Lithos 118 (2010) 119–130

Contents lists available at ScienceDirect

Lithos

j ourna l homepage: www.e lsev ie r.com/ locate / l i thos

H2O- and temperature-zoning in magma chambers: The example of the Tufo Giallodella Via Tiberina eruptions (Sabatini Volcanic District, central Italy)

M. Masotta a, M. Gaeta a,b,⁎, F. Gozzi a, F. Marra b, D.M. Palladino a, G. Sottili a

a Dipartimento di Scienze della Terra, Sapienza-Università di Roma, P.le Aldo Moro, 5, 00185, Rome, Italyb Istituto Nazionale di Geofisica e Vulcanologia, Via di Vigna Murata 605, 00143, Rome, Italy

⁎ Corresponding author. Dipartimento di Scienze dellRoma, P.le Aldo Moro, 5, 00185, Rome, Italy. Tel.: +34454729.

E-mail address: mario.gaeta@uniroma1.it (M. Gaeta)

0024-4937/$ – see front matter © 2010 Elsevier B.V. Aldoi:10.1016/j.lithos.2010.04.004

a b s t r a c t

a r t i c l e i n f o

Article history:Received 21 December 2009Accepted 7 April 2010Available online 24 April 2010

Keywords:Sabatini Volcanic DistrictPhonoliteVolatile zoningMagma chamberDecompression crystallization

Textural and chemical variations of juvenile clasts are widely observed in pyroclastic deposits. In particular,the co-existence of whitish, pumiceous, and dark grey, scoriaceous, juvenile clasts has been observed inmany eruptive units of well-known volcanoes (i.e., Somma–Vesuvius, Vulsini, Colli Albani, Stromboli). Herewe report the example of the Tufo Giallo della Via Tiberina (TGVT) pyroclastic succession, which comprisestwo eruptive units emplaced at ca. 561 and 548 ka, during the early explosive activity of the Sabatini VolcanicDistrict (SVD; Roman Province, central Italy). TGVT deposits, as well as underlying pyroclastic products (FAD,ca. 582 ka), are characterized by coexisting whitish pumice and black–grey scoria clasts showing commonphonolitic composition but different textural features: white pumice is highly vesicular, vitrophyric, andcontains scarce, N50 µm-sized, feldspar and clinopyroxene crystals, while black–grey scoria is poorlyvesicular, highly crystallized, and contains diffuse leucite phenocrysts. The latter records crystallizationunder H2O-undersaturated conditions, as opposed to the vitrophyric texture of white pumice indicatinghigher temperature and H2O concentration. On these grounds, a thermally and H2O-zoned pre-eruptivesystem has been modelled for the phonolitic magma chambers feeding the early SVD events, in which whitepumice and black–grey scoria represent the inner and peripheral portions of the reservoirs, respectively.Extensive leucite+clinopyroxene crystallization in the H2O-undersaturated, peripheral portions of thereservoirs, resulted in water flux toward the inner zones, where the higher temperature and increasing H2Ocontent acted to delay crystallization in the white pumice-feeder magma. The withdrawal of white pumice atthe eruption onset produced decompression of the peripheral magma, triggering black–grey scoria eruptionduring the late phases of explosive events.

a Terra, Sapienza-Università di9 06 49914916; fax: +39 06

.

l rights reserved.

© 2010 Elsevier B.V. All rights reserved.

1. Introduction

Variations in colour, composition and texture of juvenile clasts arefrequently observed in the deposits of large explosive eruptions. Inparticular, the co-existence, with changing proportions, of white anddark juvenile components is reported in several eruptive units of well-known volcanoes, such as Somma–Vesuvius (i.e., 79 A.D. plinianeruption, Carey and Sigurdsson, 1987; Cioni et al., 1995; Pomici diBase eruption, Bertagnini et al., 1998), Phlegrean Fields (i.e., CampanianIgnimbrite, Signorelli et al., 1999), and Vulsini (i.e., Sovana eruption,Vezzoli et al., 1987; Palladino and Taddeucci, 1998; Landi, pers. com.).Changing juvenile features are often explained either by the composi-tional zoning of the magma chambers, as described for the 0.2 Ma CãoGrandephonolitic eruption (CapeVerde Islands,Mortensenet al., 2009),or by pre-eruptive mixing processes of different magma batches, as

proposed by Suzuki and Nakada (2007) for the 16th century ADphonolitic eruption of Haruna Volcano (Japan). Cioni et al. (1995)interpreted the juvenile zoning of Vesuvius 79 A.D. eruption as acombination of syn-eruptive mixing processes in a thermally andcompositionally layered magma chamber, ranging from phonolite totephri-phonolite. In the aforementioned examples of zoned eruptions,changes in juvenile clast chemistry during a single event areaccompanied by transitions from vitrophyric–aphyric to highly crystal-line textures, from themost (upper) and the least (lower) differentiatedportions of the reservoir, respectively.Magma zoning commonly resultsin low-crystallinity andH2O-rich, silicic top portions (Blake, 1984; Blakeand Ivey, 1986a; Dunbar andHervig, 1992;Wallace et al., 1999;Wark etal., 2007). However, the pre-eruptive H2O-saturation of the differenti-ated magma portions feeding aphyric juvenile material is notnecessarily related to a crystal fractionation process (e.g., Miller andMittlefehldt, 1984). Actually, the lowdensity contrast betweenmelt andfeldspar phases makes it unlikely that crystal–melt separation wouldoccur in a phonolitic magma chamber.

On the other hand, syn-eruptive decompression–crystallizationprocesses (Mastrolorenzo and Pappalardo, 2006; Humphreys et al.,

120 M. Masotta et al. / Lithos 118 (2010) 119–130

2008; Brophy, 2009) may explain the formation of highly crystallinejuvenile components, provided that crystal growth rates duringdecompression paths be compatible with the eruptive timescales.Alternatively, the transition from aphyric K-foiditic magma to highlycrystalline tephriphonolitic magma in the Villa Senni Eruption (ColliAlbani, central Italy) is explained by Freda et al. (1997) bymeans of anisobaric volatile-zoning model of the magma chamber, induced by aCO2 flux from carbonate wall rocks.

Here we address the case where significant changes in texturalfeatures of juvenile clasts are not accompanied by bulk compositionalvariations, as it is observed in different examples of major trachy-phonolitic explosive eruptions from the Roman Province (e.g.,the above mentioned Sovana eruption in the Vulsini District). Up tonow, as compared to chemically zoned examples, compositionallyhomogeneous eruptions have been poorly investigated, althoughchanging textural features may highlight the effects of differentH2O concentrations and thermal gradients in variably crystallizedmagma portions. In this regard, the occurrence of both subaphyric–vitrophyric white pumice and highly crystalline black–grey scoria inthe juvenile component represents the most intriguing feature of theTufo Giallo della Via Tiberina (hereafter TGVT) eruptions of theSabatini Volcanic District (SVD; central Italy) and is the subject of thepresent work. We provide a new petrologic model to explain the co-existence of different kinds of juvenile clasts in relatively largeexplosive eruptions, based on a volatile-zoned magma system, asobtained by integrated new stratigraphic and petro-chemical data forthe TGVT, in the light of phase equilibria of the phonolitic feedermagma, gained by MELTS simulations.

2. Geological setting

The SVD is one of themain volcanic districts of the Roman Province(Peccerillo, 2005), covering a ca. 1800 km2-wide area NW of the cityof Rome (Fig. 1). The SVD produced voluminous explosive eruptions,which often overlapped in time with those of the neighbour volcanic

Fig. 1. Sketch map of the Quaternary Sabatini Volcanic District (SVD), Roman Province. L4) inferred source areas during the Morlupo (ca. 800–500 ka) and Southern Sabatini (ca. 50the Sacrofano Caldera are also shown.

district of Colli Albani (Palladino et al., 2001; Marra et al., 2004). TheSVD areal volcanism started at ca. 800 ka, and can be summarized inthree main periods of activity: 1) the early activity period, from asource area located around the town of Morlupo (eastern sector ofSVD), encompasses the whole eruptive activity between the “First AshFall Deposits” (ca. 800–580 ka; Karner et al., 2001) and the GrottaRossa Pyroclastic Sequence (514±3 ka; Karner et al., 2001), includingthe TGVT pyroclastic succession we are dealing with in this paper;2) the intermediate activity period comprises the main eruptions ofthe Tufo Rosso a Scorie Nere (449±1 ka, Karner et al., 2001) and theTufo Giallo di Sacrofano (285±1 ka, Karner et al., 2001), attributed tothe Southern Sabatini (Sottili et al., 2004) and Sacrofano source areas,respectively. During this period, themajor pyroclastic flow eruption ofthe Tufo di Bracciano (ca. 320 ka, Sottili et al., submitted) also tookplace; 3) the most recent activity period was mainly located in thenorthern sectors of Bracciano caldera and between the latter and theSacrofano caldera. In the early stages of this period (250–200 ka,Sottili et al., submitted), the Tufo di Pizzo di Prato and the Tufo diVigna di Valle (de Rita et al., 1993) pyroclastic flow eruptionsoccurred, followed by the development of several coalescent hydro-magmatic centres, mainly in the 130–90 ka time span (Sottili et al.,submitted).

Compositionally, SVD volcanics encompass a wide spectrum ofpotassic rock-types, ranging from trachybasalts to trachytes andphonolites (e.g., de Rita et al., 1993; Campobasso et al., 1994;Conticelli et al., 1997; Sottili et al., 2004).

In this frame, the TGVT groups a series of widespread pyroclasticdeposits emplaced in the eastern sector of the SVD, with estimatedvolume of about 10 km3 (DRE), which can be attributed to at least twomain eruptive events, i.e., the Lower Tufo Giallo della Via Tiberina(561±1 ka, Karner et al., 2001) and the Upper Tufo Giallo della ViaTiberina (551±5-547±5 ka, Karner et al., 2001; broadlycorresponding to the Tufo Giallo della Via Tiberina of Mattias andVentriglia, 1970 and the Sacrofano lower pyroclastic flow unit of deRita et al., 1993).

egend: 1) SVD volcanics; 2) outcrop area of TGVT eruption products; 3) caldera rim;0–400 ka) activity periods. The volcano-tectonic depression hosting Bracciano Lake and

121M. Masotta et al. / Lithos 118 (2010) 119–130

The lower unit (LTGVT) consists of a massive, light grey to paleyellow, pyroclastic flow deposit, containing abundant whitish pumiceand dark grey scoria lapilli and diffuse accretionary lapilli, reaching amaximum thickness of 8 m in paleovalleys. A 3 cm-thick ash layer,associatedwith a 4 cm-thick pumice lapilli layer, is often recognized atthe base of the deposit. The upper unit (UTGVT) comprises foursubunits ((a–d); Fig. 2). The lowermost subunit (a) is a massive,accretionary lapilli-bearing, yellow ash deposit containing diffusewhite pumice and scarce grey scoria, lava and non-metamorphosedsedimentary lithic fragments, reaching a maximum thickness of 20 mat the Calcata locality (Fig. 1). Subunit (b) is texturally andpetrographically similar to subunit (a), although less thick (8–10 m).Subunit (c), the thinnest one (max 5 m-thick), displays alternatingplanar and cross-laminated lapilli and ash layers, containing whitepumice lapilli and subordinate dark grey scoria. Finally, the uppermostsubunit (d) consists of a lithic-rich breccia deposit, containingcentimeter-sized granular inclusions and lava lithic clasts anddispersed feldspar, leucite and clinopyroxene crystals in greyish ashmatrix. The juvenile component is characterized by the prevalence ofblack–grey scoria lapilli over whitish pumice lapilli.

Notably, two kinds of juvenile clasts coexist throughout the wholeTGVT pyroclastic succession, i.e. whitish pumice and dark grey–blackscoria (Fig. 2). The abundance of dark grey–black scoria versuswhitishpumice increases upward in individual eruptive units (i.e., lacking

Fig. 2. Stratigraphy of the study pyroclastic succession in the Morlupo area, including FAD(phenocrysts+microcrysts, on a vesicle-free basis) contents, and degree of vesicularity of

intervening temporal breaks). The same juvenile clast distributionis also observed in the underlying pyroclastic deposit (hereon FAD;582±1 ka; part of the “First Ash Fall Deposits” of Karner et al., 2001),which will be also considered for comparison.

3. Analytical methods

Pumice and scoria clasts and cognate granular inclusions from thewhole TGVT eruptive succession were investigated in order tocharacterize textural and geochemical features of the juvenilecomponent. Microprobe analyses on glasses and mineral phaseswere performed at the CNR-Istituto di Geologia Ambientale eGeoingegneria (Rome, Italy) by a Cameca SX-50 EMP, equipped withfive wavelength-dispersive spectrometers, using 15 kV acceleratingvoltage, 15 nA beam current, 10 µm beam diameter and 20 s countingtime. Major elements of bulk samples were determined on glass beadsat the XRF Laboratory of the Dipartimento di Scienze della Terra(Sapienza-Università di Roma).Matrix effects formajor elementswerecorrected by using the method of Franzini et al. (1972). Thin sectionimages of juvenile clasts were obtained by a Jeol FE-SEM 6500Fequipped with an energy dispersive microanalysis system, at theIstituto Nazionale di Geofisica e Vulcanologia (INGV), Rome. Imageanalysis was performed by the free software package ImageJ (ImageProcessing and Analysis in Java; http://rsb.info.nih.gov/ij/).

and TGVT, showing upsection location of analysed samples. Leucite and total crystaljuvenile clasts, from modal analyses in thin section, are also reported.

122 M. Masotta et al. / Lithos 118 (2010) 119–130

4. Petrographic and compositional features

The analysed juvenile clasts were collected throughout the TGVTsuccession, as well as the underlying FAD (Fig. 2). Based on colour anddegree of porphyricity, two juvenile clast types can be distinguished:white pumice (glass≥50 vol.%, up to 80–90 vol.%) and black–greyscoria (glass≤30 vol.%). Centimeter-sized white pumice clasts occurthroughout the deposits, reaching the maximum concentration in thelower portions of the FAD, LTGVT and UTGVT eruptive units.Conversely, centimeter-sized black–grey scoria clasts mostly occurin the middle–upper portions of the above divisions, reaching themaximum concentration at the top of FAD, LTGVT and UTGVT (i.e.,subunit (d)). Subunit (d) is also characterized by the occurrence ofcognate, mafic granular inclusions made up of clinopyroxene, leuciteand dark mica, and occasional sanidine-rich, clinopyroxene-bearing,granular inclusions, phonolitic in composition (similar to thosereported by Facchinelli and Gaeta, 1992 for other SVD deposits).

4.1. Juvenile clast textures

White pumice (Table 1, Fig. 3) is characterized by subaphyrictexture (degree of porphyricity b10 vol.%), vitrophyric groundmass(glass≥50 vol.%) and moderate vesicularity (up to 50 vol.% in thinsection) and contains scarce clinopyroxene, sanidine, plagioclase andoxide phenocrysts (∅N50 µm). In particular, submillimeter-sized,euhedral to subhedral, clinopyroxene occurs as either single pheno-crysts or glomerocrysts associated with plagioclase and oxides.Sanidine is millimeter- to submillimeter-sized, round-shaped, andoften glass-embayed. Very small-sized (∅b30 µm) leucite typicallyoccurs in glassy vesicle septa.

Black–grey scoria (Table 1; Fig. 3) shows low vesicularity (b25 vol.%in thin section), highly crystalline groundmass (glass≤30 vol.%), andabundant phenocrysts (∅N50 µm; up to 35 vol.%). Leucite, the mostabundant phenocryst phase, is often characterized by crown-likepoikilitic texture (Fig. 3). Clinopyroxene (second in order of abundance)shows two crystal populations: a submillimeter-sized, euhedral tosubhedral, green- to deep-green-coloured, phenocryst population,similar to that occurring in white pumice, and a millimeter-sized,anhedral and colourless, population of xenocrystic nature (also seeSection 5). Scarce sanidine is similar in shape as in white pumice.Plagioclase occurs either in glomerocrysts with clinopyroxene, or asmicrocrysts in the groundmass (Fig. 3). Rare phenocrysts of dark micaand apatite are also present.

4.2. Bulk and glass compositions

White pumice and black–grey scoria show common bulk composi-tions, which essentially plot in a relatively narrow area in thephonolite field (with minor scatter into the trachyte field) of the TAS

Table 1Textural features of FAD and TGVT juvenile clasts determined by SEM image analysis.

Sample E.U. Clast type Glass (vol.%)a Ves. (vol.%) Lct (vol.%)a

MG-11 UTGVT BGS b10 20 35MG-11/2 UTGVT BGS 25–30 25 25MG-10/2 UTGVT WP 75–80 50 5–10MG-10 UTGVT WP 80–85 50 5–10MG-2 UTGVT WP 45–50 25 10–15MG-1 UTGVT WP N90 45 b5MG-7 LTGVT BGS 25–30 20 25MG-29 LTGVT WP 75–80 30 5–10MG-6 LTGVT WP 55–60 40 5–10FAD13-B FAD BGS 25–30 25 30FAD13 FAD WP 85–90 50 5–10

E.U.: eruptiveunit; BGS: black–grey scoria;WP:whitepumice;Ves: vesicularity; Lct: leucite.a Measured on a vesicle-free basis.

diagram (Table 2, Fig. 4). Nevertheless, the compositions of interstitialglasses slightly differ in the two kinds of juvenile components (Fig. 4,Table 3): in particular, glass in black–grey scoria is more variable incomposition than in white pumice and depleted in SiO2 and K2O(Fig. 5), as an effect of the abundant crystallization of leucite (20–35 vol.%). In this regard, mass balance calculations were performedon the phonolitic glass composition of the K2O-richer white pumice(MG-29, Table 3), by subtracting variable amounts of leucite. Thecrystallization of leucite, accompanied by minor amounts of plagio-clase±sanidine±clinopyroxene, as observed in black–grey scoriagroundmass (Fig. 3), roughly reproduces the glass trend in variationdiagrams (Fig. 5).

4.3. Mineral chemistry

Clinopyroxenes show high Ca contents (Table 4), wide ranges ofSiO2 and mg# (Fig. 6), and plot along the Di-Hd joint of the pyroxenequadrilateral. In particular, euhedral to subhedral clinopyroxenephenocrysts show similar compositions in white pumice and black–grey scoria and are characterized by low SiO2 contents and mg#values (Fig. 6), whereas clinopyroxene xenocrysts in black–grey scoriaare Mg-richer (mg#=64–86) like those in cognate, granular inclu-sions (Fig. 6; Table 1EA in Electronic Appendix).

Concerning feldspar compositions (Table 5), in both white pumiceand black–grey scoria, plagioclase is characterized by An-richcompositions (88–95%), while variably sized, round-shaped, sanidinecrystals share Or-rich compositions (88–95%). Noteworthy, in somecases sanidine shows higher BaO contents (up to 2.50 wt.%.), possiblysuggesting near-liquidus crystallization.

Compared to nearby Colli Albani (Freda et al., 1997; Gaeta et al.,2006), leucite is enriched in SiO2 (57–59 wt.%) and Na2O (0.40–0.70 wt.%) and depleted in FeOtot (0.30–0.50 wt.%); dark mica isclassified as biotite, for its Mg/Fe ratiob2. Fe–Ti-oxides show commonTi-magnetite composition in white pumice and black–grey scoria(Table 6).

5. Pre-eruptive conditions (T, P, XH2O)

Clinopyroxene is considered as the main phase driving thedifferentiation of the Roman Province magmas, starting from aninferred K-basaltic to trachybasaltic parental magma (Conte et al.,2009; Gaeta et al., 2009). If we assume a poorly differentiated SVDrock-type as possible parental composition (i.e., as represented bythe San Celso phonotephritic lava, Conticelli et al., 1997), we estimate60–70 vol.% crystallization (Olb5+Cpx40+Lct20+Splb5) of the initialmass to obtain a phonolitic composition matching that of TGVTjuvenile clasts.

The occurrence of high-mg# clinopyroxene xenocrysts in black–grey scoria confirms the role of this mineral phase in the early stagesof the differentiation process leading to the pre-eruptive phonoliticmagma chamber feeding TGVT eruptions. Our petrological model ofthe TGVT magma chamber assumes that the differentiation processesstarting from the parental composition should have occurred atdeeper levels than the emplacement depth of the phonolitic magmabody, i.e., in the “Intermediate Storage System” (sensu Scandone et al.,2007). In fact, the TGVT phonolite lacks evidence of mingling/mixingwith a less differentiated magma. The association of late-eruptedblack–grey scoria with cognate granular inclusions (representative ofa solidification front, sensu Marsh, 1995), as well as the occurrence ofCpx+Pl glomerocrysts and diopsidic xenocrysts in black–grey scoria,consistently indicates that black–grey scoria represents peripheral,cooler, regions of the magma reservoirs. Thus, the late eruption ofblack–grey scoria, following the tapping of white pumice from theinner parts of the magma chambers, would be compatible with aconcentric shell withdrawal model (Blake, 1981; Spera, 1984). Thewithdrawal of different portions of the reservoirs in the course of

Fig. 3. SEM images of TGVT juvenile clasts (sample location in Fig. 2): different crystal and vesicle contents in coexisting white pumice (sample MG-29; a) and black–grey scoria(sample MG-7; b); crown-like poikilitic texture of leucite in black–grey scoria (sample MG-11/2; c); high-magnification of plagioclase microcrysts in the highly crystallinegroundmass of black–grey scoria (sample MG-11; d).

123M. Masotta et al. / Lithos 118 (2010) 119–130

individual eruptive events confirms the remarkable homogeneity ofthe TGVT phonolitic magma chambers. In this regard, the commonoccurrence of sanidine-rich, phonolitic granular inclusions in SVDpyroclastic deposits (e.g., Facchinelli and Gaeta, 1992) may record thesolidification of the non-erupted parts of phonolitic magma chambers.

In order to evaluate T, P and XH2O conditions in the pre-eruptivephonolitic magma system, we applied the clinopyroxene–liquidequilibrium model (Putirka, 2008) and compared phase relationshipsobserved in juvenile clasts with those obtained by MELTS (Ghiorsoand Sack, 1995) simulations. The latter also resulted in agreementwith phase relationships obtained by equilibria experiments onSomma–Vesuvius phonolite (Shea et al., 2009), thus we are fairly

Table 2Bulk chemical composition of FAD and TGVT juvenile clasts, determined by X-rayfluorescence.

E.U. FAD LTGVT UTGVT

Sample FAD13 FAD13-B MG-29 MG-21 MG-30 MG-2 Mg-10/2 MG-11/2

Type WP BGS WP WP BGS WP WP BGS

SiO2 56.01 55.24 53.79 54.69 52.82 55.78 54.28 54.80TiO2 0.48 0.54 0.41 0.46 0.52 0.49 0.47 0.45Al2O3 19.26 19.13 18.30 19.34 18.49 19.55 19.32 19.54Fe2O3 3.61 4.16 2.92 3.43 4.33 3.77 3.44 3.24MnO 0.25 0.25 0.24 0.25 0.25 0.26 0.25 0.25MgO 0.60 0.61 0.55 0.60 1.13 0.67 0.63 0.47CaO 3.20 3.86 4.68 3.81 4.74 3.83 4.09 3.48Na2O 3.09 5.70 3.90 4.87 4.79 5.83 4.33 4.29K2O 9.50 5.05 7.89 7.91 6.18 6.27 8.55 8.91P2O5 0.12 0.16 0.11 0.12 0.19 0.14 0.12 0.11LOI 3.78 5.00 6.41 4.04 5.62 2.89 4.30 3.56Total 99.90 99.70 99.20 99.52 99.06 99.48 99.78 99.10

E.U.: eruptive unit; BGS: black–grey scoria; WP: white pumice.

confident with the application of MELTS code to the present casestudy. Temperature estimates obtained through the Putirka (2008)model are water- and pressure-independent in the low-pressurerange (Pb500 MPa) hypothesized for the TGVT shallow magmasystem. The high-mg# clinopyroxene xenocrysts were not consideredin the calculations because they were not in equilibrium with thephonolitic melt (KdFe/MgN0.40) (in fact, these xenocrysts are from anearly stage of magma evolution). From Cpx–Liq equilibria, using theinterstitial glass composition for the liquid, we would obtaintemperatures of 920–950±20 °C and 950–970±20 °C for whitepumice and black–grey scoria, respectively. However, we note thathigher MgO contents in the melt result in increased temperatureestimates from the Putirka (2008) equations. Thus, we relate theslightly higher temperature value obtained for black–grey scoria

Fig. 4. Bulk (XRF) and glass (EMP) compositions of FAD and TGVT juvenile clasts in theTAS diagram. WP = white pumice; BGS = black–grey scoria.

Table 3Average composition of glasses in FAD and TGVT juvenile clasts, determined by EMP analyses, normalized to 100 on a H2O-free basis.

E. U. FAD LTGVT UTGVT

Sample FAD13 FAD13-B MG-29 MG-7 MG-1 MG-2 MG-10 MG-10/2 MG-11 MG-11/2

Type WP BGS WP BGS WP WP WP WP BGS BGS

σ (4)a σ (3)a σ (5)a σ (6)a σ (3)a σ (5)a σ (5)a σ (6)a σ (6)a σ (5)a

SiO2 58.01 0.32 58.02 0.17 57.74 0.22 57.76 0.87 57.02 0.13 57.86 0.63 56.73 0.11 58.08 0.48 54.96 0.93 57.98 0.27TiO2 0.48 0.04 0.59 0.01 0.40 0.05 0.59 0.06 0.52 0.03 0.58 0.03 0.56 0.04 0.50 0.04 0.68 0.03 0.50 0.04Al2O3 19.75 0.20 19.23 0.09 20.23 0.08 19.22 0.35 19.96 0.14 20.18 0.17 19.56 0.14 20.78 0.19 19.02 0.42 20.87 0.21FeO 3.20 0.11 4.19 0.06 2.62 0.06 4.43 0.87 3.36 0.14 3.49 0.15 3.86 0.12 3.32 0.15 5.33 0.37 3.33 0.21MnO 0.15 0.05 0.12 0.02 0.20 0.02 0.17 0.03 0.17 0.04 0.21 0.04 0.17 0.06 0.21 0.05 0.23 0.03 0.18 0.02MgO 0.55 0.02 0.79 0.03 0.29 0.02 0.68 0.07 0.48 0.03 0.53 0.03 0.73 0.02 0.52 0.03 1.08 0.38 0.49 0.03CaO 4.22 0.65 4.70 0.15 3.55 0.21 5.25 0.95 4.12 0.12 4.11 0.35 4.83 0.26 4.05 0.18 5.93 0.92 4.08 0.19Na2O 3.28 0.08 3.43 0.20 4.27 0.19 4.02 0.62 4.37 0.28 4.64 0.37 3.88 0.09 3.04 0.32 4.49 0.19 3.07 0.23K2O 9.83 0.10 8.32 0.67 9.76 0.41 7.05 0.98 9.27 0.39 7.85 0.49 8.89 0.36 8.76 0.56 6.99 0.40 8.81 0.38SO2 0.16 0.03 0.17 0.04 0.35 0.16 0.23 0.08 0.21 0.05 0.20 0.03 0.24 0.04 0.18 0.02 0.42 0.07 0.18 0.05F 0.22 0.07 0.26 0.10 0.45 0.03 0.39 0.23 0.37 0.02 0.15 0.09 0.32 0.06 0.42 0.12 0.55 0.08 0.36 0.12Cl 0.06 0.02 0.07 0.01 0.11 0.01 0.07 0.04 0.08 0.02 0.09 0.02 0.09 0.02 0.09 0.02 0.09 0.01 0.08 0.01P2O5 0.09 0.02 0.11 0.02 0.03 0.02 0.14 0.03 0.07 0.04 0.11 0.04 0.14 0.02 0.05 0.04 0.23 0.04 0.07 0.03Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00 100.00Totalb 94.26 94.56 93.86 96.54 95.50 95.17 94.06 95.11 94.36 96.01

E.U.: eruptive unit; BGS: black–grey scoria; WP: white pumice.a 1σ standard deviation; number of analyses in parentheses.b WDS-EMP total.

Fig. 5. Chemical variations of FAD and TGVT glasses. Black arrows show the differentiation paths resulting frommass balance calculations, starting from the K2O-richer white pumiceglass composition (MG-29, Table 3), by subtracting increasing amounts of leucite (i.e., 10, 20, 30 wt.%, as represented by bar ticks).

124 M. Masotta et al. / Lithos 118 (2010) 119–130

Table 4EMP analyses of clinopyroxenes in TGVT juvenile clasts.

E. U. LTGVT UTGVT

Sample MG-6 MG-6 MG-7 MG-7 MG-1 MG-1 MG-2 MG-2 MG-10 MG-10 MG-11 MG-11 MG-11a MG-11a

Type WP WP BGS BGS WP WP WP WP WP WP BGS BGS BGS BGS

Core Rim Core Rim Core Rim Core Rim Core Rim Core Rim Core Rim

SiO2 46.80 45.46 47.35 47.36 42.76 44.77 42.03 43.93 44.82 45.51 44.57 44.75 52.85 54.00TiO2 1.04 1.42 1.17 1.23 2.80 1.83 2.31 1.97 1.43 1.33 1.49 1.34 0.56 0.53Al2O3 8.32 8.38 7.53 7.52 9.95 8.20 10.53 9.44 9.05 8.34 9.58 9.14 3.41 2.68FeO 9.01 10.76 9.92 9.68 13.70 12.76 13.80 12.87 10.88 10.50 12.40 11.72 4.35 3.90MnO 0.24 0.33 0.25 0.29 0.49 0.70 0.38 0.65 0.29 0.24 0.25 0.34 0.10 0.05MgO 10.07 9.33 10.63 10.56 6.94 8.00 7.04 7.80 9.12 9.65 8.32 8.38 15.25 16.03CaO 22.69 22.51 23.04 22.88 22.55 23.04 22.68 22.68 22.64 22.87 23.13 23.24 24.12 24.13Na2O 0.37 0.33 0.27 0.31 0.60 0.56 0.44 0.50 0.34 0.30 0.35 0.34 0.18 0.16K2O 0.02 0.04 0.03 0.05 0.02 0.06 0.00 0.00 0.09 0.02 0.03 0.00 0.02 0.02Cr2O3 0.02 0.00 0.03 0.03 0.02 0.00 0.00 0.00 0.07 0.04 0.02 0.00 0.00 0.00Total 98.55 98.56 100.20 99.90 99.82 99.93 99.21 99.84 98.71 98.77 100.14 99.23 100.84 101.49

Formula on the basis of 6 oxygensSi 1.780 1.742 1.775 1.780 1.642 1.708 1.621 1.678 1.714 1.736 1.690 1.710 1.922 1.947Ti 0.030 0.041 0.033 0.035 0.081 0.053 0.067 0.056 0.041 0.038 0.042 0.038 0.015 0.014Al IV 0.220 0.258 0.225 0.220 0.358 0.292 0.379 0.322 0.286 0.264 0.310 0.290 0.078 0.053Al VI 0.153 0.120 0.108 0.113 0.092 0.076 0.100 0.103 0.122 0.111 0.118 0.122 0.069 0.061Fe3+ 0.036 0.083 0.071 0.061 0.150 0.155 0.178 0.143 0.109 0.098 0.133 0.117 −0.008 −0.025Fe2+ 0.251 0.261 0.239 0.243 0.290 0.252 0.267 0.268 0.239 0.237 0.260 0.258 0.141 0.142Mn 0.008 0.011 0.008 0.009 0.016 0.023 0.012 0.021 0.009 0.008 0.008 0.011 0.003 0.002Mg 0.571 0.533 0.594 0.592 0.397 0.455 0.405 0.444 0.520 0.549 0.471 0.478 0.827 0.862Ca 0.924 0.924 0.925 0.921 0.928 0.942 0.937 0.928 0.928 0.935 0.940 0.952 0.940 0.932Na 0.027 0.025 0.019 0.022 0.045 0.041 0.033 0.037 0.025 0.022 0.026 0.025 0.012 0.011K 0.001 0.002 0.001 0.002 0.001 0.003 0.000 0.000 0.004 0.001 0.001 0.000 0.001 0.001Cr 0.001 0.000 0.001 0.001 0.001 0.000 0.000 0.000 0.002 0.001 0.001 0.000 0.000 0.000

E.U.: eruptive unit; BGS: black–grey scoria; WP: white pumice.a Xenocrysts.

125M. Masotta et al. / Lithos 118 (2010) 119–130

magma to the specific MgO-enriched, K2O-depleted, glass composi-tion (Fig. 5, Table 3). In order to avoid over-estimate of black–greyscoria crystallization temperature from Cpx–Liq modelling, we

Fig. 6. Plots of the magnesium number (mg#) versus SiO2 (a) and Di+Hd (molar)(b) for clinopyroxenes from TGVTwhite pumice and black–grey scoria (data in Table 4).Clinopyroxene crystals from cognate granular inclusions (Table 1EA) are also reportedin comparison with the field of clinopyroxene xenocrysts in black–grey scoria (BGS;Table 4).

considered the bulk-chemistry of juvenile clasts, instead of interstitialglass composition, as the best representative of the chemicalcomposition of the phonolitic magma at the time of clinopyroxenecrystallization. In this case, we obtained crystallization temperaturesof 890–920 °C and 925–940 °C for black–grey scoria and whitepumice, respectively, which better agree with the inferred magmachamber geometry (and related temperature distribution).

Pressure conditions cannot be determined directly by Cpx–Liqequilibria, because of the high uncertainty of clinopyroxene geobaro-metry at low pressure (Putirka et al., 1996; Nimis and Ulmer, 1998;Putirka, 2008). However, based on the water content in glasses ofaphyric white pumice (4–5 wt.%), as determined by the “differencefrom 100” method (i.e., based on the difference to 100% of the totalobtained by EMP analyses of glasses, with accuracy of ca. 0.5 wt.% H2O,Devine et al., 1995; Thomas, 2000), a minimum pressure of 100–150 MPa can be estimated. This P value is consistent with phaserelationships inferred from petrographic features and MELTS simula-tions, showing clinopyroxene (±spinel) as the only phase of thewhite pumice in equilibrium with liquid at T=920–940 °C andP=100 MPa (Fig. 7). The estimates of temperature and water contentare also consistent with the Ca-rich (AnPl=90–95%, Table 5)plagioclase composition (following Sisson and Grove, 1993; Takagiet al., 2005; Hamada and Fujii, 2007), as well as the application of two-feldspars system models (Ghiorso, 1984; Green and Usdansky, 1986;Elkins and Grove, 1990), all indicating a crystallization temperaturebelow the 1000 °C isotherm, in a H2O-saturated system.

The H2O content in the interstitial melt of black–grey scoriaindicates a minimum pressure value similar to that deduced for whitepumice. Nevertheless, according to phase relationships, the abundantleucite crystallization (Fig. 3) implies that the phonolitic black–greyscoria magma remained H2O-undersaturated (H2Ob4 wt.%) whilecooling down to T=890–920 °C. On the other hand, the absence ofleucite phenocrysts in white pumice implies that the correspondingmagma remained above the stability field of leucite, until it reachedT=920–940 °C.

Table 5EMP analyses of feldspars and leucite in TGVT juvenile clasts.

Plagioclase Sanidine Leucite

E. U. LTGVT UTGVT LTGVT UTGVT LTGVT UTGVT

Sample MG-7 MG-7 MG-1 MG-2 MG-10 MG-6 MG-6 MG-7 MG-7 MG-10 MG-10 MG-11 MG-7 MG-11

Type BGS BGS WP WP WP WP WP BGS BGS WP WP BGS BGS BGS

Core Rim Core Rim Core Rim Core Rim

SiO2 48.10 47.48 47.57 48.63 47.46 66.75 66.46 64.91 65.01 63.92 63.78 65.90 58.40 56.55TiO2 0.02 0.00 0.05 0.05 0.01 0.00 0.00 0.05 0.06 0.11 0.11 0.08 0.06 0.04Al2O3 33.05 33.03 32.51 32.44 33.54 18.09 18.55 19.33 19.52 19.11 19.14 18.78 23.66 23.78Cr2O3 0.01 0.03 0.03 0.05 0.01 0.00 0.03 0.00 0.00 0.04 0.00 0.03 0.06 0.00MgO 0.02 0.01 0.02 0.02 0.02 0.01 0.00 0.02 0.01 0.00 0.01 0.00 0.00 0.01CaO 17.29 17.60 17.54 16.76 17.23 0.34 0.34 0.41 0.45 0.45 0.44 0.36 0.01 0.05MnO 0.00 0.03 0.00 0.00 0.03 0.00 0.00 0.04 0.03 0.01 0.00 0.01 0.00 0.08FeO 0.63 0.61 0.58 0.58 0.49 0.06 0.08 0.17 0.23 0.17 0.20 0.19 0.37 0.43BaO 0.00 0.05 0.00 0.00 0.01 0.00 0.00 1.81 1.87 1.60 1.61 0.23 0.09 0.13Na2O 1.34 1.25 1.34 1.60 1.23 0.84 0.86 0.86 0.85 0.94 0.96 0.95 0.55 0.66K2O 0.36 0.26 0.19 0.30 0.20 14.76 14.84 13.39 13.49 13.58 13.45 14.50 16.90 19.17Total 100.82 100.35 99.83 100.43 100.23 100.86 101.16 100.99 101.52 99.93 99.70 101.03 100.10 100.90

E.U.: eruptive unit; BGS: black–grey scoria; WP: white pumice.

126 M. Masotta et al. / Lithos 118 (2010) 119–130

T and H2O pre-eruptive conditions for both white pumice andblack–grey scoria feeder magmas, as deduced by phase relationshipsfrom petrographic observations and petrological modelling, aresummarized in Fig. 7.

6. Discussion

Different from the famous Vesuvius 79 A.D. event and other majorexplosive eruptions showing along-stratigraphy chemical variationsof juvenile clasts, the TGVT succession does not provide evidence ofchemically zoned magma chambers. Thus, the observed juveniletextural changes in TGVT should be related to the physical processes,such as decompression and volatile exsolution, controlling the lateevolution of the phonolitic magma reservoir at shallow crustal level.

A key question arising from the textural features of TGVT juvenileclasts, in particular for white pumice, is how the phonolitic magmareached pre-eruptive H2O-saturation. Although it is well known thatvolatiles dissolved in silicate melts play a fundamental role incontrolling magma evolution (e.g., Huppert and Woods, 2002), thecauses of volatile concentration are often poorly constrained.

Table 6EMP analyses of Fe–Ti-oxides and biotite in TGVT juvenile clasts.

Oxide Biotite

E. U. LTGVT UTGVT LTGVT UTGVT

Sample MG-7 MG-7 MG-2 MG-2 MG-7 MG-7 MG-11 MG-11

Type BGS BGS WP WP BGS BGS BGS BGS

Glom. Glom.

SiO2 0.06 0.32 0.09 0.60 36.32 36.02 36.74 36.12TiO2 6.53 6.16 6.03 5.69 5.06 4.44 3.99 4.39Al2O3 3.02 4.75 4.09 4.07 15.33 15.01 15.80 15.66Cr2O3 0.05 0.05 0.04 0.04 0.03 0.08 0.00 0.08Fe2O3 52.45 50.96 52.84 52.13 nd nd nd ndFeO 33.60 32.86 32.88 33.06 15.16 13.81 12.89 12.96MgO 1.40 2.31 2.06 2.00 13.40 14.70 15.55 14.76CaO 0.03 0.16 0.05 0.21 0.02 0.07 0.03 0.06MnO 1.49 1.00 1.10 1.09 0.21 0.14 0.13 0.15BaO 0.00 0.00 0.00 0.00 1.59 1.36 1.10 1.35Na2O 0.04 0.03 0.04 0.17 0.20 0.30 0.16 0.45K2O 0.02 0.05 0.00 0.08 8.95 9.10 9.16 9.19Total 98.69 98.65 99.22 99.14 96.27 95.03 95.55 95.17

E.U.: eruptive unit; BGS: black–grey scoria; WP: white pumice; glom: glomerocrysttexture; n.d.: not determined.

In our case study, since a low amount of phenocrysts (typicallyeven b5 vol.%) occurs in white pumice, the achievement of H2O-saturation in the magma by mere crystal fractionation seems unlikely,as it is difficult to imagine significant crystal removal by a settlingmechanism fast enough to produce the low porphyricity of whitepumice magma. Indeed, by assuming an initial content of 2 wt.% H2O,the crystallization of 50 vol.% of anhydrous phases (Cpx7Pl10Sa30Spl3according to MELTS simulations) would be required to reach thesaturation value of 4 wt.% at 100 MPa. The removal of at least 35–40 vol.% phenocryst amount is unlikely to have occurred, consideringthe low density contrast between feldspars (sanidine in particular)and the phonolitic magma (Fig. 8). Moreover, the fractionalcrystallization model (either by crystal settling or an alternativemechanism such as filter-pressing) would imply a compositionallyzoned magma system, which contrasts with the similar bulkcompositions of white pumice and black–grey scoria.

Existing models of volatile-zoned magma systems feeding siliciclarge explosive eruptions indicate higher volatile concentrations inthe upper portions of magma chambers, as a result of compositionallayering and/or recharge by mafic magmas (e.g., Blake, 1984; Blakeand Ivey, 1986a,b; Dunbar and Hervig, 1992; Wallace et al., 1999;Wark et al., 2007). In this regard, the lack of inverse chemical zoningin clinopyroxene, as well as the homogeneity of the juvenile bulkcomposition, lead us to rule out the occurrence of pre-eruptivecompositional zoning in the TGVT reservoir and/or mixing with lessevolved magmas.

Fig. 7. Temperature versus H2Omelt phase stability diagram obtained by MELTSsimulations on bulk phonolitic composition (sample MG-21; Table 3) at P=100 MPa.T, XH2O values refer to initial crystallization conditions for black–grey scoria (BGS)feeder magma and to pre-eruptive conditions for white pumice (WP) feeder magma (asdiscussed in Section 6.1).

Fig. 8. Diagram illustrating crystal settling velocity in a phonolitic melt driven by crystal–melt density contrast, by applying the Stokes law for sinking particles. Phonolite meltviscosity (104 Pa s)was calculated at900 °C andH2O content of 4 wt.%, followingGiordanoet al. (2008) and Whittington et al. (2006). Densities (ρ) and maximum growth rates (Y)of different mineral phases have been taken into account in the calculations, asfollows: ρmagma=2450 kg/m3, ρCpx=3320 kg/m3, ρPl=2620 kg/m3, ρSa=2500 kg/m3,ρLct=2420 kg/m3; YLct≈10−8 mm/s (Shea et al., 2009); YSa≈10−8 mm/s (Hammer andRutherford, 2002); YPl≈10−8 mm/s (Cashman and Marsh, 1988; Cashman, 1993);YCpx≈10−8 mm/s (Crisp et al., 1994). For example, the diagram shows that it takes ca.10 years for a plagioclase crystal to attain a size suitable for settling (unmovable–movableboundary), and ca. 25 years to reach 7.4 mm in size at 500 m settling distance. On thesegrounds, leucite and sanidine populations are considered as almost unmovable inphonolitic melts (in particular, note that leucite tends to float in these magmas).

127M. Masotta et al. / Lithos 118 (2010) 119–130

Alternatively, the emplacement of a near H2O-saturated, crystal-poor, phonolitic magma at pre-eruptive depth (100–150 MPa) isinconsistent with petrographic evidence from highly porphyritic,leucite-rich, black–grey scoria. Indeed, phase relationships of phono-litic magmas at 100–150 MPa (Fig. 7), point out an amount of leucitecomparable to that actually occurring in black–grey scoria only atH2O-undersaturated conditions.

The widening of the leucite stability field, as a result of possibledecompression of a H2O-saturated phonolitic magma cannot fullyexplain the relevant petrographic features of the black–grey scoria.Indeed, the wide range of leucite and plagioclase crystal sizes inblack–grey scoria consistently reflects a range of crystallizationconditions (Cashman, 1992). For example, decompression simula-tions (Fig. 9) show that a 50 MPa pressure drop acting on a H2O-saturated, partially crystallized (T=890–920 °C), phonolitic magma(at the initial estimated P=100 MPa) would result in up to 20 wt.%increase of solid phase. In particular, during the decompression path,the leucite and plagioclase stability fields show larger widening thansanidine and clinopyroxene (Fig. 9).

While scarce leucite microcrysts (∅b30 µm) in white pumicemay record conduit ascent of a H2O-saturated phonolitic magma, the

Fig. 9. Increasing crystal content (Δ vol.%) in a phonolitic magma as a result of aΔP=50 MPa decompression step, at different temperatures (i.e., 920, 900 and 890 °C),obtained by MELTS simulations. Note that decompression–crystallization mostlyinvolves leucite+plagioclase (Δ up to ∼18 vol.%).

occurrence in black–grey scoria of leucite+plagioclase-rich ground-mass and crown-like poikilitic textures of leucite phenocrysts(Fig. 3), well agrees with a rapid (i.e. at the eruptive time-scale)crystallization induced by decompression in the magma chamber,possibly due to the withdrawal of the white pumice-feeder magma(cf. eruption–decompression–crystallization models; Mastrolorenzoand Pappalardo, 2006; Humphreys et al., 2008; Brophy, 2009).However, a decompression event on the eruptive time-scale cannotexplain the extensive formation of larger leucite crystals(∅N100 µm) in black–grey scoria magma. Actually, considering anaverage growth rate of 2×10−8 mm/s (Shea et al., 2009), the timerequired to reach N100 µm in size (102 days) would be considerablylonger than the explosive eruption time-scale (100–1 days). On thesegrounds, the diffuse occurrence of leucite phenocrysts in black–greyscoria puts constraints to the corresponding pre-eruptive magmadomain at initial H2O-undersaturated conditions (Fig. 7).

6.1. T and H2O-zoning model

It has been discussed above how both crystal fractionation anddecompression-induced crystallization cannot adequately explain theachievement of H2O-saturation in the subaphyric white pumice-feeder magma and the diffuse occurrence of leucite phenocrysts inblack–grey scoria. Therefore, in order to explain the highly variablecrystal contents in TGVT juvenile clasts, as well as the diffuseoccurrence of aphyric juvenile material in the products of largeexplosive eruptions, we propose amodel based on a temperature- andvolatile-zoned magma system. Accordingly, T and H2O-zoning couldbe the factor controlling the degree of crystallization and phaserelationships in different portions of magma chambers.

Specifically, we recall that the presence of abundant, Lct+Cpx+dark mica-bearing, cognate granular inclusions, associated withblack–grey scoria (e.g., in subunit (d) of UTGVT), suggests the tappingof different regions of the phonolitic reservoir during the emission ofwhite pumice and black–grey scoria in the course of individualeruptive events. In particular, black–grey scoria represents late-erupted, cooler, peripheral regions of the magma system, within the

Fig. 10. Sketch of T and H2O paths for the peripheral and inner portions of the phonolitemagma reservoirs feeding early SVD eruptive events (FAD, TGVT), represented byblack–grey scoria and white pumice, respectively. Starting from H2O-undersaturatedconditions (H2O≈2 wt.% at 100 MPa and T≈1000 °C), increasing H2O concentration inthe peripheral portions (black arrow) occurred as a result of crystallization within theleucite stability field (leucite-in curve after MELTs simulations; see Fig. 7) at relativelyhigh cooling rates. On the other hand, H2O concentration in the inner (hotter) portions(white arrow) increased due to H2O flux from peripheral regions and leucite-freecrystallization. H2O-saturation conditions (H2O≈4 wt.% at 100 MPa) for black–greyscoria and white pumice magmas were eventually achieved at T≈890 and 920 °C,respectively (see Section 5 for details).

128 M. Masotta et al. / Lithos 118 (2010) 119–130

frame of magma withdrawal from a concentric shell geometry (Blake,1981; Spera, 1984). Phase relationships and crystal growth rates dataindicate that crystallization of relatively large leucite (∅N100 µm)should occur under H2O-undersatured conditions (Hammer andRutherford, 2002; Freda et al., 2008; Shea et al., 2009). Thus,phonolitic black–grey scoria magmas from the cooler, peripheralregions of the TGVT magma chambers were initially H2O-undersat-urated (H2Ob4 wt.%, at P=100 MPa and TCpx–liquidN890–925 °C,Fig. 7), allowing crystallization of leucite phenocrysts.

Fig. 11. Sketch model of the evolution of T and H2O-zoned magma chambers feeding FAD anbody at ca. 100 MPa and onset of crystallization in peripheral portions; t1: increasing H2O comigration toward inner magma chamber portions (WP); t2: achievement of pre-eruptive Hwhite pumice (WP) at eruption onset, resulting in decompression of BGS magma; t4: conexpansion leading to BGS magma fragmentation and eruption during late stages of individu

Conversely, the lack of leucite phenocrysts in white pumiceindicates H2O conditions above the leucite-in curve in the phasediagram (Fig. 7). Moreover, the presence of round-shaped, partiallyresorbed, sanidine phenocrysts, testifies that H2O-saturation wasachieved in the white pumice magma following feldspar crystalliza-tion. In particular, due to the relatively small size and low densitycontrast with the host phonolitic melt that prevented significantcrystal sinking (Fig. 8), sanidine can be considered to record changingmagma conditions.

d TGVT multiple eruptions. t0: emplacement of a H2O-undersaturated phonolite magmancentration due to crystallization of black–grey scoria (BGS) feeder magma caused H2O2O-saturation conditions; t3: tapping of inner magma portions feeding the emission ofsequent sudden crystallization, magma viscosity increase, H2O exsolution and bubbleal eruptive events.

129M. Masotta et al. / Lithos 118 (2010) 119–130

We propose that extensive magma crystallization and, conse-quently, increasing H2O concentration in the melt, at the periphery ofthe TGVT magma chamber (i.e., the black–grey scoria feeder magma),produced H2O migration toward the inner portion of the reservoir,where the higher temperature and increasing H2O content acted todelay crystallization in the white pumice-feeder magma. T and H2Opaths for the outer and inner regions of the magma reservoir, asrelated to the leucite stability field, are schematically illustrated inFig. 10.

Once a H2O concentration gradient was established between thecrystallizing peripheral regions of the magma chamber and the poorlycrystallized inner portion, the H2O mass flux was actually enhanced bythe geometry of themagma system. If we consider a roughly concentricgeometry of H2O- and thermal zoning for the TGVT magma chamber(Fig. 11), at each given point along the radial H2O-gradient the mass ofinward-migrating water would diffuse toward increasingly smallvolumes, so that the amount of dissolved water in the inner part ofmagma chamber would increase constantly, eventually leading tosaturation. In this regard, re-melting features (i.e. rounded-shape andglass-embayment) in sanidine from white pumice provide evidence ofgradually increasing H2O content that reduced the feldspars+leucitestability field in the inner portion of the magma chamber (Fig. 7).

7. Conclusions

A number of explosive volcanic eruptions involving the emplace-ment of both subaphyric–vitrophyric and highly crystalline juvenileclasts of similar bulk composition could be re-interpreted in the lightof T and H2O zoning in magma chambers. This could also account forthe eruption of large volumes of vitrophyric juvenile material, forwhich pre-eruptive H2O-oversaturation cannot be fully explained byfractional crystallization processes. The petrography of juvenile clasts,integrated with T and P estimates from phase equilibria relationships,allowed us to hypothesize T and H2O-zoned magma chambers for theTGVT eruptions. This model is based on the occurrence of extensivecrystallization in the cooler, peripheral zones of the magma chamberat initially H2O-undersaturated conditions (BGS zone in Fig. 11). Thisis considered, in a roughly concentric geometry, as a trigger for thewater flux from the peripheral toward the inner zone of the reservoir.Water migration would allow the achievement of H2O saturation andexsolution in the near-liquidus inner part of themagma chamber (WPin Fig. 11), triggering the fragmentation of a poorly crystallizedmagma, feeding the early phases of TGVT eruptions. H2O input wouldalso result in the partial resorption of early formed feldspar crystals,an endothermic process that may favour the glass transition in thesurrounding melt. We speculate that these quench domains may actas a concomitant factor toward magma fragmentation.

The withdrawal of inner magma chamber portions (WP eruptionin Fig. 11), would produce a decompression front that moves quicklyagainst the solidification front migrating from the peripheral regionsof the reservoir. This would result in extensive decompressioncrystallization (recorded in the black–grey scoria groundmass;Fig. 3), the increasing solid/melt mass ratio in turn resulting inmagma viscosity increase, H2O exsolution and bubble expansion, allfactors acting toward increasing explosive potential of the magmafeeding the black–grey scoria eruption during late phases of TGVTevents (BGS eruption in Fig. 11).

Acknowledgements

We are grateful to A.M. Conte, M. Serracino, and A. Cavallo forhelping during XRF, EMP, and SEM analyses, respectively. Construc-tive reviews of editor and two anonymous reviewers improved themanuscript. This research is part of MM PhD program and was fundedby AST 2008 “Nuove iniziative di ricerca di Ateneo Federato” ofSapienza-Università di Roma.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.lithos.2010.04.004.

References

Bertagnini, A., Landi, P., Rosi, M., Vigliargio, A., 1998. The Pomici di Base plinian eruption ofSomma–Vesuvius. Journal of Volcanology and Geothermal Research 83, 219–239.

Blake, S., 1981. Eruption from zoned magma chambers. Journal of the GeologicalSociety, London 138 (3), 281–287.

Blake, S., 1984. Volatile oversaturation during the evolution of silicic magma chambersas an eruption trigger. Journal of Geophysical Research 89 (B10), 8237–8244.

Blake, S., Ivey, G.N., 1986a. Density and viscosity gradients in zoned magma chambers,and their influence withdrawal dynamics. Journal of Volcanology and GeothermalResearch 30 (3–4), 201–230.

Blake, S., Ivey, G.N., 1986b. Magma-mixing and the dynamics of withdrawal fromstratified reservoirs. Journal of Volcanology and Geothermal Research 27 (1–2),153–178.

Brophy, J.G., 2009. Decompression and H2O exsolution driven crystallization andfractionation: development of a new model for low pressure fractional crystalli-zation in calc-alkaline magmatic systems. Contributions to Mineralogy andPetrology 157, 797–811.

Campobasso, C., Cioni, R., Salvati, L., Sbrana, A., 1994. Geology and paleogeographicevolution of a peripheral sector of the Vico and Sabatini volcanic complex, betweenCivita Castellana and Mazzano Romano (Latium, Italy). Memorie Descrittive dellaCarta Geologica d'Italia 49, 277–290.

Carey, S.N., Sigurdsson, H., 1987. Temporal variations in column height and magmadischarge rate during the 79 A.D. eruption of Vesuvius. Geological Society ofAmerica Bulletin 99, 303–314.

Cashman, K.V., 1992. Groundmass crystallization of Mount St Helens dacites, 1980–1986: a tool for interpreting shallow magmatic processes. Contributions toMineralogy and Petrology 109, 431–449.

Cashman, K.V., 1993. Relationship between plagioclase crystallization and cooling ratein basaltic melts. Contributions to Mineralogy and Petrology 113, 126–142.

Cashman, K.V., Marsh, B.D., 1988. Crystal size distribution (CSD) in rocks and thekinetics and dynamics of crystallization II: Makaopuhi lava lake. Contributions toMineralogy and Petrology 99, 292–305.

Cioni, R., Civetta, L., Marianelli, P., Metrich, N., Santacroce, R., Sbrana, A., 1995.Compositional layering and syn-eruptive mixing of periodically refilled shallowmagma chamber: the A.D. 79 Plinian eruption of Vesuvius. Journal of Petrology 36,739–776.

Conte, A.M., Dolfi, D., Gaeta, M., Misiti, V., Mollo, S., Perinelli, C., 2009. Experimentalconstraints on evolution of leucite–basanite magma at 1 and 10–4 GPa: implica-tions for parental compositions of Roman high-potassium magmas. EuropeanJournal of Mineralogy 214, 763–782.

Conticelli, S., Francalanci, L., Manetti, P., Cioni, R., Sbrana, A., 1997. Petrology andgeochemistry of the ultrapotassic rocks from the Sabatini Volcanic District, centralItaly: the role of evolutionary processes in the genesis of variably enriched alkalinemagmas. Journal of Volcanology and Geothermal Research 75, 107–136.

Crisp, J., Cashman, K.V., Bonini, J.A., Hougen, S.B., Pieri, D.C., 1994. Crystallization historyof the 1984 Mauna Loa lava flow. Journal of Geophysical Research 99, 7177–7198.

de Rita, D., Funiciello, R., Corda, L., Sposato, A., Rossi, U., 1993. Carta Geologica delcomplesso Vulcanico Sabatino. Quaderni de La Ricerca Scientifica. ProgettoFinalizzato Geodinamica Monografie Finali, vol. 11. C.N.R., Rome, Italy, p. 114.

Devine, J.D., Gardner, J.E., Brack, H.P., Layne, G.D., Rutherford, M.J., 1995. Comparison ofmicroanalytical methods for estimating H2O contents of silicic volcanic glasses.American Mineralogist 80, 319–328.

Dunbar, N.W., Hervig, R.L., 1992. Petrogenesis and volatile stratigraphy of the Bishop Tuff:evidence from melt inclusion analysis. Journal of Geophysical Research 97 (B11),15129–15150.

Elkins, L.T., Grove, T.L., 1990. Ternary feldspar experiments and thermodynamicmodels. American Mineralogist 75 (5–6), 544–559.

Facchinelli, V., Gaeta, M., 1992. Indicazioni petrogenetiche dai granati birifrangenti deiproietti sialici nelle vulcaniti alcalino potassiche dei Monti Sabatini (Lazio).Rendiconti Lincei 34, 295–310.

Franzini, M., Leoni, L., Saitta, M., 1972. A simple method to evacuate the matrix effect inX-ray fluorescence analysis. X-ray Spectrometry 1, 151–154.

Freda, C., Gaeta, M., Palladino, D.M., Trigila, R., 1997. The Villa Senni Eruption (AlbanHills, Central Italy): the role of H2O and CO2 on the magma chamber evolution andon the eruptive scenario. Journal of Volcanology and Geothermal Research 78,103–120.

Freda, C., Gaeta, M., Misiti, V., Mollo, S., Dolfi, D., Scarlato, P., 2008. Magma–carbonateinteraction: an experimental study on ultrapotassic rocks from Alban Hills (CentralItaly). Lithos 101, 397–415.

Gaeta, M., Freda, C., Christensen, J.N., Dallai, L., Marra, F., Karner, D.B., Scarlato, P., 2006.Time-dependent geochemistry of clinopyroxene from the Alban Hills (CentralItaly): clues to the source and evolution of ultrapotassic magmas. Lithos 86,330–346.

Gaeta, M., Di Rocco, T., Freda, C., 2009. Carbonate assimilation in open magmaticsystems: the role of melt-bearing skarns and cumulate-forming processes. Journalof Petrology 502, 361–385.

Ghiorso, M.S., 1984. Activity/composition relations in the ternary feldspars. Contributionsto Mineralogy and Petrology 87, 282–296.

130 M. Masotta et al. / Lithos 118 (2010) 119–130

Ghiorso, M.S., Sack, R.O., 1995. Chemical mass transfer in magmatic processes IV. Arevised and internally consistent thermodynamic model for the interpolation andextrapolation of liquid–solid equilibria in magmatic systems at elevated tempera-tures and pressures. Contributions to Mineralogy and Petrology 119, 197–212.

Giordano, D., Russell, J.K., Dingwell, D.B., 2008. Viscosity of magmatic liquids: a model.Earth and Planetary Science Letters 271, 123–134.

Green, N.L., Usdansky, S.I., 1986. Ternary-feldspar mixing relations and thermobaro-metry. American Mineralogist 71, 1100–1108.

Hamada, M., Fujii, T., 2007. H2O-rich island arc low-K tholeiite magma inferred from Ca-rich plagioclase–melt inclusion equilibria. Geochemical Journal 41, 437–461.

Hammer, J.E., Rutherford, M.J., 2002. An experimental study of the kinetics ofdecompression-induced crystallization in silicicmelt. Journal of Geophysical Research107 (B1), 1–24.

Humphreys, M.C.S., Blundy, J.D., Sparks, R.S.J., 2008. Shallow-level decompressioncrystallization and deep magma supply at Shiveluch Volcano. Contributions toMineralogy and Petrology 155, 45–61.

Huppert, H.E., Woods, A.W., 2002. The role of volatiles in magma chamber dynamics.Nature 420, 493–495.

Karner, D.B., Marra, F., Renne, P.R., 2001. The history of the Monti Sabatini and AlbanHills volcanoes: groundwork for assessing volcanic–tectonic hazards for Rome.Journal of Volcanology and Geothermal Research 107, 185–219.

Marra, F., Taddeucci, J., Freda, C., Marzocchi, W., Scarlato, P., 2004. Recurrence ofvolcanic activity along the Roman Comagmatic Province (Tyrrhenian margin ofItaly) and its tectonic significance. Tectonics 23 (TC4013).

Marsh, B.D., 1995. Solidification fronts and magmatic evolution. MineralogicalMagazine 60, 5–40.

Mastrolorenzo, G., Pappalardo, L., 2006. Magma degassing and crystallization processesduring eruptions of high-riskNeapolitan-volcanoes: evidenceof commonequilibriumrising processes in alkalinemagmas. Journal of Volcanology andGeothermal Research83, 219–239.

Mattias, P.P., Ventriglia, U., 1970. La regione vulcanica dei Monti Sabatini. Memoriedella Società Geologica Italiana 9, 331–384.

Miller, C.F., Mittlefehldt, D.W., 1984. Extreme fractionation in felsic magma chambers: aproduct of liquid-state diffusion or fractional crystallization? Earth and PlanetaryScience Letters 68 (1), 151–158.

Mortensen, A.K., Wilson, J.R., Holm, P.M., 2009. The Cão Grande phonolitic fall depositon Santo Antão, Cape Verde Islands. Journal of Volcanology and GeothermalResearch 179, 120–132.

Nimis, P., Ulmer, P., 1998. Clinopyroxene geobarometry of magmatic rocks Part 1: anexpanded structural geobarometer for anhydrous and hydrous, basic and ultrabasicsystems. Contributions to Mineralogy and Petrology 133, 122–135.

Palladino, D.M., Taddeucci, J., 1998. The basal ash deposit of the Sovana eruption(Vulsini Volcanoes, Central Italy): the product of a dilute pyroclastic densitycurrent. Journal of Volcanology and Geothermal Research 87, 233–254.

Palladino, D.M., Gaeta,M., Marra, F., 2001. A large K-foiditic hydromagmatic eruption fromthe early activity of the Alban Hills Volcanic District, Italy. Bulletin of Volcanology 63,345–359.

Peccerillo, A., 2005. Plio-Quaternary Volcanism in Italy. Springer-Verlag, BerlinHeidelberg. (365 pp.).

Putirka, K.D., 2008. Thermometers and barometers for volcanic systems. Reviews inMineralogy and Geochemistry 69 (1), 61–120.

Putirka, K.D., Johnson, M., Kinzler, R., Longhi, J., Walker, D., 1996. Thermobarometry ofmafic igneous rocks based on clinopyroxene–liquid equilibria, 0–30 kbar. Con-tributions to Mineralogy and Petrology 123, 92–108.

Scandone, R., Cashman, K.V., Malone, S.D., 2007. Magma supply, magma ascent and thestyle of volcanic eruptions. Earth and Planetary Science Letters 253, 513–529.

Shea, T., Larsen, J.F., Gurioli, L., Hammer, J.E., Houghton, B.F., Cioni, R., 2009. Leucitecrystal: surviving witnesses of magmatic processes preceding the 79 AD eruption atVesuvius, Italy. Earth and Planetary Science Letters 281, 88–98.

Signorelli, S., Vagelli, G., Francalanci, L., Rosi, M., 1999. Origin of magmas feeding thePlinian phase of the Campanian Ignimbrite eruption, Phlegrean Fields (Italy):constraints based on matrix–glass and glass-inclusion compositions. Journal ofVolcanology and Geothermal Research 91, 199–220.

Sisson, T.W., Grove, T.L., 1993. Experimental investigations of the role of H2O in calc-alkaline differentiation and subduction zone magmatism. Contributions toMineralogy and Petrology 113, 143–166.

Sottili, G., Palladino, D.M., Zanon, V., 2004. Plinian activity during the early eruptivehistory of the Sabatini Volcanic District, Central Italy. Journal of Volcanology andGeothermal Research 135, 361–379.

Sottili, G., Palladino, D.M., Marra, F., Jicha, B., Karner, D.B., Renne, P. Geochronology ofthe most recent activity in the Sabatini Volcanic District, Roman Province, centralItaly. Journal of Volcanology and Geothermal Research, submitted.

Spera, F.J., 1984. Some numerical experiments on the withdrawal of magma fromcrustal reservoirs. Journal of Geophysical Research 89 (B10), 8222–8236.

Suzuki, Y., Nakada, S., 2007. Remobilization of highly crystalline felsic magma byinjection of mafic magma: constraints from the middle sixth century eruption atHaruna Volcano, Honshu, Japan. Journal of Petrology 48 (8), 1543–1567.

Takagi, D., Sato, H., Nakagawa, M., 2005. Experimental study of a low-alkali tholeiite at1–5 kbar: optimal condition for the crystallization of high-An plagioclase inhydrous arc tholeiite. Contributions to Mineralogy and Petrology 149, 527–540.

Thomas, R., 2000. Estimation of water contents of granite melt inclusions by confocallaser Raman microprobe spectroscopy. American Mineralogist 85, 868–872.

Vezzoli, L., Conticelli, S., Innocenti, F., Landi, P., Manetti, L., Palladino, D.M., Trigila, R.,1987. Stratigraphy of the Latera Volcanic Complex: proposals for a newnomenclature. Periodico di Mineralogia 56, 89–110.

Wallace, P.J., Anderson Jr., A.T., Davis, A.M., 1999. Gradients in H2O, CO2, and exsolvedgas in large-volume silicic magma system: interpreting the record preserved inmelt inclusions from the Bishop Tuff. Journal of Geophysical Research 104 (B9),20097–20122.

Wark, D.A., Hildreth, W., Spear, F.S., Cherniak, D.J., Watson, E.B., 2007. Pre-eruptionrecharge of the Bishop magma system. Geology 35 (3), 235–238.

Whittington, A., Richet, P., Linard, Y., Holtz, F., 2006. The viscosity of hydrous phonolitesand trachytes. Chemical Geology 174, 209–223.