

04P

b¼

15�5

37

60�5

5%

;2

08P

b/2

04P

b¼

37�6

65

60�6

4%

(all

un

cert

ain

tie

s1r

).E

xce

sso

ve

rb

lan

kw

as

ass

ign

ed

toin

itia

lco

mm

on

Pb

,u

sin

gth

ea

ve

rag

eis

oto

pic

com

po

siti

on

of

coe

xis

tin

gsa

nid

ine

feld

spa

r:2

06P

b/2

04P

b¼

17�2

44

60�0

4%

;2

07P

b/2

04P

b¼

15�5

49

60�0

6%

;2

08P

b/2

04P

b¼

38�3

05

60�0

8%

.(f

)E

rro

rsa

re2r

,p

rop

ag

ate

du

sin

gth

ea

lgo

rith

ms

of

Sch

mit

z&

Sch

oe

ne

(20

07

).(g

)C

alc

ula

tio

ns

are

ba

sed

on

the

de

cay

con

sta

nts

of

Ja

ffe

ye

ta

l.(1

97

1).

20

6P

b/2

38U

an

d2

07P

b/2

06P

ba

ge

sco

rre

cte

dfo

rin

iti

Journal of Petrology Advance Access published October 29, 2016 · 2019. 10. 25. · ice blocks and the associated ash-flow tuff, including zir-con morphology, trace element chemistry,

Documents