Repositorio Institucional de Documentos

2014 42 Patricia Larrea Márquez The magmatic evolution of Graciosa and Corvo oceanic islands, Azores Archipielago Departamento Director/es Ciencias de la Tierra Widom, Elisabeth França, Zilda Lago San José, Marceliano Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Patricia Larrea Márquez THE MAGMATIC EVOLUTION OF GRACIOSA AND CORVO OCEANIC ISLANDS, AZORES ARCHIPIELAGO Director/es Ciencias de la Tierra Widom, Elisabeth França, Zilda Lago San José, Marceliano Tesis Doctoral Autor 2014 Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA Departamento Director/es Director/es Tesis Doctoral Autor Repositorio de la Universidad de Zaragoza – Zaguan http://zaguan.unizar.es UNIVERSIDAD DE ZARAGOZA PATRICIA LARREA MÁRQUEZ TESIS DOCTORAL 2014 UNIVERSIDAD DE ZARAGOZA THE MAGMATIC EVOLUTION OF GRACIOSA AND CORVO OCEANIC ISLANDS, AZORES ARCHIPELAGO THE MAGMATIC EVOLUTION OF GRACIOSA AND CORVO OCEANIC ISLANDS, AZORES ARCHIPELAGO PATRICIA LARREA MÁRQUEZ PhD THESIS 2014 UNIVERSIDAD DE ZARAGOZA DEPARTAMENTO DE CIENCIAS DE LA TIERRA GRUPO DE INVESTIGACIÓN GEOTRANSFER Supervisors Dr. Marceliano Lago San José (Universidad de Zaragoza, Spain) Dr. Zilda França (Universidade dos Açores, Portugal) Dr. Elisabeth Widom (Miami University, USA) Nalgumas ilhas, onde o vulcanismo se extinguiu há muito e onde as formas, profundamente gastas pela erosão, perderam a sua estrutura primitiva, só o observador entendido se aperceberá de cómo elas se formaram, emergindo das águas pela acumulação sucessiva de materiais eruptivos. António de Brum Ferreira (1968) Gracias a los responsables y técnicos del servicio de microsonda electrónica del Centro Nacional de Microscopía Electrónica de la Universidad Complutense de Madrid, y del laboratorio IBERCRON de la Universidad del País Vasco UPV/EHU, por las facilidades y ayudas prestadas durante estos años. Asimismo, agradezco al Dr. José Luis Fernández Turiel y a la Dra. Marta Rejas su ayuda e implicación, y el invitarme a realizar la preparación de muestras para análisis geoquímico por ICP-MS en el labGEOTOP del Institut de Ciències de la Terra Jaume Almera - CSIC. Además, agradecer a los miembros del Departamento de Geoquímica, Petrología y Prospección Geológica de la Universidad de Barcelona el hacerme un huequito en su despatx de doctorands siempre que he pasado unos días por allí. Gracias también al Dr. Jan R. Wijbrans (VU Universiteit Amsterdam) por supervisar el proyecto de datación de mi tesis y acogerme en el laboratorio de argón. En especial, quiero agradecer a Roel Van Elsas, técnico responsable del laboratorio de separación mineral, su atención durante el proceso de preparación de muestras, ya que su buen humor y su dedicada atención hizo muy cálida mí llegada al laboratorio. Además, durante la etapa analítica conté con la valiosa ayuda y asistencia técnica de Christel Bontje, Bertram Uunk y de la Dr. Klaudia Kuiper (VU Universiteit Amsterdam). Quisiera agradecer muy especialmente la ayuda prestada por Dave Kuentz, técnico responsable del Isotope Geochemistry & Mass Spectrometry Laboratory de la Miami University (OH, USA). Liz y Dave, junto con su grupo de investigación (con especial mención a Fara y a George) y el resto de integrantes del Depatment of Geology and Environmental Sciences hicieron que mi estancia de cuatro meses en el laboratorio fuera estupenda y súper productiva. ¡Nos vemos pronto! Además, la colaboración con otros grupos de investigación y la asistencia a congresos internacionales a lo largo de estos años, me ha llevado a conocer a personas excepcionales y grandes expertos en el vulcanismo de las Azores como el Dr. Victor Hugo Forjaz (Observatório Vulcanológico e Geotémico dos Açores), la Dra. Luisa P. Ribeiro (emepc, Governo de Portugal) y el Dr. Vittorio Zanon (Universidade dos Açores), quien además se ofreció a revisar, muy amablemente, esta memoria en sus últimas etapas de creación. Sin financiación pública, esta tesis no habría sido posible. Por ello, me gustaría agradecer al Gobierno de Aragón la concesión de mi beca y contrato pre-doctorales (B023/10). Asimismo, al Ministerio de Ciencia e Innovación la financiación del proyecto de investigación CGL2011-27477, y a la Fundação Luso-Americana para o Desenvolvimento por el proyecto de investigación “Identificação dos perigos vulcânicos do archipélago dos Açores através da caracterização geoquímica das lavas emitidas pelos respectivos vulcões”. Quiero destacar también las ayudas recibidas para las estancias concedidas por el Programa CAI Europa de Investigación (CAI - CONAID CB29/11 y CB12/12) y la Universidad de Zaragoza, así como el apoyo económico del Gobierno de Aragón al grupo de investigación consolidado Geotransfer. Durante la realización de esta Tesis he recibido el apoyo, la ayuda y los ánimos de muchos compañeros de promoción y del máster (Universidad de Zaragoza), de compañeros del curso Erasmus 2008/2009 en Cardiff University y de becarios del Departamento de Ciencias de la Tierra. Deseo recordar, en especial, a Esther, Rocío, Galo, Pablo, Claire, Alba y Tere que, en diferentes períodos, han sido personas muy importantes en mi vida. Gracias a las estancias de investigación en Ámsterdam y USA, tuve la oportunidad de conocer a personas muy especiales que se han convertido en grandes amigos. Gracias a Marina y a todo su entorno de físicos locos por hacerme sentir como en casa (espero que Azores esté contigo por muchos años), y a Marta y a Sechu por acogerme como invitada en su casa y pasar unos días fantásticos de turismo por el país de los tulipanes. Al otro lado del Atlántico, tuve la suerte de compartir mi tiempo con Nancy (y todo el Zoology team), Manu, Sandra, Ferrán y las chicas de Oviedo ¡muchas gracias por ser tan auténticos! Quiero dedicar también un agradecimiento muy especial a mis amigas Eli, Agar, Elena e Isa. Por estar siempre dispuestas a pasar un buen rato, escuchar mis aventurillas y apoyarme en todo momento, aunque me encuentre a cientos o miles de kilómetros. En especial, quiero agradecer a mis padres y a mi hermana toda la confianza y todo el cariño que me dan día a día. Sin su apoyo e incondicional ayuda jamás hubiera podido llegar hasta aquí. También quiero tener un recuerdo especial y un agradecimiento al resto de las Márquez (¡incluyendo a los chicos!), a mi tía Leo (siempre orgullosa de mis logros) y a Teo - Jesús junto con el resto de la familia Ferrer, que recientemente se han unido de forma oficial a mi gran familia. Por último, quería dedicar esta Tesis Doctoral a Ismael, mi otra mitad, por recibirme siempre con una sonrisa, y por haber hecho de este 2013 un año mágico. Zaragoza, 5 de Diciembre de 2013 RESUMEN El archipiélago de las Azores se encuentra localizado en el Océano Atlántico, entre las latitudes 37º y 40º N y las longitudes 25º y 31º O. Está constituido por nueve islas y varios montículos submarinos construidos sobre la plataforma de las Azores. Dos de estas islas han sido objeto de estudio en la presente Tesis Doctoral: Graciosa, al este de la dorsal centro-atlántica, y Corvo, al oeste de la misma. Dado que los estudios previos sobre estas islas son escasos y se han limitado a la descripción geológica y petrológica de las diferentes unidades volcánicas, la presente investigación aborda un completo estudio de la petrología, la mineralogía y la geoquímica de las distintas unidades volcánicas, así como de los procesos magmáticos ocurridos. En la Isla de Graciosa se han estudiado diferentes coladas de lava y xenolitos de gabros (alcalinos y subalcalinos), de sienitas y de dunitas, pertenecientes a los complejos volcánicos de Serra das Fontes, Serra Branca y Vitória - Vulcão Central. Se ha demostrado que las coladas de lava de los tres complejos volcánicos, junto con los gabros de composición alcalina (cumulados) y las sienitas (frozen liquids), están relacionados mediante un proceso de cristalización fraccionada polibárica, que comenzó en una cámara magmática situada a unos 15 km de profundidad. Además, la alternancia temporal de rocas de composición básica y ácida se explica a través de la existencia de varias recargas magmáticas del sistema. Por el contrario, los gabros de composición subalcalina, que son descritos por primera vez en el archipiélago de las Azores, se interpretan como cumulados relacionados con un proceso de fraccionación de fundidos altamente refractarios en niveles más someros (~ 3 km). En la Isla Corvo, se han muestreado también todas las unidades volcanoestratigráficas (Pre-, Siny Post-caldera), que incluyen coladas de lava, diques y xenolitos de gabro. Tanto algunas de las coladas, como algunos de los diques, presentan antecristales no cogenéticos que fueron incorporados al fundido previamente a la erupción. Se ha puesto de manifiesto, por primera vez en Azores, que la composición geoquímica de roca total está altamente influenciada por la acumulación de estos antecristales, lo que produce un enmascaramiento de los procesos magmáticos de diferenciación que han dado lugar a la formación de estas rocas. Por ello, para identificar los procesos que han dado lugar a la formación de la isla tan solo se han considerado rocas sin antecristales. De este modo, las coladas de lava y los diques sin antecristales, junto con los xenolitos de gabro (cumulados) están relacionados mediante un proceso de cristalización polibárica, que comenzó en una cámara magmática situada a unos 15 km de profundidad y en la que tuvieron lugar procesos de recarga constantes con fundidos más profundos y de gran carga cristalina. Por lo tanto, se propone que la Isla de Corvo ha sido formada a partir de un complejo sistema magmático en el que los diferentes fundidos y su carga de antecristales provienen de diferentes profundidades y han estado en constante interacción a los largo de toda la evolución de la isla. La Isla de Graciosa ha sido estudiada desde un punto de vista isotópico, para adecuar el conocimiento de la isla con las demás islas del Rift de Terceira. La edades 40Ar/39Ar han permitido establecer la evolución temporal de la isla a lo largo del Pleistoceno y el Holoceno, desde los 1056 ± 28.0 ka hasta los 3.9 ± 1.4 ka, en contraste con las edades previas obtenidas por K-Ar y 14C (620 ka a 2 ka). Las composiciones isotópicas de Sr-NdPb han permitido la caracterización de la fuente del manto de los fundidos, señalando una mezcla entre un componente empobrecido de tipo MORB y un componente enriquecido de tipo HIMU. Estos datos son similares a la composición de manto recientemente redefinida como FOZO, la cual parece estar presente en la signatura isotópica de la mayor parte de islas oceánicas. ABSTRACT The Azores archipelago is located in the Atlantic Ocean, between 37º - 40º N latitude and 25º - 31º W longitude, and comprises nine ocean islands and several seamounts built on the Azores submarine plateau. Two remote and poorly studied islands from this archipelago have been investigated in the present PhD Thesis: Graciosa and Corvo, which are located to the east and west of the Mid-Atlantic Ridge, respectively. Overall, previously published studies mainly focused on the geological and petrographic description of the different volcanic units found in both islands. In addition, the small size of these islands with straightforward volcanostratigraphical sequences, make both of them ideal for an in-depth study of the magmatic processes controlling their evolution. Accordingly, a comprehensive study of the petrology, mineralogy and geochemistry of the different magmatic products found across the volcanic units, has allowed for the geochemical characterization of the units and the investigation of the magmatic processes. In the case of Graciosa Island, lava flows, gabbroic xenoliths (alkaline and subalkaline), syenites and dunites found across the Serra das Fontes, Serra Branca and Vitória - Vulcão Central volcanic complexes have been considered. Lava flows from the three volcanic complexes, together with alkaline gabbros (cumulates) and syenites (frozen liquids), are related by polybaric fractional crystallization starting at ~ 15 km depth. The alternation of basic and evolved products is related to replenishments of the magma chamber with a more primitive melt. Subalkaline gabbro xenoliths, found for the first time in the Azores Islands, are related to fractionation of highly refractory melts at shallower levels (~ 3 km). In Corvo Island, representative samples of the complete volcanostratigraphic sequence (Pre-, Synand Post-caldera stages) including lava flows, dikes and cumulate gabbroic xenoliths have been studied. Some lava flows and dikes contain non-cogenetic antecrysts, which have been reincorporated into the host lava before eruption. It has been noted for the first time in the Azores, that the accumulation of antecrysts strongly affects the composition of the rocks, and obscures the identification of magmatic processes responsible for the differentiation of the volcanic products. Accordingly, only antecryst-free samples are considered to identify magmatic processes involved in the formation of the island. Corvo antecryst-free lava flows, dikes and cumulate xenoliths can be related by polybaric fractional crystallization starting at ~ 15 km, and several primitive melt magma chamber replenishments. These results suggest that the formation of Corvo Island is related to a complex magmatic plumbing system in which magmas and their crystal cargo, derived from variable depths, have interacted and been tapped throughout the evolution of the island. In addition, Graciosa Island has been further studied to make data on this island comparable to those obtained in the other two islands of the Terceira Rift (São Miguel and Terceira). 40Ar/39Ar age determinations on Graciosa samples have constrained the Pleistocene and Holocene temporal evolution of the island. These new ages, spanning the whole eruptive record of the island, range from 1056 ± 28.0 ka to 3.9 ± 1.4 ka, whereas prior published K-Ar and 14C ages range from 620 - 2 ka. The oldest age provides a new minimum age for the volcanism on the island, contributing to the overall knowledge on the temporal evolution of the Azores archipelago. Moreover, the Sr-NdPb isotopic composition of Graciosa samples has enabled the characterization of its magma source in the context of the Azores archipelago. These data point to mixing between depleted MORB and enriched HIMU mantle components. They also resemble the composition of the redefined FOZO mantle component, which has been recently proposed as the commonest mantle end-member in ocean island basalts. PART I: INTRODUCTION Chapter 2: Azores Regional Setting Searle 1980; Vogt and Jung, 2004; Fig. 2.1). It comprises the volcanic Islands of Graciosa, Terceira and São Miguel, the João de Castro seamount, and the non-volcanic basins located among them (e.g., Vogt and Jung, 2004; Georgen and Sankar, 2010). Except from the East Azores Fracture Zone and the eastern part of São Miguel, which are E-W striking tectonic structures, all other tectonic structures located to the East of the MAR run parallel to the Terceira Rift (e.g., the Islands of Faial, Pico and São Jorge are elongated and align parallel just south of the Terceira Rift; Fig. 2.1). In contrast, the Islands of Flores and Corvo, the only two located to the west of the MAR, lie on a NNESSW trending ridge that is subparallel to the MAR. Moreover, the structural influence of the MAR transform faults is reflected by the near linear E-W northern and southern coastlines of Flores (Fig. 2.1) together with a westward displacement of Corvo from Flores of about 1 cm/year (Baptista et al., 1999). Different geodynamic models have been proposed to explain the boundaries between the North American, Eurasian and African plates. Overall, the boundary between the American and the Eurasian - African plates is believed to be defined by the MAR. In contrast, there is disagreement about the structures defining the Eurasian and African plate boundary in the segment between the MAR and the Gloria Fault. A classical model suggests that this boundary is defined by the Terceira Rift, forming a Ridge-Ridge-Ridge type triple junction (e.g., Krause and Watkins, 1970; Udías and Arroyo, 1972; Buforn et al., 1988). Other authors (McKenzie, 1972; Laughton and Whitmarsh, 1974; Searle, 1980; Ribeiro, 1982) suggest that the Eurasian and African plate boundary corresponds to a leaky-transform fault passing through São Jorge submarine channel and continuing south of São Miguel Island as far as the Azores - Gibraltar Fracture Zone. Forjaz (1983) proposes the existence of an Azorean microplate limited by the East Azores Fracture Zone in the south, by the MAR in the west and by the North Azores Fracture Zone and the Terceira Axis in the north; he proposes that the strain generated by the MAR and the Terceira Axis produces a compressive regimen oblique to the Gloria Fault. In contrast, geophysical studies (e.g., Luis et al., 1994; Miranda and Luis, 1995) covering the MAR and oceanic crust created ~ 10 Ma ago, 9 Part I: Introduction suggest that the volcanism and the tectonic regime of the archipelago are controlled by the movement of the plates and the Azores plateau; i.e., a progressive migration from the East Azores Fracture Zone to the Princess Alice Fracture Zone before the last 10 Ma, from the Princess Alice Fracture Zone to the Açor Fracture Zone in the last 10 Ma, and currently from the Açor Fracture Zone to the Faial Fracture Zone. Moreover, they suggest that the recent migration of the triple junction might have triggered the formation of the youngest Islands of Faial and Pico. More recently, Marques et al. (2013) have determined GPS velocities implying that oblique WSW-ENE extension between the African and Eurasian plates is accommodated across a series of horsts and grabens. These include the Faial - Pico volcanic ridge, which moves mostly with the African plate; Terceira Island, which moves mostly with the Eurasian plate; and São Jorge Island, whose motion is intermediate between that of the African and Eurasian plates. As a conclusion they argue against a rigid or semi-rigid Azores microplate, and instead suggest that the African - Eurasian plate boundary at the longitude of the Azores is diffuse and represented by a ca. 140 km wide zone of deformation. 2.2 The Azorean ages Studies of the magnetic anomalies in the Atlantic Ocean and kinematic reconstruction models discussed in the previous section have shed light on the temporal evolution of the Azores plateau (Fig. 2.2; e.g., Searle, 1980; Luis et al., 1994; Cannat et al., 1999; Gente et al., 2003). Moreover, 40Ar/39Ar ages have been recently obtained from submarine samples dredged from the plateau (Beier, 2006) providing important temporal constraints. Searle (1980) suggested an age of ~ 53 Ma for the easternmost part of the Azores plateau, based on the age of the oldest magnetic anomaly in the area (Fig. 2.2). In contrast, the 40Ar/39Ar age of ~ 39 Ma obtained by Beier (2006) in a sample from the Monaco Bank south of São Miguel Island (Fig. 2.2), indicates that the first intraplate magmas of the Azorean plateau were formed ca. 14 Ma after the formation of the lithosphere. Further to the west, the age of the lithosphere close to the MAR has been estimated to be 10 Ma based on magnetic lineations (Cannat et al., 1999; Fig. 2.2). 10 Chapter 2: Azores Regional Setting Fig. 2.2 - Bathymetry of the Azores Plateau, modified after the GEBCO world map 2013 (www.gebco.net) with the general geotectonic setting (as in Fig. 2.1; note the different color for each tectonic plate) and the most remarkable ages of the Azores archipelago reported in millions of years (Ma). The preferred ages of each island are reported in white (see section 2.2; STM: Santa Maria - Feraud et al., 1981; SM: São Miguel - Abdel-Monem et al., 1975; T: Terceira - Calvert et al., 2006; G: Graciosa - Larrea et al., 2014; SJ: São Jorge - Hildenbrand et al., 2008; F: Faial - Hildenbrand et al., 2012; P: Pico - Demand et al., 1982; FL: Flores - Azevedo and Ferreira, 2006; C: Corvo - Azevedo et al., 2003. Lithospheric ages (in black) are from Searle (1980), Luis et al. (1994) and Cannat et al. (1999). Submarine samples dredged from the plateau in yellow were obtained by Beier (2006); PS - Ponta Sul Ridge; FB - Formigas Bank; MB - Monaco Bank; AB - Alcatraz Bank; PR - Pico Ridge; CTB - Condor de Terra Bank; PAB - Princess Alice Bank; FR - Faial Ridge. Moreover, Luis et al. (1994) presented an aeromagnetic survey covering both sides of the MAR up to the anomaly 5 (~ 10 Ma). It is located to the west of São Jorge Island and beneath Faial Island (to the east of the MAR), and to the west of Corvo and Flores Islands (to the west of the MAR). In contrast, 40Ar/39Ar ages calculated in submarine elevations of the Azores plateau reveal that most part of the plateau was formed from 6 to 4.9 Ma (Beier, 2006; Fig. 2.2), which is coincident to the age range proposed for the formation of Santa Maria Island and the Formigas Bank (see below). However, during the last 1.5 Ma, abundant volcanism formed most of the islands and numerous seamounts predominantly situated along the Terceira Rift and the Faial - Pico lineament. 11 Part I: Introduction Santa Maria is the easternmost and oldest island of the Azores archipelago (Fig. 2.2). Some authors propose a range of volcanic activity from 5.5 to 3 Ma (Feraud et al., 1980, 1981; Salgueiro, 1991), whereas Abdel-Monem et al. (1975) suggests older ages (8 - 4 Ma). This large variation has been a cause of debate (see discussion in Feraud et al., (1981) and Serralheiro (1993)) as these age calculations are considered erroneous due to the presence of excess argon (see also the problems reported by Carracedo (2011) in the ages obtained by Abdel-Monem et al. (1972) in the Canary Islands). Moreover, ostracods from Santa Maria Island have been categorized as Messinian in age (Meireles et al., 2012; stage from the Late Miocene ranging from 7.246 to 5.322 Ma). The subaerial formation of São Miguel Island was initiated at ~ 4 Ma according to the K/Ar ages obtained by Abdel-Monem et al. (1975), although this age might be overestimated due to excess argon commonly found in young basalts (see above). This unit has been recently dated by 40Ar/39Ar given a younger age of 0.88 Ma (Johnson et al., 1998), in accordance with the reverse polarity paleomagnetic field directions of late Matayuma age. However, the sampled analyzed by Johnson et al. (1998) is a dike intruding the oldest unit of the Nordeste Complex (V.H. Forjaz, pers. com.). Therefore, due to the uncertain ages estimated for this oldest unit, the age of the subaerial formation of São Miguel remains unknown, although it has to be older than 0.88 Ma (Fig. 2.2). Terceira (3.52 Ma) and Graciosa (2.5 Ma) Islands were considered the oldest islands of the Central Group, based on K/Ar ages in Terceira (Ferreira and Azevedo, 1995) and Rb/Sr apparent ages for Graciosa (White et al., 1976). However, more accurate and recent 40Ar/39Ar studies report younger ages for these islands (Fig. 2.2): < 0.39 Ma for Terceira (Calvert et al., 2006) and < 1.05 Ma for Graciosa (Larrea et al., 2014; see chapter 5). São Jorge Island has been recently dated by 40Ar/39Ar showing that the first subaerial volcanic phase started at ~ 1.3 Ma (Hildenbrand et al., 2008; Ribeiro, 2011; Fig. 2.2). Pico and Faial Islands lack 40Ar/39Ar accurate age data. The oldest K/Ar ages are 0.73 Ma and 0.27 Ma respectively, showing the youth of the Faial - Pico lineament in comparison to the other islands of the Central Group (Fig. 2.2). 12 Chapter 2: Azores Regional Setting The islands located to the west of the MAR have also been dated (Fig. 2.2). The maximum K/Ar age obtained for Flores Island is 2.16 Ma (Azevedo and Ferreira, 2006), while the oldest age obtained in Corvo Island is 0.7 Ma (Azevedo et al., 2003). Comparing these ages with those of the plateau where they are rooted, it is clear that these islands formed after the plateau. Overall, the geographic distribution of the eruptive centers reveals that the oldest ages are those located most distant to the MAR (Santa Maria Island and Monaco Bank seamount; Fig. 2.2). Nevertheless, the temporal evolution of the archipelago does not appear to be compatible with a conventional hotspot model, in which volcanism is controlled by plate migration with time over a fixed plume, as proposed for other oceanic islands (e.g., Hawaii; Wilson, 1963; Morgan, 1972; DePaolo and Manga, 2003), but rather may be partially controlled by complex extensional tectonics related to the Azores triple junction (e.g., Searle, 1980; Luis et al., 1998; Lourenço, 2007). In order to better constrain the temporal evolution of the Azores archipelago and the potential magmatic and tectonic controls, it is critical to obtain additional and improved age constraints on the islands. Accordingly, a detailed 40Ar/39Ar dating project has been carried out in Graciosa Island to better constrain the age of the different volcanological units and reconstruct its volcanic evolution (Larrea et al., 2014; see chapter 5). The volcanism in the Azores archipelago is still active, with 26 historical eruptions reported; 12 of those were subaerial and located in the Islands of São Miguel, Terceira, São Jorge, Pico and Faial (França et al., 2003). The most recent submarine eruptions were the Capelinhos in 1957 - 1958 to the west of Faial Island, and Serreta in 1998 - 2000, which occurred 8.5 km to the NW of Terceira Island (Forjaz et al., 2001; Gaspar et al., 2003). Present-day activity is recognized in the Islands of São Miguel, Terceira, Graciosa, Faial, Pico and Flores (Baxter et al., 1999; Cruz et al., 1999; Ferreira and Oskarsson, 1999; Nunes et al., 2001; Cruz and França, 2004) as hot water springs, lowtemperature fumarolic degassing fields and soil degassing emissions (e.g., Viveiros et al., 2009). 13 Part I: Introduction 2.3 The mantle plume beneath the Azores archipelago The Azores archipelago has been considered as an example of a hotspot interacting with a mid-ocean ridge. A variety of geochemical studies and geophysical observations point out the importance of a mantle plume on the formation of the Azores plateau and the nearby MAR: long-wavelength observations along the MAR, including ridge-axis elevation (Anderson et al., 1973; Vogt, 1976; Le Douaran and Francheteau, 1981; Gente, 1987; Thibaud et al., 1998), geoid topography (Bowin et al., 1984; Cazenave et al., 1992) and geochemical anomalies (Schilling, 1975; White and Schilling, 1978; Bougault and Treuil, 1980; Yu et al., 1997; Dosso et al., 1999). This mantle plume arrived to the vicinity of the MAR ~ 40 Ma ago (see section 2.2), and is characterized by a slow upwelling rate of ~ 3 - 4 cm/y (Sleep, 1990; Bourdon et al., 2005). It has a buoyancy flux of 1.2 mg/s (Sleep, 1990), which produces small amounts of melts if compared to the unusually strong and long-lived Hawaiian plume, with an estimated plume buoyancy flux of 6.5 - 8.7 mg/s (White and McKenzie, 1989; Davies, 1999). The present location of the Azores hot spot is not exactly known. However, it is clearly located to the east of the MAR, and probably centered underneath Terceira Island, as suggested by He and Ne isotope ratios (Madureira et al., 2005) and U-series systematics (Bourdon et al., 2005). Geochemical data on submarine lavas (Schilling, 1975; White et al., 1976; Beier, 2006) reveals that the Azores plateau is alkaline in composition. Accordingly, the composition of the subaereal volcanisms is mainly alkaline, despite few punctual tendencies for transitional basalts; e.g., Furnas Volcanic Complex in São Miguel Island (Rodrigues et al., 1995) and certain lava flows in Pico (França et al., 1995) and Terceira Islands (Madureira, 2006). A significant isotopic variability characterizes the archipelago, and it is interpreted as a result of small-scale heterogeneities in the mantle source (e.g., França et al., 2006b; 14 Chapter 2: Azores Regional Setting Beier et al., 2007, 2008, 2010; Madureira et al., 2011). Most islands from the archipelago show a common isotopic array (linear trend in Sr-Pb, Nd-Pb and Pb-Pb isotopic systems) which suggests the involvement of at least two mixing end-members including a depleted MORB component (related to the nearby MAR) and an enriched mantle component (see chapter 6; Larrea et al., under review). In contrast, São Miguel Island presents a distinct isotopic composition characterized by extremely high 207Pb/204Pb, 208Pb/204Pb and 87Sr/86Sr ratios and low 143Nd/144Nd ratios extensively discussed elsewhere (e.g., White et al., 1979; Turner et al., 1997; Widom et al., 1997; Beier et al., 2007; Elliott et al., 2007; França et al., 2010). The isotopic composition of the MAR in the vicinity of the Azores plateau is enriched compared with the N-MORB isotopic composition found in other segments of the MAR (e.g., Schilling, 1975; White and Schilling, 1978; Bourdon et al., 1996; Yu et al., 1997; Dosso et al., 1999). This enrichment might be linked to the isotopic composition of the Azores mantle plume which is affecting the MAR isotopic composition. Recently, a finitefrequency seismic tomography study in the Azores region (Yang et al., 2006) proposed a model of plume-ridge interaction in which the plume conduit is deflected to the southwest (Fig. 2.3), and contributes mantle material and excess melt to the MAR. On the other hand, Foulger (2010) pointed out the impossibility of carrying out surfacewave tomography to resolve bodies smaller than ~ 100 km across and thus, detecting narrow structures such as plumes, so that he considers ambiguous the interpretation of Yang et al. (2006). Adam et al. (2013) carried out a highly resolved tomography model to characterize mantle convection in the upper mantle beneath the Azores region. These authors found two distinct upwellings in the shallow mantle that converge into a single anomaly at a deeper level (Fig. 2.3). One of the mantle upwellings is proposed to be centered among Faial, Pico, São Jorge, and Graciosa Islands, and the other one is located between Terceira and São Miguel Islands. These authors suggest that the two reservoirs may account for the mantle heterogeneity recognized for the Azores plume, explaining the difference in isotopic signatures observed through the islands and, in particular, the 15 Part I: Introduction Fig. 2.3 - Depth cross section along the Terceira Rift showing the tomography model of Yang. et al. (2006) (in Adam et al., 2013). unique isotopic ratios in São Miguel Island lavas. However, it must be taken into account that the isotopic signature of Terceira lavas is much closer to the isotopic compositions of Graciosa and São Jorge lavas than to the distinct isotopic signature found in São Miguel lavas, so this hypothesis is not completely in agreement with geochemical data. Other authors remark that in the Azores archipelago there are neither flood basalt implying precursory uplift, nor time-progressive volcanism, nor evidence for high source temperature (e.g., Foulger, 2010). These data disagree with the theory of a mantle plume located beneath the Azores. They suggest that the magmatism was enhanced by astenospheric upwelling beneath the ridge branches, and also by the long-term tectonic instability and migration of major fault zones at the Azores (e.g., Searle, 1980; Madeira and Ribeiro, 1990; Gente et al., 2003; Luis and Miranda, 2008). In agreement to these works, a recent study by Neves et al. (2013) hypothesizes that linear volcanic ridges are the result of magma emplacement into pre-existing damaged lithosphere, without the need of enhanced magmatic pulses related to mantle plume activity. They suggest that the construction of the islands is dominated by episodic fissural eruptions whose size and frequency depend on melt availability (anomalous amount of melt production over a relatively long period) and tectonic stress. 16 Chapter 3: Materials and Methods CHAPTER 3: Materials and Methods 3.1 Field work and sampling strategy This PhD Thesis comprises the magmatic evolution of Graciosa and Corvo Islands. After a careful review of the material sampled previously by members of the research group and the studies published previously on both islands (e.g., Gaspar, 1996; Almeida, 2001; Dias, 2001), two field campaigns were planned in 2011 - 2012. The main aim of the field work was the revision of the volcanostratigraphy (see chapters 4 and 7), and a thorough sampling of the different units. All diverse magmatic products found in both islands were sampled, with special emphasis on those not sampled previously. Moreover, field observations on the petrology, composition and emplacement characteristics of the outcrops were used to sample the least altered material in each case. A summary of samples are presented in Tables 3.1 and 3.2. These tables include the sampled unit, the campaign in which the sample was collected, and the location, UTM coordinates, classification, texture of the samples, as well as the methods applied in each case. All of the samples were studied macroand microscopically, and the most representative samples were selected for geochemical determinations. 17 Part I: Introduction Table 3.1 - Summary of samples studied from Graciosa Island and methods applied (see chapter 4). Unit Subunit Sample Campaign Location UTM Coordinates Classification Texture Thin Section EMP WR Isotopes Ar ages Vulcão Central lavas VCU-g GRZF5 1 South of Canada Longa 26S 0413670 4319206 Basalt porphyritic x x x VCU-g GRZF19 2 Quarry. Canada da Chicha 26S 0414049 4320557 Basalt porphyritic x x x x VCU-f GRZF16 1 North of Beco 26S 0415266 4319817 Benmoreite microporphyritic x x x x x VCU-d GRZF1 1 Baía do Folga 26S 0413669 4319206 Hawaiite porphyritic x x x x x VCU-b GRZF6 1 Enxudreiro 26S 0414756 4318450 Trachyte microporphyritic x x x VCU-b GRZF7 1 Enxudreiro 26S 0414756 4318451 Basalt porphyritic x x x VCU-b GRZF14 1 Carapacho 26S 0416894 4318520 Basalt porphyritic x x x x x VCU-b GRZF15 1 Portela 26S 0417178 4320652 Basalt porphyritic x x VCU-b GRZF18 2 Baía do Folga 26S 0413376 4319314 Hawaiite porphyritic x x x VCU-b GRZF36 2 Praia town 26S 0415628 4322545 Mugearite microporphyritic x x x VCU-b GRZF37 2 Enxudreiro 26S 0414960 4318607 Trachyte microporphyritic x VCU-b GRZF38 2 Enxudreiro 26S 0414960 4318607 Basalt porphyritic x Vitória lavas VU-c GRZF2 1 Pico Timão 26S 0413698 4321723 Basalt microporphyritic x x x VU-c GRZF3 1 Quarry. Praia town 26S 0415701 4323383 Basalt microporphyritic x x x x x VU-c GRZF21 2 Feteira 26S 0413059 4322312 Basalt microporphyritic x x VU-a GRZF11 1 Pico do Barcelo 26S 0409099 4325197 Basalt microporphyritic x x VU-a GRZF12 1 Serra Branca Windmills 26S 0410857 4321152 Basalt porphyritic x x x VU-a GRZF4 1 Quitadouro 26S 0414611 4323676 Basalt porphyritic x x VU-a GRZF9 1 Airport 26S 0410809 4327836 Basalt microporphyritic x x VU-a GRZF10 1 Da Barca lighthouse 26S 0409239 4327560 Basalt microporphyritic x x x VU-a GRZF17 1 Ponta Pesqueira 26S 0413479 4327011 Basalt porphyritic x x x VU-a GRZF22 2 Quitadouro 26S 0415369 4324445 Basalt porphyritic x x x x x VU-a GRZF23 2 Quitadouro 26S 0415369 4324445 Basalt porphyritic x x VU-a GRZF24A/ 24B 2 Quitadouro 26S 0415369 4324445 Basalt porphyritic x x VU-a GRZF25 2 Quitadouro 26S 0415369 4324445 Basalt microporphyritic x x x VU-a GRZF26 2 Quitadouro 26S 0415369 4324445 Basalt porphyritic x x x VU-a GRZF27 2 Quitadouro 26S 0415374 4324445 Basalt porphyritic x x x x x VU-a GRZF30 2 Baía da Vitória cliff 26S 0408606 4326149 Basalt porphyritic x x x VU-a GRZF31 2 Baía da Vitória cliff 26S 0408606 4326149 Basalt microporphyritic x x x VU-a GRZF32A 2 Baía da Vitória cliff 26S 0408606 4326149 Basalt porphyritic x x VU-a GRZF32B 2 Baía da Vitória cliff 26S 0408606 4326149 Basalt porphyritic x x 18 Chapter 3: Materials and Methods Moreover, some previously studied lava flows were re-sampled for dating purposes (out of the scope of this PhD Thesis). 3.2 Laboratory work The samples were first studied macroscopically to select areas of interest for microscopic examination. Thin sections of the samples (30 µm thickness) were prepared by the Servicio General de Apoyo a la Investigación of the Universidad de Zaragoza (Spain). They were studied under the petrographic microscope of the Servicio General de Apoyo a la Investigación of the Universidad de Zaragoza (Spain) to describe the petrography of the samples (mineral assemblage, mineral modes and texture). The petrographic study led to the selection of representative thin sections for mineral analyses and whole rock chemical analyses. The mineral chemistry study involved mineral classification and a careful evaluation of the different crystal populations. This was particularly relevant in Corvo Island, where three main populations were defined: phenocrysts, crystals co-genetic with their magmatic host; antecrysts, co-magmatic crystals which are recycled one or several times before inclusion in the host magma; and microcrysts, which represent small co-genetic crystals which nucleate and grow rapidly on decompression and eruption. Mineral major element compositions were determined on polished thin sections by electronic microprobe (EMP) at the Centro Nacional de Microscopía Electrónica of the Universidad Complutense de Madrid (Spain) (Fig. 3.3). Lava flow, dike and xenolith samples were carefully selected for whole rock chemistry considering the nature of the rocks Fig. 3.3 - Scanned thin sections (2.5 cm width) of porphyritic (left) and microporphyritic (right) textures of Corvo I sland rocks. Whole rock compositions are strongly influenced by the volume fractio n and type of macrocrysts (see chapter 8). Circles indicate areas analyzed by EMP. 25 Part I: Introduction (composition - primitive or evolved, and texture - porphyritic, microporphyritic, trachytic or glassy for the lavas and dikes, and cumulate, equigranular or inequigranular textures for xenoliths). The selected samples were crushed in a manganese steel jaw-crusher and milled in an agate vibrating cup mill at the Servicio General de Apoyo a la Investigación of the Universidad de Zaragoza (Spain). These samples were analyzed by different techniques and in different laboratories according to the field work campaign in which they were collected (see details in chapters 6 and 8). Major and trace element concentrations of lava flow and xenoliths samples from Graciosa Island were determined in two different laboratories (see chapter 6 for analytical details and procedures). Samples of the first fieldwork campaign1 (Table 3.1) were analyzed at the IBERCRON laboratory of the Universidad del País Vasco (Spain) by ICP-MS. In contrast, samples collected during fieldwork campaign-2 (Table 3.2) were analyzed by ICP-MS at the labGEOTOP laboratory of the Institut de Ciències de la Terra Jaume Almera (Spain). Samples from Corvo were determined in three different laboratories. The major and trace element concentrations of lava flow samples collected in fieldwork campaign-1 (Table 3.2) were determined the same year at the ACTLAB laboratory (Canada), using an ICP-OES for major elements and ICP-MS for trace elements. Whole rock analyses for dikes (campaign-1) and xenoliths (campaign-2) were carried out in 2006 (Table 3.2). Major element contents were analyzed by X-Ray Fluorescence Spectrometry at the Servicio de Espectrometría y Difracción de rayos X of the Universidad de Oviedo, and trace elements were analyzed at the Centro de Instrumentación Científica of the Universidad de Granada by ICP-MS (see chapter 8 for analytical details and procedures). Sample preparation and analysis for Sr, Nd and Pb isotope ratios were performed at the Isotope Geochemistry & Mass Spectrometry Laboratory of the Miami University (Ohio, USA) during a research stay (August - December 2012) under the supervision of Prof. Dr. Elisabeth Widom (Fig. 3.4; see details in chapter 6). Samples were crushed in an alumina jaw-crusher and powdered in a high-purity alumina mixer mill. The isotopic compositions of Sr, Nd and Pb were measured by thermal ionization mass spectrometry (TIMS). 26 Chapter 3: Materials and Methods Fig. 3.4 - Sampling preparation for isotopic analysis. Process of extraction and purification of (a) Sr and (b) Nd carried out at the Geochemistry & Mass Spectrometry Laboratory of the Miami University. 40Ar/39Ar geochronology was carried out on feldspars and groundmass separates at the Argon Geochronology laboratory of the VU University Amsterdam (The Netherlands). Two research stays were done under the supervision of Prof. Dr. Jan R. Wijbrans. The separates were prepared at the Mineral Separation Laboratory of the VU University Amsterdam during the first research stay (May 2011); each sample was crushed, sieved and separated (groundmass or feldspar crystals) using heavy liquids and Frantz® magnetic separator if necessary. The separated fractions were purified by HNO3-leaching and hand-picking under a binocular microscope (Fig. 3.5). The final separates together with laboratory standard packages were sent for irradiation to the Cd-lined RODEO facility of the High Flux Reactor (HFR) at Petten (The Netherlands). Fig. 3.5 - Photomicrographs of a 400 - 500 µm homogeneous groundmass (left) and plagioclase (right) grain separates after being selected by hand-picking prior to irradiation for Ar dating. 27 Part I: Introduction Experiments on Ar isotope ratios were carried out during the second research stay (February - March 2012) to obtain crystallization ages. The separates were first preheated and then analyzed with a CO2 laser heating system by incremental heating (groundmass) or total fusion (feldspars). The purified gas fractions were analyzed with an Automated Gas Extraction System (AGES) noble gas quadrupole mass spectrometer (see details in chapter 5). 3.3 Data treatment Mineral and whole-rock geochemical data were mainly treated with ad-hoc built spreadsheets. Structural formulae were calculated for each EMP spot analysis on minerals for classification and modeling purposes. The different crystal populations (xenocrysts, antecrysts, phenocrysts, and microcrysts) were determined after a careful evaluation of mineral-melt equilibrium (olivine, clinopyroxene and amphibole; see chapters 6 and 8). Recalculations and normalizations were applied to whole rock compositions as appropriate. Trace element fractional crystallization models were calculated from whole-rock compositions and mineral modes. A large geochemical dataset was constructed with new and previously published isotopic data from the Azores archipelago for the discussion section in chapter 6. The major element fractional crystallization models were carried out using the MELTS algorithm (Ghiorso and Sack, 1995; Asimow and Ghiorso, 1998; http://melts.ofmresearch.org). MELTS is a software package designed to facilitate thermodynamic modeling of phase equilibria in magmatic systems and quantify natural processes and test petrological and mineralogical hypotheses (see chapters 6 and 8). The contribution and effect of antecrysts to the whole rock composition of porphyritic rocks has been quantified for the first time in the Azores archipelago (Larrea et al., 2013; see chapter 8) using the MINSQ software (Herrmann and Berry, 2002). It is a least squares regression model that calculates the proportions of constituent phases in a rock according to the composition of the phases and the whole rock (see chapter 8). 28 Chapter 3: Materials and Methods The ArArCALC2.5 dedicated data reduction software (Koppers, 2002; http://earthref.org/tools/ararcalc/) was used for the calculation of 40Ar/39Ar ages. Previously published K/Ar and 14C ages were compiled and discussed together with new Ar ages to constrain the temporal evolution of Graciosa Island. 3.4 Presentation of results Petrological, mineralogical and geochemical data have been interpreted and discussed to characterize individual volcanological units and investigate the magmatic evolution of Graciosa and Corvo Islands. Moreover, the ages obtained in Graciosa have allowed to better constrain the temporal evolution of this island. This Thesis contains material published, accepted or currently under review for publication in peer-reviewed journals:  Chapter 5: Larrea, P., Wijbrans, J.R., Galé, C., Ubide, T., Lago, M., França, Z. and Widom, E. (2014). 40Ar/39Ar constraints on the temporal evolution of Graciosa Island, Azores (Portugal). Bulletin of Volcanology, doi: 10.1007/s00445-014-0796-8.  Chapter 6: Larrea, P., Galé, C., Ubide, T., Widom, E., Lago, M. and França, Z. (under review). Magmatic evolution of Graciosa island (Azores, Portugal). Journal of Petrology.  Chapter 8*: Larrea, P., França, Z., Lago, M., Widom, E., Galé, C. and Ubide, T. (2013). Magmatic processes and the role of antecrysts in the genesis of Corvo island (Azores archipelago, Portugal). Journal of Petrology 54, 769-793. * The geochemical dataset of this chapter is available in the global geochemical database GEOROC (http://georoc.mpch-mainz.gwdg.de/georoc/) 29 Part I: Introduction In addition, some of the results have been also presented in numerous national and international meetings:  Larrea, P., Widom, E., Galé, C., Ubide, T., Lago, M. and França, Z. (2013). New Sr-NdPb isotopic data on Graciosa island lavas (Azores). Mineralogical Magazine 77 (5), 1547. Goldschmidt Conference (poster).  Larrea, P., Galé, C., Ubide, T., Widom, E., Lago, M., França, Z. and Tierz, P. (2012). Magmatic Evolution of the Western Azores Islands (Corvo and Flores). Abstract DI51A-2353, AGU 2012 Fall Meeting (poster).  Larrea, P., Lago, M., França, Z., Widom, E., Galé, C., Ubide, T., Arranz, E. and Tierz, P. (2012). The influence of antecrysts on the whole-rock composition of lava flows and dykes from Corvo Island (Azores). Geophysical Research Abstracts 14, EGU2012-115. EGU General Assembly (poster).  Larrea, P., Lago, M., França, Z., Widom, E., Galé, C., Ubide, T., Tierz, P. and Sanz, T. (2012). Equilibrios mineral-fundido en las lavas de la Isla Graciosa (Azores, Portugal): inferencias sobre la etapa proto-isla. Geogaceta 53, 77-80.  Larrea, P., França, Z., Lago, M., Widom, E., Galé, C., Ubide, T., Arranz, E., Forjaz, V.H., Ribeiro, L. and Pueyo, O. (2011). Origin of lavas and xenoliths of Graciosa volcanic island (Açores, Portugal). Abstract #1921. IUGG General Assembly (poster).  Larrea, P., Lago, M., França, Z., Widom, E., Galé, C., Ubide, T., Forjaz, V.H., Pueyo, O., Arranz, E. and Tierz, P. (2011). Magma chamber recharge in Corvo volcanic Island (Açores, Portugal). Geophysical Research Abstracts 13, EGU2012-6867. EGU General Assembly (poster). 30 Chapter 3: Materials and Methods  Larrea, P., Lago, M., França, Z., Widom, E., Galé, C., Ubide, T. and Arranz, E. (2010). Cumulate origin of the alkaline xenoliths from Graciosa Island (Açores, Portugal): geochemical modelling. Geogaceta 49, 75-78.  Larrea, P., Lago, M., França, Z., Widom, E., Galé, C., Ubide, T. and Pueyo, O. (2010). Mineralogy of gabbroic xenoliths in the Graciosa Island Vulcão Central Unit (Azores, Portugal). Macla 13, 133-134. Spanish Mineralogical Society Meeting (oral).  Larrea, P., Lago, M., França, Z., Widom, E., Galé, C., Ubide, T. and Arranz, E. (2010). Geochemistry of gabbroic xenoliths in the Graciosa Island Vulcão Central Unit (Azores, Portugal). Macla 13, 131-132. Spanish Mineralogical Society Meeting (oral). 31 PART II: STUDY OF GRACIOSA ISLAND Chapter 4: Graciosa Island: A Review Fig. 4.5 - Description of the volcanostratigraphic sequence of Graciosa Island. Equivalence between Gaspar (1996) subunits and those used in the present study. 41 Part II: Study of Graciosa Island Fig. 4.6 - Geological map of Graciosa Island modified from Gaspar and Queiroz (1995). Contour interval: 50 m. Volcanostratigraphic units are described in Fig. 4.5. hawaiites; Gaspar, 1996; Almeida, 2001; Larrea et al., under review). Feraud et al. (1980) obtained a K-Ar age of 620 ± 120 ka, whereas preliminary Rb-Sr data (White et al. 1976) suggest older ages of ca. 2.5 Ma (Table 4.1). The Serra Branca volcanic complex (hereafter referred to as Serra Branca) was formed by the evolution of the Serra das Fontes magmatic system, which eventually produced a central composite volcano (Gaspar, 1996). Most of the products of Serra Branca have been eroded and covered by younger volcanic rocks (Vitória - Vulcão Central volcanic complex rocks), so at the present time this complex is only exposed in the W coast (Serra Branca massif; Fig. 4.7), in the central plateau, and in isolated outcrops on the coast to the north of Quitadouro volcano (Fig. 4.6). It is mainly composed of evolved products (Gaspar, 1996; Almeida, 2001; Larrea et al., under review), including minor lava flows 42 Chapter 4: Graciosa Island: A Review (SB-a subunit in Figs. 4.5 - 4.6) and highly explosive volcaniclastic deposits (pyroclastic flows as pumice fall deposits, ignimbrites, surges and lahars; SB-b and SB-c subunits in Figs. 4.5 - 4.6). Two samples from Serra Branca were analyzed by K-Ar (Feraud et al., 1980), giving an age range from 350 ± 40 to 260 ± 20 ka (Table 4.1). Fig. 4.7 - Graciosa Island panoramic view taken from the Vulcão Central northern flank, showing Serra das Fontes and Serra Branca massifs, Quitadouro volcano and Pico Timão and related youngest lava flow. The Vitória - Vulcão Central volcanic complex is the most recent and comprises two units: Vitória and Vulcão Central. The Vitória Unit was formed after a period of volcanic inactivity in Graciosa, allowing for the erosion of the former central composite volcano (Gaspar, 1996). Vitória Unit volcanism was characterized by weakly explosive volcanism (isolated scoria and spatter cones; VU-b, d subunits in Figs. 4.5 - 4.6) and the emission of their associated lava flows (VU-a, c subunits in Figs. 4.5 - 4.6), which formed the present Vitória Unit NW platform (Gaspar, 1996; Fig. 4.8). Fig. 4.8 - Panoramic view of the Vitória Unit NW platform. Note the isolated monogenetic cones distributed throughout the Vitória Unit platform. 43 Part II: Study of Graciosa Island In contrast, the formation of the Vulcão Central Unit started to the south of the oldest volcanic structures. Pre-caldera volcanism was initially strombolian, forming monogenetic volcanoes that alternated with minor surtseyan eruptions (Maund, 1985; Gaspar, 1996; VCU-a subunit in Figs. 4.5 - 4.6 and Fig. 4.9). Afterwards, volcanism became associated with a main eruptive center, forming a stratovolcano (the Vulcão Central). This period was characterized by alternating lava flows (VCU-b, d, subunits in Figs. 4.5 - 4.6 and Fig. 4.9) and associated pyroclastic deposits (VCU-c subunit in Figs. 4.5 - 4.6) that occasionally became more explosive (plinian to sub-plinian pumitic deposits and lahars). Fig. 4.9 - Photographs of Graciosa Vulcão Central Unit: (a) pre-caldera VCU-a subunit, (b) Vulcão Central panoramic view, showing the pre-caldera Baia da Folga lava flow bearing xenoliths (VCU-d subunit), the caldera and its associated hydromagmatic sequence (VCU-e subunit), and the postcaldera volcanic dome (VCU-f subunit) and lava lake flow (VCU-g subunit). Afterwards, the caldera forming event was characterized by highly explosive hydromagmatic eruptions (syn-caldera stage), which produced the present caldera and the most extensively exposed pyroclastic sequence (products of stratovolcano draining during a particularly large and violent eruption; Maund, 1985; Fig. 4.9). This hydromagmatic sequence consist of airfall, base surge and lahar deposits (VCU-e subunit 44 Chapter 4: Graciosa Island: A Review Table 4.1 - Compilation of the available geochronological data of Graciosa Island prior to this PhD Thesis. It includes K/Ar ages from Feraud et al. (1980), 14C ages from Maund (1985) and Gaspar (1996) and an RbSr apparent age calculated by White et al. (1976). Ages are plotted in Fig. 4.10 according to their location as described in each work. Author Method Age Sample Location Feraud et al. (1980) K/Ar 620 ± 120 Ka Basaltic lava Caminho da Igreja; 400 m NNE from Guadalupa 350 ± 40 ka Trachytic lava 250 m W from Canada Longa 260 ± 40 ka Trachytic lava 30 m NW from Serra Branca top White et al. (1976) Rb/Sr 2. 5 Ma Rb -Sr apparent age - Gaspar (1996) 14 C 12200 ± 90 y Wood fragment Fonte de Areia Maund (1985) 14 C 12740 ± 70 y Wood fragment Beira Mar ; 15 m below cliff top in soil 29660 ± 580 y Wood fragment Caldeira, midway between Baía Engrade and Baía Homizados 31350 ± 280 y Wood fragment Baía Homizados, 40 m below cliff top in basalt scoria 31407 ± 300 y Peat / wood Baía Engrade, 50 m below cliff top in soil horizon in Figs. 4.5 - 4.6; Maund, 1985). The post-caldera stage was mainly effusive, forming a lava dome (VCU-f subunit in Figs. 4.5 - 4.6 and Fig. 4.9) on the caldera southern flank and a lava lake (VCU-g subunit in Figs. 4.5 - 4.6 and Fig. 4.9) in the caldera. This lava lake overspilled the rim at two points before partially draining out (Maund, 1985), allowing the subsequent formation of two small intracaldera pyroclastic cones (VCU-h subunit in Figs. 4.5 - 4.6). Compositionally, the Vulcão Central Unit comprises a wide range of compositions (basaltic to trachytic), whereas the Vitória Unit is mostly basaltic (Gaspar, 1996; Almeida, 2001; Larrea et al., under review). The Vitoria Unit has not been dated before. In contrast, several 14C ages were determined in peat and wood fragments found within pyroclastic deposits (Maund, 1985; Gaspar, 1996; Table 4.1) from the Vulcão Central Unit. These ages range from 31 ± 0.3 to 12.2 ± 0.09 ka, but only some of them (mentioned below) were used to constrain the temporal evolution of the island. Maund (1985) based on his knowledge of the geology of the island, indicates that most of the Vulcão Central Unit was formed from 50 to 10 ka (excluding the earliest and inaccessible deposits, which are probably older than 50,000 years). 45 Part II: Study of Graciosa Island Fig. 4.10 - Geological map of Graciosa Island simplified from Gaspar (1996), showing geochronological data from Feraud et al. (1980), Gaspar (1996) and Maund (1985) (ages given in thousands of years); see Table 4.1. Contour interval: 50 m. Gaspar (1996) mentioned a pumice deposit found within the Vitória Unit volcanic sequence, which seems to be related to the Vulcão Central plinian activity (12.2 ± 0.09 ka; Gaspar, 1996). Following this assumption, Gaspar (1996) suggested that most volcanic products of the Vitória Unit are older than the caldera-formation event. In addition, the Pico Timão - Vitória subunit (VU-c, d in Figs. 4.5 - 4.6) has been proposed as the most recent eruption on Graciosa Island (Walker unpublished data, in Maund (1985); Fig. 4.7). These authors correlate a thin pumice horizon lying below a Pico Timão lava flow (VU-c) with a pumice bed that originated from a plinian eruption on Faial Island at approximately 2 ka. 46 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island CHAPTER 5: 40Ar/ 39Ar CONSTRAINTS ON THE TEMPORAL EVOLUTION OF GRACIOSA ISLAND 5.1 Introduction Most of the available ages in Azorean Islands were determined using K-Ar (lava samples) and 14C (peat, charcoal and carbonized wood fragments) dating techniques. Only a few recent studies of São Miguel (Johnson et al., 1998), Terceira (Beier, 2006; Calvert et al., 2006; Gertisser et al., 2010) and São Jorge (Ribeiro, 2011) have used the more recent 40Ar/39Ar dating method. In order to better constrain the temporal evolution of the Azores archipelago and the potential magmatic and tectonic controls, it is critical to obtain high quality age data. Previous studies on the petrology and geochemistry of the volcanism on Graciosa Island included K-Ar and 14C datings (Feraud et al., 1980; Maund, 1985; Gaspar, 1996) and provided ages from 620 to 2 ky for the volcanic activity on the island (see chapter 4). In contrast, Rb-Sr data yielded ages up to 2.5 Ma for the formation of the island (White et al., 1976). In order to better constrain the age of the individual volcanological units and reconstruct the volcanic evolution of the island, a new 40Ar/39Ar dating campaign has been carried out. For this purpose lava samples from all of the recognized volcanological units and a genetically related gabbro cumulate (Larrea et al., under reviewer; chapter 6) have been analyzed. The new 40Ar/39Ar ages, together with an evaluation of the previously published ages, enable a new geochronological characterization of the development of the volcanism on Graciosa Island. 47 Part II: Study of Graciosa Island 5.2 Samples and dating methodology 5.2.1 Location and description of samples A total of 10 samples were selected for dating using the 40Ar/39Ar method (UTM coordinates in chapter 3, Table 3.1). These samples are categorized according to the volcanostratigraphy described in chapter 4. The sample set comprises 9 lava flows from the three different volcanic complexes and 1 gabbro xenolith (Fig. 5.1). One lava flow was analyzed from the Serra das Fontes shield volcano. The oldest part of the volcanic complex is exposed in inaccessible cliffs. Therefore, the oldest accessible Serra das Fontes lava flow was collected on the SW flank of the Serra das Fontes massif (GRZF8 from SF unit; Fig. 5.1) in order to better define the age of this volcanic complex. A sample from the same outcrop was previously dated by the K-Ar method (Feraud et al., 1980; Fig. 4.10). One sample from the Serra Branca volcanic complex (GRZF20 from SB-a subunit; Fig. 5.1) was analyzed. The evolution of the Serra das Fontes - Serra Branca magmatic system produced a central composite volcano (Gaspar, 1996; see chapter 4). Therefore, GRZF20 sample was collected in the central plateau outcrop as it represents the oldest accessible part of this volcanic complex and allows to better define the maximum age of the Serra Branca volcanic complex. The oldest age (350 ± 40) obtained by Feraud et al. (1980) within Serra Branca was sampled from the same outcrop (Fig. 4.10). The other seven samples were collected from the youngest volcanic complex because lava flows from the Vitória and Vulcão Central units have not been dated before. The Vitória Unit was formed by coalescence of volcanic products from different scoria and spatter cones. As a result, it is difficult to ascertain the precise area in which the formation of the NW platform started. Nevertheless, the NE coast was selected because the volcanic succession is well exposed and accessible. Lava flows from the bottom and 48 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island Fig. 5.1 - Geological map of Graciosa Island (modified from Gaspar and Queiroz (1995)) indicating the sample locations from this study (lava flows: white and black circles; xenolith: white and black star). Contour interval: 50 m top of a cliff sequence were selected for analysis (GRZF22 and GRZF27 samples respectively, from VU-a subunit; Fig. 5.1) in order to define the range of ages for the formation of the Vitória platform in the NE coast. In contrast, the volcanostratigraphy of the Vulcão Central Unit is well defined (Maund, 1985; Gaspar, 1996) and a systematic 49 Part II: Study of Graciosa Island sampling of the volcanic sequence from bottom to top of the unit is easier. The oldest subunit (VCU-a; Fig. 5.1), related to the initial monogenetic volcanoes, is only exposed in the southernmost extreme of the island and was not accessible during the fieldwork campaign owing to severe sea conditions. However, lava flow samples from the other Vulcão Central stratovolcano subunits were collected (Fig. 5.1). Pre-caldera samples from the VCU-b (GRZF14) and VCU-d (GRZF1) subunits were sampled in the coastline (Fig. 5.1) in order to obtain ages for these lower stratovolcano subunit. The syn-caldera hydromagmatic sequence was not sampled because these pyroclastic products are not appropriate for Ar dating purposes. However, two post-caldera samples were analyzed, delimiting the age of the caldera forming event. The post-caldera samples belong to the VCU-f lava dome subunit (GRZF16) and the VCU-g lava lake subunit (GRZF19). In addition, the most recent lava flow from the Pico Timão eruption center (GRZF3 from VU-c; Fig. 5.1) was sampled, as it may represent the most recent eruption within Graciosa Island (Walker unpublished data, in Maund (1985); see chapter 4). One gabbro xenolith hosted in a Vulcão Central lava flow (VCU-d; Fig.5.1) has also been analyzed (GRZF1x1). It has been interpreted as a magma chamber cumulate, formed during fractional crystallization of Graciosa basaltic magmas within the crust (Larrea et al., under review; chapter 6). This age is important to better constrain the magmatic evolution of the island. Thorough petrographic studies were performed on all samples selected for 40Ar/39Ar dating (see chapter 6). Most lava samples have porphyritic (Fig. 5.2A) and microporphyritic (Fig. 5.2B) textures. They are characterized by the presence of mmto cm-sized crystals of olivine, clinopyroxene, feldspars and less abundant amphibole and Fe-Ti oxides, embedded in a holocrystalline to hypocrystalline groundmass. The groundmass is composed of microcrysts of olivine, feldspars, clinopyroxene and Fe-Ti oxides. The microcrystalline groundmass was selected for 40Ar/39Ar analysis in all of these samples. In contrast, sample GRZF20 from Serra Branca is more evolved and exhibits a trachytic texture (Fig. 5.2C). It consists of isolated < 1 cm alkali feldspar 50 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island radiogenic argon, an inverse isochron age of 96.3 ± 32.3 ka was obtained for this lava sample. Sample GRZF27 (groundmass) produced relatively flat age and K/Ca spectra (Fig. 5.4). The high temperature steps yield irregular ages and a decrease in K/Ca that might be related to contamination of the groundmass separate with low K/Ca phases supplying inherited argon. Only consecutive steps that had similar K/Ca results were considered for the age calculation. The inverse isochron treatment of the data showed the low radiogenic argon content of the sample and the existence of excess argon. Hence, the inverse isochron age of 45.7 ± 22.0 ka is considered the most accurate age for this groundmass separate. Sample GRZF14 (groundmass) produced an age spectrum with abundant atmospheric argon in the first step and inherited argon in the last steps (Fig. 5.5). The steps selected for the plateau age are relatively homogeneous K/Ca values, whereas the sharp decrease of the K/Ca values in the last steps denotes the presence of contaminant phases. Excess argon was detected by the inverse isochron treatment of the data, so the 56.3 ± 21.1 ky value is considered the most likely age for this sample. Sample GRZF1 (groundmass) yielded an age spectrum with abundant atmospheric argon in the first two steps and extraneous argon in the last steps (Fig. 5.5). Even though the selected steps for the plateau age are fairly flat, the associated K/Ca values are heterogeneous, probably due to a change in the groundmass composition. Moreover, the inverse isochron treatment of the data confirmed excess argon in the samples. Consequently, the well-defined isochron age of 58.8 ± 19.2 ka is preferred. Sample GRZF1x1 was dated using 7 total fusion experiments of several plagioclase crystals. The result of each experiment is represented as a step in Fig. 5.6. The 40Ar/39Ar experiment yielded a slightly disturbed age spectrum, but most values are within the range of analytical error (Fig. 5.6). The K/Ca values are homogeneous, indicating the purity of the separate. 57 Part II: Study of Graciosa Island Fig. 5.4 - 40Ar/39Ar incremental heating results for samples GRZF22 and GRZF27. Explanation is as in Fig. 5.3. Fig. 5.5 - 40Ar/39Ar incremental heating results for samples GRZF14 and GRZF1. Explanation is as in Fig. 5.3. 58 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island All of the 39Ar released was considered for calculating the plateau and the inverse isochron ages, which were very similar. No excess argon was found as the 40Ar/36Ar intercept was concordant with atmosphere values. Hence, the plateau age of 865.6 ± 61.1 ka is preferred due to its lower MSWD value (Table 5.2). Sample GRZF16 (groundmass) produced an age spectrum with abundant atmospheric argon in the first step and probably excess argon in the last steps as indicated by the homogeneous K/Ca spectrum (Fig. 5.6). The isochron treatment of the data indicated the presence of excess argon, as the 40Ar/36Ar intercept was higher than the atmosphere value. Accordingly, the well-defined inverse isochron age of 47.6 ± 16.6 ka is considered the most accurate estimate of the age of the sample. Fig. 5.6 - 40Ar/39Ar single fusion results for sample GRZF1x1 and incremental heating results for sample GRZF16. Explanation is as in Fig. 5.3. Groundmass samples GRZF19 and GRZF3 belong to the youngest lava flows on the island. They have low amounts of radiogenic argon, probably due to their low K2O content, and therefore are near the limit of the technique (Wijbrans et al., 2011). In both samples, the age spectrum was somewhat disturbed but ca. 90 % of the 39Ar released 59 Part II: Study of Graciosa Island was considered for calculating the plateau and the inverse isochron ages (Fig. 5.7). Steps with the lowest K/Ca were probably related to certain contamination of the groundmass separate that may have supplied extraneous argon, so they were omitted for the age calculations. Excess argon was found as the 40Ar/36Ar intercepts were higher than atmosphere values and consequently, the inverse isochron age was preferred in both cases. Therefore, GRZF19 and GRZF3 are inferred to be 11.1 ± 3.7 ka and 3.9 ± 1.4 ka old, respectively. Fig. 5.7 - 40Ar/39Ar incremental heating results for samples GRZF19 and GRZF3. Explanation is as in Fig. 5.3. 60 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island 5.4 Discussion 5.4.1 40Ar/39Ar constraints on the temporal evolution of Graciosa Island The obtained 40Ar/39Ar crystallization ages are plotted in Graciosa Island geological map (Fig. 5.8), and in the chronostratigraphic chart (Fig. 5.9) together with previous K-Ar ages dated by Feraud et al. (1980). Serra das Fontes volcanic complex The stratigraphically oldest accessible lava flow yielded a 40Ar/39Ar age of 1056.9 ± 28.0 ka (GRZF8 – SF unit). Another sample from the same outcrop was previously analyzed by K-Ar (Feraud et al., 1980), giving a younger age of 620 ± 120 ka (Fig. 5.9). Given the large error reported by these authors, the new result is considered more accurate. In accordance, the oldest accessible subaerial part of the island has an age of 1056.9 ± 28.0 ka and this is therefore a minimum age for the formation of the Serra das Fontes volcanic complex. Serra Branca volcanic complex Two previous K-Ar ages for Serra Branca were obtained by Feraud et al. (1980) (Fig. 5.9). These authors collected samples from two different sites, yielding ages of 260 ± 20 ka (Serra Branca massif) and 350 ± 40 ka (central plateau outcrop). In order to better constrain the oldest age of the unit, the oldest sample was analyzed from the central plateau outcrop (GRZF20 – SB-a subunit). The 40Ar/39Ar age of 433.5 ± 12.9 ka is slightly older than the age reported by Feraud et al. (1980). The Serra Branca volcanic complex represents an evolved stage of the Serra das Fontes magmatic system, forming the Serra das Fontes - Serra Branca composite volcano 61 Part II: Study of Graciosa Island Fig. 5.8 - Geological map of Graciosa Island modified from Gaspar and Queiroz (1995), indicating the calculated 40Ar/39Ar ages (given in thousands of years) for the studied samples (lava flows: white and black circles; xenolith: white and black star). Contour interval: 50 m. Legend as in Fig. 5.1. (Gaspar, 1996; Larrea et al., under review). In accordance, magmatic activity related to the Serra Branca evolved products started ca. 600 ky after the first subaerial basaltic volcanism within the island (Fig. 5.9). Vitória Unit The Vitória Unit has been dated for the first time. Lava flows from the bottom (GRZF22 – VU-a subunit) and top (GRZF27 – VU-a subunit) of a well exposed volcanic succession in the NE coast produced 40Ar/39Ar ages of 96.3 ± 32.3 ka and 45.7 ± 22.0 ka, respectively (Fig. 5.9). These results indicate that the formation of Vitória may have started after a 62 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island Fig. 5.9 - Chronostratigraphic chart for Graciosa Island volcanism. Circles (lava flows) and star (xenolith) represent the 40Ar/39Ar ages from this study. Vertical lines represent 1σ uncertainties associated with our calculated radiometric ages when larger than the symbol size. Squares represent K/Ar data from Feraud et al. (1980). long period of volcanic inactivity on the island (ca. 160 ky based on the youngest age - 270 ± 20 ka - proposed by Feraud et al. (1980) for the Serra Branca volcanic complex; Fig. 5.9), during which the older volcanic edifices were eroded as suggested by Gaspar (1996). The Pico Timão volcanic eruption belongs to the Vitória Unit. This eruption formed the most recent lava flow (GRZF3 – VU-c subunit) within the island (Fig. 5.9). Previous authors proposed that this subunit was younger than 2 ka because it is located over a thin pumice horizon related to a plinian eruption on Faial Island that occurred at ca. 2 ka (Walker unpublished data, in Maund (1985)). The 40Ar/39Ar dating experiment yielded an age of 3.9 ± 1.4 ka for the eruption of the lava flow. Therefore, the lava flow seems to be slightly older than previously inferred. The pumice horizon therefore may not be related to the plinian eruption in Faial Island, but rather came from elsewhere. 63 Part II: Study of Graciosa Island Vulcão Central Unit The Vulcão Central volcanostratigraphy was established based on field observations (Maund, 1985; Gaspar, 1996). The oldest part of the unit (VCU-a subunit) comprises monogenetic volcanoes formed to the south of the Serra das Fontes – Serra Branca oldest volcanic complexes. This subunit was not accessible during the sampling campaign, so its age remains unknown, but lava flows from the rest of the Vulcão Central subunits were dated by 40Ar/39Ar. Pre-caldera samples GRZF14 and GRZF1 were taken at the bottom of VCU-b and VCU-d subunits. They have 40Ar/39Ar ages of 56.3 ± 21.1 ka and 58.8 ± 19.2 ka, respectively (Fig. 5.9), similar to the oldest age of ca. 50 ka proposed by Maund (1985). Therefore, the formation of the Vulcão Central stratovolcano may have started prior to 50 ka. Post-caldera sample GRZF16 (VCU-f subunit) belongs to an evolved dome located on the southwestern flank of the Vulcão Central stratovolcano. Field evidence suggests that this subunit formed subsequent to the formation of the caldera and its related hydromagmatic sequence (Maund, 1985; Gaspar, 1996). The age of 47.6 ± 16.6 ka obtained for the evolved dome therefore requires an older age for the formation of the caldera (syn-caldera stage; Fig. 5.9). A previous estimate for the age of the calderaforming event suggested a relatively young age of 12.74 ± 0.070 ka , based on 14C dating of a carbonized wood fragment found within the caldera hydromagmatic sequence (VCU-e subunit) (Maund, 1985; see chapter 4). Given that this age was based on a single analysis of a wood fragment for which the location and description are vague, the new 40Ar/39Ar age of 47.6 ± 16.6 ka for the evolved dome sample is considered to represent a more reliable estimate of the minimum age for this subunit. According to Maund (1985), subsequent effusion of basaltic magmas into the caldera resulted in the formation of a lava lake. Lava filled the caldera up to a height of 240 m and overspilled the rim at two points before partially draining out (Maund, 1985). Sample GRZF19 (VCU-g subunit) was collected from the lava flow overspill, and gave an age of 11.1 ± 3.7 ka for this event (Fig. 5.9). The last eruptions of the Vulcão Central Unit 64 Chapter 5: 40Ar/39Ar constraints on the temporal evolution of Graciosa Island formed two small pyroclastic cones situated within the caldera; these were not considered in this study due to the lack of associated lava flows (VCU-h subunit). Finally, a gabbroic xenolith (GRZF1x1) hosted in the GRZF1 lava flow (VCU-d subunit) from the Vulcão Central Unit yielded a 40Ar /39Ar age of 865.6 ± 61.1 ka (Fig. 5.9). This age is significantly older than that of the host lava (58.8 ± 19.2 ka) and suggests that the formation of the gabbro was linked to the Serra das Fontes rather than to the Vulcão Central Unit (Fig. 5.9). These results indicate that gabbro xenolith, interpreted to represent a magma chamber cumulate (Larrea et al., under review) was formed by fractional crystallization of the Serra das Fontes basaltic magmas and subsequently carried to the surface by the younger GRZF1 lava flow. 5.4.2 Comparison of Graciosa temporal evolution with other oceanic islands New 40Ar/39Ar ages on Graciosa Island volcanic products have constrained the Pleistocene and Holocene temporal evolution of the subaerial volcanism (1056 ± 28.0 - 3.9 ± 1.4 ka). According to these new ages, Graciosa Island is contemporaneous with São Jorge, and older than Terceira and the southern Islands of Pico and Faial (see chapter 2). Overall, this Central Group is contemporaneous with the Western Group (Corvo and Flores Islands) but younger than the Eastern Group. The age of São Miguel remains unclear, while the easternmost island of Santa Maria is clearly the oldest (see chapter 2). If the age of the Azores archipelago is compared to well-studied Hawaii and Canary Islands, Azores is clearly the youngest archipelago. However, the formation of Graciosa is contemporaneous with the youngest islands of these archipelagos: the western group of islands from the Canary Islands (El Hierro and La Palma; Carracedo, 2011), and the southeastern islands of Hawaii (East Maui, Kaho’olowe and Hawai’i; McDonald et al., 1983). Four clearly differentiated stages are observed in the evolution of oceanic islands (e.g., Walker, 1990; Carracedo, 2011). These stages are (1) the submarine stage (frequently poorly studied) which comprises up to 80 % of the volume of an island, under the sea level, and the subsequent subaerial stages: (2) the shield stage, (3) the erosional stage 65 Part II: Study of Graciosa Island and (4) the rejuvenated stage. These stages can be recognized in the temporal evolution defined for Graciosa Island, although the submarine stage is not exposed. The shield stage comprised the formation of the Serra das Fontes shield volcano (starting at least at 1.05 Ma) and its evolution to the Serra Branca composite volcano. Therefore, this stage lasted ca. 850 ky (Fig. 5.9). The erosional stage was characterized by volcanic inactivity and erosion of the Serra das Fontes - Serra Branca volcanic edifices and lasted ca. 110 ky (Fig. 5.9). Currently, the island is in the rejuvenated stage, represented by the youngest and most extensive volcanic complex on the island (Vitória - Vulcão Central). On the NW part of the island, this stage started prior to ~ 100 ka as monogenetic cones that built the Vitória Unit NW platform. In contrast, in the SE part of the island, this stage started prior to 60 ka. Volcanism was initially associated with monogenetic cones and later became associated with a main eruptive center forming the Vulcão Central stratovolcano. Present-day activity continues as a boiling mud pool in the Furna do Enxofre (caldera cave) and hot water springs at the Termas do Carapacho and the Baía dos Homiziados. Overall, the evolutionary stages defined in Graciosa Island are slightly different than those of Hawaii and the Canary Islands. Each stage appears to be shorter in time in Graciosa, as together the shield, erosional and rejuvenated stages lasted only slightly more than 1 Myr. For example, Gran Canaria Island shows a temporal evolution similar to Graciosa but it formed over ca. 16 Myr (Schminke, 2004). Similarly, El Hierro and La Palma Islands have similar ages to Graciosa but are still in the shield stage (Carracedo, 2011). The erosional stage in Graciosa was very short (ca. 110 ky) but highly destructive. For instance, Gomera Island has an age of 10 Ma and is within the erosive phase since 4 Ma. In Graciosa Island, the rejuvenated stage has been extremely active, generating most of the current surface of the island as the Vitória NW platform and Vulcão Central SE stratovolcano, although present-day activity is scarce within the island. The rejuvenated stage is also particularly active in the older, eastern and central Canary Islands which have experienced three or more magmatic phases. In contrast, the volume of rejuvenated lavas in Hawaii archipelago is very small (Schminke, 2004). 66 Chapter 6: Magmatic evolution of Graciosa Island Fig. 6.1 - Simplified geological map of Graciosa Island modified from Gaspar and Queiroz (1995) indicating the sample sites from this study. Contour interval: 50 m. the results, the highest detection limits were considered for each element. Whole rock analyses can be found in Appendix - Table II. Nine lavas and 6 xenoliths were selected for Sr, Nd and Pb isotope analyses (Appendix - Table II). Sample preparation and analysis were performed at the Isotope Geochemistry 73 Part II: Study of Graciosa Island & Mass Spectrometry Laboratory of the Miami University (Ohio, USA). Samples were crushed in an alumina jaw-crusher and powdered in a high-purity alumina mixer mill. Approximately 200 mg of powdered samples were used to extract and purify Sr, Nd and Pb. Samples were digested overnight in Savillex® Teflon beakers with concentrated HF (4 ml) and HNO3 (2 ml) on a hotplate at 80 ºC. After drying, 0.5 - 1 ml of HNO3 was added to beakers, to break up the sample cake; samples were evaporated to dryness on a hotplate at 80 ºC (this step was repeated). After dryness, 1 ml of concentrated HCl was added to the samples to break up insoluble fluorides for at least 1 hour. In order to avoid undissolved materials, the samples were redissolved in 5 ml of ultrapure E-pureTM H2O, leaving them on the hot plate at 80 ºC overnight. After total dissolution, samples were evaporated to dryness on the hot plate. The Sr, Nd and Pb purified fractions were all extracted from the same digested sample. Pb was first processed using anion exchange columns with BioRad AG1-X8 100 - 200 mesh resin and HBr and HNO3. The Sr was recovered using a Eichrom Industries Sr-Resin and HCl and HNO3. The sample residue was passed through AG-50W-X8 cation resin to remove iron, and then LN-Spec columns and HCl was used for Nd purification. The isotopic compositions of Sr, Nd and Pb were measured by thermal ionization mass spectrometry (TIMS). Sr and Nd isotope ratios were corrected for fractionation using 86Sr/88Sr = 0.1194 and 144Nd/146Nd = 0.7219. Measurements of NBS 987 Sr standard and La Jolla Nd standard show a long-term reproducibility of 87Sr/86Sr = 0.71024 + 0.00002 (2 SD) and 143Nd/144Nd = 0.511846 + 0.000007 (2 SD), respectively, which represent the external reproducibilities of Sr and Nd isotopic measurements. Pb isotope ratios were corrected for fractionation by 0.1 % (206Pb/204Pb), 0.097 % (207Pb/204Pb) and 0.102 % (208Pb/204Pb) per amu, based on deviations of measured ratios in NBS 981 Pb standard from values in Todt et al. (1996). Long-term external reproducibility (2 SD) for 206Pb/204Pb, 207Pb/204Pb and 208Pb/204Pb is + 0.015, 0.020, and 0.062, respectively. Procedural blanks were better than 0.1 ng for Sr, 0.1 ng for Nd and 0.2 ng for Pb. 74 Chapter 6: Magmatic evolution of Graciosa Island 6.3 Petrology The composition of sampled lavas ranges from basalt to trachyte following the TAS classification by Le Bas et al. (1986) (Fig. 6.2). Basalts and hawaiites are the most abundant rock compositions in Serra das Fontes (48.09 - 52.47 wt. % SiO2 and 3.58 - 3.92 wt. % total alkalis (Na2O + K2O)) and the Vitória - Vulcão Central (46.18 - 55.94 wt. % SiO2 and 1.97 - 5.81 wt. % total alkalis) volcanic complexes, apart from some more evolved rocks (mugearite - trachyte; 57.12 - 59.81 wt. % SiO2 and 4.68 - 6.29 wt. % total alkalis) in the Vulcão Central Unit. On the contrary, Serra Branca is entirely trachytic in composition (> 65.44 wt. % SiO2 and > 6.37 wt. % total alkalis). These rocks are characterized by porphyritic, microporphyritic and trachytic textures (Table 3.1). Fig. 6.2 - Total Alkalis vs. Silica (TAS) diagram for Graciosa Island lava flows (after Le Bas et al. (1986)). Pc: picrobasalt; U1: tephrite; B: basalt; S1: hawaiite, S2: mugearite; S3: benmoraite; T: trachyte. * The SiO2 content of the samples analyzed in the LabGEOTOP laboratory has been estimated as SiO2 = 100 - Σ (total oxide content + LOI) given the analytical procedure in this laboratory (see AppendixTable I-II). The porphyritic rocks (Fig. 6.3A) are characterized by the presence of 15 - 50 % large crystals (2 - 10 mm) embedded in a holocrystalline to hypocrystalline groundmass. 75 Part II: Study of Graciosa Island According to crystal size, the term macrocryst is used for crystals larger than 2 mm, whereas those smaller than 2 mm are referred to as microcrysts. The macrocryst assemblage comprises subhedral to anhedral olivine, clinopyroxene, feldspars and less abundant amphibole and Fe-Ti oxides. The groundmass is composed of microcrysts of olivine, feldspars, clinopyroxene and Fe-Ti oxides. The microporphyritic rocks (Fig. 6.3B) are holocrystalline to hypocrystalline and are composed of microcrysts of feldspars, olivine, clinopyroxene and Fe-Ti oxides with a bimodal size distribution (Fig. 6.3B); most microcrysts are smaller than 0.5 mm, although crystals of 0.5 - 2 mm are also recognized. Amphibole microcrysts are also present in the evolved rocks (benmoreites and trachytes). In some microporphyritic samples are observed larger (> 1 mm) isolated olivine, clinopyroxene and feldspar crystals that are highly corroded and have rounded rims. Samples from Serra Branca exhibit a trachytic texture. They consist of isolated < 1cm tabular feldspar crystals embedded in a groundmass predominantly composed of sub76 Chapter 6: Magmatic evolution of Graciosa Island Fig. 6.4 - Xenolith sampling sites in Graciosa Island. (a) Baía da Folga sampling area; (b) distribution of the xenoliths at this site; (c) Enxudreiro outcrop; (d) Ponta da Pesqueira site; (e) Quitadouro scoria cone; (f) sampled quarry; note its location in (e). parallel, minute (< 0.5 mm), acicular to tabular feldspar microcrysts, with scarce clinopyroxene and opaque microcrysts (Fig. 6.3C). The xenoliths are 3 to 20 cm long and show sub-rounded to angular shapes and a mediumto coarse-grained plutonic textures. They present sharp contacts with the host lavas and occur without any apparent preferred orientation in the different outcrops (Baía da Folga, Enxudreiro, Ponta da Pesqueira and Quitadouro; Fig. 6.4). These 77 Part II: Study of Graciosa Island Fig. 6.5 - (a) Syenite (GRENx-1) hand specimen and (b) photomicrograph of the syenite in crosspolarized transmitted light; (c) dunite sample (GRQUx1) after sawing and as hand specimen found in the sampling site, within the inset; (d) photomicrograph of the dunite (GRQUx1) in cross-polarized transmitted light. xenoliths are classified according to Le Maitre (2002) as syenite, dunite and gabbro (Table 3.1). The latter comprises two types: alkaline and subalkaline gabbros (see mineral chemistry and whole rock chemistry sections). Two syenites were collected from a lahar deposit at Enxudreiro outcrop (Figs. 6.4C and 6.5A). They occur as an inequigranular (0.1 - 2 mm), allotriomorphic mosaic composed of major alkaline feldspar (~ 90 %), amphibole (~ 7 %), Fe-Ti oxides (~ 2 %), and pyroxene, quartz, biotite, sphene and zircon as accessory minerals (Fig. 6.5B). Two dunite samples were found at the Quitadouro outcrop (Figs. 6.4E-F and 6.5C). One of them was collected from a pyroclastic layer of the scoria cone and the other one was found within a lava flow. These samples have a phaneritic texture constituted by an equigranular mosaic of ~ 90 % olivine crystals and minor clinopyroxene crystals (~ 5 %). Cr-rich spinel crystals (~ 5 %) are located along grain boundaries or included in olivine 78 Chapter 6: Magmatic evolution of Graciosa Island Fig. 6.6 - (a) Alkaline gabbro (GRZF1x13) hand specimen; (b) photomicrograph of the alkali gabbro (GRZF1x1) in cross-polarized transmitted light; (c) subalkaline gabbro (GRQUx4) hand specimen as found in the sampling site; (d) photomicrograph of the subalkaline gabbro (GRQUx4) in cross-polarized transmitted light. grains. They have fractures but do not present infiltrating glass. The dunite found in the lava flow is smaller (< 2 cm) and shows reddish iron oxide alteration within some olivine crystals and fractures, so this sample is only used for petrographic and mineral chemistry analyses. The dunite found in the pyroclastic layer shows a mm-thick, brownish rim in contact with the host rock. This rim is composed of fine-grained olivine and feldspar crystals in a palisade arrangement (Fig. 6.5D). Alkaline gabbros were found in Baía da Folga and Ponta da Pesqueira outcrops (Figs. 6.4A-B and 6.6A). They have been described by Larrea (2010) and Larrea et al. (2010). They display a mediumto coarse-grained orthocumulate texture and are composed of variable proportions of feldspars (10 - 80 %), clinopyroxene (8 - 35 %), amphibole (5 - 70 %), olivine (0 - 20 %), Fe-Ti oxides (3 - 20 %) and apatite (up to 3 % in some samples; Fig. 6.6B). Euhedral to subhedral feldspars, olivine and clinopyroxene are the main cumulus phases. Amphibole is the principal intercumulus phase in all the cases; when it develops large poikilitic crystals (up to 2 cm), the xenolith texture can be described as 79 Part II: Study of Graciosa Island heteroadcumulate. Some alkaline xenoliths are permeated by the host lava (5-15 % of the total volume fraction; Fig. 6.6A); in these cases, a discontinuous reaction rim between the xenoliths and the host basalt has been recognized. Where the feldspars, olivine or pyroxene crystals are in contact with the host basalt, a clear contact appears; in contrast, a thick rim is observed between amphibole phases and the host basalt (see section 6.4). Subalkaline gabbros are only present in the Vitória Unit outcrops (Ponta da Pesqueira and Quitadouro; Figs. 6.4D-F and 6.6C). They consist of olivine (5 - 20 %), clinopyroxene (10 - 70 %), feldspars (8 - 80 %) and accessory opaque minerals (< 1 %). These xenoliths present inequigranular textures characterized by a clear bimodal distribution of sizes. Anhedral to subhedral crystals (1 - 2 mm in size) are surrounded by a framework of polygonal grains smaller than 0.5 mm (Fig. 6.6D). 6.4 Mineral chemistry All mineral phases and crystal types present in lava flows and xenoliths were analyzed. Major element compositions were determined by electron microprobe and are accessible in Tables 6.1 - 6.7 of the electronic supplement. 6.4.1 Olivine Olivine compositions are plotted in Fig. 6.7A. The macrocryts and microcrysts of lava flows show normal zoning, with decreasing forsterite (Fo89-60 and Fo85-55, respectively) and NiO (< 0.36 wt. %) contents, and increasing MnO (0.01 - 0.86 wt. %) and CaO (0.13 - 0.51 wt. %) concentrations from core to rim. In the gabbroic xenoliths, olivine crystals are characterized by a normal compositional zoning with Fo84-74, < 0.21 wt. % NiO, 0.11 - 0.38 wt. % MnO and < 0.13 wt. % CaO in the subalkaline gabbros, and Fo78-54, < 0.10 wt. % NiO, 0.05 - 1.37 wt. % MnO and 0.02 - 0.20 wt. % CaO in the alkaline group. The olivine crystals in dunites are weakly zoned Fo92-90 and have high NiO (0.40 - 0.19 wt. %), and low MnO (0.08 - 0.20 wt. %) and CaO (< 0.05 wt. %) concentrations. 80 Chapter 6: Magmatic evolution of Graciosa Island Fig. 6.7 - (a) Olivine CaO (wt. %) vs. Fo# contents from Graciosa lava and xenolith analyzed crystals. (b) Clinopyroxene TiO2 (wt. %) vs En# diagram ; Evol-Microcryst: microcryst analyzed within trachytes. 6.4.2 Clinopyroxene Clinopyroxene compositions are shown in Fig. 6.7B. Macrocrysts and microcrysts of clinopyroxene from the lava flows are classified as augite and diopside (Morimoto et al., 1988; Fig. 6.8). Clinopyroxene macrocrysts (Wo42-46 En49-37 Fs10-16) and microcrysts (Wo3648 En49-34 Fs15-13) analyzed in the basaltic to hawaiitic lava flows define a variation trend with 0.33 - 3.63 wt. % TiO2 and 1.05 - 10.34 wt. % Al2O3 contents. In contrast, clinopyroxene in the evolved rocks (Wo44-43 En40-20 Fs16-36), classified as augitehedenbergite, forms a clearly distinct trend characterized by high SiO2 (> 50 wt. %) and the lowest TiO2 (< 0.63 wt. %) and Al2O3 (< 1.43 wt. %) contents, similarly to trachytic rocks from Sete Cidades volcano from São Miguel (Beier et al., 2006). Clinopyroxene crystals from the gabbroic and syenite xenoliths are classified as augite and diopside. In the alkaline gabbros, clinopyroxene compositions display an alkaline trend in the Ti vs Ca+Na (per formula unit) diagram by Leterrier et al. (1982) (not shown). Clinopyroxene in subalkaline gabbros is more primitive (Wo35-48 En52-44 Fs13-8), with 0.13 - 0.95 wt. % TiO2 and 1.58 - 4.68 wt. % Al2O3, than clinopyroxene in alkaline gabbros (Wo45-47 En46-38 Fs9-15; 0.37 - 2.68 wt. % TiO2 and 1.34 - 6.95 wt. % Al2O3). Clinopyroxene crystals in the syenites are barely zoned Wo45 En42-41 Fs13-14 with intermediate TiO2 (1.36 - 1.41 wt. %) and Al2O3 (4.28 - 4.44 wt. %) contents. The clinopyroxene crystals in dunites are the most primitive (Wo46-51 En50-46 Fs3) with low TiO2 (0.13 - 0.31 wt. %) and Al2O3 (1.31 - 2.38 wt. %) contents. 81 Part II: Study of Graciosa Island Fig. 6.8 - Clinopyroxene classification diagram wollastonite - enstatite - ferrosilite (Morimoto et al., 1988). 6.4.3 Feldspars The macrocrysts and microcrysts in lava flow samples range from bytownite to anorthoclase (An82-0 Ab17-66 Or0-34). The subalkaline xenolith crystals range from anorthite to bytownite (An98-76 Ab2-24 Or0-0), whereas the alkaline gabbro crystals are bytownite to andesine (An76-31 Ab23-67 Or0-3). Syenite xenoliths are characterized by the presence of major alkaline feldspar and minor Na-rich plagioclase (An16-0 Ab81-22 Or3-78). The ternary classification diagram Ab-An-Or for feldspars is shown in Fig. 6.9. 6.4.4 Amphibole In general, amphibole crystals are scarce in Graciosa lava flows. However, they are a common macrocryst phase in the basaltic lava flow that carries the Baía da Folga alkaline gabbros (GRZF1 sample) and they appear as microcrysts in the evolved rocks. Amphibole shows quite varied compositions, classified as kaersutite, magnesiotaramite and riebeckite depending on their Ca and Na contents (Leake et al., 1997; Fig. 6.10). 82 Chapter 6: Magmatic evolution of Graciosa Island Fig. 6.13 - Isotope systematics of the studied Graciosa lavas and xenoliths (error bars smaller than symbol size when not shown). Mantle component fields based on data compiled in GEOROC (http://georoc.mpchmainz.gwdg.de/georoc) and PETDB (http://www.earthchem.org/petdb)for Atlantic MORB (MORB), Saint Helena (HIMU), Samoa and Society (EM2) and Tristan da Cunha, Gough and Pitcairn (EM1). Previous published data from Azores are from Davies et al. (1989), Hawkesworth et al. (1979), Turner et al. (1997), Widom et al. (1997), França (2000), Beier et al. (2006), França et al. (2006b), Beier et al. (2007), Elliott et al. (2007), Beier et al. (2008, 2010), Millet et al. (2009), Madureira et al. (2011), Ribeiro (2011), Beier et al. (2013) and Genske et al. (2012). 89 Part II: Study of Graciosa Island 6.6 Discussion 6.6.1 Origin of lava crystal populations The macrocrysts and microcrysts found in Graciosa lavas are olivine, clinopyroxene, plagioclase, amphibole and Fe-Ti oxides. The macrocrysts show either more primitive or similar compositions to microcrysts. Nevertheless, they define a common evolution trend (e.g., Fig. 6.7 and Fig. 6.9) pointing to a comagmatic origin for these crystals. The macrocrysts could potentially be classified as cogenetic phenocrysts or non-cogenetic antecrysts with the host groundmass (Jerram and Martin (2008) and references therein). The existence of antecrysts in the volcanic products from other Azorean Islands (Corvo and Flores) and their influence on whole rock chemistry have been investigated by Larrea et al. (2012, 2013). Mineral-melt equilibrium diagrams are useful to decipher the origin of comagmatic crystals (Rhodes et al., 1979; Fig. 6.14). Equilibrium curves respond to the iron-magnesium distribution coefficient of the mineral: 0.30 ± 0.03 for olivine (Roeder and Emslie, 1970), 0.26 ±0.05 for clinopyroxene (Akinin et al., 2005) and 0.38 for amphibole (LaTourrette et al., 1995). Most analyzed olivine, clinopyroxene and amphibole macrocryst cores plot within or just below the equilibrium curve, so they are phenocrysts in equilibrium with the rock in which they are included. Microcryst and phenocryst rims fall below the equilibrium curve due to progressive fractionation of the magma. On the other hand, some macrocryst cores plot noticeably above the equilibrium curve, so they are non-cogenetic crystals (antecrysts) in disequilibrium with their host rock. Two types of antecrysts are recognized: (1) large, isolated, highly corroded crystals in disequilibrium with the host rock, which are embedded in a finegrained groundmass (olivine, clinopyroxene and amphibole antecrysts from samples GRZF8, GRZF16, GRZF6 and GRZF13) and (2) large crystals with cores in disequilibrium overgrown by rims in equilibrium with the host rock and equivalent to the microcrysts (clinopyroxene from sample GRZF14). Some clinopyroxene crystals analyzed in trachytes from Serra Branca and Vulcão Central Unit appear to be in disequilibrium; however, the Fe-Mg distribution coefficients used for the equilibrium diagram may be not applicable to the low Mg# content of these rocks [Mg# = MgO/(FeO+MgO)·100 and FeO = 0.9·Fe2O3T]. 90 Chapter 6: Magmatic evolution of Graciosa Island 6.6.2 Magmatic differentiation in Graciosa Island Differentiation of Graciosa magmas by fractional crystallization: genesis of lavas, alkaline xenoliths and syenites All of the studied lavas from Graciosa Island fall within a single liquid line of descent, pointing to a comagmatic origin and a single fractionating process involved in their genesis. In addition, the presence of alkaline xenoliths with cumulate textures potentially indicates their origin as magma chamber cumulates (e.g., Shamberger and Hammer, 2006) and therefore, magmatic differentiation controlled by fractional crystallization. Moreover, olivine Fo78-54 and CaO contents (0.02 - 0.20 wt. %) in the alkaline gabbros indicate that they were not formed in the mantle (e.g., Stormer, 1973; Jurewicz and Watson, 1988). Fractional crystallization has been evaluated as a potential mechanism for magmatic differentiation using both major and trace elements, aiming to establish (a) the conditions of magma evolution and fractionation (crystallization pressure and temperatures), (b) the extent of fractional crystallization needed to reproduce the studied lava compositions, (c) the composition of the fractionated mineral assemblage controlling the fractional crystallization process and (d) a comparison between these fractionated phases and the studied xenoliths. A major element model was carried out using the MELTS algorithm (Ghiorso and Sack, 1995; Asimow and Ghiorso, 1998). Sample GRZF12 has a MgO content of 10.58 % and a Ni content of 200 ppm and the olivine and clinopyroxene macrocrysts are in equilibrium with the bulk rock composition (see Fig. 6.14). Therefore, sample GRZF12 represents the naturally occurring major element composition closest to a primary magma (~ 12 wt. % MgO and 200 - 300 ppm Ni contents have been proposed for Azores primitive melts (Beier et al., 2006, 2008; Madureira et al., 2011) and can be used as the starting composition for the model. The MELTS fractional crystallization model was run with cooling steps of 5 ºC under isobaric conditions at different pressures (1000, 700 and 500 MPa) and under polybaric conditions (starting at 500 with dP/dT = 1). 91 Part II: Study of Graciosa Island Fig. 6.14 - Mg# in mineral vs. Mg# in whole rock [(Mg# = MgO/(FeO+MgO) · 100 and FeO = 0.9 · Fe2O3 T ] for olivine (a), clinopyroxene (b) and amphibole (c). The black curves represent the range of equilibrium compositions between mineral and melt. Evol-Microcryst: microcryst analyzed within an evolved rock. 92 Chapter 6: Magmatic evolution of Graciosa Island Table 6.1 - Parameters of the MELTS modeling runs; the grey row indicates the best fit MELTS run. Starting P (MPa) % H2O dT (ºC) dP (Mpa) dP/dT Buffer fO2 QFM Liquidus T (ºC) Final P (MPa Final T (ºC) initial logfO2 Final logfO2 Fractionated phases 1000 2 5 0 - 0 1315 1000 900 -6.41 -11.61 Cpx, Amp, Pl, Opx, Opq, Ap, Grt 1000 1 5 0 - 0 1345 1000 900 -6.13 -11.61 Cpx, Amp, Pl, Opx, Opq, Ap, Grt, Wht 1000 0.5 5 0 - 0 1375 1000 error at 1105 -5.87 -8.65 Cpx, Amp, Pl, Opx, Opq, Grt, Wht 700 2 5 0 - 0 1285 700 900 -6.9 -11.89 Ol, Cpx, Amp, Pl, Opx, Opq, Ap, Grt 700 1 5 0 - 0 1305 700 900 -6.71 -11.89 Ol, Cpx, Amp, Pl, Opx, Opq, Ap, Grt 700 0.5 5 0 - 0 1330 700 900 -6.48 -11.89 Ol, Cpx, Amp, Pl, Opx, Opq, Ap, Wht 500 2 5 0 - 0 1275 500 990 -7.14 -12.08 Ol, Cpx, Amp, Pl, Opq, Ap, Grt 500 1 5 0 - 0 1285 500 900 -7.04 -12.08 Ol, Cpx, Amp, Pl, Opx, Opq, Ap 500 0.5 5 0 - 0 1305 500 error at 1165 -6.85 -8.32 Ol, Cpx, Pl, Opq 500 2 5 4 0.8 0 1280 48 715 -7.09 -16.39 Ol, Cpx, Amp, Pl, Opx, Opq, Ap, Btt 500 1 5 4 0.8 0 1285 76 760 -6.99 -15.29 Ol, Cpx, Amp, Pl, Opx, Opq, Ap 500 0.5 5 4 0.8 0 1305 12 700 -6.8 -16.81 Ol, Cpx, Amp, Pl, Opq, Ap, sph Mineral abbreviations are according to Whitney and Evans (2010) Different initial water contents (0.5, 1 and 2 % H2O) were considered according to the presence of hydrous amphibole (Eggler, 1972; Ustunisik and Kilinc, 2011). Based on redox condition estimates by Mallmann and O'Neill (2009) for OIBs, oxygen fugacity was fixed relative to the QFM buffer. These authors show that OIBs record a very restrict range of relative oxygen fugacity between QFM and QFM+1. The parameters used in these MELTS runs are summarized in Table 6.1. The major element compositions of Graciosa lavas are best modeled by a MELTS polybaric fractionation process (500 - 10 MPa) starting at a pressure of 500 MPa (ca. 15 km depth) with 0.5 % of H2O content, for which conditions the liquidus temperature is 1305 ºC. The residual melt compositions obtained from this polybaric fractional crystallization model thoroughly reproduce the compositions of Graciosa lava flows (Fig. 6.15), with the following fractionated phases: olivine, clinopyroxene, plagioclase, Fe-Ti 93 Part II: Study of Graciosa Island Fig. 6.15 - Major element contents plotted vs. MgO contents for Graciosa lavas. The best-fit MELTS-1 polybaric fractional crystallization model (dotted lines) was obtained at an initial pressure of 500 MPa with decreasing steps of 5 MPa and cooling steps of 5 ºC with a water content of 1 % and considering GRZF12 sample as the initial composition. oxides, apatite and an additional Ti-rich ferromagnesian phase. The latter, with low Si and high Al and Ti contents, is interpreted as amphibole (kaersutite), because the MELTS thermodynamic database does not include this Ti-rich amphibole (M.S. Ghiorso, pers. com.). The fractionated mineral assemblage and the composition of these minerals in the range of temperature from ~ 1235 to 800 ºC (MELTS-1; Table 6.2) is broadly consistent with that of the alkaline xenoliths and lavas, suggesting that these alkaline 94 Chapter 6: Magmatic evolution of Graciosa Island xenoliths are the fractionated solids (cumulates) formed during the fractional crystallization process. In accordance, their bulk composition deviates from the composition of the lava flows, due to the accumulation of olivine, clinopyroxene, plagioclase, kaersutite, opaque minerals and apatite. For instance, the positive Ti anomaly can be related to the accumulation of kaersutite, Ti-rich clinopyroxene and opaque minerals, complementary to the lava flows, some of which are depleted in Ti (negative anomaly). Similarly, the variable P contents can be linked to the presence or absence of apatite (Fig. 6.12). Moreover, the composition of the fractionating mineral phases occurring below 900 ºC in the MELTS-1 model (feldspars, clinopyroxene, amphibole, sphene and Ti-magnetite) resembles the mineral compositions of the syenite sample (Table 6.2). Therefore, it is important to determine whether these syenites are cumulates like the alkaline xenoliths or if they represent frozen liquid compositions (residual magma completely solidified within the magma chamber). Their bulk composition is quite similar to the trachytic samples (Fig. 6.12), except for a glaring Zr-Hf depletion in the syenites. In addition, the Eu/Eu* anomaly (EuN/(SmN+GdN)/2) in the syenites is negative (Eu/Eu* < 1) and similar to that of the trachytes (Appendix - Table II), implying that there is minimal accumulation of alkali feldspar. According to these data, the syenites represent a frozen bulk trachyte liquid, similar to the syenite nodules studied by Widom et al. (1993) in Agua de Pau Volcano on São Miguel Island. The low Zr, Hf, Y and Lu, yet similar La concentrations of the Graciosa syenites relative to trachytes indicates that there was likely incomplete digestion of syenite zircons during sample dissolution by acid digestion. The fractional crystallization process was further appraised using trace element data, considering those trace elements for which partition coefficients are available for all fractionated phases. The trace element model was carried out by applying the fractional crystallization equation of Rayleigh (1896). Due to changes in the proportions of fractionating mineral assemblage during fractional crystallization, at least two different stages were needed to reproduce the whole range of Graciosa sample compositions. The partition coefficients used in both models were taken from previous studies in alkaline rocks with similar compositions and are listed Table 6.8 of the electronic 95 Part II: Study of Graciosa Island Table 6.2 - Composition of the obtained fractionated phases in MELTS-1 and MELTS-2 models, and those of the studied xenoliths and lavas for comparison . Mineral Phase MELTS-1 Alkaline gabbros Syenites Lavas MELTS-2 Subalkaline gabbros Olivine Fo# 86 - 53 78 - 54 - 89 - 55 89 - 71 84 - 74 SiO2 40.3 - 35.3 38.3 - 34.7 - 40.8 - 34.0 40.8 - 37.9 40.5 - 37.9 Clinopyroxene Mg# 75 - 9* 75 - 59 66 - 64 78 - 25 79 - 58 85 - 74 TiO2 3.9 - 0.1 2.6 - 0.4 1.4 - 1.3 3.6 - 0.1 0.5 - 1.5 0.1 - 1 Kaersutite Mg# 75 - 9* 58 - 43 30 - 9 57 - 14 - - Feldspars An 70 - 4 76 - 31 16 - 0 82 - 0 85 - 45 98 - 76 Al2O3 31.6 - 19.5 33.4 - 24.9 22.9 - 18.4 34.2 - 18.1 34.2 - 27.2 36.24 - 30.6 *Mg# values for the ferromagnesian phases (including clinopyroxene and kaersutite) estimated by MELTS supplement. In the trace element model-1 (Fig. 6.16), the GRZF4 lava sample was used as starting point because it has the least enriched trace element composition and therefore may represent the least fractionated magma. The starting lava used for the trace element model is different from the one used for the major element MELTS model (GRZF12) because the latter is more enriched in trace elements, despite being slightly less evolved in major and minor element composition. Fractionated mineral phases were chosen according to the most primitive alkaline xenolith mineral modes and in accordance to the MELTS-1 model estimates (27 % olivine, 21 % clinopyroxene, 33 % amphibole, 12 % plagioclase and 7 % ilmenite; see discussion above). If the results of the trace element model-1 are compared to the composition of the Graciosa lavas, a good agreement is observed for samples from Serra das Fontes (50-70 % fractional crystallization) and Vitória - Vulcão Central volcanic complex (10-70 % fractional crystallization). However, no agreement is observed between the model and the most evolved samples (trachytes from Serra Branca and Vulcão Central Unit; Fig. 6.16). Accordingly, a second model (rare earth element (REE) model-2) with a mineral fractionating assemblage similar to the mineral modes of the most evolved alkaline xenoliths (3.5 % clinopyroxene, 45 % amphibole, 35 % plagioclase, 1.5 % apatite and 15 % ilmenite; Fig. 6.17) was carried out to derive the trachyte sample compositions. In this 96 Chapter 6: Magmatic evolution of Graciosa Island Fig. 6.16 - Primitive mantle (McDonough and Sun, 1995) normalized trace element patterns for the Rayleigh fractional crystallization model-1 starting from sample GRZF4 with 10 % crystallization steps (grey dashed lines, from 100 % melt to 10 % melt) compared to the lava samples from Graciosa Island. The fractionating assemblage is composed of 27 % olivine, 21 % clinopyroxene, 33 % amphibole, 12 % plagioclase and 7 % ilmenite. Partition coefficients used to calculate the model are listed in Table 6.8 of the electronic supplement. 97 Part II: Study of Graciosa Island Fig. 6.17 - Primitive mantle (McDonough and Sun, 1995) normalized REE patterns for the fractional crystallization model-2 starting from the composition obtained for 90 % of residual melt in model-1. Serra Branca and Vulcão Central Unit trachytes are shown and compared to the 10 % crystallization intervals (grey dashed lines). For this model the fractionating mineral phases are 3.5 % clinopyroxene, 45 % amphibole, 35 % plagioclase, 1.5 % apatite and 15 % ilmenite. The REE-s included are those with available partition coefficients for all fractionated mineral phases (Table 6.8 of the electronic supplement). case, the starting point was the residual melt resulting at 10 % of fractional crystallization of sample GRZF4 in the trace element model-1. Crystallization rates from 40 % to 80 % are needed to reproduce the composition of the trachytic samples. As a conclusion, polybaric fractional crystallization is the dominant process controlling the differentiation of Graciosa magmas, producing lava flows that range from basalt to trachytes (syenites as frozen liquids), and magma chamber cumulates represented by alkaline gabbros. Primary magma compositions were likely very similar to samples GRZF12 and GRZF4. Origin of the subalkaline xenoliths The existence of subalkaline gabbros has been identified for the first time within the Azores archipelago. They are composed of olivine, clinopyroxene, plagioclase and scarce Cr-rich spinel. These mineral phases are characterized by much more primitive compositions than those of the alkaline gabbros (Fig. 6.7 and Fig. 6.9). However, the composition of the olivine (Fo84-74) rules out their mantle origin (e.g., Stormer, 1973; Jurewicz and Watson, 1988). The bulk rock composition of the subalkaline group is more primitive in major elements and less enriched in all incompatible elements relative to the alkaline gabbros (Figs. 6.11 - 6.12). Furthermore, the bulk composition of the 98 Appendix APPENDIX Table I Analytical procedures in the two different laboratories used for Graciosa Island whole rock analyses. (1) IBERCRON laboratory, Universidad del País Vasco, Spain (www.ikerkuntza.ehu.es) http://tiny.cc/s8r05w Samples were prepared by alkaline fusion with LiBO2 in Pt-Au crucibles, followed by acid dissolution of the melt. The fusion process is as follows: 250 mg of sample and 500 mg of flux are put into the crucible with three to four drops of LiBr solution. The mixture is fused. The melted glass (in fact, dissolution of the mixture in the flux) is poured automatically onto a weighed polypropylene beaker containing 100 ml HNO3 1N, with a few drops of HF to ensure stability of the HFSE. The acid solution is stirred ca. 10 min to ensure total dissolution. This primary solution is diluted gravimetrically to ca. 1:6500 in a mixture of HNO3 0.32N and very diluted HF, and spiked with In (50 μg/m) and Bi (10 μg/m) standard solutions (see Mendinabeitia et al., 2008). The samples were run for major oxides and selected trace elements using a mass spectrometer with inductivelycoupled plasma source (ICP-MS) X7 by Thermo Scientific updated to XSeries 2. 105 Part II: Study of Graciosa Island (2) LabGEOTOP laboratory, Institut de Ciències de la Terra Jaume Almera, Spain (www.ija.csic.es) http://tiny.cc/eds05w Samples were prepared by acid digestion in closed environment; 100 mg of sample and 10 ml acid mixture of 5 ml of HF, 2.5 ml of HNO3 and 2.5 ml of HClO4 was added to a PFA EvapoCean® vessels. The vessels were then closed and heated at 135 ºC during 24 hours. After total dissolution of all samples, the vessels were cooled to room temperature. Each vessel was connected to a PFTE EvapoCean® elbow plus vessel and heated in the EvapoCean® heating block at 120 ºC overnight to evaporate and collect acids. The residue was dissolved in 1 ml of HNO3 and brought to a final volume of 100 ml of MiliQ H2O. Clear solutions were obtained in all cases. A procedural blank solution was also prepared. The samples were run for major oxides (excluding SiO2 due to the digestion type) and selected trace elements using a high resolution mass spectrometer with inductively-coupled plasma source (HR-ICP-MS) Element XR byThermo Scientific. 106 Appendix Detection Limits of the analyzed elements: Element (1) IBERCRON (2) LabGEOTOP SiO2 0.03 % - TiO2 0.0003 % 0.001 Al2O3 0.083 % 0.006 Fe2O3 T 0.0033 % 0.006 MnO 0.0001 % 0.0001 MgO 0.0007 % 0.002 CaO 0.0107 % 0.004 Na2O 0.0011 % 0.007 K2O 0.0006 % 0.003 P2O5 0.0064 % 0.0003 Rb 0.716 ppm 0.444 ppm Sr 0.088 ppm 1.282 ppm Ba 0.08 ppm 7.634 ppm Sc 0.243 ppm 0.04 ppm V 1.072 ppm 0.412 ppm Cr 20.052 ppm 0.321 ppm Co 1.033 ppm 0.181 ppm Ni 20.099 ppm 2.291 ppm Cu 10.092 ppm 8.416 ppm Zn 0.932 ppm 1.849 ppm Y 0.04 ppm 0.061 ppm Nb 1.056 ppm 0.089 ppm Zr 1.084 ppm 0.246 ppm Hf 0.103 ppm 0.005 ppm Pb - 0.683 ppm U 0.006 ppm 0.012 ppm Th 0.011 ppm 0.005 ppm La 0.021 ppm 0.037 ppm Ce 0.022 ppm 0.067 ppm Pr 0.01 ppm 0.008 ppm Nd 0.065 ppm 0.034 ppm Sm 0.013 ppm 0.01 ppm Eu 0.0001 ppm 0.007 ppm Gd 0.012 ppm 0.013 ppm Tb 0.0001 ppm 0.002 ppm Dy 0.012 ppm 0.009 ppm Ho 0.0001 ppm 0.002 ppm Er 0.009 ppm 0.005 ppm Tm 0.022 ppm 0.001 ppm Yb 0.004 ppm 0.004 ppm Lu 0.0001 ppm 0.002 ppm 107 Part II: Study of Graciosa Island Table II Major (wt. %) and trace (ppm) element composition of the studied lavas and xenoliths from Graciosa Island. Unit/Site Serra das Fontes Serra Branca Vitória Unit Sample GRZF8 GRZF28 GRZF29 GRZF13 GRZF20 GRZF2 GRZF3 GRZF4 Lava Lava Lava Lava Lava Lava Lava Lava Texture Microporph. Microporph. Microporph. Trachytic Trachytic Microporph. Microporph. Porphyritic Campaigne 1 2 2 2 2 1 1 1 wt. % SiO2* 48.09 52.47* 50.84* 65.44 66.89* 48.80 48.84 47.28 TiO2 3.99 3.78 3.58 0.47 0.57 2.58 2.61 2.42 Al2O3 15.57 14.54 15.61 16.57 15.19 16.05 15.81 15.32 Fe2O3T 12.70 11.03 12.42 3.94 4.34 10.74 10.55 10.20 MnO 0.19 0.20 0.21 0.17 0.26 0.16 0.16 0.15 MgO 4.39 3.47 3.07 0.38 0.33 7.38 6.66 9.59 CaO 9.50 8.24 6.83 0.80 0.58 10.00 9.55 11.10 Na2O 3.58 3.91 3.80 7.45 6.34 3.25 3.53 2.70 K2O 1.44 1.41 1.30 4.56 4.93 1.21 1.33 0.83 P2O5 0.64 0.74 0.65 0.06 0.11 0.48 0.66 0.39 TOTAL 100.09 99.84 98.37 99.85 99.60 100.65 99.70 99.98 LOI -0.28 0.16 1.62 0.28 0.40 -0.25 -0.55 -0.26 Mg# 27 25 21 9 7 43 41 51 (ppm) Rb 28 35 33 78 119 30 33 16 Sr 574 647 526 37 12 492 491 492 Ba 393 404 494 832 86 372 401 258 Sc 22 18 2 3 4 24 22 28 V 369 265 146 5 1 262 261 283 Cr 20 11 <DL 20 < DL 210 170 440 Co 30 28 18 1 < DL 33 31 39 Ni 20 < DL < DL 20 0 80 60 140 Cu 40 39 34 10 < DL 30 30 40 Zn 110 102 187 110 127 80 90 70 Y 35.2 41.9 41.2 37.1 32.4 24.3 25.8 19.5 Nb 65 78 77 143 265 50 55 30 Zr 304 365 361 753 1315 212 236 141 Hf 7.0 8.7 7.5 16.0 29.8 4.8 5.4 3.3 Pb - 2.4 2.9 8.0 6.4 - - - U 1.38 1.70 1.29 1.59 5.26 1.13 1.26 0.61 Th 4.35 6.20 5.65 11.40 19.91 3.69 4.12 1.99 La 47.2 50.0 47.4 76.4 122.4 33.2 37.4 21.4 Ce 94.8 110.7 97.8 162.0 292.2 67.9 76.0 45.5 Pr 11.6 12.7 13.6 17.5 28.2 7.9 8.8 5.5 Nd 48.1 53.9 55.3 64.9 95.3 33.0 36.5 24.2 Sm 10.1 11.1 11.2 11.8 16.3 7.1 7.7 5.5 Eu 3.14 3.54 3.39 2.97 2.00 2.31 2.45 1.87 Gd 9.23 11.33 11.57 9.24 15.40 6.55 6.93 5.14 Tb 1.36 1.42 1.52 1.41 1.71 0.98 1.05 0.78 Dy 7.38 8.30 8.92 7.78 9.56 5.35 5.72 4.20 Ho 1.36 1.26 1.34 1.45 1.22 0.99 1.06 0.78 Er 3.67 3.72 3.99 4.04 3.79 2.65 2.83 2.10 Tm 0.49 0.61 0.61 0.58 0.67 0.36 0.39 0.29 Yb 3.00 3.76 3.51 3.91 4.56 2.23 2.40 1.79 Lu 0.43 0.56 0.52 0.66 0.67 0.34 0.36 0.27 Eu/Eu* 0.97 0.95 0.90 0.84 0.38 1.01 1.00 1.06 87Sr/86Sr 0.70358 - - - 0.70390 - 0.70336 - 143Nd/144Nd 0.512922 - - - 0.512883 - 0.512954 - 206Pb/204Pb 19.695 - - - 19.419 - 19.479 - 207Pb/204Pb 15.603 - - - 15.623 - 15.588 - 208Pb/204Pb 39.105 - - - 39.165 - 39.038 - * Due to the analytical procedure of the samples analyzed in the LabGEOTOP laboratory (acid digestion; see Appendix - Table I), the SiO2* content of these samples has been estimated as SiO2 = 100 - Σ (total oxide content + LOI). Fe2O3T: total Fe expressed as Fe2O3; LOI: Loss On Ignition; < DL: below the detection limit. Mg# = MgO/(FeO+MgO)·100 where FeO = 0.9·Fe2O3T; Eu/Eu* = (EuN/((EuN+GdN)/2). Microporph.: Microporphytitic; Orthocum.: Orthocumulate; Heteroadcum.: Heteroadcumulate; Equigran.: Equigranular; Inequigran.: Inequigranular; Alk. Gabbro: Alkaline gabbro; Subalk. Gabbro: Subalkaline gabbro. 108 Appendix Unit/Site Vitória Unit Sample GRZF9 GRZF10 GRZF11 GRZF12 GRZF17 GRZF21 GRZF22 GRZF23 Lava Lava Lava Lava Lava Lava Lava Lava Texture Microporph. Microporph. Microporph. Porphyritic Porphyritic Microporph. Porphyritic Porphyritic Campaigne 1 1 1 1 1 2 2 2 wt. % SiO2* 47.50 47.25 48.16 47.81 46.92 48.98 50.12* 50.67* TiO2 2.97 2.94 3.06 2.45 2.55 2.84 2.81 2.92 Al2O3 15.48 15.29 15.62 14.04 15.19 15.80 14.39 14.31 Fe2O3T 10.67 10.89 11.23 10.21 10.60 10.82 10.05 10.02 MnO 0.15 0.15 0.16 0.15 0.15 0.15 0.16 0.14 MgO 8.11 9.21 6.66 10.58 9.94 7.98 8.00 8.66 CaO 10.58 10.78 9.61 10.91 10.20 9.53 9.66 9.16 Na2O 3.05 2.98 3.25 2.51 2.82 2.82 2.66 2.53 K2O 0.97 0.89 1.16 1.00 0.95 0.81 0.89 0.68 P2O5 0.44 0.41 0.59 0.42 0.52 0.37 0.58 0.36 TOTAL 99.92 100.79 99.51 100.09 99.84 100.15 99.38 99.50 LOI -0.01 -0.43 -0.07 -0.08 -0.06 -0.15 0.62 0.49 Mg# 45 48 39 53 51 45 46 48 (ppm) Rb 20 18 24 19 17 29 24 20 Sr 532 515 574 453 518 473 607 465 Ba 273 247 323 281 276 339 299 248 Sc 27 28 22 30 27 1 26 1 V 267 270 257 273 247 116 188 135 Cr 250 290 140 550 400 315 196 224 Co 38 41 35 42 42 27 43 30 Ni 100 130 60 200 140 139 126 161 Cu 40 40 30 60 30 40 46 43 Zn 80 80 110 80 70 118 80 128 Y 23.8 22.3 29.0 21.6 24.2 27.4 27.6 23.9 Nb 41 36 55 42 41 47 49 42 Zr 223 197 275 215 196 192 226 186 Hf 5.1 4.6 6.1 4.8 4.5 4.2 5.7 4.0 Pb - - - - - 1.6 1.4 1.3 U 0.96 0.84 1.44 0.91 0.91 0.87 0.94 0.70 Th 2.96 2.56 3.82 2.92 3.13 3.38 3.52 2.57 La 29.1 26.2 39.2 27.6 27.9 27.9 31.2 23.3 Ce 62.5 57.1 83.6 59.2 59.6 58.7 70.5 50.3 Pr 7.6 7.0 10.1 7.1 7.2 7.4 8.6 6.7 Nd 33.0 30.3 42.9 30.3 30.9 29.8 37.0 27.1 Sm 7.2 6.8 9.0 6.4 6.9 6.7 7.8 5.7 Eu 2.35 2.20 2.85 2.06 2.22 2.04 2.66 1.79 Gd 6.50 6.21 8.30 5.79 6.83 6.84 7.74 5.69 Tb 0.96 0.92 1.20 0.86 1.03 0.90 0.97 0.79 Dy 5.19 4.91 6.39 4.64 5.47 5.39 5.62 4.62 Ho 0.94 0.89 1.16 0.84 0.98 0.82 0.83 0.69 Er 2.46 2.36 3.05 2.26 2.60 2.43 2.38 2.05 Tm 0.33 0.32 0.40 0.31 0.35 0.38 0.37 0.31 Yb 2.05 1.98 2.49 1.95 2.16 2.18 2.38 1.79 Lu 0.31 0.30 0.38 0.30 0.33 0.33 0.35 0.28 Eu/Eu* 1.03 1.01 0.99 1.01 0.97 0.91 1.03 0.95 87Sr/86Sr - - - - - - 0.70338 - 143Nd/144Nd - - - - - - 0.512945 - 206Pb/204Pb - - - - - - 19.633 - 207Pb/204Pb - - - - - - 15.592 - 208Pb/204Pb - - - - - - 39.131 - 109 Part II: Study of Graciosa Island Unit/Site Vitória Unit Sample GRZF24 GRZF25 GRZF26 GRZF27 GRZF30 GRZF31 GRZF32A GRZF32B Lava Lava Lava Lava Lava Lava Lava Lava Texture Porphyritic Microporph. Porphyritic Porphyritic Porphyritic Microporph. Porphyritic Porphyritic Campaigne 2 2 2 2 2 2 2 2 wt. % SiO2* 47.54* 55.94* 50.09* 47.66* 48.81* 49.02* 46.18* 48.6* TiO2 2.60 2.44 2.76 2.97 3.04 3.10 3.28 3.22 Al2O3 16.13 13.31 15.52 14.42 13.59 14.17 14.56 13.55 Fe2O3T 10.69 9.39 10.06 10.77 10.46 10.31 11.16 10.66 MnO 0.16 0.14 0.14 0.16 0.14 0.15 0.15 0.15 MgO 8.29 6.84 7.57 9.22 9.50 10.04 10.14 10.14 CaO 9.36 7.46 9.47 10.63 10.49 10.25 10.58 10.17 Na2O 3.15 3.01 3.01 2.22 2.14 2.23 1.95 2.02 K2O 0.87 0.93 0.80 0.53 0.70 0.58 0.43 0.51 P2O5 0.51 0.37 0.43 0.46 0.41 0.32 0.34 0.32 TOTAL 99.35 99.88 99.91 99.08 99.34 100.21 98.80 99.39 LOI 0.64 0.11 0.08 0.92 0.66 -0.21 1.19 0.60 Mg# 46 44 45 48 50 51 50 51 (ppm) Rb 25 32 22 8 16 12 7 12 Sr 565 492 558 616 629 470 514 497 Ba 349 399 284 270 252 245 270 258 Sc 2 1 1 31 31 1 1 1 V 120 174 201 317 402 142 117 160 Cr 258 275 245 322 166 375 366 471 Co 28 26 27 51 50 33 35 33 Ni 118 132 113 171 191 209 206 202 Cu 41 37 41 46 56 44 53 59 Zn 136 131 133 77 78 139 143 154 Y 29.5 27.8 26.5 25.2 24.8 23.6 26.4 25.7 Nb 46 52 45 44 44 39 41 42 Zr 232 261 200 209 220 169 198 188 Hf 4.9 5.3 4.3 5.5 5.6 4.1 4.3 4.2 Pb 1.9 2.1 0.9 1.2 1.5 1.3 1.2 1.8 U 0.92 1.09 0.83 0.64 0.78 0.51 0.50 1.63 Th 3.36 3.96 2.91 3.00 3.04 2.17 2.41 2.37 La 32.2 33.2 26.6 26.3 25.9 23.3 25.9 24.2 Ce 70.7 70.3 56.4 58.9 60.3 49.3 56.4 54.6 Pr 8.9 9.0 7.2 7.6 7.6 6.7 7.6 7.1 Nd 37.2 35.3 28.8 32.2 32.0 27.8 31.2 29.8 Sm 7.8 7.4 5.8 7.1 7.0 6.1 7.0 6.7 Eu 2.44 2.31 1.80 2.44 2.33 1.91 2.11 2.06 Gd 7.94 7.52 5.59 7.07 6.86 6.29 7.06 6.82 Tb 1.02 0.98 0.74 0.88 0.87 0.85 0.93 0.91 Dy 6.00 5.68 4.41 5.26 5.06 5.01 5.44 5.28 Ho 0.87 0.84 0.69 0.79 0.75 0.75 0.79 0.76 Er 2.57 2.54 2.06 2.20 2.16 2.19 2.37 2.28 Tm 0.39 0.38 0.32 0.35 0.35 0.33 0.35 0.33 Yb 2.28 2.25 1.98 2.17 2.15 1.93 2.04 2.00 Lu 0.34 0.35 0.31 0.31 0.32 0.29 0.30 0.29 Eu/Eu* 0.94 0.93 0.95 1.04 1.02 0.93 0.90 0.92 87Sr/86Sr - - - 0.70340 - - - - 143Nd/144Nd - - - 0.512944 - - - - 206Pb/204Pb - - - 19.601 - - - - 207Pb/204Pb - - - 15.596 - - - - 208Pb/204Pb - - - 39.124 - - - - 110 Appendix Unit/Site Vitória Unit Vulcão Central Unit Sample GRZF33 GRZF34 GRZF35 GRZF1 GRZF5 GRZF6 GRZF7 GRZF14 Lava Lava Lava Lava Lava Lava Lava Lava Texture Porphyritic Microporph. Porphyritic Porphyritic Porphyritic Microporph. Porphyritic Porphyritic Campaigne 2 2 2 1 1 1 1 1 wt. % SiO2* 47.76* 50.55* 49.25* 51.06 47.34 59.81 48.38 49.89 TiO2 3.05 3.27 3.10 2.77 3.18 1.01 2.42 2.38 Al2O3 14.39 15.11 13.73 17.12 16.60 17.47 16.61 16.65 Fe2O3T 10.95 11.35 10.32 10.61 10.33 6.08 9.72 9.96 MnO 0.16 0.16 0.15 0.17 0.14 0.15 0.14 0.16 MgO 9.69 6.56 9.22 4.17 7.01 1.15 6.87 6.08 CaO 10.70 9.16 10.45 8.14 10.70 3.22 10.45 9.06 Na2O 2.19 2.94 2.15 4.22 3.03 6.22 3.18 3.64 K2O 0.60 0.84 0.69 1.61 1.03 3.41 1.03 1.25 P2O5 0.34 0.48 0.41 0.55 0.46 0.31 0.40 0.56 TOTAL 99.86 100.48 99.52 100.42 99.83 98.84 99.21 99.63 LOI 0.13 -0.48 0.48 -0.06 0.77 0.42 -0.15 -0.27 Mg# 49 39 49 30 42 17 43 40 (ppm) Rb 14 26 14 31 19 81 19 27 Sr 520 577 568 542 600 307 527 547 Ba 259 336 271 452 297 794 310 380 Sc 1 2 33 19 24 5 24 20 V 161 139 250 253 287 30 265 230 Cr 334 130 370 < DL 160 < DL 190 120 Co 33 27 48 23 35 4 34 28 Ni 193 76 180 < DL 70 < DL 70 60 Cu 53 35 57 20 30 < DL 30 20 Zn 142 162 70 100 80 90 80 80 Y 26.7 32.0 25.9 28.9 22.9 31.9 20.6 24.8 Nb 39 59 45 52 43 82 36 41 Zr 189 265 220 272 210 452 169 194 Hf 4.1 5.8 5.7 6.3 4.8 9.8 4.0 4.5 Pb 1.5 1.9 1.4 - - - - - U 0.61 1.19 0.58 1.08 0.84 2.50 0.76 1.04 Th 2.41 3.71 3.03 3.82 2.85 9.13 2.47 3.23 La 25.3 38.2 27.7 37.8 30.3 57.1 25.4 31.5 Ce 54.6 81.2 62.0 78.1 64.3 105.0 53.5 64.6 Pr 7.3 10.4 7.8 9.1 7.7 11.6 6.4 7.7 Nd 29.5 42.8 33.3 38.8 33.3 43.9 27.5 33.1 Sm 6.8 9.1 7.2 8.3 7.1 8.3 6.0 7.1 Eu 2.04 2.73 2.42 2.70 2.32 2.50 2.06 2.45 Gd 6.77 9.22 7.33 7.61 6.47 7.19 5.58 6.67 Tb 0.89 1.19 0.92 1.15 0.96 1.12 0.84 1.01 Dy 5.13 6.89 5.30 6.37 5.12 6.45 4.53 5.35 Ho 0.78 1.01 0.79 1.18 0.94 1.23 0.83 0.97 Er 2.28 3.02 2.26 3.14 2.49 3.49 2.22 2.58 Tm 0.34 0.45 0.35 0.43 0.33 0.51 0.30 0.35 Yb 1.99 2.57 2.16 2.76 2.00 3.48 1.83 2.22 Lu 0.30 0.38 0.32 0.43 0.30 0.56 0.27 0.35 Eu/Eu* 0.91 0.90 1.00 1.01 1.02 0.96 1.07 1.07 87Sr/86Sr - - - 0.70343 - - - 0.70341 143Nd/144Nd - - - 0.512926 - - - 0.512957 206Pb/204Pb - - - 20.096 - - - 20.090 207Pb/204Pb - - - 15.659 - - - 15.626 208Pb/204Pb - - - 39.520 - - - 39.426 111 Part II: Study of Graciosa Island Unit / Site Vulcão Central Unit Baía da Folga Sample GRZF15 GRZF16 GRZF18 GRZF19 GRZF36 GRZF 1x1 GRZF 1x2 GRZF 1x5 Lava Lava Lava Lava Lava Alk. gabbro Alk. gabbro Alk. gabbro Texture Porphyritic Microporph. Porphyritic Porphyritic Microporph. Heteroadcum. Heteroadcum. Orthocum. Campaigne 1 2 2 2 2 1 1 1 wt. % SiO2* 48.93 57.12 49.68* 48.33* 55.74* 44.85 41.80 44.93 TiO2 2.37 1.58 2.91 3.56 2.45 2.35 3.84 3.75 Al2O3 16.11 16.47 17.65 16.21 14.91 17.14 13.14 21.01 Fe2O3T 10.31 6.96 10.05 10.12 9.35 9.73 14.54 9.32 MnO 0.16 0.14 0.16 0.15 0.19 0.19 0.29 0.08 MgO 6.97 2.97 3.89 7.04 2.81 7.25 10.48 3.38 CaO 9.43 5.36 9.15 10.46 6.60 10.83 10.82 13.29 Na2O 3.45 4.59 3.88 2.98 4.78 3.55 2.51 2.96 K2O 1.10 2.59 1.27 0.97 1.89 0.53 0.46 0.34 P2O5 0.52 0.37 0.74 0.47 1.05 0.58 0.56 0.62 TOTAL 99.36 98.15 99.43 100.34 99.82 97.04 98.47 99.71 LOI -0.31 0.33 0.57 -0.34 0.18 0.14 0.31 0.01 Mg# 42 32 30 43 25 45 44 28 (ppm) Rb 24 74 31 26 49 12 10 12 Sr 527 348 653 627 598 670 488 833 Ba 342 493 400 296 536 358 281 186 Sc 22 10 19 26 15 18 15 15 V 241 119 144 239 112 200 246 304 Cr 140 50 17 147 < DL 125 140 < DL Co 33 16 27 42 16 30 35 20 Ni 80 20 23 105 < DL 91 101 < DL Cu 20 10 42 45 23 35 36 30 Zn 90 90 83 86 106 118 157 117 Y 23.9 26.8 32.7 25.5 44.7 40.4 44.2 15.4 Nb 37 74 55 49 78 48 51 15 Zr 176 390 256 239 357 112 122 71 Hf 4.2 8.7 6.3 5.9 8.7 3.3 3.6 1.8 Pb - 6.0 2.2 1.4 2.7 - - - U 0.87 3.35 1.22 0.98 1.80 0.26 0.20 0.29 Th 2.76 10.40 4.40 3.17 6.30 0.95 0.80 1.13 La 28.4 51.5 36.6 30.5 53.5 31.2 25.7 14.4 Ce 59.6 95.0 82.2 66.6 115.8 73.3 65.2 32.1 Pr 7.2 9.9 10.0 8.0 13.6 10.6 10.0 4.1 Nd 30.9 36.6 41.6 33.8 59.9 44.9 45.4 18.7 Sm 6.8 7.0 8.7 7.0 12.5 9.9 10.9 4.2 Eu 2.38 2.00 2.96 2.47 3.99 3.19 3.32 1.56 Gd 6.39 6.19 8.68 7.10 12.52 8.65 9.49 3.60 Tb 0.95 0.94 1.12 0.91 1.53 1.43 1.60 0.55 Dy 5.11 5.20 6.34 5.14 8.83 8.01 9.02 3.16 Ho 0.92 0.98 0.99 0.77 1.36 1.47 1.64 0.56 Er 2.48 2.80 2.84 2.24 3.90 3.83 4.23 1.37 Tm 0.34 0.40 0.46 0.37 0.64 0.52 0.55 0.18 Yb 2.11 2.77 2.80 2.24 3.93 3.04 3.25 1.04 Lu 0.32 0.45 0.42 0.33 0.59 0.42 0.45 0.14 Eu/Eu* 1.08 0.91 1.02 1.05 0.96 1.03 0.97 1.19 87Sr/86Sr - 0.70337 - 0.70341 - 0.70340 - - 143Nd/144Nd - 0.512964 - 0.512937 - 0.512957 - - 206Pb/204Pb - 19.863 - 19.863 - 20.008 - - 207Pb/204Pb - 15.608 - 15.611 - 15.610 - - 208Pb/204Pb - 39.234 - 39.270 - 39.341 - - 112 Appendix Unit / Site Baía da Folga Enxudreiro Sample GRZF 1x6 GRZF 1x8 GRZF 1x10 GRZF 1x13 GRZF1X19 GRZF1X22 GRZF1X24 GRENX-1 Alk. gabbro Alk. gabbro Alk. gabbro Alk. gabbro Alk. gabbro Alk. gabbro Alk. gabbro Syenite Texture Orthocum. Orthocum. Orthocum. Orthocum. Orthocum. Orthocum. Orthocum. Inequigran. Campaigne 1 1 1 1 2 2 2 2 wt. % SiO2* 47.32 42.55 46.36 42.39 40.6* 37.15* 44.77* 69.63* TiO2 1.07 5.14 2.39 4.69 4.12 5.92 3.09 0.36 Al2O3 16.69 15.16 17.78 16.73 17.76 14.29 22.15 14.12 Fe2O3T 9.43 12.59 9.66 13.33 19.04 17.18 9.98 2.82 MnO 0.11 0.17 0.12 0.13 0.15 0.16 0.09 0.10 MgO 6.74 6.64 3.87 7.70 5.22 11.70 4.35 0.31 CaO 15.00 10.97 13.05 9.36 10.11 10.71 12.24 1.55 Na2O 2.15 2.69 3.23 2.93 2.69 2.40 2.86 6.68 K2O 0.47 0.45 0.29 0.52 0.41 0.47 0.18 3.74 P2O5 0.03 4.57 1.97 0.27 0.12 0.10 0.08 0.09 TOTAL 99.03 100.95 98.73 98.08 100.26 100.14 99.84 99.47 LOI 0.37 1.39 0.36 0.54 -0.26 -0.14 0.15 0.53 Mg# 44 36 30 39 23 43 32 11 (ppm) Rb 9 12 14 11 8 5 1 85 Sr 762 704 894 611 615 475 912 184 Ba 152 282 249 239 178 229 129 651 Sc 16 17 17 17 < DL < DL < DL < DL V 321 355 239 390 487 430 186 3 Cr 66 < DL < DL < DL < DL < DL 72 < DL Co 28 33 18 38 35 39 19 2 Ni 30 < DL < DL < DL < DL < DL 23 < DL Cu 37 43 23 48 51 60 29 < DL Zn 112 170 122 165 196 238 123 58 Y 16.1 51.2 32.1 27.4 17.4 34.8 8.5 18.0 Nb 9 48 27 33 19 29 11 100 Zr 51 135 106 105 76 90 37 36 Hf 1.6 3.6 2.6 2.9 2.0 2.7 1.0 1.0 Pb - - - - 0.8 1.2 < DL 3.7 U 0.11 0.48 0.58 0.33 0.20 0.13 0.09 3.27 Th 0.43 1.75 1.79 1.07 0.82 0.50 0.12 15.94 La 6.1 40.4 32.4 13.8 9.9 9.9 5.0 54.9 Ce 17.3 92.8 71.3 34.7 21.7 27.4 11.4 108.5 Pr 2.2 14.0 10.2 5.1 3.1 4.6 1.6 9.5 Nd 12.2 61.3 44.4 24.1 14.6 25.1 7.9 30.3 Sm 3.4 13.4 9.3 6.2 3.8 7.7 2.2 4.7 Eu 1.39 3.79 3.03 2.04 1.38 2.46 1.13 1.27 Gd 3.12 11.87 8.10 5.55 3.65 7.32 2.11 5.38 Tb 0.53 1.90 1.24 0.96 0.55 1.16 0.31 0.58 Dy 3.26 10.25 6.50 5.51 3.22 6.79 1.85 3.55 Ho 0.58 1.86 1.17 1.01 0.52 1.09 0.29 0.53 Er 1.42 4.68 2.87 2.55 1.43 2.86 0.80 1.70 Tm 0.19 0.60 0.35 0.34 0.22 0.43 0.12 0.31 Yb 1.13 3.46 2.09 2.00 1.24 2.26 0.65 1.96 Lu 0.15 0.47 0.29 0.28 0.18 0.33 0.09 0.27 Eu/Eu* 1.27 0.90 1.04 1.04 1.11 0.99 1.58 0.76 87Sr/86Sr 0.70338 - - - - - - - 143Nd/144Nd 0.512966 - - - - - - - 206Pb/204Pb 20.032 - - - - - - - 207Pb/204Pb 15.621 - - - - - - - 208Pb/204Pb 39.393 - - - - - - - 113 Part II: Study of Graciosa Island Unit/Site Enxudreiro Ponta Pesqueira Quitadouro Sample GRENX-2 GRZF17A GRZF17B GRZF17D GRZF17C GRZF17E GRQUX1 GRQUX3 Syenite Subalk. gabbro Subalk. gabbro Subalk. gabbro Alk. gabbro Alk. gabbro Dunite Subalk. gabbro Texture Inequigran. Inequigran. Inequigran. Inequigran. Inequigran. Orthocum. Equigran. Inequigran. Campaigne 2 1 2 2 2 2 2 2 wt. % SiO2* 71.3* 49.1* 49.03* 50.61* 39.4* 45.3* 49.63* 45.01* TiO2 0.19 0.27 0.26 0.11 7.16 4.36 0.05 0.35 Al2O3 15.80 17.36 18.51 23.05 15.41 16.29 0.82 9.02 Fe2O3T 2.44 4.17 4.07 1.83 15.89 11.08 9.35 8.12 MnO 0.06 0.07 0.07 0.04 0.16 0.12 0.14 0.13 MgO 0.15 12.52 8.67 5.68 6.90 6.87 39.31 21.80 CaO 0.86 14.80 17.36 15.67 13.13 13.07 0.81 15.03 Na2O 5.63 0.98 1.45 1.60 1.53 1.90 0.03 0.48 K2O 3.37 0.04 0.07 0.08 0.12 0.15 0.02 0.03 P2O5 0.04 ‹LD 0.01 0.01 0.04 0.05 0.01 0.03 TOTAL 99.88 99.32 99.54 98.73 99.79 99.25 100.20 100.06 LOI 0.11 0.48 0.46 1.27 0.21 0.75 -0.20 -0.06 Mg# 6 76 70 77 32 40 82 74 (ppm) Rb 104 1 ‹LD ‹LD 2 2 ‹LD ‹LD Sr 70 230 297 309 728 761 2 110 Ba 279 25 12 24 65 91 3 13 Sc < DL 15 49 26 42 32 13 < DL V 1 123 83 89 915 211 63 121 Cr < DL 1029 227 170 < DL 327 4122 1048 Co < DL 30 31 22 59 46 114 51 Ni < DL 213 52 64 20 113 1524 347 Cu ‹LD 40 37 < DL 68 57 25 20 Zn 27 25 19 16 81 57 48 39 Y 28.1 6.4 4.8 2.2 13.1 11.5 0.6 5.4 Nb 136 < DL 2 1 18 18 < DL 1 Zr 24 19 8 8 86 77 3 14 Hf 0.8 0.6 0.3 0.2 2.8 2.4 < DL 0.4 Pb 2.7 - 0.7 0.7 0.8 ‹LD - ‹LD U 3.90 0.05 0.15 0.15 0.15 0.13 - 0.02 Th 21.83 0.35 0.49 0.32 0.39 0.43 - 0.08 La 89.2 1.7 1.4 1.4 6.2 8.1 0.6 1.4 Ce 154.8 5.8 3.3 3.0 15.2 15.6 1.0 4.0 Pr 14.9 0.3 0.4 0.3 2.2 2.4 0.1 0.5 Nd 46.3 2.4 2.1 1.3 11.0 11.6 0.6 2.7 Sm 7.1 0.8 0.7 0.3 3.1 2.9 0.1 0.8 Eu 0.58 0.34 0.38 0.28 1.26 1.32 0.05 0.27 Gd 8.76 0.60 0.79 0.40 3.09 2.98 0.14 0.84 Tb 0.94 0.14 0.13 0.06 0.44 0.39 0.02 0.14 Dy 5.53 1.14 0.85 0.40 2.54 2.30 0.12 0.91 Ho 0.78 0.22 0.15 0.07 0.41 0.35 0.02 0.16 Er 2.52 0.49 0.42 0.19 1.08 0.97 0.07 0.46 Tm 0.40 0.08 0.07 0.03 0.17 0.15 0.02 0.07 Yb 2.39 0.51 0.45 0.21 1.05 0.94 0.11 0.44 Lu 0.34 0.07 0.06 0.03 0.15 0.14 0.02 0.07 Eu/Eu* 0.23 1.42 1.54 2.31 1.23 1.35 1.11 0.98 87Sr/86Sr - - 0.70356 - 0.70338 - 0.70351 - 143Nd/144Nd - - 0.513011 - 0.512947 - 0.512946 - 206Pb/204Pb - - 19.513 - 18.710 - 19.706 - 207Pb/204Pb - - 15.595 - 15.597 - 15.598 - 208Pb/204Pb - - 38.981 - 38.424 - 39.142 - 114 Chapter 7: Corvo Island: A Review The island has an oval shape with a N-S elongated morphology (Fig. 7.2). The caldera centered on the north side of the island, locally known as O Caldeirão (Fig. 7.3), is the most prominent feature of Corvo landscape, comprising the 90 % of the island total surface area. This volcanic caldera has an elliptical shape, with a maximum diameter of 2.3 km and a depth of 300 m. Inside the caldera are several cinder and spatter cones (20 - 30 m in height) giving rise to small lakes, peat bogs and islets. Fig. 7.3 - Corvo Island “O Caldeirão” showing the caldera arc, some dikes and the intracaldera cones, lakes, peat bogs and islets. The island shows abrupt, high and steep cliffs (Fig. 7.4A-B), especially in the north, eastern and western coasts due to the strong wave erosion together with tectonic processes. These processes caused occasional landslides, which uncovered many basaltic dikes that cut the oldests volcanic units (Fig. 7.4A). The southern part of the island is characterized by a flat platform, known as fajã, where the single town within the island is located (Vila Nova do Corvo; Fig. 7.4C). This fajã was formed by at least two lava flows from a secondary cone located on the caldera southern flank (França et al., 2002). Volcanotectonic studies of Corvo (Dias, 2001) and Flores (Azevedo and Ferreira, 1999, 2006) Islands showed that the main structural direction (WNW-ESE) observed throughout the eastern islands of the Azores archipelago are not dominant in these western islands. Azevedo and Ferreira (2006) proposed that Flores and Corvo Islands are the emerged tops of a single larger edifice, built on a 10 Ma oceanic crust (Luis et al., 1994). Their N-S elongated shapes and their tectonic setting suggest an 121 Part III: Study of Corvo Island Fig. 7.4 - Photographs of Corvo Island: (a) western coast, (b) eastern coast and (c) the fajã and Vila Nova do Corvo town. evolution linked to the Mid-Atlantic Ridge (Azevedo and Ferreira, 1999, 2006; França et al., 2006a). Moreover, the westward displacement of Corvo from Flores at about 1 cm/year, together with the near-linear E-W northern and southern coastlines of Flores, reflect the structural control of the Mid-Atlantic Ridge and associated transform faults (Baptista et al., 1999). 7.2 History and previous studies A non-official exploration during the period of the late 13th century mentioned obscure islands within the Atlantic Ocean. As a result of this first exploration, maps such as the Genoves Atlas Medici (1351) were published, and the first reference to an Insula Corvi Marini (island of the Marine Crow) in a seven island archipelago appeared. Later, on the Atlas Catalán of 1375 the Islands of Corvo and Flores were also named as the Insula Corvi Marini. The first sighting of Corvo Island by a Portuguese navigator (Diogo de Teive) might have occurred in 1452. Given its small size, this island was not occupied till the 16th century when a group of slaves from Cape Verde were sent to Corvo to farm the land and breed cattle. A permanent settlement was not possible until 1580, when some settlers from Flores increased the number of inhabitants. 122 Chapter 7: Corvo Island: A Review The first geological work published on Corvo Island was the geological map (Zbyszewski et al., 1967) edited by the Portuguese Geological Survey. This map was the result of the first fieldwork campaign on the island led by Prof. Zbyszewski during the 60’s, when five volcanic units were proposed for the evolution of the island. At the beginning of the 21st century, earth scientists carried out a new field campaign on Corvo Island in order to publish the 1:10,000 geological map, together with a geomorphologic and neotectonic characterization of the island. These studies led to the establishment of a new volcanostratigraphical sequence (Dias, 2001; Azevedo et al., 2003). Concurrently, a Portuguese research team guided by Dr. Zilda França proposed a straightforward volcanostratigraphic succession based on the caldera-formation event (França et al., 2002), which in contrast to what was previously published (see section 7.3). Scarce studies deal with the magmatic processes related to the evolution of Corvo Island. A few contributions focus on geochemical and isotopic data (França et al., 2003, 2006a) and anisotropy of magnetic susceptibility magma flow directions (AMS; Pueyo et al., 2006). In addition, there are some papers that focus on other Azorean Islands but compare the data with those from Corvo (White et al., 1976; White et al., 1979; Lemarchand, 1987; Beier et al., 2010). Recently, Genske (2012) studied the western Azores mantle source including Corvo and Flores Islands and using Sr-Nd-Pb-Hf-Os isotope systematic. These authors incorporated their new results within a broader assessment of the mantle source across the plateau. The present PhD Thesis represents the first in-depth investigation of magmatic processes based on petrology, mineralogy and major and trace element data (Larrea et al., 2012; chapter 8). 7.3 Volcanostratigraphy The volcanostratigraphy of Corvo has been studied by previous authors (Dias, 2001; França et al., 2002; Azevedo et al., 2003). These authors proposed two different volcanostratigraphic successions based on different criteria. Dias (2001) and Azevedo et 123 Part III: Study of Corvo Island Fig. 7.5 - Geological map of Corvo Island modified from Dias (2001). Contour interval: 50 m. al. (2003) followed the same criteria used for the establishment of the volcanostratigraphy of Flores island (e.g., Azevedo, 1998; Azevedo et al., 2006), which is mainly based on petrologic criteria. They recognized two main volcanic complexes on the island (Fig. 7.5): a) the Basal Complex (CB) associated to submarine volcanism and the formation of the proto-island, which comprises lava flows and associated pyroclasts, breccias and palagonized tuffs, and b) The Upper Complex (CS) formed by subaerial 124 Chapter 7: Corvo Island: A Review Fig. 7.6 - Morpho-volcanic scheme of Corvo Island modified from França et al. (2002). Contour interval: 50 m. volcanism and divided into three volcanic units - lower (CS1), intermediate (CS2), and upper (CS3) units. The lower (CS1) and intermediate (CS2) units comprise basaltic and hawaiitic lava flows and associated pyroclasts, with major presence of pyroclasts in the CS2 unit. The upper unit (CS3) is entirely formed by pyroclastic deposits covering most of the total surface area of the island. Moreover, these authors dated three samples by K/Ar ages in order to obtain the ages of the different units. However, only two of the 125 Part III: Study of Corvo Island samples yielded reliable age results according to field observations (Dias, 2001); 0.71 ± 0.49 Ma for the upper part of the CS1 unit and 0.43 ± 0.34 Ma for the lower part of the CS2 unit. In contrast, França et al. (2002) proposed a straightforward volcanostratigraphic succession based on the caldera-formation event that includes three main units: (1) Precaldera, (2) Syn-caldera, and (3) Post-caldera units (Fig. 7.6). The Pre-caldera unit (1) is formed by submarine volcanism (proto-island) (1a) followed by two subaerial phases divided into the Lower (1b) and Upper (1c) subunits. The submarine volcanism (1a) is represented by hydromagmatic deposits of submarine pyroclasts exposed in the SW and N cliffs. The Lower unit (1b) comprises a sequence of pahoehoe and aa lava flows that crop out mainly in the oriental cliff of the island; this sequence includes a primitive shield volcano, strombolian deposits, the upper lava flows (Fig 7.7) and the associated dikes. The Upper unit (1c) comprises secondary cones (Coroa do Pico and Morro da Fonte), buried cones (Ribeira do Feno and Ribeira do Cerrado das Vacas) and their related lava flows. Fig. 7.7 - Pre-caldera Upper unit (1c): lava flows and associated pyroclastic deposits from Ribeira do Feno quarry. The Syn-caldera unit (2) includes plinian to sub-plinian pumiceous deposits (Fig. 7.8), lahars, surges and other pyroclastic flows associated with the stratovolcano collapse and the caldera formation. 126 Chapter 7: Corvo Island: A Review Fig. 7.8 - Syn-caldera unit (2): (a) Fine-grained pyroclastic deposits and (b) pumitic deposit associated to the stratovolcano collapse and the caldera formation event. The Post-caldera unit (3) comprises the most recent lava flow (Pão de Açúcar; Fig. 7.9), the intracaldera pyroclastic and spatter cones (e.g., Montinho do Queijo, Cachimbo and Malaguetas) and the Cortinhas scoria cone and associated lava flows. Fig. 7.9 - Pão de Açúcar: the most recent lava flow from the Post-caldera unit (3). In Fig. 7.10 an attempt to compare both volcanostratigraphic successions is presented. The volcanostratigraphy by Dias (2001) is mainly based on the type of volcanic products and is closer to a lithostratigraphical characterization. In contrast, the volcanostratigraphy by França et al. (2002) is based on the volcanic events and their relative chronostratigraphy. Some minor similarities are found. Submarine volcanism on the island is indicated by both authors as the CB unit (Dias, 2001) and the Pre-caldera 127 Part III: Study of Corvo Island Fig. 7.10 - Comparison between the volcanostratigraphic succesions of Corvo by Dias (2001) and França et al. (2002). Unit abbreviations as explained in the text. unit (1a) (França et al., 2002). Moreover, pyroclastic, spatter and scoria cones are recognized in the CS3 unit by Dias (2001) and the Post-caldera unit (3) by França et al. (2002). On the contrary, the (1b) and (1c) Pre-caldera units by França et al. (2002) are linked with difficulty to CS1 and CS2 units by Dias (2011). Additionally, the pumiceous deposits and pyroclastic flows associated to the caldera-forming event of the Syn-caldera (2) unit described by França et al. (2002) are omitted by Dias (2001). It is likely that these materials were not properly identified by Dias (2001) and wrongly included within the CS2 - pyroclastic subunit (Fig. 7.5 and Fig. 7.10). As a consequence, the volcanostratigraphy proposed by Dias (2001) is not fully accepted by the scientific community, as it presents uncertainties and is not supported by the K/Ar ages obtained by himself. Consequently, the samples collected on Corvo Island (see chapter 8) are categorized according to the volcanostratigraphy proposed by França et al. (2002). 128 Chapter 8: Magmatic Processes and the Role of Antecrysts in the Genesis of Corvo Lavas CHAPTER 8: MAGMATIC PROCESSES AND THE ROLE OF ANTECRYSTS IN THE GENESIS OF CORVO ISLAND LAVAS 8.1 Introduction The aim of this study is to investigate the magmatic processes controlling the evolution of Corvo Island volcanism. Its small size and the existence of a single main eruptive center make it an ideal island for an in-depth study. This chapter presents detailed petrographic and geochemical data for representative samples of the complete volcanostratigraphic sequence (Pre-, Synand Post-caldera stages) including lava flows, dikes and cumulate gabbroic xenoliths hosted in a Pre-caldera lava flow. Based on the relative volcanostratigraphic positions of the studied samples, a comprehensive model that establishes a genetic link between the petrologically diverse materials is proposed. 8.2 Samples and analytical methods All of the diverse igneous materials in Corvo (lava flows, dikes and xenoliths) were sampled for petrological and geochemical studies. Samples are categorized according to the volcanostratigraphy proposed by França et al. (2002) (see chapter 7). The three different units of the island were sampled in several locations (Fig. 8.1; UTM coordinates in chapter 3, Table 3.2). A total of 39 lava flow samples were collected, and are representative of the volcanic products of the whole island. Most samples were 129 Part III: Study of Corvo Island Fig. 8.1 - Morpho-volcanic scheme of Corvo Island and sampling sites (modified from França et al. (2006a)). Contour interval: 50 m. collected from the Pre-caldera (22 samples) and Post-caldera (11 samples) stages owing to their relatively mafic compositions, which potentially preserve petrogenetic information regarding mantle source and melting processes. Most samples belong to the Pre-caldera stage because it accounts for ca. 90 % of the volcanic products of the island. Pumice and evolved lavas were sampled from the Syn-caldera stage (6 samples) in order to evaluate the complete evolutionary trend of the magmatic plumbing system. 130