No description available.

The processes associated with magma injection at shallow depths within the crust have been the topic of many geophysical studies, some investigating the seismicity that accompanies volcanic activity and others attempting to map the subsurface extent and geometry of the resulting magma bodies. The aim of this study is to obtain a better understanding of these processes by investigating the nature of seismic signals that accompany volcanic eruptions and by seismically imaging a magma body beneath a mid-ocean ridge, both located on, or adjacent to Iceland. The seismic phenomena associated with the 1996 Vatnajӧkull subglacial eruption in central Iceland, have been studied using data recorded by both temporary (HOTSPOT) and permanent (SIL) seismic networks. These networks comprise 60 broadband and short-period three-component seismographs and cover most parts of the country. Two very active volcanic systems, Bárdarbunga and Grimsvӧtn, are situated underneath the Vatnajokull ice cap. The volcanoseismic signals recorded there were categorised according to their waveform shape and frequency content, into three groups: (a) low-frequency events (1-2 Hz); (b) mixed-frequency events (1-4 Hz); and (c) volcanic tremor. The eruption was preceded by intense seismic activity which began with a = 5.6 earthquake located at the Bárdarbunga volcanic system. The epicentres of the earthquake swarm that followed the M(_w), = 5.6 event initially delineated the Bárdarbunga caldera rim and then migrated towards Grimsvӧtn, to a place where a fissure was later observed. Pre-eruptive tremor started at least two days before the eruption as a harmonic signal around five narrow frequency bands (0.5-0.7, 1.6, 2.2, 2.8 and 3.2 Hz). Co-eruptive tremor started as a broadband, continuous signal which evolved into low-amplitude background tremor interrupted by high-amplitude, cigar-shaped bursts. Further analysis revealed that continuous tremor and the cigar-shaped bursts had all the characteristics of low- dimensional chaotic signals. Geophysical and geochemical evidence suggest that a lateral migration of magma from Bárdarbunga facilitated the rupture of the roof of a magma chamber, situated at the fissure area, which subsequently erupted as tephra on the glacier. The second phase of the RAMESSES (Reykjanes Ridge Axial Melt Experiment: Structural Synthesis from Electromagnetic and Seismics) experiment involved the acquisition of multichannel seismic reflection data from 39 along- and across-axis lines shot over the magmatically active 57º 45'N axial volcanic ridge. The data from one along-axis line were processed using a variety of techniques that mainly aimed at reducing the large amount of coherent noise present, a result of scattered energy at the rough seabed. The final processed section revealed a number of reflection events that could be interpreted as intra-crustal reflections, originating from the interface between pillow lavas and sheeted dykes, and from the top part of a thin melt lens.

A técnica da Distribuição do Tamanho de Cristais - DTC (Crystal Size Distribution - CSD), que relaciona a densidade de cristais com a distribuição do tamanho, foi aplicada à população de plagioclásio de diques máficos do Enxame de Cabo Frio - Búzios (RJ). Os diques possuem larguras variáveis, de alguns centímetros a 20 metros, e orientação em torno de N45E. A textura dos diques é geralmente fina, localmente microporfirítica e intergranular no centro dos diques mais largos. Bordas resfriadas de alguns centímetros de largura são frequentes nos contatos com as rochas metamórficas encaixantes. Foram estudados dois diques na Praia das Conchas com espessura de 0,8m e 8,2m e, um outro na Praia da Lagoinha, com 2m de largura. As amostras foram coletadas junto às margens (~10 cm do contato) e no centro dos diques. O tamanho médio dos cristais de plagioclásio varia de 0,07 a 0,13 mm na borda dos diques mais finos (<= 2 m de largura) e de 0,09 a 0,20 mm na borda do dique mais largo. No centro do dique de Lagoinha e no dique largo de Praia das Conchas o tamanho de plagioclásio é da ordem de 0,19 ± 0,02 mm e 0,60 ± 0,07 mm, respectivamente. As DTCs nas bordas dos diques, independentemente de sua largura, mostraram um padrão tipicamente encurvado, e que tem sido atribuído na literatura como evidência para misturas de magmas com populações de cristais de tamanhos distintos. No entanto, no centro do dique largo (8,2m) de Praia das Conchas, a DTC é log-linear consistente com uma cristalização magmática simples. A química mineral mostrou que os cristais maiores (precoces) de plagioclásio apresentam um teor mais elevado em An (bytownita-labradorita) que os cristais menores (tardios) da matriz (labradorita-andesina). Além disso, a olivina é mais rica em Fo na borda que no centro do dique e, respectivamente, o piroxênio mais enriquecido em Ca. Esses resultados indicam que as margens resfriadas são mais máficas que o centro sugerindo uma evolução química normal com o resfriamento do magma. Portanto, as DTCs encurvadas provavelmente refletem taxas de cristalização heterogêneas possivelmente induzidas pela despressurização durante a ascensão do magma basáltico seguida de rápido resfriamento. O padrão da DTC log-linear no centro do dique de 8,2m de largura é atribuído ao maior tempo de residência do magma que favoreceria os processos de difusão química e re-equilíbrio textural. Os cálculos da taxa de resfriamento utilizando a inclinação da DTC permitiram estimar que o centro do dique largo da Praia das Conchas estaria completamente cristalizado (~ 900 °C) em torno de 73 dias. O estudo da Orientação Preferencial de Forma (OPF) de plagioclásio mostrou que a petrotrama tende a isotrópica nas margens dos diques com largura menor que 2 metros, o que poderia refletir uma rápida cristalização de plagioclásio por despressurização. Quando a trama é localmente definida, como no dique largo da Praia das Conchas, a lineação de plagioclásio é subhorizontal sugerindo que o fluxo magmático moveu-se predominantemente na lateral do dique. / The method of Crystal Size Distribution (CSD), which relates crystal density with size distribution, has been applied on the plagioclase population of the Mafic Dyke Swarm of Cabo Frio-Búzios (RJ). The dykes are NE-trending with widths from a few centimetres to 20m. The texture is generally fine grained and locally microporphyritic and intergranular at the center of the larger dykes. Chilled margins of a few centimetres in width are common at contacts with the metamorphic basement. Two dykes of 0.8m and 8.2m in width of the Conchas Beach and another of 2m in width at the Lagoinha Beach have been studied. Samples were collected at the margins (~10cm from the contact) and at the center of the dykes. The average characteristic size of the plagioclase crystals varies from 0.07 to 0.13mm at the margins of the narrow dykes (<=2m of width) and from 0.09 to 0.20mm at the margins of the large dyke. At the center of the Lagoinha and Conchas dykes the plagioclase size varies from 0.19 ±0.02mm and 0.60±0.07mm respectively. The CSDs at the dyke margins, irrespective of the dyke width, are typically concave-up, and in literature such patterns have been attributed as evidence of magma mixing with distinct crystal populations. However, at the center of the largest dyke (8.2m) of Conchas Beach, the CSD is log-linear, consistent with simple steady-state crystallization pattern. The mineral chemistry shows that the plagioclase phenocrysts have a high An content (bytownite-labradorite) than the groundmass grains (labradorite-andesine). At the margins olivine is richer in Fo than at the center, and respectively, pyroxene is richer in Ca. These results indicate that the chilled margin is more mafic than the center suggesting a normal chemical evolution in a cooling magma. Nevertheless, the concave-up CSDs probably depict heterogeneous crystallization rates possibly induced by depressurization during the ascent of the basaltic magma followed by rapid cooling. The log-linear CSD pattern at the center of the Conchas dyke (8.2m width) is attributed to a higher residence time of the magma which favors the processes of chemical diffusion and textural re-equilibration. The calculated cooling rates using the CSD slope enables us to estimate that the larger dyke of Conchas would be completely crystallized (at ~900oC) in 73 days. The study of the Shape Preferred Orientation (SPO) in plagioclase shows an isotropic petrofabric at the margins of the dykes <=2m, which could reflect a rapid crystallization of plagioclase by depressurization. When the fabric is defined, as in the larger Conchas Beach dyke, the plagioclase lineation is subhorizontal, suggesting that the magma flow was predominantly lateral to the dyke plane.

L'activité des volcans andésitiques, tels que le Mont St Helens (États-Unis), Montserrat (Antilles) ou encore le Merapi (Indonésie), alterne entre des périodes relativement calmes, avec coulées de lave et formation d'un dôme, et des événements explosifs parfois très violents. Prévoir les transitions entre ces deux régimes est essentiel pour assurer la sécurité des populations voisines, mais demeure actuellement un vrai défi. Or les données expérimentales et les observations de terrain montrent que l'explosivité du magma est étroitement liée à son contenu en gaz. L'objectif de cette thèse est d'améliorer notre compréhension de l'évolution de ce contenu en gaz et de son influence sur l'activité volcanique, en nous appuyant sur des simulations numériques, l'analyse de données expérimentales ainsi que sur l'interprétation de données de déformation enregistrées au Merapi.Une part importante de ce travail réside dans le développement et l'amélioration de modèles d'écoulement en 2D pour prendre en compte le dégazage dans la partie supérieure du conduit, en régime transitoire. Nous présentons un modèle d'écoulement du gaz en temps qui tient compte des pertes en gaz aux bords du conduit et à sa sortie, selon les conditions présentes dans la roche encaissante et le dôme. Nous proposons également une adaptation des modèles de conduit permettant de coupler complètement l'écoulement du gaz avec celui du magma pour étudier l'évolution des conditions dans le conduit en régime transitoire. À partir de simulations de l'évolution du dégazage lors de l'emplacement d'un dôme, nous identifions les para-mètres contrôlant les pertes en gaz. Nos résultats montrent que ces pertes sont extrêmement sensibles à l'évolution de la perméabilité du magma et des gradients de pression autour du conduit en réponse au poids du dôme. La perméabilité du dôme a quant à elle peu d'influence. Au cours de la croissance du dôme, les pertes en gaz diminuent en profondeur. En haut du conduit, la pression du gaz augmente de quelques dizaines de MPa. Ces effets sont associés à une augmentation de l'explosivité du magma et de l'aléa volcanique en cas d'effondrement du dôme.Bien que la perméabilité du magma exerce un fort contrôle sur la perte de gaz, comme l'ont montré nos résultats, son évolution dans le conduit est peu contrainte. Les lois de perméabilité utilisées actuellement ne sont pas en accord avec l'ensemble des mesures réalisées sur des échantillons de magmas riches en silice. Dans le but d'améliorer notre compréhension du développement de la perméabilité dans le conduit, nous avons cherché à éclaircir le lien entre perméabilité, conditions d'écoulements, et caractéristiques géométriques du réseau de bulles connectées. Nous proposons une formulation du seuil de percolation, moment exact où le magma devient perméable compatible avec un grand nombre d'échantillons naturels et expérimentaux. Nous présentons aussi une nouvelle loi de perméabilité en accord avec la plupart des observations existantes, que nous avons intégrée à notre modèle 2D de dégazage. Nos résultats montrent qu'en fonction du nombre de bulles dans le magma et de la distribution de leurs tailles, l'importance des pertes en gaz et par conséquent les conditions d'écoulement dans le conduit varient d'effusives à explosives.Enfin, afin d'évaluer l'utilité des données de déformation pour suivre l'évolution des conditions d'écoulement, nous utilisons des modèles d'écoulement simples couplés à de la déformation élastique en 3D pour retrouver la déformation observée au sommet du Merapi peu avant l'éruption de 2006. Bien que ces modèles permettent de mieux comprendre les déplacements observés, le peu de données, associé à la complexité géologique et rhéologique du sommet, ainsi qu'à celle des processus physiques intervenant dans le conduit font qu'il est difficile de contraindre les conditions d'écoulement grâce à la déformation dans ce cas précis. / At silicic volcanoes, such as Mount St Helens (United States), Montserrat (British West Indies), or Merapi (Indonesia), periods of relative quiescence, with lava flows and dome emplacement, alternate with explosive, sometimes very violent events. Forecasting the effusive/explosive transitions, which is essential for the safety of nearby populations, remains currently a real challenge. However, experimental as well as field observations provide evidence that magma gas content is a major clue for understanding explosivity. This thesis, based on numerical simulations, experimental samples analysis, as well as on the interpretation of ground deformation data recorded at Merapi volcano, aims at improving our understanding of gas loss evolution, and its impact on the eruptive regime.A major part of this work consisted in developing and improving 2D axisymmetric conduit flow models for integrating gas loss in transient conditions. We provide a time-dependent model for gas flow in the upper conduit, that accounts for gas loss both at the conduit walls and at its top, depending on conditions in the surrounding rock and dome. We also propose an adaptation of conduit flow models allowing for full coupling between magma and gas flow in 2D that should be used to further investigate flow conditions evolution during transient regimes. From time-dependent gas flow simulations in the case of an effusive dome emplacement, we identify controlling parameters for gas loss. Our results provide evidence that gas loss is extremely sensitive to the evolution of magma permeability and of pressure gradients around the conduit due to dome loading, whereas, contrary to the common idea, dome permeability has almost no influence. Along with dome growth, gas loss decreases at depth, thus causing an increase in the magma gas content. At the top of the conduit, this results in an increase in gas pressure by a few tens of MPa, thus increasing the likelihood of magma explosivity and hazard in the case of a rapid decompression due to dome collapse.Although magma permeability plays a major role for gas extraction, as revealed by our results, its evolution within the conduit is poorly constrained. Currently used permeability laws fail in reassembling the whole dataset of permeability measurements from natural and experimental silicic samples. In order to improve our understanding of permeability development in the conduit, we worked on linking permeability and flow conditions with geometrical parameters that characterise the connected bubble network, based on experimental samples analysis. We propose an expression for the percolation threshold, i.e. the very moment when magma becomes permeable, that succeeds in classifying a wide dataset of natural and experimental samples. We also develop a new permeability law that reassembles most of the existing observations, and implement it within our gas flow 2D model. Results show that depending on the number of bubbles within the magma and on their size distribution, gas loss and then magma flow conditions evolve from effusive to explosive conditions.Eventually, we evaluate the applicability of monitoring flow conditions from observed ground deformation by using simplified conduit flow models, coupled with elastic deformation in 3D, to interpret ground deformation recorded in the near field at Merapi a few days before the 2006 eruption. Although conduit flow models provide important clues for interpreting observed displacements, the sparsity of field observations together with the complexity of the volcano summit geology, rheology and processes happening in the conduit make it very complex to constrain flow conditions from observed deformation.

A técnica da Distribuição do Tamanho de Cristais - DTC (Crystal Size Distribution - CSD), que relaciona a densidade de cristais com a distribuição do tamanho, foi aplicada à população de plagioclásio de diques máficos do Enxame de Cabo Frio - Búzios (RJ). Os diques possuem larguras variáveis, de alguns centímetros a 20 metros, e orientação em torno de N45E. A textura dos diques é geralmente fina, localmente microporfirítica e intergranular no centro dos diques mais largos. Bordas resfriadas de alguns centímetros de largura são frequentes nos contatos com as rochas metamórficas encaixantes. Foram estudados dois diques na Praia das Conchas com espessura de 0,8m e 8,2m e, um outro na Praia da Lagoinha, com 2m de largura. As amostras foram coletadas junto às margens (~10 cm do contato) e no centro dos diques. O tamanho médio dos cristais de plagioclásio varia de 0,07 a 0,13 mm na borda dos diques mais finos (<= 2 m de largura) e de 0,09 a 0,20 mm na borda do dique mais largo. No centro do dique de Lagoinha e no dique largo de Praia das Conchas o tamanho de plagioclásio é da ordem de 0,19 ± 0,02 mm e 0,60 ± 0,07 mm, respectivamente. As DTCs nas bordas dos diques, independentemente de sua largura, mostraram um padrão tipicamente encurvado, e que tem sido atribuído na literatura como evidência para misturas de magmas com populações de cristais de tamanhos distintos. No entanto, no centro do dique largo (8,2m) de Praia das Conchas, a DTC é log-linear consistente com uma cristalização magmática simples. A química mineral mostrou que os cristais maiores (precoces) de plagioclásio apresentam um teor mais elevado em An (bytownita-labradorita) que os cristais menores (tardios) da matriz (labradorita-andesina). Além disso, a olivina é mais rica em Fo na borda que no centro do dique e, respectivamente, o piroxênio mais enriquecido em Ca. Esses resultados indicam que as margens resfriadas são mais máficas que o centro sugerindo uma evolução química normal com o resfriamento do magma. Portanto, as DTCs encurvadas provavelmente refletem taxas de cristalização heterogêneas possivelmente induzidas pela despressurização durante a ascensão do magma basáltico seguida de rápido resfriamento. O padrão da DTC log-linear no centro do dique de 8,2m de largura é atribuído ao maior tempo de residência do magma que favoreceria os processos de difusão química e re-equilíbrio textural. Os cálculos da taxa de resfriamento utilizando a inclinação da DTC permitiram estimar que o centro do dique largo da Praia das Conchas estaria completamente cristalizado (~ 900 °C) em torno de 73 dias. O estudo da Orientação Preferencial de Forma (OPF) de plagioclásio mostrou que a petrotrama tende a isotrópica nas margens dos diques com largura menor que 2 metros, o que poderia refletir uma rápida cristalização de plagioclásio por despressurização. Quando a trama é localmente definida, como no dique largo da Praia das Conchas, a lineação de plagioclásio é subhorizontal sugerindo que o fluxo magmático moveu-se predominantemente na lateral do dique. / The method of Crystal Size Distribution (CSD), which relates crystal density with size distribution, has been applied on the plagioclase population of the Mafic Dyke Swarm of Cabo Frio-Búzios (RJ). The dykes are NE-trending with widths from a few centimetres to 20m. The texture is generally fine grained and locally microporphyritic and intergranular at the center of the larger dykes. Chilled margins of a few centimetres in width are common at contacts with the metamorphic basement. Two dykes of 0.8m and 8.2m in width of the Conchas Beach and another of 2m in width at the Lagoinha Beach have been studied. Samples were collected at the margins (~10cm from the contact) and at the center of the dykes. The average characteristic size of the plagioclase crystals varies from 0.07 to 0.13mm at the margins of the narrow dykes (<=2m of width) and from 0.09 to 0.20mm at the margins of the large dyke. At the center of the Lagoinha and Conchas dykes the plagioclase size varies from 0.19 ±0.02mm and 0.60±0.07mm respectively. The CSDs at the dyke margins, irrespective of the dyke width, are typically concave-up, and in literature such patterns have been attributed as evidence of magma mixing with distinct crystal populations. However, at the center of the largest dyke (8.2m) of Conchas Beach, the CSD is log-linear, consistent with simple steady-state crystallization pattern. The mineral chemistry shows that the plagioclase phenocrysts have a high An content (bytownite-labradorite) than the groundmass grains (labradorite-andesine). At the margins olivine is richer in Fo than at the center, and respectively, pyroxene is richer in Ca. These results indicate that the chilled margin is more mafic than the center suggesting a normal chemical evolution in a cooling magma. Nevertheless, the concave-up CSDs probably depict heterogeneous crystallization rates possibly induced by depressurization during the ascent of the basaltic magma followed by rapid cooling. The log-linear CSD pattern at the center of the Conchas dyke (8.2m width) is attributed to a higher residence time of the magma which favors the processes of chemical diffusion and textural re-equilibration. The calculated cooling rates using the CSD slope enables us to estimate that the larger dyke of Conchas would be completely crystallized (at ~900oC) in 73 days. The study of the Shape Preferred Orientation (SPO) in plagioclase shows an isotropic petrofabric at the margins of the dykes <=2m, which could reflect a rapid crystallization of plagioclase by depressurization. When the fabric is defined, as in the larger Conchas Beach dyke, the plagioclase lineation is subhorizontal, suggesting that the magma flow was predominantly lateral to the dyke plane.

A travers deux champs d'étude, nous nous intéressons au couplage mécanique entre le magma et l'encaissant, utilisant des méthodes numériques. Tout d'abord, nous étudions l'influence de perturbations de contraintes sur les réservoirs magmatiques, avec comme exemple deux volcans sous-glaciaires (Islande). Au volcan Grímsvötn, notre modèle montre que les vidanges, du lac sous-glaciaire (jökulhlaup) présent dans la caldera, peuvent déclencher une éruption comme ce fut le cas en 2004, en favorisant la rupture du réservoir magmatique. L'effet est cependant faible, ce qui implique que le système magmatique doit déjà être proche des conditions de rupture avant que le jökulhlaup se produise. Au volcan Katla, notre modèle indique que les conditions de rupture sont favorisées en été durant la fonte du glacier Mýrdalsjökull. Les changements de contrainte de Coulomb montrent aussi une plus forte probabilité de séismes durant la même période, résultats en accord avec la sismicité enregistrée sous le Mýrdalsjökull. Il existe une modulation à la fois de l'activité volcanique et sismique au Katla, en relation avec la variation saisonnière de la charge glaciaire. Dans un second temps, nous travaillons sur les écoulements de magma dans les conduits andésitiques. Améliorer nos connaissances sur la dynamique du magma durant son ascension est nécessaire, car les processus dans le conduit volcanique semblent gouverner l'évolution de l'activité éruptive de ces volcans. Des précédents modèles ont montré que la viscosité du magma augmente dans la partie supérieure du conduit lors de l'écoulement, ce qui cause la formation d'un plug visqueux. Mais la relation entre la mise en place du plug et les signaux précurseurs, telles que la déformation ou la sismicité, n'est pas totalement établie. A partir de nos modèles de plug, nous trouvons que les déplacements de surface sont contrôlés par la géométrie du conduit et du plug ainsi que le contraste de viscosité entre le plug et la colonne de magma. Nous montrons que l'évolution de la taille du plug est une hypothèse possible pour expliquer les rapides transitions inflation/subsidence observées à la surface des volcans andésitiques. / Through two different applications, we focus on the mechanical coupling between magma and host rocks, using numerical method. First, we study the influence of stress perturbations on shallow magma chambers, with the exemple of two subglacial volcanoes (Iceland). A variation in the stress field acts to modify the magma pressure within the reservoir as well as the failure conditions to initiate an intrusion at the reservoir wall. At Grímsvötn volcano, subglacial lake discharges (so-called "jökulhlaup") often occur in relation to eruptions. Our models show that jökulhlaup promote the failure of the magma reservoir and thus trigger eruptions, as observed for 2004 eruption. The triggering effect is small, so magmatic system must be already pressurized and close to failure before the discharge of the lake occurs. At Katla volcano, our models indicate that reservoir failure is highest in summer period when the ice load at Mýrdalsjökull icecap is reduced. Coulomb stress calculation predicts also an increase of earthquake occurrence at the caldera rim at the same time. A seasonal modulation of volcanic and seismic activity at Katla related to the loading/unloading of the icecap can exist. Secondly, we focus on magma conduit flow for andesitic volcanoes. Improving our knowledge of magma dynamic during ascent is a challenge, because processes occurring in the conduit seem to govern the temporal evolution of eruptive activity. Previous models showed that magma viscosity increases during flow in the upper part of the conduit, which causes the emplacement of a viscous plug. But the relationship between the plug emplacement and precursors signals, such as ground deformation or seismicity, is not yet fully described. From our plug flow model, we find that surface deformation is mainly controlled by the geometry of the conduit, the ratio between the plug length and the total conduit length and the viscosity contrast between the plug and the magma column. We show that the evolution of the plug size is a possible explanation for rapid transition between ground inflation and ground subsidence observed on andesitic volcanoes before extrusion.