Le phénomène de dégazage permanent observé sur le volcan Erebus s’accompagne d’une variation cyclique de la composition des gaz et du niveau de son lac de cratère que nous nous sommes proposé de modéliser en partant de l’hypothèse que ces fluctuations sont causées par l’arrivée de batch de magma naissant à faible profondeur, ascendant dans un conduit à travers duquel percole un flux continu de gaz dont l’origine est plus profonde. Nous avons tout d’abord montré par simulation numérique que la vigueur de la convection observée en surface ne pouvait être expliquée par la seule convection thermique d’un mélange liquide-cristaux. Si une alimentation continue en magma et gaz dans un système ouvert permet de simuler un comportement pulsatif de la surface, cet apport doit être suffisamment important pour que les changements de vitesse de surface ainsi générés puissent être appréciés. Le conduit doit avoir un diamètre suffisant large pour assurer la pérennité de la convection et maintenir le magma au-dessus de sa température de transition vitreuse. La présence de cristaux accélère la convection et améliore l’efficacité du transfert de chaleur entre les régions inférieures et supérieures du système magmatique ; ces cristaux se déposent dans le fond de la chambre pour former une couche de forte concentration d’une dizaine de mètres d’épaisseur. L’introduction de deux batch de magma d’origines différentes a permis de mettre en évidence combien leur composition et le comportement du lac de lave étaient sensibles à la température à laquelle ces batch sont générés. D’autre part, un batch ne contribuera au budget de dégazage dans des proportions consistantes avec les observations que si un seuil de flottabilité suffisant a été atteint. Dans le cas contraire, la migration d’une quantité de magma nécessaire à ce dégazage ne pourrait se faire que dans un conduit de très grand diamètre. Les paramètres physiques de la croute qui se forme lors du refroidissement de la surface du lac de lave, tels que son épaisseur et sa perméabilité, influent sur sa capacité à se déformer sous la pression qu’elle piège et à permettre un dégazage de type effusif. Cette pression conditionne à son tour la porosité du magma en surface et éventuellement le degré de dégazage du magma qui redescend vers la chambre magmatique. Cette étude nous a permis de mieux appréhender les mécanismes associés à une éruption de régime effusif et constitue une étape dans la compréhension de la transition à un régime explosif, préoccupation majeure des centres de surveillance volcaniques. / The observed phenomenon of permanent degassing of the volcano Erebus is accompanied by a cyclic variation of gas composition and level of its crater lake that we proposed to model in the assumption that these fluctuations are caused by the arrival of batch of magma originated at shallow depth, ascending in a conduit through which percolates a continuous flow of gas whose origin is deeper. We first showed that the observed effect of convection on the surface could not be explained solely by thermal convection of a liquid-crystals mixture. If a continuous supply of magma and gas in an open system simulates a pulsating behavior of the surface, this contribution must be large enough for the changes in surface velocity thus generated can be appreciated. The conduit diameter should be sufficient large to ensure the sustainability of convection and maintain the magma above its glass transition temperature. The presence of crystals convection accelerates and improves the efficiency of heat transfer between the upper and lower regions of the magmatic system; these crystals settle at the bottom of the chamber to form a layer of high concentration of about ten meters thickness. The introduction of two batches of magma from different depths of nucleation helped to highlight how their composition and the behavior of the lava lake were sensitive to the temperature at which they are generated. On the other hand, a batch will contribute to the degassing budget in proportions consistent with the observations only if a threshold of sufficient buoyancy was achieved. Otherwise, the migration of the amount of magma required for the degassing will only occur in a conduit of very large diameter. The physical parameters of the crust forming as the surface of the lava lake cools down, such as thickness and permeability, affect its ability to deform under pressure and to allow the trapped pressure to degas under an effusive regime. This pressure in turn influences the porosity of the magma at surface and possibly the degree of degassing of the magma being recirculated toward the magma chamber. This study allowed us to better understand the mechanisms associated with an effusive eruption regime and is a step in understanding the transition to an explosive regime, a major concern for the volcanic monitoring centers.

In this dissertation, I use seismic imaging and waveform modeling methods to investigate melt migration processes and the structure of the magma plumbing system beneath the East Pacific Rise (EPR) and Endeavour segment of the Juan de Fuca Ridge, respectively. This work begins by studying shallow mantle reflections beneath the EPR. I find the amplitude versus offset and waveform characteristics of the reflections to be consistent with a sub-horizontal dunite channels located up to 20 km off-axis. The depth of the dunite channels correlate with patterns of mantle melt delivery and the predicted base of the thermal lithosphere, suggesting the channels are thermally controlled and may have formed in situ via dissolution by focused flow at the base of the lithosphere. This interpretation is consistent with field observations in ophiolites and numerical modeling of melt-focusing channels. The three-dimensional velocity structure of the Endeavour segment is then investigated to identify how patterns of mantle melt delivery influence the segment-scale distribution of crustal melt and crustal accretion. The results from this study indicate that the mantle magmatic system is skewed relative to the ridge-tracking crustal magmatic system and that this skew exerts primary control on magmatic, tectonic, and hydrothermal activity at the Endeavour segment. In regions where mantle melt delivery is axis-centered, mantle-derived melts are efficiently transported from the mantle to the crust, resulting in frequent crustal melt replenishment, associated seismogenic cracking, and enhanced crustal melt content that drives vigorous hydrothermal activity. Conversely, sites of off-axis melt delivery are characterized by less efficient vertical melt transport, resulting in infrequent crustal melt injection and hence, reduced crustal melt content and hydrothermal activity. Next, I focus on how along-axis variations in magma replenishment modulate crustal permeability and the intensity of hydrothermal circulation. Using full-waveform inversion, I show that sites of localized magma replenishment to the axial magma lens, along with induced seismogenic cracking, coincide with enhanced permeability. I conclude that the frequency of magma injection governs hydrothermal circulation patterns and heat flux at mid-ocean ridges. This dissertation includes previously published and unpublished coauthored material.

<p dir="ltr">To better predict volcanic behavior, we must understand the processes that occur in the underlying magma reservoirs. This thesis contains three chapters that work together to better understand processes that occur in sub-volcanic reservoirs. Chapter 2 is a study of an ancient, coupled volcanic-plutonic system to determine the link between the volcanic and plutonic parts of the system. The IXL-Job Canyon magmatic system is an ~28-29 Ma system, which shows a rapid transition between eruption of tuffs and lava flows to construction of an upper-crustal pluton, via incremental emplacement. The system experienced an eruptive hiatus during and after pluton construction, until the eruption of a newly identified, younger, rhyolitic tuff. This work suggests that the absence of volcanic activity at the surface does not mean that the underlying magmatic plumbing system is also inactive. Chapter 3 compiles existing U-Pb zircon ID-TIMS data for upper-crustal, silicic magmatic systems, to determine the size and frequency of magmatic increments that accumulate to build up these systems. A Monte Carlo-based model is used to investigate the underlying distributions of the increment size and time between increments, and results in sizes and inter-event times that follow an exponential distribution. This work helps guide how we can try to introduce broadly generalizable complexities into thermal models of such systems. Chapter 4 focuses on organic interferences, a common issue that impacts the speed and quality of U-Pb and Pb-Pb data collected on TIMS instruments. We share two techniques used at the Purdue Radiogenic Isotope Geology Lab to first reduce and then avoid any residual organic interferences. These techniques help shorten analytical times, increasing throughput, and provide a means to reduce uncertainties on our measurements, since the presence of organic interferences can bias and increase the uncertainties on U-Pb dates.</p>