egs002 801. - university of washington€¦ · title: egs002 801..840 created date: 3/15/2012...

The Crustal Magma Storage System of Volca¤ nQuizapu, Chile, and the Effects of MagmaMixing on Magma Diversity

PHILIPP RUPRECHT1*, GEORGE W. BERGANTZ1,KARI M. COOPER2 AND WES HILDRETH3

1UNIVERSITY OF WASHINGTON, DEPARTMENT OF EARTH AND SPACE SCIENCES, BOX 351310, SEATTLE, WA 98195, USA2UNIVERSITY OF CALIFORNIA, DEPARTMENT OF GEOLOGY, ONE SHIELDS AVENUE, DAVIS, CA 95616, USA3US GEOLOGICAL SURVEY, 345 MIDDLEFIELD RD, MENLO PARK, CA 94025, USA

RECEIVED MARCH 3, 2011; ACCEPTEDJANUARY 3, 2012ADVANCE ACCESS PUBLICATION FEBRUARY 4, 2012

Crystal zoning as well as temperature and pressure estimates from

phenocryst phase equilibria are used to constrain the architecture of

the intermediate-sized magmatic system (some tens of km3) of

Volca¤ n Quizapu, Chile, and to document the textural and compos-

itional effects of magma mixing. In contrast to most arc magma sys-

tems, where multiple episodes of open-system behavior obscure the

evidence of major magma chamber events (e.g. melt extraction,

magma mixing), the Quizapu magma system shows limited petro-

graphic complexity in two large historical eruptions (1846^1847

and 1932) that have contrasting eruptive styles. Quizapu magmas

and peripheral mafic magmas exhibit a simple binary mixing rela-

tionship. At the mafic end, basaltic andesite to andesite recharge

magmas complement the record from peripheral cones and show the

same limited range of compositions.The silicic end-member compos-

ition is almost identical in both eruptions of Quizapu.The effusive

1846^1847 eruption records significant mixing between the mafic

and silicic end-members, resulting in hybridized andesites and

mingled dacites. These two compositionally simple eruptions at

Volca¤ n Quizapu present a rare opportunity to isolate particular as-

pects of magma evolutionçformation of homogeneous dacite

magma and late-stage magma mixingçfrom other magma chamber

processes. Crystal zoning, trace element compositions, and crystal-size

distributions provide evidence for spatial separation of the mafic

and silicic magmas. Dacite-derived plagioclase phenocrysts (i.e.

An25^40) show a narrow range in composition and limited zonation,

suggesting growth from a compositionally restricted melt.

Dacite-derived amphibole phenocrysts show similar restricted com-

positions and furthermore constrain, together with more mafic

amphibole phenocrysts, the architecture of the magmatic system at

Volca¤ n Quizapu to be compositionally and thermally zoned, in

which an andesitic mush is overlain by a homogeneous dacitic

magma that is the source for most of the 1846^1847 and 1932 erupted

magmas. Dacite formation is best explained by mineral^melt separ-

ation (crystal fractionation) from an andesitic mush, which is

inferred to have thermally and compositionally buffered the dacite

magma thereby keeping it at relatively low crystallinity (530 vol.

%). The dominant cause of compositional diversity is melt separ-

ation. Back-mixing of mush (i.e. crystals with signatures of growth

both in the andesitic mush and in the dacite magma) into the overly-

ing dacite magma is rarely observed. Recharge events that increase

crystal and magma diversity in the dacite magma are limited to an

episode of mafic recharge and mixing just prior to the 1846^1847

eruption, where evidence for magma mixing is present on all scales.

Chamber-wide mixing was incomplete (mixing efficiency of

�0·53^0·85) as flow lobes vary significantly in composition along

the proposed mixing array. Estimates of viscosity variations during

the course of magma mixing suggest that mixing dynamics and the

degree of magma interaction on all scales were established at the be-

ginning of the recharge event.

KEY WORDS: mineral^melt separation; magma homogeneity; magma

mush; recharge; magma mixing; magma reheating

I NTRODUCTIONHeterogeneity and chemical disequilibrium on the crystaland outcrop scale are common in intermediate to evolved

*Corresponding author. Present address: Lamont^Doherty EarthObservatory of Columbia University, 61 Route 9W, Box 1000,Palisades, NY10964, USA. E-mail: [email protected]

� The Author 2012. Published by Oxford University Press. Allrights reserved. For Permissions, please e-mail: [email protected]

JOURNALOFPETROLOGY VOLUME 53 NUMBER 4 PAGES 801^840 2012 doi:10.1093/petrology/egs002 at U

niversity of Washington on D

ecember 23, 2015

http://petrology.oxfordjournals.org/D

ownloaded from

http://petrology.oxfordjournals.org/

magmas and are a testimony to their complex assemblyhistories and open-system behavior (e.g. Blundy &Cashman, 2008; Streck, 2008). Petrographic complexity inandesitic to dacitic magmas evolves over time through anintricate interplay of: (1) magma chamber tapping duringeruption (Sparks et al., 1977; Eichelberger & Izbekov,2000); (2) addition of and mixing with magma batchesfrom deeper sources (Clynne, 1999; Annen et al., 2006;Kent et al., 2010); (3) mineral^melt segregation leading tothe formation of magma mushes and cumulates(McKenzie, 1985; Bachmann & Bergantz, 2004; Hildreth,2004); (4) the entrainment of those mushes and cumulatesas well as xenoliths from the surrounding country rockinto the magma (e.g. DePaolo, 1981; Bacon & Lowenstern,2005; Davidson et al., 2007). The combination of these pro-cesses leads to dynamic magma plumbing systems thatconstitute the feeders for volcanic eruptions. Whereassome of these processes occur at shallow depths in anupper crustal reservoir, the bulk of the silicic magma mayoriginate (1) in the mantle (e.g. Straub et al., 2011), (2) inthe lower crust (e.g. Hildreth & Moorbath, 1988; Annenet al., 2006) or (3) in an upper crustal mush zone(Hildreth, 1981, 2004; Bachmann et al., 2002). In the caseof upper-crustal (within �10 km of the Earth’s surface)dacite formation, the evolved magma is envisioned to seg-regate within mush zones (sponge-like rigid crystal net-works of �40^60 vol. % crystallinity with interstitialmelt) in the shallow crust. These shallow crystal-rich sys-tems may be fed by more mafic magmas that undergocrystal fractionation^differentiation, which may result inchemically and/or thermally zoned magma^mush columns(e.g. Hildreth, 1981; Druitt & Bacon, 1989; Mandevilleet al., 1996; Costa & Singer, 2002; Browne et al., 2006).Differentiation during mineral^melt separation may aug-ment the chemical and thermal zonation of the plumbingsystem (Brophy, 1991; Bachmann & Bergantz, 2004; Dufek& Bachmann, 2010). Mineral^melt segregation can resultin mush zonation with sharp chemical and rheologicalboundaries, whereas magma mixing during recharge ofmore mafic magma and/or incomplete homogenizationduring sluggish magma convection would spread outthose boundaries over more diffuse areas (Oldenburget al., 1989; Jellinek & Kerr, 1999; Ruprecht et al., 2008). Inthis study we use crystal zoning as well as other petrologi-cal data to constrain the architecture of anintermediate-size silicic magmatic system (some tens ofkm3) and the textural and compositional effects of magmamixing in a historically active silicic system.Although some of the textural complexity that is found

in igneous rocks may occur during mineral^melt segrega-tion (Holness et al., 2007), the dominant process causingchemical disequilibrium is mixing and mingling ofmagma batches within the mush zone and during eruption.Mixing and mingling manifests itself typically by the

presence of mafic enclaves and the juxtaposition of textur-ally and compositionally distinct crystal populations.Even magma batches that mix cryptically (i.e. the mixingof small pulses with similar anhydrous bulk compositionsbut varying volatile content or temperature) can beresolved petrographically (Dungan, 1987; D’Lemos, 1996;Humphreys et al., 2006; Ruprecht & Wo« rner, 2007; deSilva et al., 2008). The sequence and dynamics of themixing process may remain elusive where only the finalmixing product is preserved. Theoretical and dynamicmodels may be used to quantify the time-dependentmixing dynamics in such cases (Sparks & Marshall, 1986;Jellinek & Kerr, 1999; Ruprecht et al., 2008).The complexity recorded in the petrography of inter-

mediate to evolved magmas provides evidence for the op-eration of multiple processes in shallow plumbing systems,but it complicates the isolation of single processes and ob-scures their respective signatures. Some silicic, crystal-poor, arc magmatic systems appear to be less dominatedby open-system behavior (e.g. Chaite¤ n: Castro &Dingwell, 2009; Lara, 2009) as are some volcanic systemswith short eruptive histories (e.g. Huaynaputina: Thouretet al., 1999; Adams et al., 2001; de Silva et al., 2008;Quizapu: Hildreth & Drake, 1992; Ruprecht, 2009). Thelatter provide an opportunity to study specific magmaticprocesses in isolation. We use the eruptive products ofVolca¤ n Quizapu, Chile, to isolate specific processesthat lead to arc magma complexity. At Volca¤ nQuizapu different eruptive products provide insight intothe different stages of magma mixing during solid^liquiddisaggregationçmingling and complete hybridization.We document how the crystals are chemically and phys-ically affected during mixing, how those mixing pro-cesses may imprint themselves in completely hybridizedmagmas as they erupt at the surface, and how both homo-geneity and heterogeneity arise within the Quizapusystem.

VOLCAN QU IZAPUçGEOLOGICAL OVERV IEWVolca¤ n Quizapu is situated along the Holocene volcanicfront of the Southern Andean Volcanic Zone within alarge cluster of volcanic vents dominated by the stratovol-canoes Descabezado Grande and Cerro Azul (35^368S)(Fig. 1; Hildreth & Drake, 1992). Volca¤ n Quizapu is asmall volcanic cone on the northern flank of Cerro Azulwith an �700m wide crater. The eruptive history ofVolca¤ n Quizapu consists almost entirely of eruptions in1846^1847 CE (common era) and in 1932 CE. Minor an-desitic activity (small explosions) occurred during 1907^1931. Recent volcanism in the Quizapu area is bimodal.Five mafic peripheral cinder cone vents and larger com-posite polygenetic cones contrast with the mainly dacitic

JOURNAL OF PETROLOGY VOLUME 53 NUMBER 4 APRIL 2012

802

at University of W

ashington on Decem

ber 23, 2015http://petrology.oxfordjournals.org/

Dow

nloaded from


Fig. 1. Overview of the geology and topography of the area aroundVolca¤ n Quizapu. (a) Regional map showingVolca¤ n Quizapu with respectto the Andean Cordillera. Location map is modified from GeoMapApp. (b) Simplified geological map of Quizapu and its vicinity. (c)Shaded relief map with the flow field of the 1846^1847 eruption outlined. VQ, Volca¤ n Quizapu; CA, Cerro Azul; LR, La Resolana; VDG,Volca¤ n Descabezado Grande. The shaded relief map was rendered from an ASTER satellite image (https://wist.echo.nasa.gov/api/). Sample lo-cations are displayed (white squares, eruption of 1846^1847; black circles, eruption of 1932; black squares, mafic magmas from peripheralcones and from the Quizapu deposits; white circles, mafic magmas from the 1932 eruption).

RUPRECHT et al. MAGMA STORAGE SYSTEM,VOLCA¤ N QUIZAPU

803

at University of W

ashington on Decem


Dow

nloaded from


character of the volumetrically larger stratovolcanoes.Towards the east and the north the silicic character of thevolcanic cluster continues with the Pleistocene CalabozosCaldera and the quasi-historical Mondaca lava flow. TheCalabozos Caldera erupted more than 1000 km3 of rhyoda-cite to dacite magma in at least two caldera-forming erup-tions during the last 300 kyr (Hildreth et al., 1984;Grunder & Mahood, 1988). The rhyodacitic Mondacalava flow fills in the Lontue¤ valley just north ofDescabezado Grande. Historical accounts suggest that iterupted in �1760 CE (J. A. Naranjo, personal communica-tion). Small-volume, mafic cinder cones, stratocones, andmaars are ubiquitous throughout the area with the highestdensity of centers west of Quizapu crater (Hildreth &Drake, 1992).The two major eruptions of Volca¤ n Quizapu are some of

the largest historical eruptions in the entire Andean cordil-lera (Domeyko, 1850; Hildreth & Drake, 1992). In thewinter of 1846^1847 Quizapu produced 4^5 km3 of siliciclavas that covered about 50 km2 (Fig. 1). In contrast, the1932 eruption was a plinian eruption of similar volume,which dispersed ash and lapilli regionally across largeparts of the South American continent (�23^408S). Thegeology of the area and the deposits of the 1846^1847 effu-sive eruption were described by Domeyko (1850). Hildreth& Drake (1992) and Ruprecht (2009) provided more de-tailed petrographic descriptions and geochemical data forthe Quizapu volcanic system. Most notably, the 1846^1847eruption is characterized by extensive physical mixing asa consequence of mafic^silicic magma interaction, whichis dominantly preserved as mingled dacites containingmafic enclaves with crenulate and cuspate margins. Insome cases the recharge magmas caused complete hybrid-ization and mixing at the crystal scale. The 1932 magmasshow little mingling and hybridization. The work ofHildreth & Drake (1992) on the two eruptions led to therecognition of Volca¤ n Quizapu as an example of effusive^explosive transitions in silicic magmatic systems. Ruprecht& Bachmann (2010) showed that this effusive^explosivetransition is directly tied to magma recharge and reheat-ing. The 1846^1847 and 1932 eruptions of the Quizapu vol-canic system, with the presence and absence ofvolumetrically significant andesitic recharge, respectively,have been studied to evaluate the effect of a single magmamixing event on plagioclase residence times recorded by226Ra^230Th disequilibria in a silicic storage system.Plagioclase separates suggest storage for thousands ofyears prior to eruption, during which a single episode ofmagma mixing and reheating owing to mafic rechargemay have changed the crystal growth and dissolution con-ditions and modified the average crystal ages by morethan 1000 years (Ruprecht & Cooper, 2012).

ANALYT ICAL METHODSWhole-rock major and trace element analyses were per-formed at the GeoAnalytical Laboratory at WashingtonState University following the procedure of Johnson et al.(1999) and at the US Geological Survey laboratory inLakewood, Colorado (principal analyst D. F. Siems) fol-lowing the methods discussed by Baedecker (1987) andBacon & Druitt (1988). Mingled samples were carefullycrushed to 1cm, and mafic enclaves (45mm) in the daciteend-member were removed prior to wavelength-dispersiveX-ray fluorescence (XRF) determinations of whole-rockcompositions. The andesite end-member composition wasestimated by analyzing separates of mafic enclaves.Quantitative mineral analyses as well as image acquisi-

tions of textural and compositional maps were performedat the University of Washington using a JEOL 733 four-wavelength-dispersive spectrometer electron microprobe.Quantitative analyses were performed at 15 kV and 10^15 nA with a focused (Fe^Ti oxides, pyroxenes, olivines)or defocused (amphibole, 3 mm; plagioclase, 5 mm) beam.Counting times for most phases and elements were 40 s onpeak and 20 s on the background. Sodium mobility wasnot observed when Na was measured first. Minor elementsin plagioclase required longer counting times: 40^120 s forSr, 40^200 s for Ba, 90^150 s for Fe, and 50^300 s for Mg.Counting times were shorter for general mineral charac-terization and longer for mineral traverses to obtainhigher count rates. Rim analyses with apparent high Feconcentrations probably reflect secondary fluorescence(Longhi et al., 1976; Sugawara, 2001) and are not used inthis study. Semi-quantitative backscattered-electronimages of selected thin sections were obtained at 15 kVand15^40 nA. Long counting times (0·5^2ms on each pixel)at high spatial resolution (4^10 mm per pixel) provide themeans to distinguish between phases as well as betweensodic and calcic plagioclase as a consequence of varyingmean atomic number.Laser ablation inductively coupled plasma mass spec-

trometry (LA-ICP-MS) analyses on selected plagioclasecrystals were conducted at Oregon State University usinga NewWave DUV on 193 nm ArF eximer laser. Ablationwas carried out in a He atmosphere and the ablated mater-ial was analyzed using a VG PQ ExCell quadrupoleICP-MS system. Laser spots for plagioclase crystals wereeither 50 or 80 mm. Spot analyses were made on the samelocations as the preceding microprobe analyses. Calciummeasured by electron microprobe was used as an internalstandard. The analytical conditions and procedures havebeen further described in detail by Kent et al. (2004, 2007).Crystal size distributions (CSDs) were obtained on 13

thin sections. Crystal size data were evaluated with theCSD software provided by Higgins (2000); crystal aspectratios were approximated using the method described byMorgan & Jerram (2006).


804

at University of W

ashington on Decem


Dow

nloaded from


SUMMARY OF WHOLE -ROCKGEOCHEMISTRYNew whole-rock analyses of Quizapu volcanic rocks extendthe compositional range reported by Hildreth & Drake(1992) (Fig. 2; Tables 1 and 2) particularly for the effusive1846^1847 eruption, which shows a wider compositionalrange than was previously reported.The tight cluster of da-citic compositions reported for the bulk of the 1846^1847eruption by Hildreth & Drake (1992) is shown to extendsignificantly and continuously into the andesitic field.Basaltic andesite and dacite samples from the 1846^1847eruption define mixing trends (Fig. 2) for major and traceelements, with the exception of Ni and Cr (Ruprecht &Cooper, 2012). Lavas of the 1846^1847 eruption can bedivided into three groups: (1) andesite recharge magmaspresent as mafic enclaves with varying SiO2 contents upto 58wt % (e.g.VQ-22A; Table 1); (2) hybridized magmas(e.g. VQ-02; Table 1), which form the low-SiO2 end of therange of evolved magmas at �61wt % and show a transi-tion into (3) mingled dacites with varying degreesof recharge magma and a maximum SiO2 content of

�68wt % (e.g.VQ-06 and VQ-08D; Table 1). In compari-son with the compositionally heterogeneous effusive 1846^1847 eruption, the 1932 plinian eruption was relativelycompositionally homogeneous. The 1932 magma includessmall contributions of andesite magmas that erupted as ini-tial and terminal scoria. A minor volume fraction of andes-itic magma, in the form of a single thin horizon ofmingled andesitic pumice, is present in the dominantly da-citic plinian eruptive phase (Hildreth & Drake, 1992;Ruprecht & Bachmann, 2010). These compositions areoverrepresented in Fig. 2 as a result of a sampling bias to-wards rare compositions, which resulted from a samplingstrategy designed to fully characterize the mafic end-member. Differences in the whole-rock geochemistry ofthe two major eruptions of Quizapu (1846^1847 and 1932)define systematic compositional variations for the maficmagmas and illustrate that the dacitic end-members ofthe two eruptions are compositionally identical. Data forperipheral mafic cones (Los Hornitos and La Resolana)and the Mondaca Flow lie along the Quizapu mixingarray (Fig. 2), suggesting that magma diversity in the

Fig. 2. Binary mixing array in K2O^SiO2 space for samples in the vicinity of Quizapu.The 1932 eruption shows a restricted, primarily daciticcomposition, whereas the 1846^1847 lavas extend over a large range of compositions.The Mondaca Flow is another young lava flow in the vicin-ity of Quizapu extending past the regional mixing trend defined by mafic magmas related to recharge as well as peripheral cinder cones anddacitic magmas of Quizapu.


805

at University of W

ashington on Decem


Dow

nloaded from


Table 1: Major element compositions of Quizapu eruptive products and peripheral volcanic rocks1

Sample SiO2 TiO2 Al2O3 FeO* MnO MgO CaO Na2O K2O P2O5 H2O Total (dry)

1846–1847 eruption

Hybrid andesites, mingled dacites, homogeneous dacites

VQ-02 61·37 0·71 16·71 5·08 0·11 2·52 5·18 4·48 2·30 0·18 98·6

VQ-04D 64·67 0·59 15·80 3·61 0·09 1·34 3·15 4·95 2·98 0·15 97·3

VQ-06 65·89 0·56 15·62 3·09 0·09 0·95 2·48 5·17 3·19 0·15 97·2

VQ-08D 67·59 0·53 15·17 3·07 0·08 1·08 2·06 5·18 3·61 0·10 98·5

VQ-09 61·56 0·71 16·71 5·00 0·11 2·40 5·11 4·54 2·34 0·18 98·6

VQ-10 60·78 0·72 16·44 4·93 0·11 2·51 4·69 4·56 2·37 0·18 97·3

VQ-11 65·71 0·56 15·52 3·19 0·09 0·99 2·50 5·10 3·18 0·15 97·0

VQ-12 66·58 0·57 15·66 3·14 0·09 0·95 2·54 5·17 3·24 0·15 98·1

VQ-14 65·80 0·56 15·49 3·09 0·09 0·95 2·43 5·12 3·20 0·15 96·9

VQ-15 61·28 0·71 16·40 4·79 0·11 2·39 4·53 4·63 2·44 0·17 97·4

VQ-16 60·75 0·69 16·35 4·82 0·11 2·33 4·89 4·49 2·34 0·17 96·9

VQ-22D 63·29 0·64 16·00 3·96 0·10 1·70 3·57 4·85 2·77 0·16 97·0

VQ-24D 65·27 0·60 15·83 3·59 0·09 1·31 3·08 5·00 3·03 0·15 98·0

VQ-26D 63·88 0·64 16·07 4·08 0·10 1·70 3·83 4·85 2·78 0·16 98·1

VQ-35 66·82 0·57 15·73 3·16 0·09 0·96 2·53 5·20 3·24 0·15 98·5

VQ-36 66·63 0·57 15·80 3·23 0·09 0·97 2·63 5·16 3·20 0·15 98·4

VQ-39 65·99 0·60 15·93 3·51 0·10 1·12 2·90 5·12 3·07 0·17 98·5

VQ-44D 64·73 0·63 15·96 3·94 0·10 1·54 3·51 4·94 2·91 0·16 98·4

Q-152 61·39 0·74 16·93 5·29 0·11 2·61 5·46 4·62 2·24 0·20 99·3

Q-153 61·11 0·75 16·97 5·40 0·11 2·69 5·57 4·58 2·20 0·20 99·3

Q-154 65·78 0·63 16·13 3·79 0·10 1·42 3·31 5·25 3·03 0·18 99·3

Q-156 67·54 0·58 15·83 3·21 0·09 0·94 2·49 5·43 3·32 0·16 99·4

Q-157 66·89 0·60 16·05 3·35 0·09 1·09 2·71 5·39 3·22 0·22 99·2

Q-158 62·01 0·72 16·77 5·09 0·11 2·53 5·19 4·64 2·37 0·24 99·1

Q-159 64·91 0·66 16·25 4·12 0·10 1·73 3·79 4·99 2·83 0·22 100·1

Q-161 65·69 0·63 16·13 3·87 0·10 1·45 3·37 5·20 2·98 0·17 99·5

Q-163 64·84 0·66 16·32 4·12 0·10 1·71 3·78 5·01 2·84 0·22 99·4

Q-164 61·91 0·72 16·82 5·10 0·11 2·52 5·19 4·64 2·34 0·24 99·5

Q-165 62·52 0·71 16·73 4·90 0·11 2·29 4·93 4·78 2·44 0·20 99·7

Q-166 66·58 0·61 16·08 3·59 0·10 1·12 2·89 5·34 3·10 0·20 99·2

Q-167 61·85 0·72 16·87 5·11 0·11 2·48 5·19 4·66 2·35 0·26 99·5

Q-168 67·32 0·58 15·93 3·23 0·09 1·01 2·59 5·35 3·27 0·22 99·1

Q-169 67·58 0·59 15·84 3·22 0·09 0·96 2·43 5·38 3·32 0·20 99·4

Q-170 63·04 0·71 16·63 4·76 0·11 2·27 4·38 4·93 2·58 0·20 99·3

Q-171 62·46 0·74 16·74 4·92 0·11 2·46 4·62 4·82 2·50 0·24 99·6

Q-172 67·77 0·58 15·83 3·10 0·08 0·92 2·33 5·38 3·40 0·21 99·5

Q-174 62·59 0·72 16·72 4·88 0·11 2·27 4·82 4·81 2·48 0·20 99·5

Q-175 68·09 0·56 15·32 3·18 0·08 1·11 2·07 5·40 3·63 0·15 99·7

Q-176 68·39 0·55 15·27 3·09 0·98 1·04 1·95 5·40 3·68 0·15 99·3

Q-177 67·35 0·59 15·91 3·26 0·09 0·98 2·56 5·43 3·28 0·17 99·5

Q-178 62·65 0·73 16·74 4·84 0·10 2·37 4·55 4·87 2·52 0·22 99·7

Q-179 66·75 0·59 16·10 3·47 0·09 1·07 2·80 5·40 3·13 0·19 99·4

Q-180 66·94 0·59 16·09 3·47 0·09 1·06 2·68 5·27 3·18 0·22 99·2

Q-181 67·00 0·59 16·01 3·41 0·09 1·04 2·70 5·34 3·19 0·22 99·2

(continued)


806

at University of W

ashington on Decem


Dow

nloaded from


Table 1: Continued


Q-182 67·72 0·57 15·79 3·14 0·09 0·92 2·47 5·42 3·32 0·17 99·3

Q-183 68·07 0·57 15·77 3·11 0·09 0·86 2·28 5·40 3·33 0·12 99·4

Q-184 67·02 0·59 15·99 3·38 0·09 1·03 2·71 5·40 3·19 0·20 99·3

Q-186 66·74 0·60 16·01 3·47 0·09 1·18 2·81 5·28 3·19 0·23 99·8

Q-187 61·14 0·74 16·99 5·33 0·11 2·70 5·55 4·58 2·21 0·25 100·3

Q-188 62·19 0·72 16·83 5·00 0·11 2·41 5·05 4·67 2·39 0·23 100·1

Q-198 65·30 0·62 16·20 4·12 0·09 1·51 3·47 4·77 2·95 0·21 99·2

Q-199 65·30 0·63 16·10 4·16 0·10 1·48 3·47 4·83 2·94 0·22 99·2

Q-200 64·70 0·64 16·20 4·34 0·10 1·64 3·70 4·77 2·85 0·21 99·2

Andesitic inclusions

VQ-04A 54·92 0·81 17·53 6·78 0·13 4·08 7·85 3·77 1·38 0·18 97·4

VQ-22A 53·85 1·00 17·95 7·61 0·13 4·76 7·95 3·81 1·19 0·21 98·5

VQ-24A 56·78 0·79 17·14 6·39 0·13 3·56 7·05 3·98 1·65 0·19 97·7

VQ-26A 57·32 0·80 17·30 6·27 0·12 3·60 7·21 3·97 1·68 0·19 98·5

VQ-44A 56·92 0·81 17·29 6·53 0·12 3·75 7·37 3·89 1·58 0·19 98·4

Q-154i 56·71 0·85 17·89 6·87 0·13 3·73 7·64 4·02 1·55 0·21 99·1

Q-160-i 56·74 0·83 17·78 6·79 0·13 3·94 7·68 3·91 1·55 0·26 99·6

Q-162 56·51 0·84 17·77 6·95 0·13 3·93 7·79 4·04 1·44 0·21 99·4

Q-173 57·18 1·00 17·97 7·28 0·14 3·08 6·48 4·56 1·53 0·39 99·5

Q-195 53·20 1·08 18·00 9·42 0·15 4·34 8·31 3·81 1·03 0·28 99·6

Q-198i 58·00 0·79 17·30 6·83 0·12 3·64 7·06 3·96 1·64 0·24 99·6

Q-199i 56·50 0·81 17·50 7·11 0·13 3·83 7·44 3·71 1·56 0·23 98·8

1932 eruption

Dense dacites and dacitic pumices

VQ-17 66·30 0·53 15·46 2·94 0·09 0·87 2·30 5·16 3·25 0·14 97·0

VQ-20 66·74 0·53 15·51 2·88 0·09 0·85 2·28 5·15 3·30 0·14 97·5

VQ-21 66·85 0·55 15·64 3·04 0·09 0·93 2·40 5·14 3·29 0·14 98·1

VQ-23 66·64 0·54 15·60 2·96 0·09 0·88 2·34 5·19 3·26 0·14 97·6

VQ-31 66·22 0·57 15·60 3·14 0·09 0·91 2·47 5·13 3·21 0·15 97·5

VQ-37A 66·50 0·54 15·47 2·97 0·09 0·87 2·38 5·13 3·27 0·14 97·4

VQ-37C 66·76 0·53 15·38 2·90 0·09 0·84 2·28 5·15 3·33 0·14 97·4

VQ-37D 66·41 0·54 15·50 2·94 0·09 0·86 2·36 5·15 3·28 0·14 97·3

VQ-40 66·94 0·52 15·45 2·87 0·08 0·81 2·27 5·15 3·34 0·13 97·6

VQ-41 66·42 0·55 15·71 3·09 0·09 0·93 2·53 5·15 3·22 0·14 97·8

‘Brown band’, initial and terminal scoria

Q-85 55·26 0·94 17·95 7·20 0·12 4·61 7·78 4·11 1·42 0·20 0·17 99·9

Q-91 61·89 0·73 17·06 5·03 0·11 2·63 4·83 4·75 2·42 0·16 0·36 99·0

Q-92 59·84 0·82 17·17 5·87 0·11 3·21 5·72 4·62 2·05 0·18 0·24 99·7

Q-94 61·16 0·76 16·99 5·37 0·11 2·80 5·20 4·80 2·24 0·18 99·7

Q-95 57·59 0·98 17·86 7·17 0·13 2·91 6·36 4·72 1·59 0·30 99·3

Q-07 55·41 0·95 17·97 7·19 0·12 4·67 7·82 3·86 1·41 0·20 0·11 100·9

Q-09 59·65 0·80 17·39 5·82 0·11 3·29 5·90 4·40 2·07 0·19 0·27 100·1

Q-16 54·54 1·00 17·95 7·57 0·13 4·86 8·26 3·80 1·29 0·22 0·2 100·5

Q-17 52·57 1·04 18·62 8·03 0·13 5·35 9·12 3·55 0·98 0·23 50·1 100·7

Q-18 55·34 0·96 17·95 7·21 0·12 4·50 7·96 3·92 1·42 0·22 0·2 99·9

Q-89 52·12 1·05 18·71 8·11 0·13 5·43 9·01 3·82 0·99 0·22 0·15 99·6

(continued)


807

at University of W

ashington on Decem


Dow

nloaded from


region aroundVolca¤ n Quizapu is dominantly controlled bymixing of mafic and silicic magmas similar to theQuizapu end-member compositions.

PHENOCRYST COMPOSIT IONALAND TEXTURALCHARACTER I ST ICS OF THEQU IZAPU MAGMASThree classes of magmatic crystals are conventionally dis-tinguished on the basis of crystal-chemical and texturalcriteria, and distinct origins for these classes are implied(e.g. Bacon & Lowenstern, 2005; Charlier et al., 2005): (1)phenocrysts sensu stricto (also called ‘autocrysts’; Milleret al. 2007) are large crystals that grew from the melt inwhich they were erupted; (2) antecrysts (a term introducedby Hildreth at the ‘Longevity and Dynamics of RhyoliticMagma Systems’ Penrose Conference, 2001) grew in aliquid related to but different from the host liquid andtherefore represent crystals recycled from earlier gener-ations of magmas within the same conduit^reservoirsystem; (3) xenocrysts that originate from assimilatedwall-rocks or from earlier distinct magma compositions.The terminology referring to crystals is complicated bythe fact that phenocrysts sensu lato and microlites are

textural terms that are also used to refer to crystal sizewhere phenocrysts and microlites refer to large (450 mm)and small (550 mm) crystals, respectively, regardless ofthe origin of the crystals. Small phenocrysts (less than afew hundred microns) are sometimes further distinguishedas microphenocrysts. We use the terminology sensu lato

unless otherwise noted or a distinction between pheno-crysts, antecrysts, and xenocrysts is made.

PlagioclasePlagioclase is the volumetrically dominant phenocrystphase in Quizapu magmas. Most lavas and pyroclastshave plagioclase phenocryst modes of 10^15 vol. % on avesicle-free basis (Table 3). OnlyVQ-08D, sampled from avolumetrically minor flow lobe in the interior of the 1846^1847 flow field, is much more crystal poor with a plagio-clase mode of 3·3 vol. %. Hybrid andesites consistentlycontain �14 vol. % plagioclase phenocrysts.Plagioclase phenocrysts are divided into six plagioclase

types on the basis of textures and chemical composition(Fig. 3a^p,Table 4). Only three of these types are volumet-rically significant (types I, III, and IV) in Quizapumagmas. However, all crystal types are described belowin similar detail, regardless of their abundance, to docu-ment the full range of textures and compositions. We

Table 1: Continued


Q-93 57·83 0·96 18·14 7·00 0·13 2·81 6·21 4·61 1·60 0·30 0·33 99·9

Q-01 57·37 1·03 17·95 7·36 0·14 3·05 6·49 4·30 1·59 0·32 0·5 99·3

VQ-18 54·27 0·97 17·91 7·21 0·13 4·77 8·03 3·75 1·29 0·21 98·5

VQ-37B 59·52 0·79 16·94 5·54 0·11 2·99 5·53 4·41 2·11 0·19 98·1

VQ-42 54·42 0·97 17·91 7·14 0·13 4·72 8·02 3·77 1·33 0·21 98·6

Peripheral cones

HOR-06-01 53·55 0·81 17·42 7·48 0·14 5·55 9·46 3·30 0·98 0·17 98·9

CLHE1 54·35 0·83 17·76 7·56 0·14 5·49 9·70 3·38 1·00 0·17 100·4

CLHE2 54·08 0·79 17·02 7·67 0·14 6·92 9·45 3·30 0·99 0·17 100·5

CLHE-TF-04 51·51 0·64 13·99 8·13 0·15 13·66 9·26 2·38 0·65 0·12 100·5

CLHW1 53·73 0·88 17·91 7·51 0·14 6·06 9·58 3·38 0·98 0·18 100·4

CLHW4 54·17 0·83 17·66 7·39 0·13 5·34 9·47 3·46 1·07 0·18 99·7

RES-06-01 54·94 0·89 17·38 7·13 0·13 5·10 7·85 3·84 1·37 0·21 98·8

CLRN2 55·01 0·92 17·80 7·39 0·14 5·24 8·31 3·83 1·30 0·22 100·1

CLRS2 54·44 0·99 18·35 7·38 0·14 4·79 8·93 3·83 1·22 0·24 100·3

Mondaca Flow

MO-07-01 69·75 0·37 14·53 2·06 0·06 0·44 1·31 4·99 3·97 0·04 97·5

MO-07-02 70·10 0·38 14·52 2·08 0·06 0·46 1·36 5·01 3·98 0·07 98·0

1Samples starting with Q were sampled by W. Hildreth during field seasons in 1993 and 2000; other samples were taken in2006 and 2007 by P. Ruprecht.Measurement uncertainties are as reported by Baedecker (1987), Bacon & Druitt (1988) and Johnson et al. (1999).


808

at University of W

ashington on Decem


Dow

nloaded from


Table2:

TraceelementcompositionsofQuizapu

eruptiveproductsandperipheralvolcanicrocks1

Sam

ple

Ni

Cr

Sc

VBa

Rb

Sr

Zr

YNb

Ga

Cu

Zn

Pb

La

Ce

Th

Nd

Sm

Eu

Gd

Tb

Tm

Yb

Lu

Co

Cs

Hf

Sb

Ta

U

1846-1847eruption

Hyb

ridan

desites,mingleddacites,homogen

eousdacites

VQ-06-02

3167

17125

468

70468

184

205

2046

6814

1840

620

VQ-06-04D

1934

1166

576

94325

238

247

1919

6317

1847

923

VQ-06-06

1218

944

608

103

282

255

248

1910

6120

2252

1124

VQ-06-08D

2645

1134

652

114

185

291

298

1815

6420

2758

1126

VQ-06-09

3669

16123

479

69461

187

227

1939

6914

1541

521

VQ-06-10

3866

14108

487

71399

194

227

2027

6814

1741

522

VQ-06-11

1726

944

605

100

275

252

248

1812

6016

2341

820

VQ-06-12

412

944

611

103

278

255

247

1812

6315

2552

1225

VQ-06-14

1319

843

613

100

271

253

248

199

6217

2648

824

VQ-06-15

53113

13102

498

71385

196

218

1930

6914

2239

519

VQ-06-16

3464

16115

475

70442

186

207

2041

6513

2040

519

VQ-06-22D

2948

1179

543

87342

222

237

1918

6416

2146

923

VQ-06-24D

2033

1062

578

97316

243

258

1916

6217

2554

1025

VQ-06-26D

822

1182

545

86361

222

246

1728

6414

2039

1021

VQ-07-35

59

943

619

104

277

257

257

1716

6417

2352

1123

VQ-07-36

818

846

617

102

294

247

247

1815

6116

2456

1025

VQ-07-39

817

856

599

98314

239

246

1914

6716

2547

1223

VQ-07-44D

719

1274

561

90340

231

246

1825

6614

2446

1021

Q-22

80

827

643

105

285

256

58

3826

5012

265·0

1·0

5·4

0·6

0·4

2·4

0·4

4·3

4·2

6·7

0·7

0·6

2·9

Q-34

93

829

638

107

277

272

62

4825

5212

255·2

1·1

5·3

0·7

0·4

2·5

0·4

4·5

5·2

6·8

0·7

0·6

2·8

Q-35

81

27619

283

264

77

52

Q-52

90

28624

288

262

74

42

Q-59a

1610

25519

378

217

620

53

Q-59b

2312

1229

534

81374

225

720

5522

449

204·7

1·1

0·7

2·4

0·3

11·4

4·0

5·5

0·5

0·5

2·3

Andesitic

inclusions

VQ-06-04A

2356

25201

318

36654

106

173

2049

779

1331

418

VQ-06-22A

4871

22205

315

30573

114

185

1948

789

1229

315

VQ-06-24A

2145

22183

358

45603

126

195

2062

7212

1529

416

VQ-06-26A

1342

22174

359

44591

128

184

1867

718

1737

418

VQ-07-44A

1542

23181

348

41606

121

193

1852

769

1731

518

Q-59c

4359

2324

336

32564

150

676

6614

304

153·9

1·1

0·6

1·8

0·2

26·8

1·5

3 ·3

0·2

0·3

1·0

(continued

)


809

at University of W

ashington on Decem


Dow

nloaded from


Table2:

Continued

Sam

ple

Ni

Cr

Sc

VBa

Rb

Sr

Zr

YNb

Ga

Cu

Zn

Pb

La

Ce

Th

Nd

Sm

Eu

Gd

Tb

Tm

Yb

Lu

Co

Cs

Hf

Sb

Ta

U

1932

eruption

Den

sedacites

anddacitic

pumices

VQ-06-17

1313

837

612

106

269

254

257

187

6019

2448

1223

VQ-06-20

1010

837

618

108

265

256

248

2011

5919

2452

1124

VQ-06-21

1012

939

612

108

274

252

248

1715

6119

2552

1125

VQ-06-23

86

736

621

105

272

255

247

1912

6120

2449

1224

VQ-06-31

511

841

616

102

278

252

256

189

6217

2250

1225

VQ-07-37A

511

840

617

104

265

256

257

1812

6117

2452

1223

VQ-07-37C

715

737

624

108

256

258

246

1720

6018

2557

1326

VQ-07-37D

48

736

620

105

266

255

246

1812

6017

2450

1223

VQ-07-40

510

736

617

107

258

257

267

1612

6118

2647

1223

VQ-07-41

512

843

609

104

283

247

256

179

6217

2249

1225

Q-02

82

630

652

108

190

286

84

4625

5114

204·1

0·9

5·3

0·7

0·3

2·5

0·4

6·1

7·0

3·3

Q-02

82

730

652

123

190

286

84

4625

5415

245·1

0·9

4·7

0·7

0·4

2·7

0·4

3·1

6·3

7·5

0·8

0·7

3·3

Q-03

1913

1226

513

86372

222

824

5721

4410

234·0

1·0

4·0

0·7

0·5

2·3

0·3

5·4

5·7

2·9

Q-03

1913

1226

513

86372

222

824

5722

4410

294·2

1·0

5·1

0·7

0·3

2·3

0·4

5·3

5·7

2·7

Q-04

120

29636

260

267

88

53

Q-05

90

728

622

106

262

260

68

4324

5012

204·9

1·0

5·2

0·6

0·4

2·5

0·4

3·8

5·2

6·6

0·7

0·6

2·9

Q-05

90

728

622

101

262

260

68

4325

4912

254·3

1·0

5·6

0·7

0·4

2·5

0·4

5·1

6·4

2·5

Q-05

90

828

622

106

262

260

68

4326

5113

254·6

1·0

4·8

0·8

0·4

2·5

0·4

5·4

6·8

3·9

Q-06

100

27659

274

267

61

48

Q-08

90

28655

277

265

81

48

Q-10

28

28594

99275

260

524

5112

264·9

1·0

5·7

0·6

0·3

2·4

0·4

4·5

5·1

6·6

0·6

0·6

2·8

Q-10

28

28594

98275

260

524

5112

264·3

1·1

53

0·7

0·4

2·6

0·4

5·2

6·8

2·7

Q-21

81

727

625

105

272

260

56

4724

4912

225·0

1·0

0·7

2·5

0·4

3·9

5·1

6·4

0·7

0·6

3·0

Q-81

83

31669

204

270

76

46

Q-82

81

29672

209

273

93

41

‘Brownban

d’,initialan

dterm

inal

scoria

Q-85

3851

23364

548

144

563

64

Q-91

2317

24534

399

205

525

55

Q-92

2837

22438

434

185

830

58

Q-94

2321

26481

415

198

927

56

Q-95

93

24432

574

155

637

56

(continued

)


810

at University of W

ashington on Decem


Dow

nloaded from


Table2:

Continued

Sam

ple

Ni

Cr

Sc

VBa

Rb

Sr

Zr

YNb

Ga

Cu

Zn

Pb

La

Ce

Th

Nd

Sm

Eu

Gd

Tb

Tm

Yb

Lu

Co

Cs

Hf

Sb

Ta

U

Q-07

3560

2122

335

42540

139

355

6015

314

193·4

1·0

4·0

0·5

0·4

1·8

0·3

1·7

3·2

1·0

Q-09

2331

1623

445

62434

182

537

5719

377

214·3

1·0

3·7

0·6

0·3

2·0

0·3

17·4

3·0

4·6

0·4

0·4

1·7

Q-16

3340

2220

338

36588

134

463

4714

304

163·9

1·0

0·5

1·7

0·2

25·6

1·4

2·9

0·2

0·3

1·3

Q-17

4368

2418

273

26599

121

366

5913

273

173·8

1·1

4·3

0·5

0·2

1·6

0·2

29·0

1·0

2·4

0·3

0·2

0·6

Q-18

3859

2120

353

40544

147

358

5915

325

163·9

1·0

0·6

1·7

0·2

2 4·2

1·8

3·1

0·2

0·3

1·1

Q-89

4664

19285

610

121

671

60

Q-93

1015

23432

581

156

629

55

Q-01

1210

24445

550

162

640

75

VQ-06-18

67110

23201

322

34587

117

176

1959

759

1529

317

VQ-07-37B

2136

16131

455

62445

173

215

1940

7312

1940

821

VQ-07-42

3863

24194

326

35577

117

184

1882

747

1333

419

Peripheralcones

HOR-06-01

32109

31241

251

21668

6915

218

9873

78

222

12

CLHE1

26105

30246

256

22680

7414

219

101

7610

921

213

2·3

CLHE2

60199

29231

258

23663

7413

318

9876

1110

243

11

CLHE-TF-04

347

1209

31210

161

14468

5713

015

8470

510

141

91·2

CLHW1

31122

28228

280

21743

7914

220

6376

57

253

151·2

CLHW4

48139

28234

276

24704

8215

420

9978

129

263

151·5

RES-06-01

4384

24198

339

32660

109

175

1873

758

1632

416

CLRN2

3265

25209

336

29691

106

173

1978

807

1431

417

2·4

CLRS2

1957

25224

328

25732

102

164

1984

8310

1637

220

0·9

MondacaFlow

MO-07-01

49

617

648

132

143

268

258

177

5121

2448

1621

MO-07-02

49

617

644

133

144

274

268

189

5020

2154

1424

1Sam

plesstartingwithQ

weresampledbyW.Hildreth

duringfieldseasonsin

1993

and2000;other

samplesweretakenin

2006

and2007

byP.Ruprech

tMeasuremen

tuncertainties

areas

reported

byBaedecker(1987),Bacon&

Druitt(1988)

andJohnsonet

al.(1999)


811

at University of W

ashington on Decem


Dow

nloaded from


emphasize here that in particular the dacite magma isdominated by a simple crystal cargo of low-anorthiteplagioclase.

Low-An plagioclase (type I)

Type I comprises the low end of the anorthite content spec-trum (An25^40, Fig. 4) and is texturally simple. Anhedralto subhedral varieties of type I are predominant in lavasfrom the 1846^1847 eruption, whereas euhedral type Iphenocrysts are characteristic of the 1932 dacite. Type Iphenocrysts (Fig. 3a^d; type Ia) show weak patchy zoningand rare melt inclusions in the core associated with prom-inent resorption surfaces. Strongly resorbed crystals andsubhedral to anhedral type Ib phenocrysts (Fig. 3e and f)that are fully pervaded with melt inclusions and melt chan-nels are present only in 1846^1847 lavas. Type Ib crystalsare rare and occur in both dacite and andesiteend-members of the 1846^1847 lavas. Type Ic crystals (Fig.3g and h) are texturally similar to type Ia, although theycontain antecrystic calcic cores (4An45) that show a sharptransition towards the rim to the typical dacite plagioclasephenocryst compositions (An25^40). They often occur

within glomerocrysts (Fig. 3g) with pyroxenes and Fe^Tioxides.

Transitional plagioclase (types II andV)

Type II and V plagioclases are antecrysts as documentedby the presence of both reverse and normal An zoning,and mostly have intermediate anorthite contents (Fig. 4).The rims are mostly similar to Low-An plagioclase.Normal zoning in type II (Fig. 3i) plagioclase is gradualfrom Low-An rims to more calcic compositions (�An60)in the interior of the crystals. It is distinct from the stepwisetransitions of type Ic. Reverse zoning occurs as sharp com-positional boundaries in zones of coeval resorption^re-growth (exposed in Fig. 3i in the center of the crystal).Whereas type IIa almost lacks the reverse zoning in thecenter of the crystal (Fig. 3i), type IIb crystals (Fig. 3j)are dominated by large Low-An cores. The overgrowthmantles of higher An content in type IIb are also followedby normal zoning towards the rims. The Low-An coresthat dominate the crystal growth history of type IIb resem-ble type Ia crystals. In a few cases, the high-An portionsof these type II crystals are entirely sieve-textured. TypeVcrystals have extensively resorbed cores in the form ofmelt inclusions and melt channels (Fig. 3o). Multiple nor-mally zoned overgrowth rims (An45^55) may occur fortype V, which sometimes are similar to type IIb over-growth rims.

High-An plagioclase (types III and IV)

Types III and IV have high anorthite content (An60^90,Fig. 4). Type III (Fig. 3k and l) plagioclase crystals are eu-hedral and calcic (An70^90) and except for the core andthe outermost rim show very little internal zoning. Corescontain melt inclusions and areas of extensive resorption.The 10^20 mm thick rims are sodic (An40) and similar incomposition to groundmass crystals. This type is ubiqui-tous in 1846^1847 lavas. Type III is present mostly assingle crystals in glass or groundmass matrix; occasionallythese crystals are associated with glomerocrysts. Type IVphenocrysts (Fig. 3m and n) are euhedral, calcic in com-position (An60^70), and show low-amplitude,high-frequency oscillatory zoning with only minor resorp-tion surfaces. Mafic enclaves are typically dominated byeither type III or type IV and rarely contain bothHigh-An types simultaneously.

Minor plagioclase types (typeVI)

Distinct textural characteristics suggest a separate classifi-cation for type VI plagioclase despite its rare occurrencein magmas from Quizapu. Type VI (Fig. 3p) shows themost pronounced disequilibrium textures. Almost theentire crystal is sieve-textured. Only a small more calcicrim is developed.Plagioclase crystals in 1846^1847 lavas range in An con-

tent from An18 to An90 (Fig. 4; Supplementary DataTable

Table 3: Plagioclase phenocryst mode for selected samples

from Quizapu

Sample Plagioclase phenocryst mode SiO2 (wt %)

1846–1847

Hybrid andesite

VQ-02 14·9 61·4

VQ-09 14·0 61·6

VQ-10 13·5 60·8

VQ-15 13·7 61·3

Mingled dacite

VQ-04D 11·6–14·4 64·7

VQ-06 11·6 65·9

VQ-08D 3·3 67·6

VQ-11 14·6 65·7

VQ-14 14·9 65·8

VQ-22D 10·9 (71% so; 29% ca) 63·3

VQ-24D 12·8 65·3

Mafic enclave

VQ-22A 14·0 (100% ca) 53·9

1932

Glassy dacite

VQ-17 11·0 61·3

so, sodic plagioclase (5An56·5); ca, calcic plagioclase(4An56·5); An cut-off was chosen following Ruprecht &Cooper (2012). Plagioclase modes were obtained via theCSDcorrections software by Higgins (2000).


812

at University of W

ashington on Decem


Dow

nloaded from


Fig. 3. Backscattered electron images of plagioclase types in magmas from the two historical eruptions of Quizapu. Low-An plagioclase(type I) is the most common type in dacites from Quizapu. High-An plagioclase (types III and IV) occur predominantly in the andesiterecharge magmas. Transitional plagioclase can be found mainly in the 1932 dacites (type II), whereas type V is rare but found in all magmasfrom Quizapu. TypeVI is also rare. The relatively limited compositional zoning in single crystals compared with common arc magmas shouldbe noted. In particular, Low-An crystals (type Ia) are characterized by little compositional zoning. (a^d) Type Ia; (e, f) type Ib; (g, h) typeIc; (i) type IIa; (j) type IIb; (k, l) type III; (m, n) type IV; (o) typeV; (p) typeVI.


813

at University of W

ashington on Decem


Dow

nloaded from


Fig. 3. Continued


814

at University of W

ashington on Decem


Dow

nloaded from


A1, available for downloading at http://www.petrology.oxfordjournals.org). Three dominant modes correspondto Low-An plagioclase (type I), and High-An plagioclase(types III and IV). The transitional plagioclases (types IIand V) have compositions between those of types I andIV. Hybridized andesites contain the entire range ofplagioclase compositions; crystals in one of those hybri-dized andesites (VQ-02) range from An18 to An89. Theplagioclase phenocrysts in the andesitic and daciticend-members (e.g.VQ-22A andVQ-06) are volumetricallymostly restricted to High-An and Low-An plagioclase,respectively. The plagioclase phenocryst population of

the 1932 dacite is also dominated by Low-An plagioclasewith a secondary population of transitional plagioclase.High-An antecrysts are almost absent in the dacitemagma of 1932. They are overrepresented as a result ofpreferential analysis (Figs 4 and 5e, f). Plagioclase analysesin terminal scoria from the 1932 eruption provide evidencethat the 1932 recharge magma contained phenocrysts simi-lar to those of the 1846^1847 recharge magma.Microlite compositions have not been analyzed quanti-

tatively, but backscattered grayscale comparison with adja-cent phenocrysts provides qualitative constraints formicrolite compositions (normal zoning from about An60

Fig. 3. Continued


815

at University of W

ashington on Decem


Dow

nloaded from


to An30). Microlites are absent in dacite ejecta from the1932 eruption, but they are abundant in mingled dacitesfrom the 1846^1847 lavas and ubiquitous in hybridized1846^1847 andesites.

Plagioclase: Fe^Mg concentrations

The use of Fe and Mg in plagioclase has been shown to bevaluable in distinguishing among compositionally differentmagma inputs (Hattori & Sato, 1996; Ginibre et al., 2002;Ruprecht & Wo« rner, 2007; Humphreys et al., 2009). Minorelement compositions may also be used to distinguish be-tween High-An and Low-An plagioclase phenocrysts ofthe Quizapu magma system (Fig. 5a^h). Different texturaltypes have distinct chemical signatures and their originis discussed further below (Fig. 5h). Low-An plagioclaseis generally lower in Fe (2000^3000 ppm) and Mg(5200 ppm), whereas High-An plagioclase has higherconcentrations of Fe (up 7000 ppm) and Mg (up to1000 ppm). Equilibrium plagioclase Mg-partitioningcurves fordacitecompositionsat8708Candandesitecompos-itions at 1100 8C pass through the Low-An and High-Anplagioclase compositions, respectively (Fig. 5b, d and f).

Transitional plagioclase crystals have correlated Fe^AnandMg^An compositions.The compositions predominant-ly fall on a linear array that connects the Low-An and theHigh-An plagioclase compositions. High-An plagioclasegrains in hybridized 1846^1847 magmas are characterizedby distinct Mg concentrations in two samples, whereasFe concentrations are identical for those two samples(Fig. 5a and b).Transitional plagioclase is rare in the hybri-dized andesites. Some crystals with transitional plagioclasecompositions show a trend of constant low Fe and Mg.The1846^1847 mingled dacites contain both High-An andLow-An plagioclase that form distinct compositional clus-ters (Fig. 5c and d). Some High-An plagioclase containintermediate Mg concentrations. This cluster of intermedi-ate Mg concentrations at �An60 does not correlate with aspecific sample or crystal. Transitional plagioclase crystalsare present and either have constant low Fe and Mg or cor-relate with An content.Plagioclase from pumice samples of the 1932 eruption

show the same range as plagioclase from the 1846^1847lavas; however, only a few crystals make up the field ofHigh-An plagioclase and these are overrepresented in the

Table 4: Plagioclase types in magmatic rocks fromVolca¤ n Quizapu

Type Characteristics Occurrence and interpretation

Ia Low-An plagioclase (An25–40), rare melt inclusions; often embayed,

weak patchy zoning

Most common phenocryst in dacite end-member

Ib Low-An plagioclase (An25–40), strongly resorbed with extensive melt

channels

Rare crystals in dacite end-member

Ic Phenocrystic Low-An plagioclase rims (An25–40) and antecrystic cores

(4An45); mostly as glomerocrysts; typically subhedral

Antecrystic cores derived from crystal mush overgrowth in the dacite

end-member

IIa Normally zoned transitional plagioclase (An25–60); outermost rims

Low-An plagioclase; cores sometimes patchy zoned Low-An plagio-

clase (An25–40) or slightly sieve-textured

Antecrystic cores derived from dacite end-member with overgrowth in

the more mafic mush

IIb Low-An cores similar to type Ia (An25–40) with patchy zones and melt

inclusions overgrown by a resorption–regrowth zone; after pro-

nounced resorption, normal zoning of transitional plagioclase towards

the rim

Rare crystals with antecrystic cores derived from dacite end-member

and overgrowth in the more mafic mush

III High-An plagioclase (An70–90); cores with large melt inclusions and re-

sorption features; mantle weakly zoned, normal zoning at the outer-

most rim towards Low-An plagioclase (�An40); occurs occasionally as

glomerocrysts

Common as large phenocryst in andesite recharge magmas, in mafic

enclaves and mingled dacite as well as hybridized andesite

IV Low-amplitude, high-frequency oscillatory zoning around An70 Common as intermediate-sized phenocrysts in andesite recharge

magmas

V Coarsely sieve-textured core overgrown by pronounced normal zoned

mantles or rims of transitional plagioclase (An45–55)

Rare crystals with antecrystic cores derived from dacite end-member

and overgrowth in the more mafic mush

VI Extensively sieve-textured plagioclase, with very small rim of high An

overgrowth

Extremely rare and found only in hybridized andesite; unclear origin


816

at University of W

ashington on Decem


Dow

nloaded from


diagram because they have been preferentially analyzed(Fig. 5e and f). By volume, the crystals within the 1932dacite magma are dominated by Low-An plagioclase, tran-sitional plagioclase, and antecrysts that are occasionallypart of glomerocrysts. Phenocrysts in the 1932 scoria haveAn contents and Fe^Mg compositions that are similar tothose of the 1846^1847 recharge magmas. Most plagioclasecrystals have high Fe (up to 1wt %) and high Mg (up to800 ppm). The few crystals that are sodic have minor-element characteristics consistent with other Low-Anplagioclase (Fig. 5g).

Plagioclase: trace element compositions

Trace element compositions of plagioclase phenocrystsfrom the 1846^1847 and 1932 eruptions were measured ongrain mounts by LA-ICPMS (Fig. 6; Supplementary DataAppendixTable A2). Thus, textural correlation with chem-istry is not possible; however, trace elements show distinct

concentrations for the Low-An and High-An plagioclasecomponents. Low-An plagioclase has elevated concentra-tions compared with High-An plagioclase of the incompat-ible elements La, Ba, and Pb. Sr concentrations are morevariable.The Low-An plagioclase trace element concentra-tions (La, Ba, Pb) fall near plagioclase equilibrium parti-tioning curves for dacite magma at 8708C (Fig. 6).Low-An plagioclase has identical trace element compos-itions for the 1846^1847 and 1932 magmas. High-Anplagioclase in the 1932 magmas has generally higher traceelement concentrations than in the 1846^1847 magmas.This difference is most evident in Pb concentrations,which are almost exlusively as high as in Low-An plagio-clase. Fe and Mg concentrations in these samples fall inthe range measured by electron microprobe (bold continu-ous and dotted line) with a small systematic shift tohigher concentrations measured by LA-ICP-MS.

Plagioclase: crystal size distributions

CSDs of plagioclase in mingled dacites, hybridized andes-ites and mafic enclaves from the 1846^1847 eruption andfrom a dense dacite block from the 1932 eruption (VQ-17)resolve distinct crystal populations and quantify crystalvolume fractions (Fig. 7; e.g. Cashman & Marsh, 1988;Higgins, 2000; Salisbury et al., 2008). We have measuredCSDs only for phenocrysts and microphenocrysts, whichprovide information about the long-term evolution of themagma. Particular crystal populations that are character-ized by relatively constant nucleation and growth ratesshould follow a linear relationship on a log^linear plot ofpopulation density vs crystal size (Marsh, 1988, 1998).The two analyzed samples that follow the linear behav-

ior expected in the case of continuous growth of a singlecrystal population are a mafic enclave from the 1846^1847eruption and a dense dacite from deposits of the 1932 erup-tion (Fig. 7a). Plagioclase crystals in mafic enclaves are onaverage smaller and have higher population densities atsizes 51·5mm than dacite-derived plagioclase, in accordwith the similar crystallinities of andesites and dacites.Other samples from the 1846^1847 lavas show concave-upCSD patterns (Fig. 7b and c). CSDs for mingled dacitesfrom the 1846^1847 lavas are mostly linear for large crystalsizes (41mm). For very large crystal sizes (44mm) this lin-earity breaks down. However, counting errors are signifi-cant in this size range because such large crystals are rarein thin section. Smaller size fractions (51mm) showhigher population densities than linear projections fromlarge crystal size, which is consistent with the addition ofandesite-derived plagioclase owing to mingling andmixing. Hybridized andesites show a very pronouncedconcave-up CSD pattern. Dacite- and andesite-derivedplagioclase populations contribute equally to the plagio-clase CSDs for the hybridized andesites.

Fig. 4. Histogram of plagioclase compositions (An mol %) forQuizapu, showing a distinct peak for Low-An plagioclase (stippled)and a broader distribution for High-An plagioclase (hatched). Itshould be noted that some plagioclase crystals are strongly zoned(Fig. 3), thus labels of different types in this figure refer only to the ma-jority of the measurements making up the respective compositionalrange. Inset shows plagioclase data for scoria and pumice. MostHigh-An plagioclase analyses are from mafic scoria in 1932. Thedacite pumice contains rare plagioclase with high An content(4An60).


817

at University of W

ashington on Decem


Dow

nloaded from


Fig. 5. (a^g) Fe and Mg vs An content (mol %) in plagioclase from Quizapu. Plagioclase phenocrysts are from 1846^1847 mingled dacites(gray squares), hybridized andesites (gray diamonds; black-rimmed: VQ-02; white-rimmed: VQ-10), 1932 dacites (white circles) and 1932 re-charge andesite (gray triangles; dark: FeO data; light: MgO data). Plagioclase compositional data from Supplementary Data AppendixTableA1. Measurement uncertainties (1SD) are smaller than the symbol size unless shown. Mg-plagioclase partitioning curves after Bindeman et al.(1998) are shown for dacite magma (continuous line; 8708C) and for andesite magma (dashed line; 11008C). Melt compositions are approxi-mated by whole-rock analyses fromTables 1 and 2 (dacite: VQ-06; andesite: VQ-22A). Uncertainties on the partitioning curves are shown asgray envelopes. Mg-partitioning curves are omitted in (g). Fe-plagioclase partitioning curves are not shown, as melt compositions are less welldetermined. The dotted field highlights the highest density of data points and corresponds to Low-An phenocrysts in both Quizapu eruptions(�An30). The continuous irregular line outlines the highest data density of the High-An phenocrysts in the 1846^1847 eruption. Both data dens-ity contours were calculated by summing the 2D-Gaussian distribution functions of single data points. The shaded area and the dashed outlinerepresent the 25% and the 5% contour of the calculated maximum data density, respectively. It should be noted that these density outlinesare not fully representive of the actual volume fractions in the magmas, because underrepresented plagioclase crystals were preferentially ana-lyzed in this study. Nonetheless, they provide a general sense of the abundance of different plagioclase crystals. (h) Interpretation of the variouscrystal populations and potential processes explaining the trace element signature.


818

at University of W

ashington on Decem


Dow

nloaded from


AmphiboleAmphibole is ubiquitous in dacites from Quizapu andmakes up about 2 vol. % of the magma (Hildreth &Drake, 1992). Most amphiboles in lavas and plinian dacites

are weakly zoned and euhedral to subhedral (Fig. 8a).Some 1846^1847 flow lobes contain amphibole with upto 50 mm thick decomposition rims composed of clinopyr-oxene, plagioclase, and titanomagnetite (Fig. 8b). The

Fig. 6. Trace element concentrations in plagioclase from Quizapu, as a function of An content. Plagioclase crystals are from the 1846^1847(gray squares) and the 1932 eruption (white circles). Plagioclase trace element compositional data are from Supplementary Data AppendixTable A2. Error bars show measurement uncertainties (1SD), where uncertainties are larger than the symbol size. Thick dotted and continuouslines are the same fields as in Fig. 5c and d, and suggest good agreement between electron microprobe and LA-ICP-MS data. Equilibrium par-titioning curves as in Fig. 5 for plagioclase in dacite magma at 8708C (continuous line) and in andesite magma at 11008C (dashed line).Partitioning curve uncertainties are only partially shown for a restricted An range for the dacite composition and low temperature (An20^60;light gray envelope) as well as for the andesite composition and high temperature (An60^100; dark gray envelope). The uncertainty curve forthe upper limit for Pb equilibrium partitioning in dacite is off the scale, ranging between 22 and 34 ppm. Melt composition is estimated fromisotope dilution measurements of glass separates from Ruprecht & Cooper (2012; Ba) and whole-rock compositions inTables 1 and 2 (La, Pb,Mg; dacite: VQ-06; andesite: VQ-22A). Relationships for partitioning coefficients are taken from Be¤ dard (2006; Ba, La, Pb) and Bindemanet al. (1998). Equilibrium curves for Sr and Fe are not shown owing to the lack of good melt composition estimates. Mg partitioning curves areidentical to those in Fig. 5. Diffusion arrows show the re-equilibration direction of calcic plagioclase in more silicic magma. The length of thearrows shows the relative variations in diffusion coefficient (long indicates fast; short indicates slow); element diffusion coefficients are fromCherniak (1995, 2002, 2003), Giletti & Shanahan (1997) and Costa et al. (2003).


819

at University of W

ashington on Decem


Dow

nloaded from


interior parts of the decomposition rims are relativelycoarse-grained (5^10 mm grain size). The exterior parts ofthe rims are more fine-grained (55 mm grain size).Hybridized andesites contain entirely decomposed amphi-bole remnants (Fig. 8c).They also show fine-grained outer-most rims, on coarse-grained interiors.

Amphibole: compositional variations

A total of 336 amphibole analyses from 29 crystals in theQuizapu dacites document the range of calcic amphibolecompositions (Fig. 9; Supplementary Data Table A3), fol-lowing the classification scheme of Leake et al. (1997), withCa ranging from 1·6 to 1·84 a.p.f.u. (atoms per formulaunit). Amphiboles vary from edenite, a magnesiohornble-nde, to magnesiohastingsite and pargasite, with a totalrange for Si of 6·25^6·85 a.p.f.u. Magnesiohornblendes areeuhedral phenocrysts in the 1846^1847 and 1932 dacites.In contrast to magnesiohornblende, pargasites and magne-siohastingsite are found either as cores surrounded bymagnesiohornblende mantles or as subhedral glomero-crysts intergrown with plagioclase, Fe^Ti oxides, and apa-tite (Fig. 8d and e). The cores and glomerocrysts aremainly pargasitic in the 1932 ejecta, but similar coresand glomerocrysts in the 1846^1847 lavas are mag-nesiohastingsite in composition with more ferric iron thanoctahedral-coordinated aluminum (Figs 8 and 9). Themagnesiohornblendes are relatively homogeneous in com-position and they co-crystallized with Low-An plagioclase(type I) in dacite magma. We assume that magnesio-hornblendes dominate the amphibole population on thebasis of textural observations (i.e. predominantly euhedralcrystals).Compositional variations occur by simple ferrous iron

and magnesium exchange as well as by the temperature-sensitive edenite exchange [TSiþ Aœ¼ TAlþ A(NaþK)],as has been observed in other magmatic systems(Bachmann & Dungan, 2002; Fig. 10). The two amphibolepopulations in the magmas from Quizapu, (1) magnesio-hornblende and (2) magnesiohastingsite and pargasite, aredistinct with respect to the Ti-Tschermak exchange(TSiþM1^M3Mn¼TAlþM1^M3Ti). Edenites show a goodcorrelation (R2

¼0·67) between TAl and M1^M3Ti (�40%accommodation of the variations in TAl). Magnesiohas-tingsites and pargasites show a weak correlation(R2¼0·27) with respect to the Ti-Tschermak exchange.

The remaining variation in TAl is accommodated by the‘plagioclase exchange’ (TSiþM4Na¼TAlþM4Ca).

Pyroxene, olivine and accessory phasesDacite magmas of the 1846^1847 and 1932 eruptions ofQuizapu contain mostly orthopyroxene (mostly at52 vol.%), whereas clinopyroxene is present predominantly as aphenocryst phase in the recharge magmas. Orthopyrox-enes in mingled dacites are frequently anhedral, suggestinglate-stage resorption. We have found a compositional

Fig. 7. Plagioclase crystal size distribution (CSD) for magmas fromVolca¤ n Quizapu. (a) Only two samples (VQ-22A and VQ-17) showlinear CSDs suggesting continuous growth of a single population. (b)Mingled dacites (lines and hatched field) from the 1846^1847 eruptionhave concave-up CSDs, predominantly as a result of the addition ofmicrolites. (c) Hybridized andesites (lines and stippled field) showstrongly curved CSDs as a result of the hybridization process. (d)Comparison of the CSDs from (a)^(c). It should be noted that theaspect ratios (VQ-06, 3·4:1·7:1; VQ-14, 2·8:1·4:1; VQ-11, 5:2·5:1; VQ-24,3·8:2·5:1; VQ-04.2, 3·2:1·6:1; VQ-04.3, 3·6:1·6:1; VQ-22, 3·8:1·7:1; VQ-09,3·2:2·2:1; VQ-02, 3·6:2·8:1; VQ-15, 3·6:2·3:1; VQ-10, 3·6:2·1:1; VQ-22,3·8:1·7:1; VQ-17, 4:1·7:1)) used for the analysis were obtained viaMorgan & Jerram (2006). The symbols are different from those inFigs 5 and 6.


820

at University of W

ashington on Decem


Dow

nloaded from


Fig. 8. Backscattered-electron images of amphiboles in the eruptive products of Quizapu. (a) Euhedral amphibole in dacite from the 1846^1847eruption. (b) Decomposition rim on euhedral amphibole from a 1846^1847 dacite. Decomposition products are clinopyroxene, plagioclase,and titanomagnetite. (c) Decomposed amphibole remnant in a hybridized andesite. (d) Euhedral edenite phenocryst with a magnesiohasting-site core from the 1932 eruption. (e) Pargasitic glomerocryst intergrown with plagioclase from the 1932 eruption. The gray scale for this imageis not linear, as the dynamic range was increased to show both amphibole and plagioclase zoning.


821

at University of W

ashington on Decem


Dow

nloaded from


range that is similar to that reported by Hildreth & Drake(1992), wherein orthopyroxenes are similar in the 1846^1847 (En65^77) and 1932 (En66^72·5) dacites. Orthopyrox-enes in the mafic encalves overlap with those in the dacites,but have slightly lower maximum and minimumMg num-bers (En62·5^71) than those in the dacite. Pyroxene compos-itions in the hybridized andesites cover the entire range(En63^77) spanned by dacites and recharge andesites. Withthe exception of crystals in a single sample from the andes-ite recharge magma (VQ-22A), most orthopyroxenes havelow Ca concentrations (Wo2^3).VQ-22A contains pigeonitemicrolites suggesting late growth at shallow depths andhigh temperatures (Longhi & Bertka, 1996). Ca contentvaries in these pigeonites in the range Wo6^23. Augiteswith an average of En44Fs14Wo42 have been analyzed onlyin the 1846^1847 lavas. Our limited dataset precludes a de-tailed comparison of pyroxene compositions between therecharge magmas and the dacite magma.Olivine is a phenocryst phase in the mafic recharge

magmas. It also occurs as xenocrysts in mingled dacitesfrom the 1846^1847 lavas and in dacites from the 1932 pli-nian ejecta. In the dacite the olivine is normally zoned.The compositions of olivine cores are more magnesianin the 1846^1847 lavas at �Fo76 compared with �Fo70 insamples from the 1932 dacite. Apatite, titanomagnetite,ilmenite, and sulfides are ubiquitous in the dacites ofQuizapu. They occur as phenocryst phases as well as min-eral inclusions in plagioclase and amphibole. Zircon wasnot reported previously (Hildreth & Drake, 1992). Wehave found zircon (550 mm) in three thin sections as wellas in several Fe^Ti oxide grain mounts of the mingled da-cites and hybridized andesites from the 1846^1847 erup-tion. These are typically associated with magnetite andamphibole crystal clots. Quartz has been observed only inplagioclase grain mounts.

PRE -ERUPT IVE MAGMATICSTORAGE CONDIT IONSMagma temperaturesThe mineral assemblages in the Quizapu volcanic rockspermit calculation of pre-eruptive magmatic temperaturesfor dacite magmas that resided in a shallow crustal cham-ber using multiple phenocryst-pair equilibrium geotherm-ometers: Fe^Ti oxide thermometry (Ghiorso & Evans,2008) and amphibole^plagioclase thermometry (Holland& Blundy, 1994). The calcic amphibole model of Ridolfiet al. (2010) has been applied as an independent test of theamphibole^plagioclase model and it simultaneously pro-vides constraints on oxygen fugacity and crystallizationpressure. Temperature estimates for the recharge andesitemagmas are also obtained from the Fe^Ti oxide thermom-eter (Ghiorso & Evans, 2008). As Fe^Ti oxides can beused for both dacite and andesite end-members, we discussthem together at the end of this section.Re-equilibration of temperature-sensitive major and

minor elements by diffusion in amphibole is very slowowing to the coupled exchange of multiple elements(Garcia-Casco et al., 2002), but Fe^Ti oxides re-equilibraterapidly.The combination of data permits the establishmentof a time^temperature evolution for magma storage condi-tions at Quizapu. Rapidly re-equilibrating Fe^Ti oxidesrecord the late-stage (days to months prior to eruption)temperature distribution in the magmatic system (Freer &Hauptman, 1978; Venezky & Rutherford, 1999), whereastemperature estimates from amphibole provide informa-tion about long-term storage conditions for these magmas.Temperature estimates using the Ridolfi et al. (2010)

model for all 336 amphibole analyses are shown in Fig. 11,and the Ridolfi et al. (2010) amphibole classification isadopted here. All analyses are consistent with the observedlow crystallinities of these magmas [i.e. the dash^dotted

Fig. 9. Amphibole compositions in magmas from Quizapu following the classification scheme of Leake et al. (1997). Most amphiboles areedenite (black squares and white circles as in Fig. 6). Old cores and amphiboles that are associated with glomerocrysts are less silica-rich andare magnesiohastingsite and pargasite (gray squares, 1846^1847; gray circles, 1932). Structural formulae of amphiboles were calculated usingthe method of 13 cations excluding Ca, Na and K (13eCNK) described by Leake et al. (1997); compositional data are reported inSupplementary Data AppendixTable A3.


822

at University of W

ashington on Decem


Dow

nloaded from


Fig. 10. Temperature-dependent exchange and Fe^Mg exchange in amphibole. Edenite exchange takes up most of the compositional variation.To a minor extent,Ti-Tschermak and plagioclase exchange contribute to the variations found in the Quizapu amphiboles. Magnesiohastingsitesand pargasites follow the same edenite exchange trend. They differ considerably for theTi-Tschermak and plagioclase exchange. Symbols as inFig. 9. Structural formulae of amphiboles were calculated using the method of 13 cations excluding Ca, Na and K (13eCNK) described byLeake et al. (1997). The arrow in (a) shows the trend for which TAl variations would be varying solely owing to varying temperature.


823

at University of W

ashington on Decem


Dow

nloaded from


line provided by Ridolfi et al. (2010) divides experimentallyconstrained high-T^low-crystallinity assemblages fromlow-T^high-crystallinity assemblages]. The pressure^tem-perature (P^T) estimates from all amphiboles show a con-tinuous range of 120^350MPa and 840^9708C. Thelow-pressure limit is sharp, whereas the high-pressurelimit is more diffuse. As shown by the error bars in Fig. 11,these high-T^low-P amphiboles lend themselves to preciseP^Testimates. Amphibole temperatures for the two erup-tions are indistinguishable from each other. The calculatedminimum temperatures are at the low end of the tempera-ture estimates by Hildreth & Drake (1992) using Fe^Tioxides.We further assess the consistency of the calculated tem-

peratures by comparing the results from the Ridolfi et al.(2010) model with those from the Holland & Blundy(1994) model (Fig. 12). We use the temperature model B of

Holland & Blundy (1994), which is appropriate for pheno-cryst assemblages lacking quartz. Consistent amphiboletemperatures between the two models are obtained for themagnesiohornblendes if the plagioclase composition isassumed to be An40.We assume slightly more calcic plagio-clase than is observed in the dacites to obtain consistenttemperature estimates for the two models. However, tem-perature estimates are still within the uncertainties whenmore sodic plagioclase (An35) is assumed, which is equiva-lent to an �208C lower calculated amphibole equilibrationtemperature. Temperatures from tschermakitic-pargasiteand magnesiohastingsite underestimate crystallizationtemperatures when the Holland & Blundy (1994) model isapplied using Low-An plagioclase. The discordance of cal-culated temperatures between the two models [Holland &Blundy (1994) and Ridolfi et al. (2010)] is most probablyan artifact of the assumed An content used in the former

Fig. 11. Crystallization conditions recorded in calcic amphibole calculated after Ridolfi et al. (2010) for Quizapu magmas (a, b:1932 eruption; c,d:1846^1847 eruption). (a and c) Pressure^temperature relationships for amphibole growth. (b and d) Temperature^water content relationshipsfor amphibole. The curves in (a)^(d) are from Ridolfi et al. (2010): the dotted curves indicate the maximum thermal stability of amphibole;the dash^dot lines divide low-T^high-crystallinity (435 vol. %) assemblages from high-T^low-crystallinity assemblages (535 vol. %) consistentwith the crystallinities observed in the Quizapu dacites; continuous lines represent changes in crystal assemblages as described by Ridolfi et al.(2010); dashed curves represent amphibole analyses that are stoichiometrically consistent. All amphiboles fall on the maximum thermal stabilitycurve (dotted line) of Ridolfi et al. (2010). Mg-hornblendes (black triangles), which represent the dacite-derived amphibole from the eruptible si-licic melt lens of the 1846^1847 and 1932 eruptions (see text for discussion), record restricted P^T^H2O crystallization conditions, whereasmush-derived amphiboles (tschermakitic-pargasite, gray diamonds; Mg-hastingsite, white squares) indicate a range in P^T^H2O.Compositional distinctions follow Ridolfi et al. (2010). Amphiboles from the 1846^1847 and 1932 eruptions show identical ranges in P,T andH2O. Error bars are dominated by the parameterization of Ridolfi et al. (2010) and P^T^H2O uncertainty estimates are shown in (a)^(d). Itshould be noted that the Ridolfi et al. (2010) uncertainties vary; pressure uncertainties are largest for high-pressure assemblages.


824

at University of W

ashington on Decem


Dow

nloaded from


model. Significant temperature variations recorded inamphibole are expected owing to the variations in thetemperature-sensitive edenite exchange (Fig. 10a). It is un-likely that the same plagioclase composition is in equilib-rium with amphiboles that grew at different temperatures.In fact, both temperature models [Holland & Blundy(1994) and Ridolfi et al. (2010)] show consistent high tem-peratures when compositions of the transitional plagio-clase ranging up to An55 are used.Dacitic to rhyodacitic calc-alkaline magmas typically

have only small amounts of clinopyroxene present, thuslimiting the use of two-pyroxene thermometry. The lackof clinopyroxene in the dacite end-member at Quizapu issimilar to other magmatic systems (Mt St. Helens:Rutherford et al., 1985; Rutherford & Hill, 1993; MtUnzen: Holtz et al., 2005; San Pedro: Costa et al., 2004;Soufrie' re Hills: Rutherford & Devine, 2003). TheQuizapu dacite may be too cold and too wet for amphiboleand clinopyroxene to coexist. Even though two pyroxenesare present in the recharge andesites, we do not calculatetwo-pyroxene equilibrium temperatures (e.g. Putirka,2008) because there is no clear textural evidence that thepyroxenes are in equilibrium.Fe^Ti oxide geothermometry has been applied to the

dacites and andesites erupted at Quizapu (Fig. 13a;

Supplementary Data Appendix Table A4; Ruprecht &Bachmann, 2010). We have obtained equilibrium pairs formost lithologies of the 1846^1847 and 1932 eruptions. Atight cluster of nine oxide pairs from two dacite samplesyields temperature estimates of 866^8858C for the 1932plinian eruption. These temperature estimates are con-sistent with previously reported two-oxide temperatures(Hildreth & Drake, 1992).Four Fe^Ti oxide pairs from the dacite end-member of

the 1846^1847 eruption yield elevated temperatures and awider range in temperatures (890^9568C) compared withthe 1932 dacite (866^8858C). Hildreth & Drake (1992) re-ported lower temperature estimates (58748C) using Fe^Tioxides for enclave-poor samples of the 1846^1847 dacite.With the exception of one Fe^Ti oxide pair that yields amagmatic temperature of 10858C, 11 pairs from the hybri-dized andesite yield temperatures of 993^10248C. Seventitanomagnetite^ilmenite pairs in a mingled dacite lavawere analyzed across an enclave^dacite interface along a1cm long transect (black and white data points in Fig. 13aand b). Temperatures for these pairs range from 1007 to10788C, where the highest temperatures are from grains inthe enclave and the lowest temperatures are found awayfrom the enclave (Fig. 13b). A second enclave from a differ-ent sample gave temperatures results that are similar to

Fig. 12. Comparison of the results from the plagioclase^amphibole thermometer of Holland & Blundy (1994) (HB94) with the amphibolethermometer of Ridolfi et al. (2010) (RRP10) assuming that amphibole is in equilibriumwith An40 plagioclase at 200MPa.The equilibrium tem-peratures calculated with the Holland & Blundy (1994) model are relatively pressure insensitive. Equilibrium temperature uncertainties owingto varying plagioclase compositions are �408C for a range of An35 to An45. Only the edenitic or magnesiohornblendes show consistent tempera-ture estimates. Tschermakitic-pargasite and magnesiohastingsite record higher temperatures for the Ridolfi et al. (2010) model. Only the edenitecrystals are in equilibrium with the observed dacite-derived plagioclase (�An40).The other amphiboles grew at higher temperatures in equilib-rium with more calcic plagioclase. Symbols as in Fig. 11.


825

at University of W

ashington on Decem


Dow

nloaded from


Fig. 13. Temperature^oxygen fugacity relationships for the Quizapu magmas. (a) Oxygen fugacity vs temperature for Fe^Ti oxide pairs fromthe 1846^1847 (circles) and 1932 magmas (squares; VQ-17,VQ-37D). The 1846^1847 samples include a mafic enclave from a recharge andesite(black; VQ-44), a hybrid andesite (gray; VQ-02), a mingled dacite (half black and white symbols; VQ-22), and a homogeneous dacite (white;VQ-06). Some Fe^Ti oxide pairs in the mafic enclave of VQ-44 show exsolution lamellae consistent with sub-solidus recrystallization (hatched).(b) Fine-scale temperature variations across the contact between the mafic enclave and the dacite host inVQ-22. BSE map from Ruprecht &Cooper (2012, fig. 2c) showing the location of the transect (gray line) along which the Fe^Ti oxide pairs were measured. (c) Oxygen fugacity cal-culated after Ridolfi et al. (2010) for amphiboles from the 1932 eruption. Temperature^oxygen fugacity ranges from Fe^Ti oxides are shown for1846^1847 mafic enclaves (field outlined by dash^dot line), 1846^1847 mingled and homogeneous dacites (gray field), and for 1932 dacites(hatched field). Magnesiohornblendes (black triangles) from the 1932 dacites show the same range as the corresponding Fe^Ti oxides. (d)Oxygen fugacity for amphiboles from the 1846^1847 eruption deposits. All amphiboles record lower temperatures than most Fe^Ti oxidesfrom the 1846^1847 deposits. Symbols in (c) and (d) as in Fig. 11.


826

at University of W

ashington on Decem


Dow

nloaded from


those for the interior of the former enclave (two pairs:10868C and 10958C), but three additional pairs inthe second enclave gave temperatures between 737 and8208C.These grains contain abundant exsolution lamellae,providing evidence that some of these Fe^Ti oxides mayhave been intermittently subsolidus and are therefore xeno-crystic (Von Gruenewaldt, 1985).

Water contentThe Ridolfi et al. (2010) model provides a formulation to es-timate the water contents of calc-alkaline magmas thatcontain amphibole (Fig. 11). Water content estimates varybetween 3·5 and 5·5wt % H2O, with a cluster around4wt %. Only the deepest, hottest, and potentially mostmafic magmas have higher water contents. Magmas from1846^1847 and 1932 are indistinguishable in terms ofwater contents derived from amphibole compositions, inaccord with the similarities in P^Tconditions.Pre-eruptive water contents for the dacite end-member

magma can also be estimated using the hygrometer ofLange et al. (2009). Assuming average glass (Supplemen-tary Data Appendix Table A5) and plagioclase compos-itions together with temperature estimates from theamphibole geothermometer we estimate dacite water con-tents of 4^6wt %, consistent with estimates using theamphibole model of Ridolfi et al. (2010). The water contentsof the 1846^1847 and 1932 eruptions appear to be indistin-guishable; this is consistent with the lack of significantamounts of quartz, which is present only under muchdrier conditions (52wt %) at �9008C (Whitney, 1988).

Oxygen fugacityOxygen fugacities in 1846^1847 lavas vary beyond the re-stricted range calculated for the 1932 eruption (Fig. 13a).Oxygen fugacity was calculated using the calibration ofGhiorso & Evans (2008). Oxygen fugacity in dacites fromthe 1932 eruption varies from 0·67 to 0·71 log units abovethe nickel^nickel oxide (NNO) buffer. Mingled dacitesfrom the 1846^1847 eruption have oxygen fugacities vary-ing between NNO þ0·74 and NNO þ0·85. Hybridized an-desites and mafic recharge magmas are characterized byrelatively reduced oxygen fugacities ranging from NNOþ0·24 to NNO þ0·53. The Fe^Ti oxide pairs that yieldlow temperatures are affected by subsolidus alteration andare the most oxidized samples (NNO þ0·94 to NNOþ1·34). Oxygen fugacity estimates using the Ridolfi et al.(2010) amphibole model are consistent with fO2 estimatesfrom the Fe^Ti oxides.The magnesiohornblende data over-lap well with the Fe^Ti oxide data, whereas tschermaki-tic-pargasite and magnesiohastingsite follow the trend ofthe more mafic magmas. At higher temperatures we ob-serve splitting into a more oxidized and a more reducedtrend, but we do not attribute this to specific magmaticprocesses (e.g. degassing, varying sources) because this

would require additional analyses to map out the potentialtrends.

ARCH ITECTURE OF THEMAGMA SYSTEMPetrological and crystal-chemical records suggest a polyba-ric evolution for the Quizapu magma system and, inparticular, the combination of petrological and geochem-ical indicators based on plagioclase and amphibole arepowerful tools with which to delineate its architecture.Amphibole pressure estimates define a range of 130^350MPa (Fig. 11). Assuming an average crustal density of2700 kgm�3 restricts the major storage magma region to�5^13 km depth.Within this depth interval the assemblageof magnesiohornblende in equilibrium with Low-Anplagioclase (type I) volumetrically dominates the magmasystem and suggests the presence of a shallow magmachamber (Figs 11 and 12) in which this crystal assemblagegrew. Pressure estimates (130^180MPa) from magnesio-hornblende restrict this magma chamber to depths of�5^7 km. Several magnesiohornblendes record occasionalpressure decreases and increases of 550MPa limited tothis depth range (Fig. 14); pressure increases require atransport mechanism that is dominated by a liquid phase(e.g. crystal settling in a mostly liquid magma and/orchamber-wide convection). These pressure excursions aremostly recorded within 100 mm of the amphibole rims,implying convective motion over an extended time inter-val, as amphibole growth rates based on CSDs may besimilar to plagioclase growth rates (Higgins & Roberge,2003) and plagioclase residence times at Quizapu are esti-mated to be thousands of years (Ruprecht & Cooper,2012). Such estimates for a convecting magma chamberwill need to be further tested given the uncertainties oncrystallization pressure using amphibole and our lack ofknowledge of amphibole crystallization kinetics.Magma from the shallow magma chamber constitutes

the bulk of the dacite magma that was erupted in 1846^1847 and 1932; most amphibole rim compositions are con-sistent with this pressure range (Fig. 14). The coexistingLow-An plagioclase is the main phenocryst phase (sensustricto) in these dacites and in the following discussion werefer to this as dacite-derived plagioclase. The long-termthermal state of this shallow chamber varies between 840and 9008C based on thermometry on magnesiohornblendefrom both eruptions and Fe^Ti oxides from the 1932eruption.The more mafic amphiboles mostly show a simple pat-

tern of high to low pressure from core to rim with noreturn to high pressure (Fig. 14). Supported by multipleamphibole crystals they define a continuous pressurerange that indicates crystal growth and differentiationover the entire depth range. The continuous pressure


827

at University of W

ashington on Decem


Dow

nloaded from


decrease towards the rims suggest that the magma plumb-ing system is organized into many tightly connected, loca-lized, small-volume melt pockets or into a highlycrystalline magmatic mush in which convective overturnis inhibited. Nonetheless, in comparison with the magne-siohornblendes only a few crystals have been measuredthat record the deeper magma evolution at Quizapu. A de-tailed delineation of potentially distinct magma lenses inaddition to the shallow storage region, or the presence ofan extended continuous crystal mush, is therefore onlyspeculative. One crystal (VQ-22 Hbl2) shows an increasein pressure during its growth at �200MPa, hinting thatadditional regions of high liquid fraction and convectiveoverturn may exist at greater depth.The tschermakitic-pargasite- and magnesiohastingsite-

bearing magmas are intermediate in their compositionwith respect to the shallow dacite magma and the rechargemagma given the estimated water contents (Lange et al.,2009) and that they are in equilibrium with transitionalplagioclase (An40^55). The inferred crystal assemblage is atestimony to the existence of such intermediate magmasat depth; however, they have not been observed ascrystal-rich erupted magmas. Evidence for crystal-richassemblages at depth comes from the presence of

glomerocrysts that are transitional in composition. Theonly erupted intermediate magmas are the hybridized an-desites that result from mixing of dacite and rechargemagma with High-An plagioclase phenocrysts. Hybridized1846^1847 magmas commonly lack transitional plagioclase(type II and V). Amphibole thermometry suggests thattschermakitic-pargasite- and magnesiohastingsite-bearingandesite temperatures reach �9708C at the base of thisquasi-continuousmagmaplumbing system.Equilibration pressures for the recharge magma are un-

known, but the thermal state and the mafic compositionof these magmas suggest an even deeper origin than thetschermakitic-pargasite- and magnesiohastingsite-bearingandesite. Their temperature (�11008C) is at the lowerlimit of primitive mantle-derived magmas that enter thecrust (Ulmer, 2001; Pichavant et al., 2002). If the rechargeandesites were stored for an extended amount of time,the high temperature suggests that they must haveresided in the lower crust to stay hot (Annen et al., 2006)before they ascended and interacted with the Quizapumagma system. The major phenocryst phase (senso stricto)of these recharge andesites is High-An plagioclase andwe refer below to these crystals as andesite-derivedplagioclase.

Fig. 14. Recorded pressure conditions after Ridolfi et al. (2010) as a function of distance from amphibole rims. Pressure estimates are averages ofall measurements within 20 mm intervals across each crystal. Uncertainties are equivalent to the error bars shown in Fig. 11, because the uncer-tainty owing to averaging is negligible (average 1SD¼ 7MPa). Dashed lines connect data points of extended crystal traverses. Other crystalswere measured only in specific zones, and data points are not connected to avoid implying a continuous pressure evolution. For crystals withlarge aspect ratios core measurements are plotted separately as the distance to rim is somewhat arbitrary depending on the exact spot location.Most amphiboles show a pressure and temperature decrease towards the rim, whereas the rims scatter tightly around a pressure estimate of150MPa (�5^6 km). Overturn and cyclical behavior is recorded in the amphibole zoning patterns and is most common between about 5 and7 km depth. It should be noted that average crustal density is assumed to be �2700 kgm�3.


828

at University of W

ashington on Decem


Dow

nloaded from


DISCUSSIONDacite homogeneity at Volca¤ n QuizapuPetrological evidence and field observations suggest the in-cremental assembly of silicic magma systems (Lipman,2007). If the Quizapu dacite magma was assembled incre-mentally by small batches of magma, the crystal- andbulk-scale homogeneity of this dacite suggests either thatthe magma generation process is continuously addingmagma of very similar composition, or that homogeneityis restored on all scales after addition of magma with di-verse compositions.Whichever process applies, it must pro-duce the narrow compositional range (An25^40; Figs 4 and5) of plagioclase in the 1846^1847 and 1932 dacites, aswell as the restricted edenitic composition of euhedralamphibole phenocrysts (Fig. 8a). Apart from the eruption-triggering hot andesite recharge, only volumetricallyminor proportions of glomerocrysts and phenocrysts(plagioclase and amphibole) with more mafic compos-itions are signs of the involvement of compositionally di-verse inputs.Magma stirring is the most effective process to attain

homogeneity. Many analog and numerical models suggestthat in crustal magma chambers in which transient re-charge or crystal-rich dense plumes initiate overturn andstirring, magma mixing is potentially incomplete and gra-dients in composition, temperature or other intensive vari-ables develop (e.g. Bergantz & Ni, 1999; Jellinek & Kerr,1999; Ruprecht et al., 2008; Huber et al., 2009). Continuedvigorous magma stirring may homogenize the melt downto the scale of a hand sample, whereas crystal zoning andother crystal-scale heterogeneities (i.e. different crystalcompositions and populations) persist over the course ofmagma mixing, as crystal dissolution is usually incomplete(e.g. Tsuchiyama, 1985) and diffusive re-equilibration istoo slow (Grove et al., 1984) to erase pre-stirring crystalcompositions. Textural and chemical heterogeneities onthe crystal scale in the form of complex zonation patternscould be indicative of episodic recharge events and conti-nuing mixing. The dacite end-member magma fromQuizapu shows little evidence for compositional gradientsas all flow lobes from the 1846^1847 eruption as wellas pumice samples throughout the 1932 eruption lack sig-nificant complex zonation patterns. Dacite-derived pheno-crysts are relatively homogeneous, which is consistentwith growth from a compositionally uniform daciticend-member magma.Despite the evidence for magma overturn documented

by amphibole zonation, such overturn may be limited andinsufficient to create the homogeneity of the Quizapu da-cites. As transient convective movement is typically toofast to be recorded in crystal zonation (Ruprecht et al.,2008), crystals tend to record the more stagnant conditionsbetween vigorous overturn events. The fact that amphi-boles record at most two episodes of changing pressure in

their zoning from core to rim (Fig. 14) suggests that anyconvection was sluggish (Marsh, 1989) and convectiveoverturn rare. Whereas convective overturn may be re-sponsible for the redistribution of crystals, it is not likelyto be the dominant control on dacitic end-member homo-geneity. Several chamber-wide overturns without anyintroduction of new heterogeneities (i.e. dense crystal-richdense plumes or low-density magma from below) wouldbe required to approach homogeneity (Bachmann &Bergantz, 2008; Huber et al., 2009).Alternatives to efficient stirring for creating homoge-

neous dacite are the addition of compositionally similarmagmas from a deeper magma processing zone (e.g.Annen et al., 2006; Straub et al., 2011), or through melt seg-regated from a subjacent andesitic mush (Brophy, 1991;Bachmann & Bergantz, 2004; Dufek & Bachmann, 2010).So-called ‘hot-zone processing’could result in wet andesiticand dacitic compositions, which may provoke crystalgrowth as the magmas degas. Magmas added throughthis mechanism would probably have textural and chem-ical heterogeneities on the crystal scale preserving evi-dence for their polybaric evolutionary paths withprogressive shallowing of crystallization. This pattern isnot observed in the Quizapu dacites. First, dacite-derivedplagioclase phenocryst zonation is minor (mostly510 Anmol %; Figs 4 and 5) and many amphibole crystals inequilibrium with the dacite-derived plagioclase show pres-sure estimates consistent with convective storage systems(41km) in the shallow crust. The tschermakitic-pargasitesand magnesiohastingsites record deeper crystallizationand some of the complexity observed in Quizapu magmamay be a result of deeper magmatic processes (Prouteau& Scaillet, 2003; Annen et al., 2006). However, amphibolecrystallization is restricted to the uppermost 13 km(5350MPa) and the bulk of the euhedral amphibolesrecord crystallization depths58 km (Fig. 14).Thus, we pro-pose that hot-zone dacite formation, and, by extension ofthe argument, also dacite formation directly from themantle (Straub et al., 2011), is not the dominant process pro-ducing dacite magma homogeneity, based on the observedcrystal-scale homogeneity and low-pressure assemblage, al-though both mechanisms may potentially be significant inother magmatic systems. In fact, the bulk (470%) of the1846^1847 and 1932 Quizapu dacite magma containsmostly weakly zoned dacite-derived plagioclase, which isinconsistent with substantial direct input of crystals fromdeep-sourced, more mafic magmas.We propose that the dacitic end-member magma

was generated by melt extraction from a predominantlyandesitic crystal mush (Fig. 15; Brophy, 1991; Bachmann &Bergantz, 2004; Hildreth, 2004; Bacon & Lowenstern,2005; Bacon & Lanphere, 2006). As the crystal mush com-pacted, melt was expelled upward and may have accumu-lated in an overlying crystal-poor magma reservoir. This


829

at University of W

ashington on Decem


Dow

nloaded from


Fig.

15.Con

ceptua

levolution

oftheQuizapu

magmasystem

.(a)

Temperature

profile

oftheplum

bing

system

andcorrespo

ndingmineral

assemblage.(b)Overalldy

namicsaredo

minated

bymeltextractionfrom

themush,

which

resultsin

aho

mogeneous

erup

tibleda

citemagma.The

nature

andgeom

etry

oftheun

derlying

andesiticmushisno

twelld

efined.C

ontinu

ousc

rystallization

over

theentire

depthrang

eha

sbeen

documented,

butitisun

likelythat

themushisfully

conn

ectedan

dcontinuo

us.R

echa

rgemagmas

form

athigh

temperature

potentially

aftersomestorage

inthelower

crust.(c)Dyn

amicsa

tthe

boun

dary

ofthemushan

dtheda

citemagmadictatetheevolutionofcomplexityon

thecrystalscale.Transitiona

lplagioclase

from

themushisconvectively

entrained.

SomeLow

-Anplagioclaseissettlin

gan

dmay

crystallize

moretran

sition

alplagioclaserimsat

high

ertemperatures.(d)Mg^

Ansystem

aticsforLow

-An(dacite-derived)plagioclase

andglom

erocryststhat

tran

sfer

from

themushinto

theda

cite.Crystalscontinue

togrow

Low

-Anplagioclasewithlow

Mgconcentrations,whereas

coresdiffu

sively

re-equ

ilibrateover

time.

(e)Tran

sition

alplagioclasethat

tran

sfersinto

themushlayerdevelops

alin

eararrayin

Mg^

Anspace.ElevatedMgconcentrations

depend

ontheexacttemperature

andpo

tentially

varying

meltcompo

sition

.Re-entrainm

entinto

theda

cite

magmamustbe

limite

d,as

nodiffu

sive

re-equ

ilibrationha

stakenplace.(f)Mg^

Ansystem

aticsforHigh-An(and

esite-derived

)plagioclase.

Extremelyrare

High-Anplagioclasewithlow

Mgconcentrations

suggests

thepresence

ofph

enocrystsfrom

previous

recharge

events

that

diffu

sively

equilib

ratedafterascent

andstoragein

moreevolvedmagma.


830

at University of W

ashington on Decem


Dow

nloaded from


extracted dacite magma is envisioned to have formed thebulk of the eruptible part of the Quizapu magma chamber.Evidence for the dynamic state of the andesite mush^

dacite magma column comes from phase relations in thedacite. The minimum temperatures (�8408C) obtainedfor wet Quizapu dacitic magmas are high compared withother dacite magmas (e.g. as low as 7708C at Mt. St.Helens: Gardner et al., 1995; Ghiorso & Evans, 2008;�760^8008C at Mt. Unzen: Venezky & Rutherford, 1999)and are consistent with relatively low crystallinity (�20vol. % crystals). The erupted dacite mineral assemblage(Low-An plagioclase and magnesiohornblende), melt com-positions, and mineral^melt equilibrium partitioning inplagioclase indicate storage conditions well above the sol-idus (Whitney, 1988) at magma temperatures of �840^9008C and �5^7 km depth (Fig. 15a, b and d). Phase rela-tions at 9008C for a more mafic magma composition atthe base of the liquid-dominated dacite magma chamberwould be consistent with 65% crystallinity for the andesitemush (Murphy et al., 2000). A quasi-rigid mush could de-velop at even higher temperatures and lower crystallinity(Marsh, 1989). The mush is the origin for transitionalplagioclase and tschermakitic-pargasite and magnesiohas-tingsite. If such an andesite mush were vertically extensiveit would thermally buffer the dacite system and limit pro-duction of the crystal-rich dacite observed in many otherarc systems (e.g. Murphy et al., 2000; Browne et al., 2006).The high temperatures above the solidus at constant bulkcomposition suggest a balance between thermal input andheat loss to the surrounding wall-rocks.The mush thereforeserves as a heat source but is too viscous to participate ex-tensively in convection of the crystal-poor dacite(Bachmann & Bergantz, 2008).Whereas a rheological bar-rier in the form of a connected mush is required for thebase of the dacite magma chamber at �7 km depth to pro-duce the dacite via melt separation from a rigid melt^crys-tal network, the vertical extent of this mush is less welldefined. The data are insufficient to answer whether themush is connected fully from 7 to 13 km under Quizapuor if distinct mush regions with ephemeral melt lensesexist in this depth range. Considering the ascent of theeruption-triggering recharge andesite, some insight maybe gained regarding the mush distribution. Upon ascentthe recharge magma either has to pass through the mushzone or if magma distribution between 7 and 13 km ismore discrete it may find an ascent path with limitedmush interaction. The fact that mafic intrusions are likelyto stall at the base of magmatic mushes (e.g. Michael,1991; Wiebe, 1996; Bachmann & Bergantz, 2008; Burgisser& Bergantz, 2011) suggests that the recharge magma didnot pass through 6 km of continuous mush. Such a path iseven less likely considering the lack of mush crystals in therecharge andesite and significant hybridization betweenrecharge and mush. Recharge magmas have only minor

potential for interaction with the mush just prior to enter-ing the major dacite magma chamber, as evidenced bythe limited presence of transitional plagioclase crystals(as documented in the 1932 recharge magmas; Fig. 5g).Nonetheless, magma storage zones must exist throughoutthis depth range to be consistent with the continuous pres-sure estimates from amphiboles.

Crystal transfer across the andesite mush^dacite interfaceThe documented entrainment of volumetrically minortransitional plagioclase with antecrystic core and pheno-crystic rims (Fig. 3i and j) and glomerocrysts (Fig. 3g) aswell as mush amphiboles (tschermakitic-pargasite andmagnesiohastingsite) provides insights into the dynamicsof the boundary region between the crystal mush anddacite magma (Fig. 15c). Such entrainment of mush crys-tals may occasionally occur in silicic systems during meltextraction and small additions of recharge magma(Claiborne et al., 2010). During melt extraction, as well asduring small amounts of recharge from the crystal mush,the melt composition (in contrast to the crystal cargo) re-mains restricted to a narrow range of dacitic to rhyodaciticcompositions (Johannes & Holtz, 1996) and is less depend-ent on the overall mush composition.The normally and reversely zoned transitional crystals

(Fig. 4; type II and V) suggest back and forth transportbetween distinct growth enivironments (varying T,pH2O, X) with intermittent resorption of the crystals. Thetrace element concentrations in plagioclase (including Feand Mg) are consistent with these growth environmentsbeing represented by an andesitic mush and a melt-richdacite magma (Fig. 15e). The linear array of transitionalplagioclase with correlated Fe and Mg concentrationswith higher An content suggest growth from a hybridmagma as Mg concentrations are elevated compared withequilibrium growth in the dacite magma (Figs 5 and 6).The fast re-equilibration of Mg in these crystals, whichwould occur within weeks to years (Costa et al., 2003;Ruprecht & Cooper, 2012) in the dacite magma, indicatesthat the entrainment of transitional plagioclase crystalsinto the dacite magma occurred recently.Glomerocrysts, type Ic plagioclase, and some amphibole

(e.g. VQ17-Hbl4, Fig. 8d) have cores that must have origi-nated from the more mafic mush, whereas the plagioclaserim compositions are identical to other dacite-derivedplagioclase crystals. Mg and Fe concentrations are consist-ent with diffusive re-equilibration of the cores (Fig. 15d),which suggests that these crystals must have been convec-tively remobilized some hundreds of years prior to erup-tion and incorporated into the dacite magma (Zellmeret al., 2003). The mush^dacite boundary is a dynamic en-vironment, potentially moving upward by crystal settlingas well as becoming eroded by convective entrainment,thereby allowing dacite-derived phenocrysts to be


831

at University of W

ashington on Decem


Dow

nloaded from


incorporated into the andesite mush. Depending on thetime that these dacite-derived crystals spend in the hotterand more mafic mush, they may partially resorb and growrims consistent with this andesitic mush. Such regrowth isusually characterized by intermediate An contents andMg concentrations, indicating that the boundary of themush is intermediate in composition and temperature.The 1846^1847 and 1932 Quizapu dacites are end-

members in terms of compositional simplicity in inter-mediate-sized (i.e. several km3) silicic magma systems.The Quizapu system may represent an end-member casein the sense of having a large, quasi-continuous mushcolumn with a homogeneous dacite magma lens, whereasother systems show more discrete zones of magma storageduring their polybaric evolution (e.g. Ridolfi et al., 2010).This is also consistent with the larger variations in mineralcompositions observed in the dacite magmas of other sys-tems (e.g. Pinatubo: Rutherford & Devine, 1996; Lassen:Clynne, 1999; Unzen: Nakada & Motomura, 1999), whichsuggest more extensive magma mixing and/or evolutionover a wider pressure range. The Quizapu dacite prior tothe 1846^1847 eruption either appears to have been wellshielded from episodic recharge events (i.e. lack of largefractions of complexly zoned crystals) or represents a re-cently developed storage system that lacks the complica-tions associated with a prolonged history of magmarecharge and magma evolution through mafic magma add-ition and fractional crystallization. Evidence for magmamixing in the 1846^1847 magmas is limited to almostsyn-eruptive processes and is discussed in the next section.

Late-stage magma mixing and thegeneration of a spectrum of magmacompositionsMagma mixing is an important process for controllingcompositional diversity in arc magmas (e.g. Gardneret al., 1995; Eichelberger et al., 2000). Late, pre-eruptivemagma mixing has been documented in many silicic sys-tems (e.g. Clynne, 1999; Scaillet & Evans, 1999; Tepleyet al., 1999; Murphy et al., 2000; Browne et al., 2006; Suzuki& Nakada, 2007; Zellmer & Turner, 2007; Kent et al.,2010).We propose for Quizapu that most of the large-scaleheterogeneity in the 1846^1847 magmas is the result of asingle episode of hot andesitic recharge that may have ori-ginated from a lower crustal storage region and passedthrough the zone dominated by the andesite mush (Fig.15b). This recharge magma, which shows some variabilityin composition and crystal cargo (Figs 2 and 5) interactedwith the dacite magma and led to the mingling andmixing present in the 1846^1847 eruption products.Evidence for a single episode of recent mixing can be ob-tained from the trace element record in theandesite-derived plagioclase. Only a few entrainedandesite-derived plagioclase crystals in the 1846^1847

eruption products show low Fe and Mg concentrations des-pite high An contents; this feature indicates that most crys-tals have not resided in the dacite magma long enough forcomplete equilibration (Fig. 5). Similar patterns exist forPb, La, and Ba; in particular, fast-diffusing Pb (Cherniak,1995) has low concentrations in the High-An plagioclase,indicating that diffusive re-equilibration with a morefelsic, cooler magmas has not occurred.The uniform character of the Quizapu dacites provides

a rare opportunity to study the effects of magma rechargeand mixing without the added complexity of extensivepast recharge records. Some magmatic systems (e.g.Soufrie' re Hills: Murphy et al., 2000; Mt. Unzen: Browneet al., 2006; Lassen Peak: Clynne, 1999) show evidence forprevious episodic recharge and this makes deciphering theeffects of a single event of mafic magma recharge on a sili-cic magma in these systems difficult. The single episode ofmagma mixing at Quizapu allows insight into the dynam-ics of magma mixing from a natural example (comple-menting our knowledge from analog and theoreticalmodels) and provides the basis for understanding the spa-tial and temporal distribution of crystal populationsduring dacite magma evolution, leading up to a singlelate-stage magma mixing event.Mingling and mixing at Quizapu have occurred on vari-

ous length scales. The local presence of both mingleddacite and hybridized andesite suggests that mixingoccurred in a transitional regime between completemixing with homogenization on the crystal scale (as aresult of similar magma viscosities for the mixingmagmas) and only partial mixing with the formation ofrigid enclaves (Sparks & Marshall, 1986). On the scale ofthe entire eruption we observe a general trend to moreevolved magmas towards the end of the 1846^1847 erup-tion, indicating incomplete large-scale mixing during over-turn and the development of chamber-wide compositionalgradients (Jellinek & Kerr, 1999). Magma chamber over-turn is driven by density inversion following intrusion ofhot recharge magma into the dacite magma system,accompanied by chilling of this hot magma against thecolder dacite magma, and vesiculation as well as crystal-lization (Bacon, 1986). This complex interplay betweencrystallinity, temperature, and degassing affects magmaviscosity, which ultimately controls magma hybridizationon all scales (Sparks & Marshall, 1986; Koyaguchi &Blake, 1991; Jellinek & Kerr, 1999).Assuming that chamber-wide gradients at Quizapu may

be approximated by a single large-scale magmatic over-turn we can use compositional variability in the differentflow lobes of the 1846^1847 eruption to constrain themixing efficiency for the Quizapu system, which reflectsthe state of the magma chamber after the overturn andprior to the eruption. The mixing efficiency of a magmasystem is the horizontally averaged vertical gradient in


832

at University of W

ashington on Decem


Dow

nloaded from


density. There can be large variations in the vertical dens-ity gradient, which is expected in time-dependent verticalchaotic mixing, yet those variations can be averaged andthe overall character of the vertical density variation canbe captured. Given the magma composition and tempera-ture, we can estimate the viscosity ratio ma/mr of the ambi-ent dacitic ma and mafic recharge magma mr during themixing episode. Dynamic simulations provide an estimatefor the Reynolds number for an overturning magma(Ruprecht et al, 2008). The simulations were performed formagmas with slightly lower viscosity than the Quizapumagmas and therefore provide an upper bound of �102

for the Reynolds number for the Quizapu mixing condi-tions.The mixing efficiency E can be obtained by adoptingthe parameterization of the experiments of Jellinek et al.(1999, fig. 10) where E is given by

E ¼

P Vmafici

Vtotalðfmafici � fmafic

mean Þ

��

��

fmaficmean

and fmafici and fmafic

mean are the fraction of mafic magma inthe ith flow lobe and the mean fraction of mafic magma,respectively. The volume of the ith flow lobe Vmafic

i andthe total volume of the 1846^1847 eruptionVtotal weigh thecontribution from the single flow lobes. Volume estimatesand an overall mass balance for the 1846^1847 eruptionhave been presented by Ruprecht & Bachmann (2010) andare used here to calculate a mixing efficiency of 0·53^0·85.Under such mixing conditions the parameterization ofJellinek et al. (1999) suggests that the viscosities of the ambi-ent and recharging magma are within one order of magni-tude of each other (10�15ma/mr5101). This is in contrast totypical magma viscosity ratios of mafic and silicic compos-itions, where4102 is a conservative estimate for the min-imum viscosity ratio of wet dacite to andesite magma(Giordano et al., 2008; Fig. 16a). A caveat in applying themodel of Jellinek et al. (1999) to the Quizapu system is thatbuoyancy flux (i.e. the mass flux from the lower boundaryof the reservoir) during recharge at Quizapu istime-dependent. Nonetheless, the model of Jellinek et al.(1999) provides the best approximation for a system ofmagma recharge, as their experiments used a constantbuoyancy flux for a restricted time.Figure 16b schematically illustrates the expected tem-

poral viscosity evolution of the system with increasingcrystallinity in the mafic magma. Melt viscosities are esti-mated using the viscosity model of Giordano et al. (2008)and measured melt compositions. The effect of crystals onthe magma viscosity is estimated using observed crystalli-nities and the model of Beckermann & Viskanta (1993).Reheating the dacite magma would lower its viscosity (nosignificant volatile exsolution is assumed for these calcula-tions, and crystal dissolution is limited), but the andesitemagma would experience a viscosity increase as it chillsagainst the dacite and starts to crystallize in response to

volatile exsolution and cooling. We would expect the vis-cosity ratio of the ambient to recharge magma to approachunity and be within the range of 10�1^101 during the earlyoverturn of the system. Thus, the overall mixing efficiencymust be established during the initiation of overturn, be-cause as continued crystallization would cause the re-charge magma to become increasingly viscous, theviscosity ratio of ambient to recharge magma would de-crease beyond 10�1. Mixing conditions in which the re-charge magma is significantly more viscous than theambient magma result in large mixing efficiencies(Jellinek et al., 1999; Fig. 16a). Such large mixing efficiencyexceeds the observed values at Quizapu, precluding pro-longed mixing.In addition to the early regulation of the overall mixing

efficiency, local hybridization is likely to occur earlyduring the overturn. A narrow range of viscosity ratiosclose to unity is required for effective hybridization, where-as large viscosity ratios lead to enclave formation (Sparks& Marshall, 1986). Thus, most of the local crystal-scale hy-bridization that follows recharge and mixing is likely tobe implemented early as well (i.e. with the onset of over-turn and mixing). This is consistent with the caveat thathybridization occurs only when the volume ratio of maficrecharge to dacite host magma is large (�0·7^1) (Sparks& Marshall, 1986). Such large contributions of andesiteare expected to be most prevalent in the beginning of themixing process, prior to the large-scale dispersal of themafic magma, and are consistent with the dominantly da-citic output towards the end of the 1846^1847 eruption.Once mafic magma is dispersed within the magma

chamber, additional small-scale mixing of mafic enclavesmay occur under solid^liquid conditions. This solid^liquiddisaggregation mechanism has been proposed for manyarc magmatic systems; for example, by Bacon & Metz(1984), Thompson & Dungan (1985) and Feeley & Dungan(1996). Those researchers argued that localized boundarylayer processes around the enclaves, rather than large-scaleoverturn, further redistribute the crystals.The andesitic re-charge magma behaves mechanically as a solid when itcrystallizes, and textural evidence suggests the slow disag-gregation of solid mafic enclaves: their cuspate contactswith the dacite host magma are snapshots of this disaggre-gation process. Moreover, the Quizapu mafic enclaveslack the gradients in crystallinity that have been observedin non-disaggregated enclaves (Browne et al., 2006).In addition to the textural and petrographic evidence for

sequential magma mixing, the crystal cargo (i.e.andesite-derived plagioclase) acts as a passive tracer ofmixing and disaggregation. The time scale of overturnand mixing is short compared with the crystal responsetime to mixing (Ruprecht et al., 2008); therefore, late stagedacite^andesite mixing is tracked by the different crystals.Andesite-derived crystals (e.g. types III and IV) are


833

at University of W

ashington on Decem


Dow

nloaded from


Fig. 16. (a) Parameterization of the magma mixing efficiency (contours) as a function of Re number and viscosity ratio of the mixing ambient(ma) and recharge (mr) magmas [modified fromJellinek et al. (1999)]. The gray field encompasses the suggested conditions of magma mixing atQuizapu inferred from the observed mixing efficiency (0·53^0·85) and ma/mr is 10

�1^101. Numerical models suggest an upper bound for the Renumber to be 102 (Ruprecht et al., 2008). As the mafic magma becomes chilled and crystallizes the mixing dynamics move towards higher (notobserved) mixing efficiencies (dashed arrow). This suggests that the overall extent of mixing at Quizapu was established early in the rechargeevent. (b) Viscosity evolution and (c) the corresponding ambient to recharge magmaviscosity ratio and their relation to the conditions for the de-velopment of textural diversity during mafic^silicic magma interaction. Both extensive local mixing (i.e. hybridization) and chamber-widemixing efficiency are set early during magma overturn.With progressing overturn ambient silicic magma to mafic recharge magma mass ratiosbecome larger and viscosity ratios decrease, resulting in mafic enclave formation and a subsequent solid^liquid disaggregation of those enclavesas the mafic recharge reaches a critical crystallinity.Viscosity estimates for the melt and the crystal^melt mixture are based on the parameteriza-tions of Giordano et al. (2008) and Beckermann & Viskanta (1993), respectively. Mafic recharge magma viscosities are calculated for two waterconcentrations (2 and 4wt %). The mafic melt composition is taken from the whole-rock composition of VQ-22A, whereas the average dacitemelt composition is taken from Ruprecht & Bachmann (2010). It should be noted that the effects of progressive degassing on the melt viscosityand potential crystal dissolution in the reheated dacite are neglected here. Degassing in the mafic recharge magmawould increase magmaviscos-ities even faster.Thus, hybrid textures and the degree of the chamber-wide mixing would develop even earlier.


834

at University of W

ashington on Decem


Dow

nloaded from


dispersed throughout the hybridized andesites, and inmingled dacites they provide evidence for the enclave dis-aggregation. Their chemical variability further suggeststhat they are gathered from different parts of the deepermagma plumbing system. A few may derive from the an-desite mush and are less mafic, whereas most are derivedfrom the decompression of the recharge magma that mayresult in sieve-textured plagioclase (Nelson & Montana,1992). Similar models of recharge and magma dynamicshave been proposed elsewhere (Kayzar et al., 2009). Weassume that most of the hybridization occurs during theoverturn close to the dacite^mush interface, because ofthe large fraction of mafic magma present, resulting in areduced effect of andesite chilling. The mingled dacites aremodified throughout the overturn, when andesitic enclavesare dispersed and further disaggregated by solid^liquidinteraction. The overturn results in the early eruption ofhybridized and mingled magmas, whereas the latermagmas become more evolved. As most of the rechargemagma left the magma system during the 1846^1847 erup-tion, we assume that the remaining magma was dacite,which was stored until a smaller recharge triggered the1932 eruption. Some precursors of minor recharge may berelated to the weak activity between 1907 and 1931, but nei-ther these nor the eruption trigger of 1932 were volumet-rically large enough to result in large-scale mixing andmingling.

Ascent processes and eruptive behaviorMicrolite growth has been shown to occur during the finalstages of magma ascent in the volcanic conduit(Geschwind & Rutherford, 1995; Hammer & Rutherford,2002; Blundy et al., 2006). However, as seen in the mingleddacites from Quizapu (and other examples: e.g. Feeley &Dungan, 1996; Humphreys et al., 2009) hybridizationduring magma mixing and mingling may result in thepresence of abundant microlites in the dacite magma thatis independent of late-stage decompression and eruption.Mafic enclaves contain abundant microlites as a conse-quence of the chilling of the andesite recharge magmaagainst the colder dacite magma. Up to 70% of theQuizapu recharge magma is composed of microlites(Ruprecht & Bachmann, 2010). Decompression would fur-ther enhance microlite growth, but a large fraction of themicrolite population was already present in the magmaplumbing system during solid^liquid disaggregation. Asmagma chemistry changes significantly for microlitesgrowing in the mafic recharge magma or in the dacitemagma, the microlite compositions provide a means to dis-tinguish between recharge-derived and late-stagedecompression-derived microlites.The observation of apparent reheating as a consequence

of decompression and crystallization proposed for arc an-desite volcanoes (Blundy et al., 2006) may in some casesrepresent hybridization at depth that is independent of

magma ascent. Blundy et al. (2006) argued that the corre-lated increase in temperature and crystallinity withdecreasing pH2O at Mt. St. Helens and Shiveluch is adirect consequence of crystal growth during magmaascent and not a result of magma mixing. The samplesexamined by Blundy et al. (2006) lack macroscopic evi-dence of magma mixing. In contrast, magma mixing isdocumented on both the macro- and micro-scale atQuizapu. Whereas phenocryst crystallinities in mingleddacites and hybridized andesites remain similar at �20%,solid^liquid disaggregation of enclaves results in the sameobservations as those presented by Blundy et al. (2006).Overall crystallinities (phenocrysts and microlites) in themixed dacites increase owing to magma mixing and theaddition of microlite-rich recharge magma. Magma tem-perature also increases significantly during mingling andmixing (Fig. 13). Although magma chamber pH2O has notbeen determined for Quizapu, the low H2O solubility(�2^3wt %) in hot recharge andesites (11008C) withplagioclase �An70^80 phenocrysts (Lange et al., 2009) com-pared with amphibole-bearing dacites (�3·5^4wt %) sug-gests that with continuing magma mixing the pH2O ofthe mixed magmas would decrease by mass balance.Therefore, magma mixing is likely to result in the samecharacteristic effects as in the decompression^reheatingmodel of Blundy et al. (2006). If hybridization is completeand macroscopic evidence for magma mixing is missing,magma recharge with mixing and reheating and conduitreheating during decompression may be indistinguishablewith respect to low pH2O and high crystallinity at hightemperatures. At Quizapu volcano magma mixing is docu-mented macroscopically. The phenocryst compositions,with calcic plagioclase microlite cores (An60), additionallyconfirm an origin from the mafic recharge magma.There is ample evidence that hot mafic recharge

magmas can provide large amounts of energy to reheat sili-cic andesite and dacite magmas in shallow crustal magmachambers. Mafic enclaves are present in many arc magmasystems (Murphy et al., 1998; Costa & Singer, 2002; Holtzet al., 2005; Zellmer & Turner, 2007), demonstrating thatlate-stage mingling and mixing is a common process inarc magmas. Magma systems for which extensive tempera-ture information exists for distinct but associated magmasshow large temperature differences (�100^2008C) betweencooler host magmas and hot recharge magmas (e.g.Montserrat: Murphy et al., 1998; Unzen: Holtz et al., 2005).Reheating is typically observed in the host dacitemagmas. Reheating has a strong effect on the viscosity ofthe dacite magma (e.g. Hess & Dingwell, 1996; Giordanoet al., 2008; Ruprecht & Bachmann, 2010) and on the dissi-pation of frictional stresses in the conduit.The simple Quizapu system with two eruptions that

manifest contrasting behavior provides evidence that re-heating of dacite magma drives the system into effusive


835

at University of W

ashington on Decem


Dow

nloaded from


eruptive behavior (Ruprecht & Bachmann, 2010). In otherarc systems, a similar temperature increase may also leadto a reduced explosive potential. Although higher magmacrystallinitiesçand, as a consequence, higher magma visc-ositiesçcharacterize these systems (e.g. Montserrat,:Murphy et al., 1998; Unzen: Holtz et al., 2005), reheatingby recharge magmas may partially offset the higher viscos-ity owing to higher crystallinity. As shown in the previoussection, and by Humphreys et al. (2009), microlites can besubstantially derived from the recharge magmas and mayonly subordinately result from decompression-induced nu-cleation and growth. Dacite magmas that are crystal-richwhen they reach the surface may initially have had muchlower crystallinities and lower viscosity, and thereforemay have undergone passive degassing early in theirascent history. Even though magma viscosity is higher formagmas with high crystallinities, bubble movementthrough the melt is enhanced owing to the lower melt vis-cosity as a result of reheating (Ruprecht & Bachmann,2010).

CONCLUSIONSCrystal zoning and geothermobarometry suggest that theQuizapu magma system was chemically and thermallyzoned. This zonation is rather abrupt, with a hot andesiticmush (recorded by the recharge magmas and their crystalcargoes) overlain by a homogeneous, cooler dacitemagma with limited complexity on the crystal scale. Incontrast to many other intermediate-sized magma systems,the plumbing system of Volca¤ n Quizapu is relativelysimple, as corroborated by the relatively simple crystal tex-tural populations. The Quizapu dacite magma chamber isa result of mineral^melt segregation, as this is the onlyprocess that can produce homogeneous magmas on bothwhole-rock and crystal scales. Amphibole geobarometrysuggests that the eruptible dacite magma was stored at�5^7 km depth and kept at relatively low crystallinityowing to heating by the underlying hot andesitic mushthat was itself buffered by latent heat addition.The amphi-bole record that suggests kilometer-scale sluggish overturnin the dacite magma chamber supports the lowcrystallinity.The limited diversity of plagioclase and amphibole

phenocrysts in the largely homogeneous Quizapu dacitessuggests that back-mixing at the andesite mush^dacitemagma interface is minimal. However, the presence of afew crystals that record growth in the homogeneousdacite and the andesite mush precludes the presence of spa-tially separated bodies.We suggest that the clearly definedmush^dacite interface is significantly disrupted onlyduring volumetrically significant recharge events. Such anevent occurred prior to the 1846^1847 Quizapu eruption.As a result of limited andesite-derived contributions to thedacite magma, the episode of late-stage magma recharge

and mixing prior to the 1846^1847 eruption provides anopportunity to unravel the local and chamber-widemixing processes. Integrating observations from the 1846^1847 eruption with analog and theoretical models formagma mixing supports the notion that the extent ofchamber-wide magma mixing is established early (i.e.during the chilling and crystallization of the rechargingmafic magma). The crystal-scale extent of mixing is estab-lished and the window for complete magma hybridizationis passed early. Once this opportunity for magma mixingis lost, continued mafic^silicic magma interaction is lim-ited to solid^liquid disaggregation.Whereas the 1846^1847 eruption is characterized by

magmamixing, the explosive1932 eruption shows very lim-ited contributions from mafic recharge, suggesting that (1)only small volumes of recharge magma remained at depth(although they were potentially the cause for the intermit-tentminor activitybetween1907 and1931), or (2) all of the re-charge magma that triggered the 1846^1847 eruption hadleft the magma plumbing system during that eruption.Nonetheless, volumetrically minor mafic scoria depositsassociated with the 1932 eruption suggest that a new smallbatch of rechargemagmawas the trigger for this event.The data presented here provide clues as to how compos-

itionally similar dacite magmas erupted first effusively in1846^1847, followed by a plinian eruption in 1932. It hasbeen proposed elsewhere (Eichelberger & Izbekov, 2000;Maksimov, 2008) that such explosive^effusive transitionsmay be the result of silicic recharge into a more mafic shal-low magma system; however, the magmatic architectureof Quizapu, derived from crystal zoning and geobarome-try, supports a magma system with more silicic compos-itions at shallower depths. The recently proposedmechanism of mafic recharge-induced reheating of silicicmagma and subsequent magma viscosity reduction leadingto enhanced degassing is more consistent with the petrog-raphy and petrology of the Quizapu magmas (Ruprecht& Bachmann, 2010).Finally, we conclude that the characteristic signatures of

decompression-related reheating are probably very similarto the signatures recorded during reheating as a result ofmagma mixing and hybridization. In the case of completehybridization, the primary indicator for distinguishing be-tween decompression^reheating and recharge^reheatingis the composition of microlites that formed late in themagma’s evolution. Calcic microlites are likely to beandesite-derived and form during chilling of maficmagma against cooler silicic magma. As the magma con-tinues to crystallize and degas the microlites may becomemore sodic. Microlites that are sodic in composition formas a result of decompression and degassing in an otherwiseclosed system. The presence of calcic microlites may begenerally indicative of reheating as a result of maficrecharge.


836

at University of W

ashington on Decem


Dow

nloaded from


ACKNOWLEDGEMENTSConstructive reviews byJames Brophy, Mike Dungan, JakeLowenstern, Shan de Silva, and HeatherWright are muchappreciated.We thank Olivier Bachmann, Carrie Brugger,Mike Dungan, Jose Antonio Naranjo, and Pablo Salas forsupport in the field and many constructive discussions onsilicic magma systems. Adam Kent and Scott Kuehnerprovided important help with LA-ICP-MS and electronmicroprobe, and Dougal Jerram shared his insights oncrystal size distributions. This is Lamont^Doherty EarthObservatory contribution 7518.

FUNDINGThe project was funded by National Science Foundationgrants EAR 0440391 (G.W.B.), EAR 0711551 (G.W.B.), andEAR 0711354 (K.M.C.).

SUPPLEMENTARY DATASupplementary data for this paper are available at Journalof Petrology online.

REFERENCESAdams, N., de Silva, S., Self, S., Salas, G., Schubring, S., Permenter, J.

& Arbesman, K. (2001). The physical volcanology of the 1600 erup-tion of Huaynaputina, southern Peru. Bulletin of Volcanology 62,493^518.

Annen, C., Blundy, J. D. & Sparks, R. S. J. (2006).The genesis of inter-mediate and silicic magmas in deep crustal hot zones. Journal ofPetrology 47, 505^539.

Bachmann, O. & Bergantz, G.W. (2004). On the origin of crystal-poorrhyolites: extracted from batholitic crystal mushes. Journal of

Petrology 45, 1565^1582.Bachmann, O. & Bergantz, G.W. (2008). Deciphering magma cham-

ber dynamics from styles of compositional zoning in large silicicash flow sheets. In: Putirka, K. D. & Tepley, F. J., III (eds)Minerals, Inclusions and Volcanic Processes. Mineralogical Society of

America and Geochemical Society, Reviews in Mineralogy and Geochemistry

69, 651^674.Bachmann, O. & Dungan, M. A. (2002). Temperature-induced

Al-zoning in hornblendes of the Fish Canyon magma, Colorado.American Mineralogist 87, 1062^1076.

Bachmann, O., Dungan, M. A. & Lipman, P. W. (2002). The FishCanyon magma body, San JuanVolcanic Field, Colorado: rejuven-ation and eruption of an upper-crustal batholith. Journal of

Petrology 43, 1469^1503.Bacon, C. R. (1986). Magmatic inclusions in silicic and intermediate

volcanic rocks. Journal of Geophysical Research 91, 6091^6112.Bacon, C. R. & Druitt, T. H. (1988). Compositional evolution of the

zoned calcalkaline magma chamber of Mount Mazama, CraterLake, Oregon. Contributions to Mineralogy and Petrology 98, 224^256.

Bacon, C. R. & Lanphere, M. A. (2006). Eruptive history and geo-chronology of Mount Mazama and the Crater Lake region,Oregon. Geological Society of America Bulletin 118, 1331^1359.

Bacon, C. R. & Lowenstern, J. B. (2005). Late Pleistocene granodioritesource for recycled zircon and phenocrysts in rhyodacite lava atCrater Lake, Oregon. Earth and Planetary Science Letters 233,277^293.

Bacon, C. R. & Metz, J. (1984). Magmatic inclusions in rhyolites, con-taminated basalts, and compositional zonation beneath the Cosovolcanic field, California. Contributions to Mineralogy and Petrology 85,346^365.

Baedecker, P. A. (1987). Methods for geochemical analysis. US Geological

Survey Bulletin 1770, , 184 p.Beckermann, C. & Viskanta, R. (1993). Mathematical modeling of

transport phenomena during alloy solidification. Applied Mechanics

Reviews 46, 1^27.Be¤ dard, J. H. (2006). Trace element partitioning in plagioclase feld-

spar. Geochimica et Cosmochimica Acta 70, 3717^3742.Bergantz, G.W. & Ni, J. (1999). A numerical study of sedimentation by

dripping instabilities in viscous fluids. International Journal of

Multiphase Flow 25, 307^320.Bindeman, I. N., Davis, A. M. & Drake, M. J. (1998). Ion microprobe

study of plagioclase^basalt partition experiments at natural con-centration levels of trace elements. Geochimica et Cosmochimica Acta

62, 1175^1193.Blundy, J. & Cashman, K. (2008). Petrologic reconstruction of mag-

matic system variables and processes. In: Putirka, K. D. &Tepley, F. J., III (eds) Minerals, Inclusions and Volcanic Processes.

Mineralogical Society of America and Geochemical Society, Reviews in

Mineralogy and Geochemistry 69, 179^239.Blundy, J., Cashman, K. & Humphreys, M. (2006). Magma heating

by decompression-driven crystallization beneath andesite volca-noes. Nature 443, 76^80.

Brophy, J. G. (1991). Composition gaps, critical crystallinity, and frac-tional crystallization in orogenic (calc-alkaline) magmatic systems.Contributions to Mineralogy and Petrology 109, 173^182.

Browne, B. L., Eichelberger, J. C., Patino, L. C.,Vogel,T. A., Dehn, J.,Uto, K. & Hoshizumi, H. (2006). Generation of porphyritic andequigranular mafic enclaves during magma recharge events atUnzenVolcano, Japan. Journal of Petrology 47, 301^328.

Burgisser, A. & Bergantz, G.W. (2011). A rapid mechanism to remobi-lize and homogenize highly crystalline magma bodies. Nature 471,212^215.

Cashman, K. V. & Marsh, B. D. (1988). Crystal size distribution(CSD) in rocks and the kinetics and dynamics of crystallization;II. Makaopuhi lava lake. Contributions to Mineralogy and Petrology 99,292^305.

Castro, J. M. & Dingwell, D. B. (2009). Rapid ascent of rhyoliticmagma at Chaiten volcano, Chile. Nature 461, 780^783.

Charlier, B. L. A., Wilson, C. J. N., Lowenstern, J. B., Blake, S., vanCalsteren, P. W. & Davidson, J. P. (2005). Magma generation at alarge, hyperactive silicic volcano (Taupo, New Zealand) revealedby U^Th and U^Pb systematics in zircons. Journal of Petrology 46,3^32.

Cherniak, D. J. (1995). Diffusion of lead in plagioclase and K-feldspar:an investigation using Rutherford backscattering and resonant nu-clear reaction analysis. Contributions to Mineralogy and Petrology 120,358^371.

Cherniak, D. J. (2002). Ba diffusion in feldspar. Geochimica et

Cosmochimica Acta 66, 1641^1650.Cherniak, D. (2003). REE diffusion in feldspar. Chemical Geology 193,

25^41.Claiborne, L. L., Miller, C. F., Flanagan, D. M., Clynne, M. A. &

Wooden, J. L. (2010). Zircon reveals protracted magma storageand recycling beneath Mount St. Helens. Geology 38, 1011^1014.

Clynne,M.A. (1999).Acomplexmagmamixingorigin for rocks eruptedin1915, Lassen Peak, California. Journal of Petrology 40,105^132.

Costa, F. & Singer, B. (2002). Evolution of Holocene dacite and com-positionally zoned magma, Volcan San Pedro, Southern VolcanicZone, Chile. Journal of Petrology 43, 1571^1593.


837

at University of W

ashington on Decem


Dow

nloaded from


Costa, F., Chakraborty, S. & Dohmen, R. (2003). Diffusioncoupling between trace and major elements and a model for calcu-lation of magma residence times using plagioclase. Geochimica et

Cosmochimica Acta 67, 2189^2200.Costa, F., Scaillet, B. & Pichavant, M. (2004). Petrological and experi-

mental constraints on the pre-eruption conditions of Holocenedacite fromVolcan San Pedro (368S, Chilean Andes) and the im-portance of sulphur in silicic subduction-related magmas. Journalof Petrology 45, 855^881.

Davidson, J., Morgan, D., Charlier, B., Harlou, R. & Hora, J. (2007).Microsampling and isotopic analysis of igneous rocks:Implications for the study of magmatic systems. Annual Review of

Earth and Planetary Sciences 35, 273^311.DePaolo, D. J. (1981). Trace element and isotopic effects of combined

wallrock assimilation and fractional crystallization. Earth and

Planetary Science Letters 53, 189^202.de Silva, S., Salas, G. & Schubring, S. (2008). Triggering explosive

eruptionsçThe case for silicic magma recharge at Huaynaputina,southern Peru. Geology 36, 387^390.

D’Lemos, R. S. (1996). Mixing between granitic and dioritic crystalmushes, Guernsey, Channel Islands, UK. Lithos 38, 233^257.

Domeyko, I. (1850). Viaje a las cordilleras de Talca i de Chilla¤ n:Segunda y Tercera Parte. In Anales de la Universidad de Chile, Vol. 7,Santiago: Del Pac|¤ fico, pp. 47^74.

Druitt, T. & Bacon, C. (1989). Petrology of the zoned calcalkalinemagma chamber of Mount Mazama, Crater Lake, Oregon.Contributions to Mineralogy and Petrology 101, 245^259.

Dufek, J. & Bachmann, O. (2010). Quantum magmatism: Magmaticcompositional gaps generated by melt^crystal dynamics. Geology

38, 687^690.Dungan, M. A. (1987). Open System Magmatic Evolution of the Taos

Plateau Volcanic Field, Northern New Mexico: II. The Genesis ofCryptic Hybrids. Journal of Petrology 28, 955^977.

Eichelberger, J. C. & Izbekov, P. E. (2000). Eruption of andesite trig-gered by dyke injection: contrasting cases at Karymsky Volcano,Kamchatka and Mt Katmai, Alaska. Royal Society of London

PhilosophicalTransactions, Series A 358, 1465^1485.Eichelberger, J. C., Chertkoff, D., Dreher, S. T. & Nye, C. J. (2000).

Magmas in collision: Rethinking chemical zonation in silicicmagmas. Geology 28, 603^606.

Feeley, T. C. & Dungan, M. A. (1996). Compositional and dynamiccontrols on mafic^silicic magma interactions at continental arc vol-canoes: evidence from Cordo¤ n El Guadal, Tatara^San PedroComplex, Chile. Journal of Petrology 37, 1547^1577.

Freer, R. & Hauptman, Z. (1978). Experimental study of magnetite^titanomagnetite interdiffusion. Physics of the Earth and Planetary

Interiors 16, 223^231.Garc|¤a-Casco, A., Torres-Rolda¤ n, R. L., Milla¤ n, G., Monie¤ , P. &

Schneider, J. (2002). Oscillatory zoning in eclogitic garnet andamphibole, Northern Serpentinite Melange, Cuba: a record of tec-tonic instability during subduction? Journal of Metamorphic Geology

20, 581^598.Gardner, J. E., Carey, S., Rutherford, M. J. & Sigurdsson, H. (1995).

Petrologic diversity in Mount St. Helens dacites during the last4,000 yearsçimplications for magma mixing. Contributions to

Mineralogy and Petrology 119, 224^238.Geschwind, C. & Rutherford, M. J. (1995). Crystallization of micro-

lites during magma ascent: the fluid mechanics of 1980^1986 erup-tions at Mount St Helens. Bulletin of Volcanology 57, 356^370.

Ghiorso, M. S. & Evans, B.W. (2008).Thermodynamics of rhombohe-dral oxide solid solutions and a revision of the Fe^Ti two-oxidegeothermometer and oxygen barometer. American Journal of Science308, 957^1039.

Giletti, B. & Shanahan, T. (1997). Alkali diffusion in plagioclase feld-spar. Chemical Geology 139, 3^20.

Ginibre, C.,Wo« rner, G. & Kronz, A. (2002). Minor- and trace-elementzoning in plagioclase: implications for magma chamber processesat Parinacota Volcano, northern Chile. Contributions to Mineralogy

and Petrology 143, 300^315.Giordano, D., Russell, J. K. & Dingwell, D. B. (2008). Viscosity of

magmatic liquids: A model. Earth and Planetary Science Letters 271,123^134.

Grove, T. L., Baker, M. B. & Kinzler, R. J. (1984). Coupled CaAl^NaSi diffusion in plagioclase feldspar: Experiments and applica-tions to cooling rate speedometry. Geochimica et Cosmochimica Acta

48, 2113^2121.Grunder, A. L. & Mahood, G. A. (1988). Physical and chemical

models of zoned silicic magmas: the Loma Seca Tuffand Calabozos Caldera, southern Andes. Journal of Petrology 29,831^867.

Hammer, J. E. & Rutherford, M. J. (2002). An experimental study ofdecompression-induced crystallization in silicic melt. Journal of

Geophysical Research 107, 2021, 24pp., doi:10.1029/2001JB000281.Hattori, K. & Sato, H. (1996). Magma evolution recorded in plagio-

clase zoning in 1991 Pinatubo eruption products. American

Mineralogist 81, 982^994.Hess, K. U. & Dingwell, D. B. (1996).Viscosities of hydrous leucogra-

nitic melts: A non-Arrhenian model. American Mineralogist 81,1297^1300.

Higgins, M. D. (2000). Measurement of crystal size distributions.American Mineralogist 85, 1105^1116.

Higgins, M. D. & Roberge, J. (2003). Crystal size distributionof plagioclase and amphibole from Soufrie' re Hills Volcano,Montserrat: evidence for dynamic crystallization^textural coarsen-ing cycles. Journal of Petrology 44, 1401^1411.

Hildreth,W. (1981). Gradients in silicic magma chambers: implicationsfor lithospheric magmatism. Journal of Geophysical Research 86,10153^10192.

Hildreth, W. (2004). Volcanological perspectives on Long Valley,Mammoth Mountain, and Mono Craters: several contiguous butdiscrete systems. Journal of Volcanology and Geothermal Research 136,169^198.

Hildreth,W. & Drake, R. E. (1992).Volcan Quizapu, Chilean Andes.Bulletin of Volcanology 54, 93^125.

Hildreth,W. & Moorbath, S. (1988). Crustal contributions to arc mag-matism in the Andes of Central Chile. Contributions to Mineralogy

and Petrology 98, 455^489.Hildreth, W., Grunder, A. L. & Drake, R. E. (1984). The Loma Seca

Tuff and the Calabozos Caldera; a major ash-flow and caldera com-plex in the Southern Andes of central Chile. Geological Society of

America Bulletin 95, 45^54.Holland, T. & Blundy, J. (1994). Non-ideal interactions in calcic

amphiboles and their bearing on amphibole^plagioclase thermom-etry. Contributions to Mineralogy and Petrology 116, 433^447.

Holness, M. B., Nielsen, T. F. D. & Tegner, C. (2007). Textural matur-ity of cumulates: a record of chamber filling, liquidus assemblage,cooling rate and large-scale convection in mafic layered intrusions.Journal of Petrology 48, 141^157.

Holtz, F., Sato, H., Lewis, J., Behrens, H. & Nakada, S. (2005).Experimental petrology of the 1991^1995 Unzen dacite, Japan.Part I: Phase relations, phase composition and pre-eruptive condi-tions. Journal of Petrology 46, 319^337.

Huber, C., Bachmann, O. & Manga, M. (2009). Homogenizationprocesses in silicic magma chambers by stirring andmushification (latent heat buffering). Earth and Planetary Science

Letters 283, 38^47.


838

at University of W

ashington on Decem


Dow

nloaded from


Humphreys, M. C. S., Blundy, J. D. & Sparks, R. S. J. (2006). Magmaevolution and open-system processes at ShiveluchVolcano: Insightsfrom phenocryst zoning. Journal of Petrology 47, 2303^2334.

Humphreys, M., Christopher, T. & Hards, V. (2009). Microlite trans-fer by disaggregation of mafic inclusions following magma mixingat Soufrie' re Hills volcano, Montserrat. Contributions to Mineralogy

and Petrology 157, 609^624.Jellinek, A. M. & Kerr, R. C. (1999). Mixing and compositional strati-

fication produced by natural convection: 2. Applications to the dif-ferentiation of basaltic and silicic magma chambers and komatiitelava flows. Journal of Geophysical Research 104, 7203^7218.

Jellinek, A. M., Kerr, R. C. & Griffiths, R. W. (1999). Mixing andcompositional stratification produced by natural convection; 1.Experiments and their application to Earth’s core and mantle.Journal of Geophysical Research 104, 7183^7201.

Johannes,W. & Holtz, F. (1996). Petrogenesis and Experimental Petrology ofGranitic Rocks. NewYork: Springer.

Johnson, D. M., Hooper, P. R. & Conrey, R. M. (1999). XRFanalysisof rocks and minerals for major and trace elements on a single lowdilution Li-tetraborate fused bead. Advances in X-ray Analysis 41,843^867.

Kayzar, T. M., Cooper, K. M., Reagan, M. K. & Kent, A. J. (2009).Gas transport model for the magmatic system at Mount Pinatubo,Philippines: Insights from (210Pb)/(226Ra). Journal of Volcanology andGeothermal Research 181, 124^140.

Kent, A. J. R., Darr, C., Koleszar, A. M., Salisbury, M. J. &Cooper, K. M. (2010). Preferential eruption of andesitic magmasthrough recharge filtering. Nature Geoscience 3, 631^636.

Kent, A. J. R., Stolper, E. M., Francis, D., Woodhead, J., Frei, R. &Eiler, J. (2004). Mantle heterogeneity during the formation ofthe North Atlantic Igneous Province: Constraints from traceelement and Sr^NdÔsÔ isotope systematics of Baffin Island pic-rites. Geochemistry, Geophysics, Geosystems 5, Q11004, doi:10.1029/2004GC000743.

Kent, A. J., Blundy, J., Cashman, K.V., Cooper, K. M., Donnelly, C.,Pallister, J. S., Reagan, M., Rowe, M. C. & Thornber, C. R.(2007). Vapor transfer prior to the October 2004 eruption ofMount St. Helens,Washington. Geology 35, 231^234.

Koyaguchi,T. & Blake, S. (1991). Origin of mafic enclaves: constraintson the magma mixing model from fluid dynamic experiments. In:Dider, J. & Barbarin, B. (eds) Enclaves in Granite Petrology.Amsterdam: Elsevier, pp. 415^429.

Lange, R. A., Frey, H. M. & Hector, J. (2009). A thermodynamicmodel for the plagioclase^liquid hygrometer/thermometer.American Mineralogist 94, 494^506.

Lara, L. (2009). The 2008 eruption of the Chaite¤ n Volcano, Chile: apreliminary report. Andean Geology 36, 125^129.

Leake, B. E.,Woolley, A. R. et al. (1997). Nomenclature of amphiboles;report of the subcommittee on amphiboles of the InternationalMineralogical Association, Commission on New Minerals andMineral Names. Canadian Mineralogist 35, 219^246.

Lipman, P.W. (2007). Incremental assembly and prolonged consolida-tion of Cordilleran magma chambers: Evidence from theSouthern Rocky Mountain volcanic Eeld. Geosphere 3, 42^70.

Longhi, J. & Bertka, C. M. (1996). Graphical analysis of pigeoniteâugite liquidus equilibria. American Mineralogist 81, 685^695.

Longhi, J., Walker, D. & Hays, J. F. (1976). Fe and Mg in plagioclase.Proceedings of the 7th Lunar Science Conference, held in Houston TX,15^19 March 1976, pp. 1281^1300.

Maksimov, A. (2008). A physicochemical model for deep degassing ofwater-rich magma. Journal of Volcanology and Seismology 2, 356^363.

Mandeville, C., Carey, S. & Sigurdsson, H. (1996). Magma mixing,fractional crystallization and volatile degassing during the 1883

eruption of Krakatau volcano, Indonesia. Journal of Volcanology andGeothermal Research 74, 243^274.

Marsh, B. D. (1988). Crystal size distribution (CSD) in rocks and thekinetics and dynamics of crystallization. Contributions to Mineralogy

and Petrology 99, 277^291.Marsh, B. D. (1989). Magma chambers. Annual Review of Earth and

Planetary Sciences 17, 439^472.Marsh, B. D. (1998). On the interpretation of crystal size distributions

in magmatic systems. Journal of Petrology 39, 553^599.McKenzie, D. (1985). The extraction of magma from the crust and

mantle. Earth and Planetary Science Letters 74, 81^91.Michael, P. J. (1991). Intrusion of basaltic magma into a crystallizing

granitic magma chamber: The Cordillera del Paine pluton insouthern Chile. Contributions to Mineralogy and Petrology 108, 396^418.

Miller, J. S., Matzel, J. E. P., Miller, C. F., Burgess, S. D. & Miller, R.B. (2007). Zircon growth and recycling during the assembly oflarge, composite arc plutons. Journal of Volcanology and Geothermal

Research 167, 282^299.Morgan, D. J. & Jerram, D. A. (2006). On estimating crystal shape for

crystal size distribution analysis. Journal of Volcanology and

Geothermal Research 154, 1^7.Murphy, M., Sparks, R., Barclay, J., Carroll, M., Lejeune, A.,

Brewer, T., Macdonald, R., Black, S. & Young, S. (1998). The roleof magma mixing in triggering the current eruption at theSoufrie' re Hills volcano, Montserrat, West Indies. Geophysical

Research Letters 25, 3433^3436.Murphy, M. D., Sparks, R. S. J., Barclay, J., Carroll, M. R. &

Brewer, T. S. (2000). Remobilization of andesite magma by intru-sion of mafic magma at the Soufrie' re Hills volcano, Montserrat,West Indies. Journal of Petrology 41, 21^42.

Nakada, S. & Motomura, Y. (1999). Petrology of the 1991^1995 erup-tion at Unzen: effusion pulsation and groundmass crystallization.Journal of Volcanology and Geothermal Research 89, 173^196.

Nelson, S. T. & Montana, A. (1992). Sieve-textured plagioclasein volcanic rocks produced by rapid decompression. American

Mineralogist 77, 1242^1249.Oldenburg, C. M., Spera, F. J., Yuen, D. A. & Sewell, G. (1989).

Dynamic mixing in magma bodies: theory, simulations and impli-cations. Journal of Geophysical Research 94, 9215^9236.

Pichavant, M., Mysen, B. O. & Macdonald, R. (2002). Source andH2O content of high-MgO magmas in island arc settings: anexperimental study of a primitive calc-alkaline basalt fromSt Vincent, Lesser Antilles arc. Geochimica et Cosmochimica Acta 66,2193^2209.

Prouteau, G. & Scaillet, B. (2003). Experimental constraints onthe origin of the 1991 Pinatubo dacite. Journal of Petrology 44,2203^2241.

Putirka, K. D. (2008). Thermometers and barometers for volcanicsystems. In: Putirka, K. D. & Tepley, F. J., III (eds) Minerals,

Inclusions and Volcanic Processes. Mineralogical Society of America

and Geochemical Society, Reviews in Mineralogy and Geochemistry 69,61^120.

Ridolfi, F., Renzulli, A. & Puerini, M. (2010). Stability and chem-ical equilibrium of amphibole in calc-alkaline magmas: an over-view, new thermobarometric formulations and application tosubduction-related volcanoes. Contributions to Mineralogy and Petrology

160, 45^66.Ruprecht, P. (2009). The time and length scales of magma mixing:

the natural laboratory of Volca¤ n Quizapu, Chile, PhD thesis,University of Washington, Seattle.

Ruprecht, P. & Bachmann, O. (2010). Pre-eruptive reheating duringmagmamixing at Quizapu volcano and the implications for the ex-plosiveness of silicic arc volcanoes. Geology 38, 919^922.


839

at University of W

ashington on Decem


Dow

nloaded from


Ruprecht, P. & Cooper, K. M. (2012). Integrating the uranium-seriesand elemental diffusion geochronometers in mixed magmas fromVolca¤ n Quizapu, Central Chile. Journal of Petrology 53, 000^000.

Ruprecht, P. & Wo« rner, G. (2007).Variable regimes in magma systemsdocumented in plagioclase zoning patterns: El Misti stratovolcanoand Andahua monogenetic cones. Journal of Volcanology and

Geothermal Research 165, 142^162.Ruprecht, P., Bergantz, G. W. & Dufek, J. (2008). Modeling of

gas-driven magmatic overturn: Tracking of phenocryst dispersaland gathering during magma mixing. Geochemistry, Geophysics,

Geosystems 9, Q07017.Rutherford, M. & Devine, J. (2003). Magmatic conditions and

magma ascent as indicated by hornblende phase equilibria andreactions in the 1995^2002 Soufrie' re Hills magma. Journal of

Petrology 44, 1433^1454.Rutherford, M. J. & Hill, P. M. (1993). Magma ascent rates from

amphibole breakdown: an experimental study applied to the 1980^1986 Mount St. Helens eruptions. Journal of Geophysical Research 98,19667^19685.

Rutherford, M. J., Sigurdsson, H., Carey, S. & Davis, A. (1985). TheMay 18, 1980, eruption of Mount St. Helens 1. Melt compositionand experimental phase equilibria. Journal of Geophysical Research

90, 2929^2947.Salisbury, M. J., Bohrson, W. A., Clynne, M. A., Ramos, F. C. &

Hoskin, P. (2008). Multiple plagioclase crystal populations identi-fied by crystal size distribution and in situ chemical data: implica-tions for timescales of magma chamber processes associated withthe 1915 eruption of Lassen Peak, CA. Journal of Petrology 49,1755^1780.

Scaillet, B. & Evans, B. (1999). The 15 June 1991 eruption ofMount Pinatubo. I. Phase equilibria and pre-eruption P^T^fO2^fH2O conditions of the dacite magma. Journal of Petrology 40,381^411.

Sparks, R. & Marshall, L. (1986). Thermal and mechanical con-straints on mixing between mafic and silicic magmas. Journal ofVolcanology and Geothermal Research 29, 99^124.

Sparks, S. R. J., Sigurdsson, H. & Wilson, L. (1977). Magma mixing: amechanism for triggering acid explosive eruptions. Nature 267,315^318.

Straub, S. M., Gomez-Tuena, A., Stuart, F. M., Zellmer, G. F.,Espinasa-Perena, R., Cai, Y. & Iizuka, Y. (2011). Formation ofhybrid arc andesites beneath thick continental crust. Earth and

Planetary Science Letters 303, 337^347.Streck, M. (2008). Mineral textures and zoning as evidence for open

system processes. In: Putirka, K. D. & Tepley, F. J., III (eds)Minerals, Inclusions and Volcanic Processes. Mineralogical Society of

America and Geochemical Society, Reviews in Mineralogy and Geochemistry

69, 595^622.Sugawara, T. (2001). Ferric iron partitioning between plagioclase and

silicate liquid: thermodynamics and petrological applications.Contributions to Mineralogy and Petrology 141, 659^686.

Suzuki, Y. & Nakada, S. (2007). Remobilization of highly crystallinefelsic magma by injection of mafic magma: constraints from themiddle sixth century eruption at HarunaVolcano, Honshu, Japan.Journal of Petrology 48, 1543^1567.

Tepley, F. J., III, Davidson, J. P. & Clynne, M. A. (1999). Magmaticinteractions as recorded in plagioclase phenocrysts of ChaosCrags, Lassen volcanic center, California. Journal of Petrology 40,787^806.

Thompson, R. A. & Dungan, M. A. (1985). The Petrology andGeochemistry of the Handkerchief Mesa Mixed MagmaComplex, San Juan Mountains, Colorado. Journal of Volcanology andGeothermal Research 26, 251^274.

Thouret, J., Davila, J. & Eissen, J. (1999). Largest explosive eruptionin historical times in the Andes at Huaynaputina volcano, A.D.1600, southern Peru. Geology 27, 435^438.

Tsuchiyama, A. (1985). Dissolution kinetics of plagioclase in the meltof the system diopside^albite^anorthite, and origin of dusty plagio-clase in andesites. Contributions to Mineralogy and Petrology 89, 1^16.

Ulmer, P. (2001). Partial melting in the mantle wedgeçthe role ofH2O in the genesis of mantle-derived ‘arc-related’ magmas. Physicsof the Earth and Planetary Interiors 127, 215^232.

Venezky, D. & Rutherford, M. (1999). Petrology and Fe^Ti oxidere-equilibration of the 1991 Mount Unzen mixed magma. Journalof Volcanology and Geothermal Research 89, 213^230.

Von Gruenewaldt, G., Klemm, D. D., Henckel, J. & Dehm, R. M.(1985). Exsolution features in titanomagnetites from massive mag-netite layers and their host-rocks of the upper zone, easternBushveld Complex. Economic Geology 80, 1049^1061.

Whitney, J. A. (1988). The origin of granite: The role and source ofwater in the evolution of granitic magmas. Geological Society of

America Bulletin 100, 1886^1897.Wiebe, R. A. (1996). Mafic-silicic layered intrusions: the role of basal-

tic injections on magmatic processes and the evolution of silicicmagma chambers.Transactions of the Royal Society of Edinburgh: EarthSciences 87, 233^242.

Zellmer, G. F. & Turner, S. P. (2007). Arc dacite genesis pathways:Evidence from mafic enclaves and their hosts in Aegean lavas.Lithos 95, 346^362.

Zellmer, G. F., Sparks, R. S. J., Hawkesworth, C. J. & Wiedenbeck, M.(2003).Magma emplacement and remobilization timescales beneathMontserrat: insights from Sr and Ba zonation in plagioclase pheno-crysts.Journal of Petrology 44,1413^1431.


840

at University of W

ashington on Decem


Dow

nloaded from


egs002 801. - university of washington€¦ · title: egs002 801..840 created date: 3/15/2012...

Documents