Artº: Magma storage and ascent conditions beneath Pico and Faial islands (Azores archipelago) A...

Magma storage and ascent conditions beneath Picoand Faial islands (Azores archipelago): A study onfluid inclusions

Vittorio ZanonCentro de Vulcanologia e Avaliac~ao de Riscos Geologicos, Universidade dos Acores, Rua M~ae de Deus, PontaDelgada, PT-9501-801, Portugal ([email protected])

Maria Luce FrezzottiDipartimento di Scienze Geologiche e Geotecnologie, University of Milan Bicocca, Milan, Italy

[1] In the islands of Faial and Pico (the Azores), uid inclusions are hosted in megacrysts of olivine(Mg#8088) and clinopyroxene (Mg#7990) in highly porphyritic lavas and in mineral assemblagesof ultramac xenoliths. Rare inclusions are contained in olivine phenocrysts (Mg#< 80) andplagioclases in poorly porphyritic lavas. Trails of late-stage inclusions are predominant over isolatedearly-stage inclusions. Almost all inclusions are re-equilibrated and the trapped uid consists of pureCO2 (Tm from 56.5 to 57.2). Rare early-stage inclusions may contain dypingite or Mg-calcite,which indicates that in earlier times some water was present along with CO2. Barometric data indicatethat CO2 inclusions in xenoliths from the two islands equilibrated at maximum pressures of 570586MPa (19.721.2 km), while in poorly porphyritic lavas from all the ssure zones at 465508 MPa(16.418.1 km). Maximum pressure values of 463 MPa (16.8 km) and 492 MPa (17 km) wererecorded for the central volcanoes of Pico and Faial, respectively. Further trapping/re-equilibrationwas recorded at 156 MPa in Faial (5.6 km), in plagioclase phenocrysts in mugearites. All thesepressures correspond to magma ponding sites and to its crystallization and can be useful for tracingthe progressive thickening of a dense transition zone, below the geophysical Moho. The ability toextract rapidly the stored magmas from these volcanic systems strictly depends on the differenttectonic styles, acting in this transition zone. Magmatic evolution in small and short-lived intracrustalreservoirs, not necessarily coaxial with main conduit system, was enhanced at the intersection ofdifferently oriented lineaments.

Components: 11,912 words, 14 gures.

Keywords: upper mantle; cumulitic xenoliths; ssure zones; basalts; storage system; underplating.

Index Terms: 1043 Fluid and melt inclusion geochemistry: Geochemistry; 1033 Intra-plate processes: Geochemistry;1036 Magma chamber processes: Geochemistry; 8415 Intra-plate processes: Volcanology; 3651 Thermobarometry: Miner-alogy and Petrology; 3615 Intra-plate processes: Mineralogy and Petrology; 3618 Magma chamber processes: Mineralogyand Petrology.

Received 5 March 2013; Revised 2 July 2013; Accepted 2 July 2013; Published 00 Month 2013.

Zanon, V., and M. L. Frezzotti (2013), Magma storage and ascent conditions beneath Pico and Faial islands (AzoresIslands): 1. A study on uid inclusions, Geochem. Geophys. Geosyst., 14, doi:10.1002/ggge.20221.

2013. American Geophysical Union. All Rights Reserved. 1

Article

Volume 00, Number 00

0 MONTH 2013

doi: 10.1002/ggge.20221

ISSN: 1525-2027

1. Introduction

[2] Volcanic eruptions in areas characterized byextensional tectonics typically originate fromssure systems, which generally follow local or re-gional weakness patterns, or from composite vol-canoes. Magmas erupted from these systems arecharacterized by different patterns of evolution,linked to different pathways to the surface [Klugelet al., 2009; Eason and Sinton, 2009; Sigmunds-son et al., 2010]. Thin mac lava ows issuedfrom ssure zones constitute the bulk of theerupted material from these volcanoes.

[3] Magma buoyancy plays a major role in the dy-namics of magma ow and storage [Wilson andHead, 1981], as it controls the depth at whichmagma reservoirs formed either in the crust oralong large lithospheric discontinuities [Ryan,1993; Ida, 1995; McLeod, 1999]. Magma buoy-ancy is related to the density difference betweenhost rocks and magma and depends on the pres-sure, temperature, and composition of both hostrocks and magma [Ryan, 1994] and also on thetectonic stress [Watanabe et al., 1999].

[4] The study of the geophysical anomalies causedby intruding magma (i.e., seismicity, deforma-tions, and gravity variations), is traditionally usedto model the location of the storage systembeneath a volcano, the dynamics of magma ascentand the mass involved. However, in the absence ofa reliable model for the composition and densityof the crust, the application of geophysical meth-ods cannot provide precise information about thelocation of magma ponding sites, as is the case ofthe Azores. An alternative way of obtaining thepressure of magma accumulation at various levelsconsists of a combination of petrological studieson magma evolution, with uid inclusion baromet-ric studies [e.g., Roedder, 1983; Belkin and DeVivo, 1993; Andersen and Neumann, 2001; Hans-teen et al., 1998; Klugel et al., 1997; Schwarz etal., 2004]. Previous studies have shown that baro-metric conditions of magma storage and ascentcan be estimated from densities of uid inclusionstrapped in phenocrysts and xenoliths [e.g., Zanonet al., 2003; Zanon and Nikogosian, 2004;Frezzotti and Peccerillo, 2004; Peccerillo et al.,2006; Klugel et al., 2005].

[5] Studies on the magmatism of the Azoresislands focused on the geochemical characteriza-tion of the mantle source [e.g., Beier et al., 2008,2010; Elliott et al., 2007; Claude-Ivanaj et al.,2001; Franca et al., 2006; Millet et al., 2009;

Moreira et al., 1999; Schaefer et al., 2002; Turneret al., 1997; Widom et al., 1997; Widom and Far-quhar, 2003]. Little is known about the shallowevolution of magmas in central eruptive systemsand associated rift zones [Self and Gunn, 1976;Metrich et al., 1981; Beier et al., 2006; Snyder etal., 2004; Snyder et al., 2007; Widom et al., 1992;Widom et al., 1993]. The few existing barometricdata are from the islands of Pico [Franca et al.,2006] and Corvo [Larrea et al., 2012], and the vol-canoes of Sete Cidades and Agua de Pau, in theisland of S~ao Miguel [Mattioli et al., 1997;Renzulli and Santi, 2000; Beier et al., 2006]. Arst attempt to dene the evolution history and thedifferent ascent paths of magmas from the ssurezones and a central volcano has been recently pro-posed for the Island of Faial [Zanon et al., 2013].

[6] In this study, barometric data were obtainedthrough the microthermometric study of the uidinclusions hosted in the lavas and xenoliths fromPico and Faial, following the approach of Zanon etal. [2003]. These two islands are the emerged topsof a larger volcanic edice, resulting from theinterplay of ssural volcanism and eruptions fromcentral volcanoes. The style and typology of mag-matism is apparently linked to the interactionbetween the activity of normal fault systems withthe main regional transextensional trend. The pres-ent data provide the rst evolutionary model ofmagma ow and storage beneath these twoislands. This model outlines how the interplaybetween tectonics and magma supply rate inu-enced the ascent and ponding of mac magmas atdepth, and the geometry of the plumbing systembeneath the different volcanic systems.

2. Geological Background

[7] The Azores archipelago consists of nine mainislands and several islets located in the AtlanticOcean and extends from northeast to southwest,between latitudes 37N and 40N (Figure 1). Mag-matism in this area presumably started in the lowerMiocene [Luis and Miranda, 2008], with the for-mation of a submarine plateau. The latter iscrossed by important tectonic structures: the Mid-Atlantic Ridge, the East Azores Fracture Zone andthe Terceira Rift. Georgen and Sankar [2010]reviewed the geodynamic importance of thesestructures and on their relationship with the mag-matism in the Azores area.

[8] The islands are the emerged tops of larger vol-canic elongated ridges that crop up from the

ZANON AND FREZZOTTI : MAGMA STORAGE BENEATH PICO AND FAIAL 10.1002/ggge.20221

2

submarine plateau; their subaerial eruptionsoccurred since at least 7.1 Ma ago in the island ofSanta Maria [Abdel-Monem et al., 1975]. A tem-poral gap of about 5 Ma separates the formation ofthese rocks from those cropping out in Floresisland, while all the other islands were apparentlyformed at least 1.3 Ma ago [Caniaux, 2005]. Sev-eral eruptions have taken place since the coloniza-tion of the archipelago (i.e., around A.D. 1427Zbyszewski, 1963) in the islands of Faial, Pico,S~ao Jorge, Terceira and S~ao Miguel. The last erup-tion was submarine and occurred in 19982001,20 km offshore Terceira.[9] The islands of Faial and Pico developed alonga WNW-ESE alignment and formed in a period of850 ka [Hildenbrand et al., 2012] and about 240ka [Madeira and Brum da Silveira, 2003], respec-tively. These islands originated from the interplaybetween ssure volcanism and eruptions fromcomposite volcanoes. Fissure volcanism occurredalong a few-kilometer-long subparallel clusters offractures, whose alignment is generally similar tothat of the other nearby islands (Figure 1). Lavaaccumulation along these fractures caused the for-mation of volcanic ridges that consist of severalrows of cinder cones. In the island of Faial, the

most developed system is the Capelo Fissure Zonein the western part of the island, where it climbsalong the western slope of the central volcano ofCaldeira. Its activity, which started presumablyabout 6 ka ago, includes also the eruptions of 1672and 19571958. The Horta Fissure Zone, in thesoutheastern part of the island is less extended(only 4 km long). The activity of this system hasnot been dated, however the presence of 16 ka oldpyroclastic fallout generated by the Caldeira Vol-cano, covering the Horta lavas, excludes magmaticactivity during the Holocene. Poorly evolved mag-mas, from basalt to hawaiites, poured out fromthese two ssure zones [Metrich et al., 1981;Zanon et al., 2013].

[10] The central volcano of Caldeira is a broad edi-ce topped by a deep circular caldera. Its mainmorphological feature is the presence of the PedroMiguel graben, which dissects the whole edicefrom WNW to ESE. This edice rises whereN150 striking tectonic systems interacted withssure systems and it is possible that this interac-tion favored the formation of short-lived andshallow-level magma reservoirs, leading to theestablishment of a centralized feeding conduit or aseries of closely spaced feeding dykes [Miranda et

Figure 1. Digital Elevation Model of the islands of Faial and Pico, showing the principal geographic fea-tures, historic eruptions, main faults (in white), and ssure systems (yellow dashed lines). Tectonic data arefrom Madeira [1998] and Madeira and Brum da Silveira [2003]. Fissure zones are labeled as follows: CFZCapelo Fissure Zone; HFZHorta Fissure Zone; PAFZPlanalto da Achada Fissure Zone. Stars indicatesampling sites. Coordinates reported in the grid must be multiplied by 10,000. The inset shows a general loca-tion map of the Azores archipelago and the main geostructural lineaments. Acronyms are as follows:MARMid-Atlantic Ridge; TRTerceira Rift ; EAFZEast Azores Fracture Zone; GFGloria Fault.


3

al., 1998]. Basalt-to-trachyte magmas [Metrich etal., 1981; Zanon et al., 2013] were erupted by thisvolcano starting from 120 ka ago [Hildenbrandet al., 2012], and its last eruptive event took place1.2 ka ago [Madeira, 1998].[11] Pico is an elongated island characterized bythe presence, in its western part, of the steep PicoMountain (2300 m asl). This 240 ka old centralvolcano [Chovelon, 1982] has an almost perfectlycircular base and numerous lateral vents which areeither isolated or arranged along small radial s-sures that punctuate its anks. Two areas of partialsector collapses are located north and south of thesummit area. On its summit, there is a small andyoung caldera, partially lled with the lavas emit-ted by a steep and tall nested hornito. A well-evident NW-SE trending fault intersects the wholeedice, passing through the summit caldera. Thisfault drove magma toward the surface during theA.D. 1718 eruption. Another eruption from thisvolcano occurred along a more southeasterlylocated radial fracture, in A.D. 1720.

[12] East of the Pico Volcano, there is the Planaltoda Achada Fissure Zone. This structure is similarto the Capelo Fissure Zone and follows the samegeographic direction, but it is considerably longer(30 km). About 170 cinder cones, some of whichhave a considerable size, constitute this ridge.The oldest lava ow unit dated here is 230 ka old[Chovelon, 1982], but magmatism continued untilthe last eruption in 1562.

[13] Both the Pico Mountain and the Planalto daAchada Fissure Zone erupted poorly evolved mag-mas, whose composition ranges from basalt tohawaiites. The only exception regards a mugeariteerupted in 1718 [Franca et al., 2006].

3. Methods

[14] Activation Laboratories (Canada) performed24 whole-rock analyses of major elements usinginductively coupled plasma (ICP). Reproducibilitywas generally better than 1%. Densities of bulkxenoliths and lava samples were measured makinguse of a MD 200s electronic densimeter and cor-rected for porosity. The density resolution was is0.001 gcm3.[15] Electron microprobe analyses of mineralphases in lava and xenolith samples were obtainedwith a JEOL JXA 8200 Superprobe, equippedwith ve wavelength-dispersive spectrometers,Energy Dispersive X-ray spectroscopy (EDS), and

cathodoluminescence detectors, (University of Mi-lan). A spot size of 1 m with a beam current of15 nA was considered for all the mineral phases,whereas a spot size of 57 m, according to theavailable surface to be analyzed, and a beam cur-rent of 2 nA were applied to glasses. Countingtimes were 30 s on the peak and 10 s on each back-ground. Natural and synthetic minerals andglasses, used as standards, were calibrated within2% at 2 standard deviation. Raw data were cor-rected applying a Phi-Rho-Z quantitative analysisprogram. The typical detection limit for each ele-ment is 0.01%. All the geochemical data are in thesupporting information (Table S3).

[16] Fluid inclusions were studied in 31 lava andsix xenolith samples. About 0.51 kg of each lavasample was coarsely crushed up to obtain the nalgrain size of the largest crystal that could be visi-ble at a macro scale. About 150/200 olivines and100/150 clinopyroxenes were embedded in epoxyand doubly polished up to attain a nal thicknessof about 12080 m. For xenoliths, 100130 mthick doubly polished wafers were prepared.

[17] Microthermometry was carried out on aLinkam MDSG600 heating-cooling stage,calibrated according to synthetic uid inclusionstandards of pure CO2 and H2O. Melting andhomogenization temperatures are reproducibleto 60.1C with heating rates in the range of0.20.5C/min.

[18] Fluid inclusions were further analyzed with aconfocal Labram multichannel spectrometer ofJobin-Yvon Ltd. at the University of Siena (Italy).An Ar laser produced the excitation line at 514.5nm. The Raman intensity was measured with apeltier-cooled carbonate compensation depth(CCD) detector. The beam was focused to a spotsize of about 12 m using an Olympus 100lens. The scattered light was analyzed using aNotch holographic lter with a spectral resolutionof 1.5 cm1 and grating of 1800 grooves/mm.Mineral identication was based on our spectraldatabase [Frezzotti et al., 2012a]. Raman identi-cation of optically hidden water (i.e.,

GPa. For CO2-H2O inclusions (H2O:CO2ratio 1:9), densities were calculated on the basisof the equation provided in Sterner and Bodnar[1991], after the application of the uid densitycorrection suggested by Hansteen and Klugel[2008]. Isochores were calculated using the soft-ware Fluids [Bakker, 2003]. The microthermo-metric data are shown in the supportinginformation (Table S4).

4. Description of the Samples

[20] Fluid inclusions were studied in the lava owsemitted from both the ssural systems and fromthe central volcanoes of the islands of Pico andFaial. Most of the samples date back to the Holo-cene age and were collected from quite well recog-nizable volcanic units.

[21] All the lavas ranged in composition from ba-salt to hawaiite and presented different degrees ofporphyricity. Three mugearites from the centralvolcanoes were collected in order to check if theevolution of magma was associated to changes inthe feeding systems. We also carefully examinedthe ultramac xenoliths found in some lava ows.

4.1. Central Volcanoes

[22] On the slopes of the central volcano of Picowe collected 16 samples from recent lava owunits, issued from isolated vents far from the sum-mit cone, or from vents located along radial s-sures connected to the main feeding system. Onelava sample issued from the summit cone was col-lected on the southwestern rim of the caldera. Allsamples except one were younger than 40 ka (sup-porting information, Table S1).

[23] Most of these samples had a very high crystalcontent and were macroscopically characterizedby the presence of a large proportion of mac phe-nocrysts ( 1.56.0 mm) and megacrysts (up to 25mm in length), totaling between 25 and 70% involume. Plagioclase was sometimes present onlyas phenocrysts of 812 mm in size. Megacrystswere commonly either widespread inside lavaows or concentrated in discontinuous layers orlenses. In a single case, they appeared as patchesof a dense network of euhedral and equidimen-sional crystals, physically linked to one another.The historical lavas had a variable degree of por-phyricity. Those collected from the southern ventsof the 1718 eruption were macroscopically charac-terized by the presence of some megacrysts of oli-

vine and clinopyroxene and few phenocrysts.Lavas issued from the northwestern vents showedno megacrysts, but, on the contrary, there werenumerous phenocrysts of plagioclase, olivine, cli-nopyroxene and many gabbroic and cumuliticxenoliths. The lava emitted in 1720 was poorlyporphyritic with few plagioclases and macphases.

[24] In Faial, the whole surface of the central vol-cano of Caldeira exhibited a thick pyroclasticcover produced during its Holocene activity. Twolavas were therefore collected from rare accessibleoutcrops along the coast and from some gullies onthe southeastern ank of the volcano. These sam-ples were poorly porphyritic lavas issued during aperiod starting from 120 ka ago. The rst oneshowed similar characteristics to the southern his-toric lavas from Pico, while the second rock hadonly few macroscopic plagioclases.

4.2. Fissure Systems

[25] All the lavas and pyroclasts collected from allthe ssure zones had in general very similar char-acteristics, i.e., they were poorly porphyritic witha phenocrysts content (olivine, clinopyroxene6plagioclase) amounting to less than 20% in vol-ume and were devoid of megacrysts.

[26] The bulk rock composition ranged from basaltto hawaiite. Some lavas with a higher crystal con-tent were collected only in the Capelo FissureZone area, where numerous phenocrysts ofolivine were scattered over the ow and did notform any evident accumulation texture. In anycase, only a lava sample that contained somexenoliths had a crystal content comparable to thatof the highly porphyritic lavas from Pico centralvolcano.

[27] In the island of Pico, seven samples were col-lected along the axial ridge of the Planalto daAchada Fissure Zone and from the vertical cliff onthe northern coast, where a succession of thinlavas is capped with the A.D. 1562 ow. With theexception of a single sample, which is older than50 ka, all the other rocks are of the Holocene orhistoric.

[28] In the island of Faial eight rocks collectedalong the ssure zones come from scoria cones orassociated lava ows and were representative ofthe whole activity which developed there. Thesamples from Capelo Fissure Zone are youngerthan 6 ka, those from Horta Fissure Zone are olderthan 11 ka.


5

4.3. Ultramafic Xenoliths

[29] Ultramac xenoliths were only found in lavasfrom Planalto da Achada Fissure Zone, in Picoand Capelo Fissure Zone, in Faial. In both cases,these lava ow units cropped out at the morpho-logical boundary between the central volcano andthe ssure zone. The ow from Planalto daAchada Fissure Zone (Pico Isl.) is older than 50ka. It crops out at the base of the northern cliff,covered by the A.D. 1562 lava, at a distance of15 km from the crater of the central volcano.However, its source vent can be reasonablylocated close to that of 1562, at a short distancefrom the eastern slopes of the central volcano, i.e.,at the boundary between a central volcano and assure zone. The host lava was poorly porphyriticand contained olivines and clinopyroxenes. Thesize of xenoliths did not exceed 34 cm.

[30] The ow from Capelo Fissure Zone (FaialIsl.) was a crystal-rich lava which cropped out onthe northern coastal cliff. This lava is not dated;however, due to the existing dating of the Capelorocks, we can assume that this ow is youngerthan 6 thousand years. Xenoliths were composi-tionally heterogeneous, with the typical size 20cm and had either a rounded or an angular shape.

5. Petrographic Characteristics

5.1. Lava and Scoria Samples

[31] All the samples showed the same petrographicand geochemical characteristics. The only differ-ence was represented by the total crystal contentthat in rocks from ssure zones did not exceed19% in volume, while in the lavas from the centralvolcano of Pico, it ranged between 25% and 70%(supporting information, Table S2). These latterlavas, known as ankaramites (i.e., dark porphyriticmac lavas with a high content of pyroxene andolivine crystals and minor amounts of plagio-clase), contained also numerous mac megacrysts.Lavas of this kind are common on many otherislands of the archipelago and have already beendescribed for the islands of Corvo [Larrea et al.,2012] and Faial [Zanon et al., 2013].

[32] Collected samples ranged in compositionfrom basalt to mugearites (Figure 2; supporting in-formation, Table S3) with mineral assemblageconstituted by olivine clinopyroxene6 plagio-clase6 oxides. The typical textures were intergra-nular and intersertal. Rarer hyaloophitic and

cryptocrystalline textures were also observed (Fig-ure 3a).

[33] Olivine phenocrysts (rarely exceeding 2 mmacross) were euhedral, subhedral, and/or skeletal ;their core Mg number [Mg# 100Mg/(MgFe)]ranged, on average, from 77 to 82 (Figure 4). Onlythe last tens of microns at their rim were Feenriched. Euhedral and zoned augites (Wo3142,En3147, Fs314) were always present in all ssurezones, with the exception of magmas from theHorta Fissure Zone (Faial Isl.), where their pres-ence was rarer. Their Mg# values ranged from 72to 86. The negative correlation between Ti and Sievidenced in the diagram of Figure 5a, indicatedthe occurrence of a common crystallization pathfor all the samples collected. The AlIV/AlVI ratiodecreased irregularly with titanium, indicating dif-ferent and variable pressures of crystallization ofclinopyroxenes from microphenocrysts to mega-crysts [Wass, 1979] (Figure 5b). Plagioclase wasubiquitous in all ssure zones while scarce orabsent in the central volcano of Pico. Crystalswere typically euhedral, tabular and with oscilla-tory or reverse zoning and with composition frombytownite to labradorite [Franca et al., 2006;Zanon et al., 2013].

[34] Olivines, clinopyroxenes and plagioclases,plus ilmenites or titanomagnetites or both, com-monly represented microphenocrysts. Amphibole( 2.5 mm) was present in the lavas emitted dur-ing the A.D. 1718 eruption from the northern seg-ment of the ssure. Apatite was frequently foundintimately associated with clinopyroxenes and/or

Figure 2. Chemical classication of collected lava samples.Open versus full symbols distinguish samples between centralvolcanoes and the ssure zones. Light-colored area representsthe entire compositional eld from existing literature [Zanonet al., 2013; Franca et al., 2006].


6

amphiboles. Nepheline was occasionally found insome basalts from Pico Volcano.

[35] The megacrysts of olivines and clinopyrox-enes (Figure 3b) were commonly euhedral or sub-hedral or anhedral, and showed disequilibriumfeatures, such as reaction rims and embayments. Itwas sometimes very difcult to discriminatebetween phenocrysts and megacrysts especiallyfor olivines, due to overlapping compositions andcrystals size.

[36] A comparison was made between the forster-ite content (Fo) in phenocrysts and in megacrystsand the composition of olivines in equilibrium

Figure 3. Microphotographs of representative samples. (a)Poorly porphyritic lava of the Horta Fissure Zone, showingtwo small and partly iddingsitized olivine dispersed amongplagioclase microphenocrysts. (b) Highly porphyritic lavasample from the Pico summit cone. It is evident the presenceof large euhedral megacrysts of olivine and clinopyroxene. (c)Wherlite sampled from a poorly porphyritic lava of the ssurezone of Pico. The presence of numerous trails of uid inclu-sions at the centre of the main is perfectly visible at the centreof the clinopyroxene.

Figure 4. Comparison between the theoretical conditions ofolivine crystallization and the measured forsterite content. (a)The variation of forsterite of megacryst (in gray) and of phe-nocrysts (in black), hosted in various samples of the twoislands. (b) The black curve represents the value of forsteritein equilibrium with coexisting silicate melt at various pres-sures. This curve was simulated in absence of water and underthe QFM buffer, with MELTS software [Ghiorso and Sack,1995], starting from a primary melt composition of FaialIsland, recalculated from the sample Fys 301, a poorly por-phyritic basalt from Horta Fissure Zone (Faial Isl.), alreadydescribed in Zanon et al. [2013]. The primary melt was recal-culated after equilibrating the sample composition with Fo89,using the geothermometers of Lee et al. [2009]. The resultingconditions of melting were used as input for fractional crystal-lization simulations (i.e., 1450C and 25 kbar). The segmentof the liquid line of descent shown is relative to the litho-sphere thickness (41.9 km) simulated beneath the island ofFaial [Dasgupta et al., 2010].


7

along the liquid line of descent of a recalculatedprimitive magma composition (Figure 4). Fo wasgenerally higher in megacrysts than in phenocrysts(Figure 4a). However, many phenocrysts alsoshowed disequilibrium conditions with the melt.For this reason, both of them should be consideredas antecrysts [Davidson et al., 2007], i.e., crys-tals, which may have formed from a different andmore primitive magma than the present host.

[37] Similar considerations were made for clino-pyroxenes. Both megacrysts and phenocrysts fol-lowed the same evolution path (Figure 5a) andwere therefore comagmatic. However, Mg# inmegacrysts was higher and less variable (i.e.,7990) and the variation of the AlIV/AlVI ratiowith Ti was limited, which evidenced higher pres-sure conditions during clinopyroxene crystalliza-

tion (Figure 5b) and made it possible to consideralso those crystals as antecrysts. However, the pa-rental link between antecrysts and phenocrystswas not relevant for the purposes of the presentwork; therefore, in the following discussion, wecontinued to distinguish megacrysts fromphenocrysts.

[38] Mugearites were subaphyric with trachytictexture and phenocrysts assemblages constitutedby numerous oligoclase-to-Na-sanidine feldspars(> 4 cm), rare olivines (> 0.15 cm) (Fo5760)and kaersutite amphiboles. The same phases werepresent as microphenocrysts together with clino-pyroxene, apatite, and oxides (both ilmenite andtitanomagnetite).

5.2. Ultramafic Xenoliths

[39] Ultramac xenoliths were coarse-grained lith-ologies made up of olivine, clinopyroxene, ortho-pyroxene, and spinel in various amounts (Figure3c). Plagioclase and plagioclase-and-amphibole-bearing lithologies were also frequent but have notbeen described here, since they did not containuid inclusions. Mantle xenoliths consisted ofthree harzburgites from Pico, and one websteriteand one wherlite from Faial.

[40] All peridotites showed porphyroclastic toequigranular textures, where olivine and orthopyr-oxene grains were present as large porphyroclasts(up to 4 mm) and polygonal grains typically< 1mm. Clinopyroxene was present as polygonalgrains in the wherlite and as interstitial grains inharzburgites. Spinel was interstitial (up to 0.5 mmin diameter) or present as inclusions in olivine.Plagioclase was absent; all studied xenoliths equi-librated in the spinel stability eld.

[41] Olivine was the dominant phase in all perido-tite samples. In harzburgites, it had Mg# rangingfrom 88 to 90.5, NiO contents ranging from 0.26to 0.47 wt% and low CaO (

processes. These were similar to ultrarefractorymantle xenoliths previously reported from S~aoMiguel [Simon et al., 2008]. In contrast, the pyrox-enite and the wherlite from Faial had compositionsthat are more fertile and appeared to representmelts crystallized or cumulated at mantle depth,but not necessarily in a magma chamber.

6. Study on the Fluid Inclusions

[43] In all the samples the uid contained in theinclusions was rich in CO2. The distribution ofuid inclusions varied considerably among thesamples. In general, they could be frequentlyfound in ultramac xenoliths, but they were foundin 1020% of megacrysts of olivine, whereasthey were scarce in 1% of phenocrysts. Theywere also found in 5% of clinopyroxene, bothmegacrysts and phenocrysts, and in some plagio-clase phenocrysts.

6.1. Petrographic Analysis

[44] On the basis of their textural characteristics,two populations of uid inclusions were observed.Early carbonic inclusions (Type I) were either iso-lated inclusions or in small and spatially denedclusters, located both inside the grains and at itsboundaries (Figure 6a). This kind of inclusions wasrarely found in few olivine and clinopyroxenegrains in highly porphyritic lavas of Pico Volcanoand in a single olivine from a poorly porphyriticlava of the Horta Fissure Zone (Faial Isl.). Theinclusions were generally rounded or rarely nega-tive crystal shaped, with a size 2030 m. Atroom temperature, Type I inclusions were singlephase (L) or might contain a vapor bubble (LV).Their texture often showed evidences of partialdensity re-equilibration, such as haloes of tiny uidinclusions (< 1 m) surrounding a main inclusioncavity, and/or short cracks radiating from themicrocavity [Viti and Frezzotti, 2000, 2001]. A fewinclusions, contained in a single olivine megacrystfrom a single sample of the volcano of Pico hosteda tiny ( 5 m) Mg-calcite [(Ca,Mg)CO3Fig-ure 7a], identied by Raman microspectroscopy byits diagnostic vibrations at 714 and 1087 cm1

[Frezzotti et al., 2012a]. In another group of inclu-sions of the same samples, the presence of dypin-gite (Mg5[(OH)(CO3)2]25H2O [Frost et al., 2008]was identied for its main Raman vibration at 1122cm1Figure 7b.

[45] Late carbonic uid inclusions (Type II) weredominant in all the studied samples, where they

Figure 6. Photomicrographs of the different types of uidinclusions hosted in olivines, and their textures. (a) Olivinefrom the 1718 eruption in Pico Island, showing a cluster ofType I uid inclusion with decrepitation features (halo of tinybubbles), marked by red arrows. (b) Olivine from a highlyporphyritic lava from the central volcano of Pico, showingnumerous Type II diffuse uid inclusions, coexisting withseveral silicate melt inclusions of different size. (c) Olivinefrom a dunite fragment from Faial Island, showing varioustrails of tiny (215 mm) Type II uid inclusions.


9

formed trails of variable length and thickness,which lined completely healed fractures, limitedor crosscutting grain boundaries (Figures 6b and6c). The inclusions were typically 30 m, usu-ally rounded (Figure 6c) or with a negative-crystal shape. At room temperature, Type IIinclusions were single phase (L) or two phases(LV), commonly liquid rich. Many inclusionsshowed evidence of re-equilibration and coex-isted with numerous silicate melt inclusions witha variable size and crystallization degree. Thesesilicate melt inclusions may contain a largeCO2-rich bubble.

6.2. Composition

[46] On cooling, CO2-rich inclusions frozebetween 68 and 92C. A further decrease oftemperature did not produce any visible phasechange. Most of the inclusions melted instantane-ously within a temperature interval between 57.1and 56.4C with most of the data at 56.6C. Clath-rates were not observed and nal homogenizationoccurred to either the liquid (ThL) or the vaporphase (ThV) at

[47] The identication of dypingite by Ramanspectroscopy in a few Type I inclusions in olivinemegacrysts suggests that some water was origi-nally present in these inclusions, and reacted withthe host olivine to form a hydrous carbonate atlater stages. The determination of the originalwater content in Type I inclusions is difcult sincethe H2O content of the magma at the time ofdegassing in not known. Different lines of evi-dence suggest that a plausible upper limit could be10 mole % [Hansteen and Klugel, 2008]. Water issubordinate in CO2-H2O uids released by basalticmagmas and generally around 10 mole % in mag-mas from intraplate settings [Dixon and Stolper,1995]. In addition, thermodynamic modeling ofthe fosteritic olivineCO2-H2O uid reaction oncooling as inclusions at 0.1 GPa suggests similarwater contents, since talc would have formedalong with carbonates in CO2-H2O inclusions withXH2O higher than 0.1 (Figure 7) [cf. Frezzotti etal., 2012b].

6.3. Density

6.3.1. Xenoliths[48] Type II inclusions homogenized to liquidphase (ThL; LV !L) from 10 to 30.4C(d 0.700.81 gcm3). Only one inclusion ho-mogenized to vapor phase (ThV; LV !V) withThV 29.4, corresponding to a density of 0.32gcm3. The histogram of homogenization temper-atures shows data from each island (Figure 8).These data formed a polymodal distribution,depending mainly on the re-equilibrated inclu-sions. A comparison of the densities of the inclu-sions in the xenoliths of the two islands is shownin Figure 9a.

6.3.2. Poorly Porphyritic Lavas From FissureZones[49] Type II uid inclusions from both the Capelo(Faial Isl.) and Planalto da Achada Fissure Zones(Pico Isl.) were in olivines and clinopyroxenes andhomogenized to the liquid phase from 16.9C to30.6C (d 0.560.81 gcm3). Those from oli-vines of Horta Fissure Zone (Faial Isl.) were bothof Type I and Type II. Type I inclusions homoge-nized to the liquid phase from 20.4 to 28.3C, cor-responding to densities of 0.77 to 0.65 gcm3. Afew re-equilibrated inclusions gave higher andscattered Th values. Type II inclusions homoge-nized to the liquid phase within a short interval oftemperature (i.e., 20.525.4C), corresponding todensities of 0.710.77 gcm3. Density valuesrelated to inclusions in ssure zones are comparedin Figure 9b.

6.3.3. Poorly Porphyritic Lavas From the CaldeiraVolcano[50] The uid inclusions found at this volcanowere only in two olivines and three plagioclasesand did not show re-equilibration textures. In oli-vine, Type I homogenized to the liquid phase from16.4 to 24.3C (d 0.720.81 gcm3). A singleType I inclusion gave a ThV of 30.9

C (d 0.42gcm3). In plagioclase from a mugearite, Type IIinclusions homogenized to the vapor phase from23.1 to 30.9C (d 0.220.42 gcm3; Figure 9).6.3.4. Highly Porphyritic Lavas From Pico[51] In all megacrysts of Pico Volcano (both clino-pyroxene and olivine) from recent eruptions TypeI uid inclusions were rare and sometimes coex-isted with Type II, silicate melt and cubic oxides.Re-equilibration features were common. Homoge-nization of Type I pure CO2 uid occurred to theliquid phase between 28.3 and 31C (d 0.520.65 gcm3). Type I H2O-CO2 uid homogenizedto the liquid phase between 24.9 and 31C, corre-sponding to a corrected density of 0.540.74gcm3 (Figure 9c).[52] Type II CO2 uid inclusions were found bothin xenocrysts and in phenocrysts and homogenizedto the liquid phase between 23 and 31C(d 0.510.74 gcm3) and to vapor phasebetween 28.5 and 31C (d 0.320.47 gcm3).

7. Discussion

7.1. Significance of Fluid Inclusion Data

[53] Fluid inclusions were present in differentkinds of crystals, linked to different periods of thevolcanism in these islands and to the rocky bodiescrossed by the ascending magmas.

[54] The density histograms (Figure 9) showedsymmetric or skewed unimodal and bimodal distri-butions to lower density values related to variousextents of uid inclusions re-equilibration proc-esses. In general, the smaller-size inclusions showthe highest density (Figure 10). This kind of inclu-sion did not undergo re-equilibration, providinguseful information on the locations of the magmareservoirs [cf. Andersen and Neumann, 2001, andreferences therein]. Based on the density versussize of inclusion plot (Figure 10) we assume thehighest uid densities as representative for baro-metric studies.

[55] The ultramac xenoliths suite preserved thedensest inclusions, (about 0.86 gcm3). The


11

maximum value of CO2 uid density in the poorlyporphyritic lavas, issued from all the ssure zones,was within the range 0.770.81 gcm3.Conversely, in plagioclase maximum density was0.42 gcm3.

[56] Type I H2O-CO2 inclusions in highly porphy-ritic lavas range from 0.54 to 0.74 gcm3.

[57] Barometric information on the magma plumb-ing system, were obtained from isochore distribution

Figure 8. Histograms of the temperatures of homogenization of uid inclusions from Faial and Pico vol-canic systems. The number of the total measurements is reported to the upper left corner of each graph. Onlyin the case of the two ssure zones of Faial, data are superimposed.


12

in the P-T space, at the assumed trapping tempera-ture, as shown in Figure 11. Trapping temperaturewere calculated using the geothermometer basedupon the clinopyroxene-liquid equilibrium [Putirka,2008] and assuming that:

[58] 1. The resulting temperatures were represen-tative of the melt evolution stage occurring

between the crystallization of olivine and that ofclinopyroxene;

[59] 2. Magma ascent was rapid enough to be con-sidered isothermal, i.e., temperatures did not sub-stantially change.

[60] The conditions of clinopyroxene-liquid equili-bration were veried according to Duke [1976].When these conditions were lacking, a thermome-ter based on the composition of the clinopyroxenewas used. The values attained mainly concernedhighly porphyritic lavas from Pico and ranged

Figure 9. Histograms of the density values found in theseinclusions. Data are grouped in the three following classes:xenoliths, poorly porphyritic lavas, and highly porphyritic lavas.

Figure 10. Comparison between inclusions size and thedensity of the trapped uid. Measurements refer to variouspopulations of inclusions trapped in different olivine crystals.Partial resetting is evidenced by the existence of various classsizes for all the density values except for the highest.

Figure 11. Summary outline of the method used to get baro-metric information. The blue and green stars indicate theintersection of the isochores of the various volcanic systemsof Pico and Faial, respectively, with the calculated eruptivetemperature of 1150C. The variations of the textural charac-teristics of the inclusions, with the occurrence of decrepitation(re-equilibration) are also shown. Acronyms as from Figure 1.


13

from 1142 to 1240C. A single value of 1145Cwas obtained for poorly porphyritic lavas fromFaial, using the clinopyroxene composition (sup-porting information, Table S5).

[61] Published thermobarometric data reportedtemperatures from 1100 to 1200C and pressuresranging from 0.6 to 1.1 GPa for Pico magmas[Franca et al., 2006]. Faial literature data, basedupon the use of a clinopyroxene-melt equilibriumthermometer, reported 11171280C (610) forpoorly porphyritic lavas from ssure zones and1015C (620) for one mugearite [Zanon et al.,2013]. On the basis of this information and inconsideration of the data calculated for the highlyporphyritic lavas of Pico, a temperature of1155C (615) was assumed for all the poorlyporphyritic lavas of the two islands and for thexenoliths contained in them, supposing the exis-tence of a thermal equilibrium between the xeno-liths and the host magma. A temperature of1100C (610) was assumed for the mugearites of

Faial. It is worth noting that a difference in tem-perature of 610C does not signicantly affectthe resulting pressure values, which vary only by65 MPa. The resulting scheme of uid pressuresis shown in Figure 12.

[62] To convert barometric data into depths it wasnecessary to include information on the depth ofthe crust-mantle transition and make someassumptions on the density of the wall rocks.Detailed geophysical data from the local earth-quake tomography indicated a crustal thickness of12.5 km beneath Faial and from 13 to 14 kmbeneath Pico [Matias et al., 2007]. Regarding den-sities, the average values of 2.8 gcm3 and3.22 gcm3 were considered as being representa-tive of the lithologies of the crust and of the shal-low mantle respectively. The rst value was theaverage density among 15 samples of basalts fromFaial, while the second was the average among 13samples of ultramac lithologies from the twoislands (see analytical appendix).

7.2. Relative Timing of CrystallizationPath and Fluid Trapping

[63] MegacrystsThese large crystals werecomagmatic with phenocrysts; however, theircompositional and textural features indicated thatthey formed in an earlier stage of evolution of themagmas of the two islands. They could be eithergenetically linked to the host magma or not. Inany case, this was not relevant for the purposes ofthe present study. In some cases, these crystalsshowed textures typical of a solidication front[Marsh, 1996]. They hosted both early (Type I)and late (Type II) stage inclusions. Type I inclu-sions were related to the degassing of the hostmelt, which these crystals equilibrated with. TypeII inclusions were linked to a further degassingphase, not necessarily related to the same magma.All these inclusions re-equilibrated to the samepressure conditions, providing information onlyabout the last ponding stage, before eruption

[64] PhenocrystsThese were fractionated by thecarrier magma at a later stage than megacrysts.They also trapped both early and late stages inclu-sions related to the degassing of the host magma,and give the same barometric information of themegacrysts.

[65] Crystals from xenolithsBoth ultramac andcumulitic assemblages were present in the mantlebeneath these islands. The inclusions they trappedwere all of Type II and were related to the

Figure 12. Arrangement of the storage systems beneatheach of the volcanic systems in a W-E transect. Deep graycolor represents the xenoliths suites of the two islands, inter-mediate gray color represents the lavas of the various ssurezones and white color represents the lavas of the centralvolcanoes.


14

degassing of the carrier magma. The resulting in-formation was related only to the depth at whichthe magmas ponded at subcrustal depths.

[66] Combining the information obtained fromuid inclusions (textures, densities, relative distri-bution), petrography (textures and constitution ofrocks and minerals) and mineral chemistry ofmegacrysts and phenocrysts provides uniqueinsights into the magmatological processes thatoccurred at the mantle-crust boundary (Figure 13).

[67] Three distinct events of uid trappingoccurred at different times. The rst eventoccurred at depth during early crystallization ofmegacrysts of olivines (Mg# 7988) and led to thetrapping of Type I CO26H2O uids (XH2O 0.1).[68] While the magma ponded for a long period,the stretching/cracking of olivine and clinopyrox-ene (both megacrysts and phenocrysts) in the pres-ence of free CO2 caused the formation of a newgeneration of uid inclusions (Type II).

[69] Olivine phenocrysts formed under CO2 satu-ration conditions of the coexisting silicate melt,generating Type I inclusions in the olivines con-tained in poorly porphyritic lavas and in the cen-tral volcano of Faial. Decompression gradientscaused the complete resetting of these inclusionsbeneath the ssure zones, as the resulting densitieswere comparable with those of Type II.

[70] Finally, only a limited number of plagioclasescontained in mugearites from the central volcanoof Faial recorded the last event of CO2 uid trap-ping/re-equilibration. The low-density Type-IIinclusions in some uids Mg-poor olivines at PicoVolcano possibly result from re-equilibration dur-ing slow magma ascent.

7.3. Magma Storage SystemThe Effectof Tectonics

[71] Overall, our results suggest pressure condi-tions of trapping and/or re-equilibration,

Figure 13. Relationships among process of crystallization of xenocrysts and phenocrysts, occurrence oftrapping and re-equilibration of CO2 inclusions and textures found. The circle represents the various subpopu-lations of mineral phases that can coexist within the same sample, just prior to ascent through the crust. Thetwo possible pathways evidenced by the dashed lines are suggestive for the central volcano and the ssurezones.


15

characteristic of the very shallow mantle, clearlydue to the common occurrence of re-equilibration.A conceptual model of the magma storage processbeneath these two islands is shown in Figure 14.

[72] The deepest uid trapping was recorded onlyin ultramac xenoliths at pressures of 570586MPa, corresponding to depths of 20 and 21 km,beneath both Faial and Pico respectively. In thepoorly porphyritic lavas of the ssure zones and ofthe central volcano of Faial, trapping/re-equilibrat-ing events occurred at 465508 MPa (a depth of1618 km), whereas at Planalto da Achada FissureZone (Pico Isl.), at a maximum pressure of 488MPa (18 km). The highly porphyritic lavas of PicoVolcano trapped uids at 463 MPa (16.8 km).

[73] A small intracrustal ponding level beneath thecentral volcano of Faial is represented by uidinclusions trapped in plagioclases at 156 MPa(5.7 km). No crustal reservoirs formed during theascent of magmas beneath the ssure zones.

[74] Geophysical studies located a series of Vp/Vsanomalies beneath the Caldeira Volcano at a depthbetween 3 and 7 km [Dias et al., 2007], whichwere interpreted as small crystallized bodies. Thisshallow-level crustal reservoir was already presentat least 120 ka ago, while the volcano was issu-

ing poorly evolved material. However, the inclu-sions hosted in the mineral phases of a mugearitesuggested that this ponding area played an impor-tant role during the geochemical evolution of themagmas. The formation of this long-lived reser-voir was probably linked to the important roleplayed by tectonics in this zone of the island [e.g.,Madeira and Brum da Silveira, 2003]. Recently,Hildenbrand et al. [2012] proposed a model of thedevelopment and evolution of the islands throughsuccessive voluminous magma emissions (edicebuilding) followed by magmatic quiescence andgraben development (edice collapsing). Accord-ing to this model, the WNW-ESE regional trans-tensional tectonics was associated with ssuresystems, low degree of magmatic evolution andrapid magma ascent, while the activity of NE-SWand NW-SE system halted the ascent of macmagmas from depth, favoring the formation ofshallow magma storage systems. In this frame-work, the recent movements of the Pedro Miguelgraben [Hildenbrand et al., 2012], the frequentseismicity along a NW-SE direction and the highdegree of the evolution of the last products issuedby the central volcano during the last 16 ka indi-cated the end of magmatic activity of the volcanoand the onset of a new phase of morpho-tectonicprocesses.

Figure 14. Overview of the distribution of the magma reservoirs and the plumbing system of the insular sys-tem Pico-Faial, in a WNW-ESE oriented prole. The information resulting from the study of the xenoliths isshown in white. The three dates located close to magma reservoirs are representative for uid inclusionsrelated to samples of historical eruptions. Blue dotted line indicates the presence of NW-SE main fault inPico. The black lined and dotted lines indicate the geophysical Moho (data from Matias et al. [2007]) and thepetrographical Moho, estimated based on uid inclusions data. Stars indicate the focal depth of historicearthquakes.


16

[75] Only the magmas erupted from the NW sectorof the Pico Volcano possibly ponded at crustaldepths, at 194 MPa (7 km), No other magmas,even those erupted from the main conduit systemrecorded a trace of this crustal step. For this rea-son, it is evident that this reservoir was probablyshort lived and with a limited size. During theA.D. 1718 eruption a mugearite magma was issuedfrom a ssure located NW of the main cone. In thepresent study, there were no abundant macphases available for uid inclusion study to testfor evidence of ponding of the mugearites. Wenote that a more detailed investigation of uidinclusions in feldspars could provide this missinginformation. However, the two analyzed macsamples, related to the eruptive phases followingthe emission of mugearite, did not pond inside thecrust. This clearly evidenced that mac magmaascended rapidly through the crust, thus prevent-ing a partial or complete resetting of the alreadytrapped inclusions.

[76] The magma emitted during the followingeruption of Pico Volcano (i.e., December 1720)was slightly more primitive than the basaltserupted 2 years before (supporting information,Table S11). This magma followed a differentascent path and did not necessarily pass throughthe central conduit. Eruptions similar to that of1720 can be compared to the deep dyke fed fromMt. Etna [Corsaro et al., 2009]. These differentascent paths are thought to be common to manyother peripheral cones of this volcano. The NW-SE fault affected the main storage system of Pico,located at a depth of 12.516 km, (i.e., well belowthe mantle-crust boundary found by geophysicalmethods). The activity rate of this tectonic featureplayed a major role in the plumbing system of thisvolcano. It could have led either to the formationof a system of deep-rooted conduits, whichfavored the rapid ascent of deep-sited magma, orcould have temporarily favored the storage ofsmall pockets of magma at crustal levels, promot-ing the onset of differentiation processes.

[77] The differential displacements caused by enechelon faulting in the ssure zones in both theislands of Pico and Faial possibly favored therapid withdrawal of small volumes of magmastored at the depth of 1618 km. This means thatthese faults probably extended at depth and that

they might have crossed the rocks existing beneaththe Moho.

7.4. Magma Supply and Consequences forUnderplating

[78] To estimate the average magma supply to thecentral volcano of Pico, to the Planalto da AchadaFissure Zone (Pico Isl.) and to the Capelo FissureZone (Faial Isl.), it was necessary to know the av-erage amount of magma emitted during well-delimited periods.

[79] A rough estimation of magma erupted wasobtained based on measured lava volumes fromspecic areas of those volcanic systems [Nunes,1999; V. Zanon, unpublished data, 2013].

[80] If melt production in the mantle did not sub-stantially change between the islands of Faial andPico [i.e., the same mantle melting degreeBour-don et al., 2005], it was necessary to hypothesizedifferent time lapses between melting events inthese areas: melt production beneath the ssurezones would have occurred sporadically and notcontinuously, as for the volcano of Pico.

[81] Regarding the second kind of information,i.e., since the ratio of intruded versus extrudedmagma volumes for oceanic seamounts is between3:1 and 5:1 [Crisp, 1984], the volume of magmaintruded beneath Pico Volcano was twice that ofthe ssure zones (supporting information, TableS6).

[82] Here it followed that:

[83] 1. The different style of tectonic setting andthe extent of faults greatly inuenced the mantlemelting process and the storage of magmas atdepth. Beneath the volcano of Pico most of themagmatic arrivals from the source area werealmost immediately withdrawn, whereas beneaththe ssure zones, magma was stored at the base ofthe crust, and remained mostly nonerupted.

[84] 2. Accumulated magmas beneath ssurezones crystallized at depth, causing the progres-sive thickening of the crust and the formation of atransition zone that extended well below the geo-physical Moho. Magma ascent through thesebodies could only occur when extensional tecton-ics opened faults at considerable depths. Thegrowth of the transition zone from 12.5 to 18 kmdeveloped starting from at least 230 ka (i.e., theoldest age available for the subaerial lavas)beneath the Planalto da Achada Fissure Zone

1Additional supporting information may be found in the onlineversion of this article.


17

(Pico Isl.). This caused an average growth rate of0.024 m/yr.

[85] 3. Crystal mush (olcpx) left behind by with-drawn melts was widespread beneath the volcanoof Pico. An ascending magma had a high probabil-ity to interact with this matter, thus changing itsoriginal crystal content and forming highly porphy-ritic liquids. This working scheme cannot beapplied also to the central volcano of Faial, due tothe lack of highly porphyritic lavas and to the samecrustal thickness of the adjacent ssure zones.

8. Conclusions

[86] According to the existing petrological modelof magma generation and evolution beneath theislands of Pico and Faial, mantle melting occurredat a pressure range of 34 GPa (95130 km)before migrating toward the surface and crystalliz-ing mac phases between 0.8 and 1.6 GPa (2650 km) [Beier et al., 2012; Zanon et al., 2013].Traditional geothermobarometry studies failed tofurther dene the evolutionary conditions of mag-mas with regard to the depth characteristics of thecrust-mantle transition.

[87] The model of magma ponding suggested herethrough the study of uid inclusions can ll in thegap and complement the overall model of the gene-sis and ascent of magmas, providing a detailed pic-ture of the processes occurring at the Moho.Magmas underwent two-step ascent through thelithosphere, separated by a ponding period close tothe Moho. During the rst ascent phase, mac mag-mas began to crystallize olivine and exolved vola-tiles, which are trapped in growing mineral phases.These ascending magmas ponded, fractionated, re-equilibrated and exolved CO2 at Moho, possiblysaturating the fractures existing in the countryrocks. Regional transextensional faulting extendedup to these rocks, driving CO2 up to the surface, asis recognized also in some degassing areas of thearchipelago [e.g., Viveiros et al., 2010].

[88] The magma ponding period was long enoughto reset completely the rst generation of uidinclusions and to allow further crystallization andcarbon dioxide degassing, with the formation of asecond generation of uid inclusions.

[89] Either rapid magma extraction or underplat-ing/fractionation was determined by existing tec-tonic regime in that area. The formation of highlyporphyritic liquids and of ultramac residua was adifferent effect of this process. An extreme degree

of crustal thickening could ultimately cause thecessation of volcanic activity. This model may beconrmed by further studies on fossil ssure sys-tems on other islands of the archipelago.

[90] The general interpretation of these data sug-gest that in volcanic systems the dynamics of thefeeding of the eruptions from ssure zones differsfrom that related to central volcanoes, since it isconnected to the existing tectonic regime. Thismeans that in an extensional tectonic environment,central volcanoes and ssure zones, coexisting inthe same area, tap completely independent mag-matic reservoirs through independent plumbingsystems.

[91] The upward migration of magma reservoirsfound at La Palma Island [Klugel et al., 2005;Galipp et al., 2006] was not observed here. On thecontrary, the storage systems migrated constantlydownward due to the increasing difculty encoun-tered by the drainage of magma and its consequentcrystallization at depth. Therefore, the mechanismsuggested here seems to be the major responsiblefactor for the thickening of lithosphere in oceanicislands, in agreement with Hansteen et al. [1998].

Acknowledgments

[92] This work has been funded by the Fundac~ao para a Ci^en-cia e Tecnologia (project PTDC/CTE-GIX/098836/2008). Vit-torio Zanon was funded by the Fundo Regional para a Ci^encia,through grant 03.1.7.2007.1 (PROEMPREGO Operational Pro-gram and Regional Government of the Azores). A. Risplen-dente and S. Poli of the Ardito Desio University of Milan(Italy) are gratefully acknowledged for their assistance duringmicroprobe analyses. Thoughtful revision by Steele-MacInnisand editorial handling by Cin-Ty Lee improved the clarity andthe quality of this manuscript.

References

Abdel-Monem, A. A., L. A. Fernandez, and G. M. Boone(1975), KAr ages from the eastern Azores group (SantaMaria, S~ao Miguel and the Formigas Islands), Lithos, 8,247254.

Andersen, T., and E. R. Neumann (2001), Fluid inclusions inmantle xenoliths, Lithos, 55(1-4), 301320.

Bakker, R. J. (2003), Package FLUIDS 1, Computer programsfor analysis of uid inclusion data and for modelling bulkuid properties, Chem. Geol., 194, 323.

Beier, C., K. M. Haase, and T. H. Hansteen (2006), Magmaevolution of the Sete Cidades volcano, S~ao Miguel, Azores,J. Petrol., 47, 13751411.

Beier, C., K. M. Haase, W. Abouchami, M.-S. Krienitz, and F.Hauff (2008), Magma genesis by rifting of oceanic litho-sphere above anomalous mantle: Terceira Rift, Azores,


18

Geochem. Geophys. Geosyst., 9, Q12013, doi:10.1029/2008GC002112.

Beier, C., S. Turner, T. Plank, and W. White (2010), Apreliminary assessment of the symmetry of source composi-tion and melting dynamics across the Azores plume,Geochem. Geophys. Geosyst., 11, Q02004, doi:10.1029/2009GC002833.

Beier, C., K. M. Haase, and S. P. Turner (2012), Conditions ofmelting beneath the Azores, Lithos, 144-145, 111.

Belkin, H. E., and B. De Vivo (1993), Fluid inclusion studiesof ejected nodules from plinian eruptions of Mt. Somma-Vesuvius, J. Volcanol. Geotherm. Res., 58, 89100.

Bonelli, R., M. L. Frezzotti, A. Peccerillo, and V. Zanon(2004), Evolution of the volcanic plumbing system of Ali-cudi (Aeolian Islands): Evidence from uid inclusions inquartz xenoliths, Ann. Geophys., 47(4), 14091422.

Bourdon, B., S. P. Turner, and N. M. Ribe (2005), Partial melt-ing and upwelling rates beneath the Azores from a U-seriesisotope perspective, Earth Planet. Sci. Lett., 239(1-2), 4256.

Caniaux, G. (2005), Analyse statistique de la frequence deseruptions volcaniques aux Acores: Contribution alevaluation des risques, Bull. Soc. Geol. Fr., 176(1), 107120.

Chovelon, P. (1982), Evolution volcanotectonique des iles deFaial et de Pico, Archipel des Acores-Atlantique Nord, the`sede docteur thesis, 193 pp, Univ. de Paris-Sud, Cent.DOrsay, Paris.

Claude-Ivanaj, C., J. L. Joron, and C. J. Alle`gre (2001), 238U230Th226Ra fractionation in historical lavas from theAzores: Long-lived source heterogeneity vs. metasomatismngerprints, Chem. Geol., 176(1-4), 295310.

Corsaro, R. A., N. Metrich, P. Allard, D. Andronico, L. Mira-glia, and C. Fourmentraux (2009), The 1974 ank eruptionof Mount Etna: An archetype for deep dike-fed eruptions atbasaltic volcanoes and a milestone in Etnas recent history,J. Geophys. Res., 114, B07204, doi:10.1029/2008JB006013.

Crisp, J. A. (1984), Rates of magma emplacement and vol-canic output, J. Volcanol. Geotherm. Res., 20(3-4), 177211.

Dasgupta, R., M. G. Jackson, and C.-T. A. Lee (2010), Majorelement chemistry of Ocean Island BasaltsConditions ofmantle melting and heterogeneity of mantle source, EarthPlanet. Sci. Lett., 289, 377392.

Davidson, J., D. Morgan, B. Charlier, R. Harlou, and J. Hora(2007), Microsampling and isotopic analysis of igneousrocks: Implications for the study of magmatic systems, Ann.Rev. Earth Planet. Sci., 35, 273311.

Dias, N. A., L. Matias, N. Lourenco, J. Madeira, F. Carrilho,and J. L. Gaspar (2007), Crustal seismic velocity structurenear Faial and Pico Islands (AZORES), from local earth-quake tomography, Tectonophysics, 445(3-4), 301331.

Dixon, J. E., and E. M. Stolper (1995), An experimental studyof water and carbon dioxide solubilities in mid-ocean ridgebasaltic liquids, Part II: Applications to degassing, J. Pet-rol., 36(6), 16331646.

Duke, J. M. (1976), Distribution of the period four transitionelements among olivine, calcic clinopyroxene and mac sili-cate liquid: Experimental results, J. Petrol., 17(4), 499521.

Eason, D. E., and J. M. Sinton (2009), Lava shields and ssureeruptions of the Western Volcanic Zone, Iceland: Evidencefor magma chambers and crustal interaction, J. Volcanol.Geotherm. Res., 186(3-4), 331348.

Elliott, T., J. Blichert-Toft, A. Heumann, G. Koetsier, and V.H. Forjaz (2007), The origin of enriched mantle beneath S~ao

Miguel, Azores, Geochim. Cosmochim. Acta, 71(1), 219240.

Franca, Z. T. M., C. C. G. Tassinari, J. V. Cruz, A. Y. Apari-cio, V. Ara~na, and B. N. Rodrigues (2006), Petrology, geo-chemistry and SrNdPb isotopes of the volcanic rocks fromPico IslandAzores (Portugal), J. Volcanol. Geotherm.Res., 156(1-2), 7189.

Frezzotti, M. L., and A. Peccerillo (2004), Fluid inclusion andpetrological studies elucidate reconstruction of magma con-duits, Eos Trans. AGU, 85(16), 157159.

Frezzotti, M. L., and A. Peccerillo (2007), Diamond-bearingCOHS uids in the mantle beneath Hawaii, Earth Planet.Sci. Lett., 262, 273283.

Frezzotti, M. L., F. Tecce, and A. Casagli (2012a), Ramanspectroscopy for uid inclusion analysis, J. Geochem.Explor., 112, 120.

Frezzotti, M. L., S. Ferrando, F. Tecce, and D. Castelli(2012b), Water content and nature of solutes in shallow-mantle uids from uid inclusions, Earth Planet. Sci. Lett.,351-352, 7083.

Frost, R. L., S. Bahfenne, and J. Graham (2008), Raman spec-troscopic study of the magnesium carbonate minerals artiniteand dypingite, J. Raman Spectrosc., 40, 855860.

Galipp, K., A. Klugel, and T. H. Hansteen (2006), Changingdepths of magma fractionation and stagnation during theevolution of an oceanic island volcano: La Palma (CanaryIslands), J. Volcanol. Geotherm. Res., 155(3-4), 285306.

Georgen, J. E., and R. D. Sankar (2010), Effects of ridge ge-ometry on mantle dynamics in an oceanic triple junctionregion: Implications for the Azores Plateau, Earth Planet.Sci. Lett., 298, 2334.

Ghiorso, M. S., and R. O. Sack (1995), Chemical transfer inmagmatic processes: IV, A revised and internally consistentthermodynamic model for the interpolation and extrapola-tion of liquid-solid equilibria in magmatic system at elevatedtemperatures and pressures, Contrib. Mineral. Petrol., 119,197212.

Hansteen, T. H., and A. Klugel (2008), Fluid inclusion thermo-barometry as a tracer for magmatic processes, in Reviews inMineralogy and Geochemistry, edited by K. Putirka and F.Tepley, Mineral. Soc. of Am., Chantilly, Virginia.

Hansteen, T. H., A. Klugel, and H.-U. Schmincke (1998),Multi-stage magma ascent beneath the Canary Islands: Evi-dence from uid inclusions, Contrib. Mineral. Petrol., 132,4864.

Hildenbrand, A., F. O. Marques, A. C. G. Costa, A. L. R.Sibrant, P. F. Silva, B. Henry, J. M. Miranda, and P. Madur-eira (2012), Reconstructing the architectural evolution ofvolcanic islands from combined K/Ar, morphologic, tec-tonic, and magnetic data: The Faial Island example(Azores), J. Volcanol. Geotherm. Res., 241-242, 3948.

Ida, Y. (1995), Magma chamber and eruptive processes at Izu-Oshima Volcano, JapanBuoyancy control of magmamigration, J. Volcanol. Geotherm. Res., 66, 5367.

Klugel, A., T. H. Hansteen, and H.-U. Schmincke (1997),Rates of magma ascent and depths of magma reservoirsbeneath La Palma (Canary Islands), Terra Nova, 9(3), 117121.

Klugel, A., T. H. Hansteen, and K. Galipp (2005), Magma stor-age and underplating beneath Cumbre Vieja volcano, La Palma(Canary Islands), Earth Planet. Sci. Lett., 236, 211226.

Klugel, A., S. Schwarz, P. van den Bogaard, K. A. Hoernle, C.C. Wohlgemuth-Ueberwasser, and J. J. Kster (2009), Struc-ture and evolution of the volcanic rift zone at Ponta de S~aoLourenco, eastern Madeira, Bull. Volcanol., 71, 671685.


19

Larrea, P., Z. Franca, M. Lago, E. Widom, C. Gale, and T.Ubide (2012), Magmatic processes and the role of antecrystsin the genesis of Corvo Island (Azores Archipelago, Portu-gal), J. Petrol., 54, 769793.

Lee, C.-T. A., P. Luf, T. Plank, H. Dalton, and W. P. Leeman(2009), Constraints on the depths and temperatures of basal-tic magma generation on Earth and other terrestrial planetsusing new thermobarometers for mac magmas, EarthPlanet. Sci. Lett., 279, 2033.

Luis, J. F., and J. M. Miranda (2008), Reevaluation of mag-netic chrons in the North Atlantic between 35N and 47N:Implications for the formation of the Azores Triple Junctionand associated plateau, J. Geophys. Res., 113, B10105,doi:10.1029/2007JB005573.

Madeira, J. (1998), Estudos de neotectonica nas ilhas do Faial,Pico e S. Jorge: Uma contribuic~ao para o conhecimento geo-dina^mico da junc~ao tripla dos Acores, PhD thesis, 428 pp.,Univ. de Lisboa, Lisbon.

Madeira, J., and A. Brum da Silveira (2003), Active tectonicsand rst paleoseismological results in Faial, Pico and S.Jorge Islands (Azores, Portugal), Ann. Geophys., 46(5), 733761.

Marsh, B. D. (1996), Solidication fronts and magmatic evolu-tion,Mineral. Mag., 60, 540.

Matias, L., N. Dias, I. Morais, D. Vales, F. Carrilho, J. Ma-deira, J. Gaspar, L. Senos, and A. Silveira (2007), The 9th ofJuly 1998 Faial Island (Azores, North Atlantic) seismicsequence, J. Seismol., 11(3), 275298.

Mattioli, M., B. G. J. Upton, and A. Renzulli (1997), Sub-vol-canic crystallization at Sete Cidades volcano, Miguel,Azores, inferred from mac and ultramac plutonic nodules,Mineral. Petrol., 60, 126.

McLeod, P. (1999), The role of magma buoyancy in caldera-forming eruptions, Geophys. Res. Lett., 26, 22992302.

Metrich, N., H. Bizouard, and J. Varet (1981), Petrology ofthe volcanic series of Fayal, Azores, Bull. Volcanol., 44,7193.

Millet, M. A., R. Doucelance, J. A. Baker, and P. Schiano (2009),Reconsidering the origins of isotopic variations in Ocean IslandBasalts: Insights from ne-scale study of S~ao Jorge Island,Azores archipelago,Chem. Geol., 265(3-4), 289302.

Miranda, J. M., et al. (1998), Tectonic setting of the Azoresplateau deduced from a OBS survey, Mar. Geophys. Res.,20, 171182.

Moreira, M., R. Doucelance, B. Dupre, and C. J. Alle`gre(1999), Helium and lead isotope geochemistry in the Azores,Earth Planet. Sci. Lett., 169(1-2), 189205.

Nunes, J. C. (1999), A actividade vulca^nica na ilha di Pico doPlistocenico Superior ao Holocenico: Mecanismo eruptivo ehazard vulca^nico, PhD thesis, 357 pp., Univ. dos Acores,Ponta Delgada.

Peccerillo, A., M. L. Frezzotti, G. De Astis, and G. Ventura(2006), Modeling the magma plumbing system ofVulcano (Aeolian Islands, Italy) by integrated uid-inclusion geobarometry, petrology, and geophysics, Geol-ogy, 34(1), 1720.

Putirka, K. (2008), Thermometers and barometers for volcanicsystems, in Minerals, Inclusions and Volcanic Processes,Reviews in Mineralogy and Geochemistry, edited by K.Putirka and F. Tepley, pp. 61120, Mineral. Soc. of Am.,Chantilly, Virginia, USA.

Renzulli, A., and P. Santi (2000), Two stage fractionation his-tory of the alkali basalt-trachyte series of Sete Cidades vol-cano (S~ao Miguel Island, Azores), Eur. J. Mineral., 12(2),469494.

Roedder, E. (1983), Geobarometry of ultramac xenolithsfrom Loihi Seamount, Hawaii, on the basis of CO2 inclu-sions in olivine, Earth Planet. Sci. Lett., 66, 369379.

Ryan, M. P. (1993), Neutral buoyancy and the structure ofmidocean ridge magma reservoirs, J. Geophys. Res., 98,22,32122,338.

Ryan, M. P. (1994), Neutral-buoyancy controlled magmatransport and storage in mid-ocean ridge magma reservoirsand their sheeted-dike complex: A summary of basic rela-tionship, inMagmatic System, edited by M. P. Ryan, pp. 97135, Academic, New York.

Schaefer, B. F., S. Turner, I. Parkinson, N. Rogers, and C.Hawkesworth (2002), Evidence for recycled Archaean oce-anic mantle lithosphere in the Azores plume, Nature, 420,304307.

Schwarz, S., A. Klugel, and C. Wohlgemuth-Ueberwasser(2004), Melt extraction pathways and stagnation depthsbeneath the Madeira and Desertas rift zones (NE Atlantic)inferred from barometric studies, Contrib. Mineral. Petrol.,147(2), 228240.

Self, S., and B. M. Gunn (1976), Petrology, volume, and agerelations of alkaline and saturated peralkaline volcanicsfrom Terceira, Azores, Contrib. Mineral. Petrol., 54, 293313.

Sigmundsson, F., et al. (2010), Intrusion triggering of the2010 Eyjafjallajkull explosive eruption, Nature, 468, 426430.

Simon, N. S. C., E. R. Neumann, C. Bonadiman, M. Coltorti, G.Delpech, M. Gregoire, and E. Widom (2008), Ultra-refractorydomains in the oceanic mantle lithosphere sampled as mantlexenoliths at ocean islands, J. Petrol., 49(6), 12231251.

Snyder, D. C., E. Widom, A. J. Pietruszka, and R. W. Carlson(2004), The role of open-system processes in the develop-ment of silicic magma chambers: A chemical and isotopicinvestigation of the Fogo A trachyte deposit, S~ao Miguel,Azores, J. Petrol., 45(4), 723738.

Snyder, D. C., E. Widom, A. J. Pietruszka, R. W. Carlson, andH.-U. Schmincke (2007), Time scales of formation of zonedmagma chambers: U-series disequilibria in the Fogo A and1563 A.D. trachyte deposits, S~ao Miguel, Azores, Chem.Geol., 239(1), 138155.

Span, R., and W. Wagner (1996), A new equation of state forcarbon dioxide covering the uid region from the triple pointtemperature to 1100 K at pressures up to 800 MPa, J. Phys.Chem. Ref. Data, 25, 15091596.

Sterner, S. M., and R. J. Bodnar (1991), Synthetic uid inclu-sions; X, Experimental determination of P-V-T-X propertiesin the CO2-H2O system to 6 kb and 700 degrees C, Am. J.Sci., 291(1), 154.

Sterner, S. M., and K. S. Pitzer (1994), An equation of state forcarbon dioxide valid from zero to extreme pressures, Con-trib. Mineral. Petrol., 117, 362374.

Turner, S., C. Hawkesworth, N. Rogers, and P. King (1997),U-Th isotope disequilibria and ocean island basalt genera-tion in the Azores, Chem. Geol., 139(1-4), 145164.

Viti, C., and M. L. Frezzotti (2000), Re-equilibration of glassand CO2 inclusions in xenolith olivine: A TEM study, Am.Mineral., 85(10), 13901396.

Viti, C., and M. L. Frezzotti (2001), Transmission electron mi-croscopy applied to uid inclusion investigations, Lithos,55(1-4), 125138.

Viveiros, F., C. Cardellini, T. Ferreira, S. Caliro, G. Chiodini,and C. Silva (2010), Soil CO2 emissions at Furnas volcano,S~ao Miguel Island, Azores archipelago: Volcano monitoringperspectives, geomorphologic studies, and land use planning


20

application, J. Geophys. Res., 115, B12208, doi:10.1029/2010JB007555.

Wass, S. Y. (1979), Multiple origins of clinopyroxenes in al-kali basaltic rocks, Lithos, 12(2), 115132.

Watanabe, T., T. Koyaguchi, and T. Seno (1999), Tectonicstress control on ascent and emplacement of magmas, J. Vol-canol. Geotherm. Res., 91, 6578.

Widom, E., and J. Farquhar (2003), Oxygen isotope signaturesin olivines from S~ao Miguel (Azores) basalts: Implicationsfor crustal and mantle processes, Chem. Geol., 193(3-4),237255.

Widom, E., H.-U. Schmincke, and J. B. Gill (1992), Processesand timescales in the evolution of chemically zoned tra-chyte; Fogo A, S~ao Miguel, Azores, Contrib. Mineral. Pet-rol., 111, 311328.

Widom, E., J. B. Gill, and H.-U. Schmincke (1993), Syenitenodules as a long-term record of magmatic activity in Aguade Pau Volcano, S~ao Miguel, Azores, J. Petrol., 34, 929953.

Widom, E., R. W. Carlson, J. B. Gill, and H.-U. Schmincke(1997), ThSrNdPb isotope and trace element evidence

for the origin of the S~ao Miguel, Azores, enriched mantlesource, Chem. Geol., 140(1-2), 4968.

Wilson, L., and J. W. Head (1981), Ascent and eruption of ba-saltic magma on the earth and moon, J. Geophys. Res., 86,29713001.

Zanon, V., and I. Nikogosian (2004), Evidence of crustal melt-ing events below the Island of Salina (Aeolian Arc, SouthernItaly), Geol. Mag., 141(4), 525540.

Zanon, V., M. L. Frezzotti, and A. Peccerillo (2003), Mag-matic feeding system and crustal magma accumulationbeneath Vulcano Island (Italy): Evidence from uid inclu-sions in quartz xenoliths, J. Geophys. Res., 108(B6), 2298,doi :10.1029/2002JB002140.

Zanon, V., U. Kueppers, J. M. Pacheco, and I. Cruz (2013),Volcanism from ssure zones and the Caldeira central vol-cano of Faial Island, Azores archipelago: Geochemicalprocesses in multiple feeding systems, Geol. Mag., 130(3),536555.

Zbyszewski, G. (1963), Les Phenome`nes Volcaniques Mod-erns dans lArchipel des Acores, vol. 47, pp. 227, Comun.dos Serv. Geol. de Portugal, Lisbon, Portugal.


21

Date post:	09-Sep-2015
Category:	Documents
Upload:	brunomeloovga
View:	10 times
Download:	1 times

Artº: Magma storage and ascent conditions beneath Pico and Faial islands (Azores archipelago) A...

Documents