No description available.

No description available.

Northeast Japan experienced an approximately constant, compressional deformation during the last 5 million years resulting from the steady subduction of the Pacific plate. Because the direction of the maximum compression axis is approximately perpendicular to the strike of the island arc, 2-D finite-element modeling can be used to examine the deformation over time of the island-arc lithosphere. The model geometry is based on geophysical and geological data, and each model run requires an assumed rheology and interplate coupling. Novel to our modeling is the ability to include erosion/deposition loading and the creation of strike-slip faults, based on a dynamically-applied fracture criterion. The criterion for acceptability is how well a model matches observed present-day topography, gravity, and seismicity patterns. Results given below are for models that satisfy this criterion. The long-term effective elastic thickness is 10 km in the inner arc, increasing to about 50 km near the trench. The effective elastic thickness in the inner arc is therefore much smaller than the about 30 km short-term elastic thickness estimated from seismological data. The viscosity of the lower crust is on the order of 1022 Pa s or less. The strength of interplate coupling off Sanriku is about two to four times greater than off Miyagi, and there is about twice as strong a coupling at greater depths. The relative strength of coupling correlates well with the observed interplate seismicity. Hence the inferred weaker coupling off Miyagi indicates a lack of seismogenic potential -- a low probability for large earthquakes in that region, not just a long return cycle. The same modeling procedure was also applied to southwest Japan. The viscosity of the lower crust is not more than 1021 Pa s, and the elas tic thickness is about 10 km. The calculated strength of interplate coupling for southwest Japan is about 1.5 times greater than for the off-Sanriku region in northeast Japan, which correlates well with the fact that there have been great (M>8) earthquakes in the Nankai Trough region, but none that large in the off-Sanriku region. / Ph. D.

2-D numerical computer models based on thermodynamic and kinematic principles have become invaluable tools for simulating geodynamic processes at these systems. Numerical models have proven effective for allowing the examination and computation of multiple factors simultaneously, providing scientists with an important resource with which to study complex systems. Previously, for instance, numerical models have been used for examining different factors involved in magma production at subduction zones and mid ocean ridges by modelling the influence and interplay of factors such as the effect of hydration and the influence of the depth of the fault between the two plates on the melting (van Keken, 2003; van Keken 2008). Additional models have explored the thermal structure of subduction zones and its relationship to the processes involved at convergent boundaries, including magma production (van Keken, 2023a). Syracuse et al. (2010) used numerical models for subduction zones, creating thermal models that examined dehydration and melting in subduction zones with a variety of slab geometries, convergence velocities, ages and structures. Still others have shown that thermal structure affects melt production, formation of arc volcanoes, dehydration, and seismicity, modelling the effects of varying slab dip, plate convergence velocity, plate age, etc. (Syracuse et al., 2010; Hayes et al, 2018). However, none have yet utilized models to systematically investigate magma production at either subduction zones or mid-ocean ridges to specifically examine both batch and fractional melting with the combination of multiple controlling factors including slab dip, convergence rate, hydration, minerology, and slab age. This project investigated the processes surrounding magma production at subduction and mid-ocean ridge systems through the creation of a numerical model and utilization of the developed model to explore the effects of a multitude of parameters on fractional and batch melting, as well as investigated the incorporation of incompatible elements, and other processes of interest in subduction and mid ocean ridge systems.

Subduction zones, where one tectonic plate descends beneath another, are the most seismically active regions on Earth and have produced the largest earthquakes and some of the most destructive tsunamis ever recorded. Significant questions remain regarding the roles both the downgoing and overriding plates play in contributing to varying styles of rupture along the main seismogenic contact between the two plates, or megathrust, where such great (Mw > 8) earthquakes are generated. In the last few decades, the scientific community has recognized how different structural and compositional properties of both plates, and in particular the hydration state of the incoming plate can contribute to variations in megathrust slip behaviors. In this thesis, I show how marine multichannel seismic (MCS) and ocean-bottom seismometer (OBS) data can be used to investigate structural controls on megathrust slip behavior including the different styles of great earthquakes and/or the generation of slow slip events. Offshore Alaska and Sumatra, we used long-streamer multichannel seismic data to create a high-resolution P-wave velocity (Vp) model of the upper oceanic crust prior to subduction. Using a differential effective medium theory, we place the first constraints on the amounts pore (free) water contained therein. Our results indicate that the uppermost oceanic crust of the incoming plates in both regions is significantly hydrated. Offshore Alaska, we show that pervasive faulting in the bending area allows seawater to penetrate into the uppermost crust. We propose that high water content in uppermost crust might contribute to observations of low coupling along the shallow plate interface in this area through the expulsion of pore fluids. Geochemical analyses of arc lavas in this segment of the Alaska subduction zone suggests significant fluid release from the downgoing crust compared to adjacent segments. Thus, we propose that during subduction, additional bending and high-temperature circulation of remaining pore fluids could further alter the upper oceanic crust that dehydrates around sub-arc depths. Offshore Sumatra, few bending-related faults are observed; however, evidence for significant and homogeneous hydration within the the uppermost crustal layer 2A (extrusives) suggests that plate bending plays a role in the shallow reopening cracks, facilitating the shallow penetration of seawater. In layer 2B (sheeted dikes) just below, our results suggest heterogeneous, yet significant, hydration that we attribute to the slow and diffuse deformation taking place in the Wharton Basin. We speculate that the large amounts of upper-crustal water carried into the Sumatra subduction zone can influence shallow slip behavior, as evidenced by recent records of a long-lasting slow slip event in the area. To further explore potential structural and compositional controls on spatial varia- tions in megathrust slip behavior in Alaska, we use OBS data to create a 3D Vp model of the Alaska Peninsula Subduction zone within a 500-by-400 km wide area with good resolution down to 20-25 km depths in both the incoming and overriding plates. Our model samples two subduction zone segments that exhibit differences in history and style of megathrust rupture. We interpret reductions in seismic velocities within the incoming plate as evidence for modest hydration of the Pacific oceanic plate resulting from a series of fracture zones and the formation of large seamounts and an associated basement swell, or platform. The bathymetry of the seamounts and platform in part modulates the distribution and lithology of subducting sediments across the margin that we propose might influence shallow slip behavior. Within the overriding North American plate, we see evidence for contrasting styles of deformation and variations in composition (i.e., rigidity) that agrees well with observed changes in plate coupling and great earthquake history. These results emphasize the importance of considering not only one, but several factors related to both the incoming and overriding plates which collectively contribute to along-strike and downdip variations in megathrust slip behavior between segments. Our final study looks at the incoming Cocos plate just before it subducts offshore Mexico beneath the North American plate. Here we jointly inverted 2D OBS and long-offset MCS data acquired parallel to the trench to derive a 270 km-long, high-resolution Vp model of the entire oceanic crust and uppermost mantle. We provide the first constraints on the quantities of both free and structural (i.e., mineral-bound) water contained within the Cocos plate outboard of the Guerrero Gap and adjacent segments of the Mexican subduction zone. The Guerrero gap hosts large slow slip events that are commonly explained through the release of water through the dehydration of altered sediments and upper oceanic crust downdip. Strikingly, our results show that while the Cocos plate is hydrated offshore Mexico, nearly all of the water is contained within the upper oceanic crust. Moreover, we see that most of the water by weight is present as free fluids in the pores and that the upper oceanic crust is only moderately altered (0.3-1.3 wt.%) compared to global averages (> 1.5-3 wt.%). While the upper crust appears hydrated everywhere across our profile, we find that ∼30% more water is subducting outboard the Guerrero seismic gap where large seamounts contribute to a thicker extrusive layer and more alteration. This, along with evidence for the subduction of seamounts in Guerrero might help explain observations of weak shallow plate coupling and a greater propensity for slow slip at greater seismogenic depths compared to adjacent segments. These results provide important new constraints on how much pore and structural water is carried in the Cocos plate offshore Mexico. We propose that global estimates of incoming structural water content are not applicable everywhere, as is commonly assumed by petrologic and thermal models. Much less structural water may be needed within the upper oceanic crust just before subduction to explain the occurrence of slow-slip events downdip in some subduction zones.

Le complexe de Bay of Islands (BOI), situé sur la côte ouest de Terre-Neuve au Canada, est l'un des systèmes de subduction fossiles les mieux préservés de la planète. Le complexe de BOI est constitué de matériel allochtone mis en place sur le continent Laurentia lors de la phase Taconienne de l'orogénèse Appalachienne. Il se compose d'une ophiolite reposant sur une semelle métamorphique. La semelle métamorphique de BOI comprend, du sommet vers la base, des granulites à grenat et clinopyroxène, des amphibolites à grenat et clinopyroxène, des amphibolites communes et des roches métasédimentaires. Deux types de protolithes peuvent être distingués parmi la portion mafique de la semelle : des protolithes cumulatifs gabbroïques avec des Mg# élevés et des compositions appauvris en éléments traces ainsi que des protolithes basaltiques avec des contenus en éléments traces similaires à des N-MORB. La datation de titanite provenant de deux amphibolites communes permet de fixer un âge minimum pour le métamorphisme à 485 ± 4 Ma. Les données détritiques U-Pb sur zircon de la partie métasédimentaire de la semelle de BOI forment un spectre d'âges allant de l'Archéen à l'Ordovicien dominé par la population la plus jeune dont le pic se situe à c. 490 Ma. Ce pic est interprété comme l'âge maximal de déposition des sédiments. L'origine de ces jeunes détritus peut être attribuée au magmatisme associé à l'orogénie Taconienne. Les données obtenues indiquent que 1) l'initiation de la subduction responsable de la formation du complexe de BOI s'est produite dans le domaine océanique possiblement à une faille de détachement avant c. 485 Ma et 2) les sédiments constituant la partie métasédimentaire de la semelle proviennent de la plaque supérieure et se sont déposés sur la plaque inférieure lorsque la subduction était déjà en cours.

En 2004 et 2007, neuf structures d'échappements de fluides (SEF), constituées de volcans de boue (VdB), d'évents de gaz (GS, gas seeps en anglais) et/ou de sources, ont été échantillonnées dans la partie émergée du Bassin de la Côte Est (BCE) de l'île nord de la Nouvelle-Zélande.L'étude granulométrique indique que la boue émise par les VdB, les roches encaissantes et les niveaux de décollement voisins sont composés d'argiles et de silts en majorité. L'étude de la fraction argileuse et de la roche totale par diffraction des rayons X (DRX) a révélé de nombreuses similitudes. Les volcans de boue, les roches encaissantes et les niveaux de décollement présentent ainsi le même assemblage minéralogique, à savoir smectite, illite, chlorite, kaolinite, quartz et feldspaths. Les proportions sont variables d'un échantillon à un autre mais le couple smectite-illite est toujours majoritaire.L'étude géochimique de la fraction solide indique que les échantillons sont riches en Si02, pauvres en Fe2O3, MgO, MnO et en alcalins, à quelques exceptions près. La composition en éléments majeurs s'organise entre un pôle argileux alcalin et un pôle carbonaté. Les spectres de terres rares sont similaires et caractéristiques des argiles ; Ils présentent également un faible degré de fractionnement, lié à la formation des carbonates. Ce sont donc les mêmes minéraux qui contrôlent la chimie des échantillons.L'étude géochimique de la phase liquide montre que l'eau impliquée dans les volcans de boue est d'origine marine essentiellement, et des réactions eau/roche similaires, notamment l'altération de smectite en illite. Cette étude a permis aussi d'obtenir une estimation de la température d'équilibre, comprise entre 60 et 110°C, impliquant une profondeur d'origine de 2 à 3 km, voire plus.L'étude géophysique indique qu'à l'aplomb des VdB et des deux sources chaudes étudiés, la croûte continentale a sensiblement la même épaisseur et que la profondeur de la croûte océanique en subduction avoisine les 20 km. A cette profondeur, la fusion de la péridotite n'est pas possible et la fusion résultante de la croûte continentale, responsable du volcanisme d'arc, non plus. Le gradient géothermique mesuré à TePuia est donc influencé par un autre phénomène, mais la modélisation de la lithosphère ne nous a pas permis de trouver lequel.Ces différentes études mettent en évidence des caractéristiques géochimiques, pétrographiques et minéralogiques communes aux volcans de boue de Nouvelle-Zélande. Les fluides impliqués dans ces structures proviendraient donc d'un même niveau source, recouvert du même assemblage sédimentaire. L'étude géophysique ne nous apporte aucune information à ce sujet mais permet cependant d'établir avec certitude que le régime thermique est le même du Nord de Hawke's Bay au Sud du BCE ; la région de TePuia est un cas particulier, peut-être influencée par le complexe volcanique de Matakaoa. / In 2004 and 2007, nine areas of fluids escapes structures (FES) with mud volcanoes (MVs), gas seeps (GSs) and springs, were sampled in the East Coast Basin (ECB) emerged part of New Zealand's north island.The grain size study indicates that the mud emitted by MVs, like surrounding rocks and decollement layers in the vicinity, is mainly composed of clays and silts. Clay fraction and whole rock XRD study revealed many similarities. Mud volcanoes, surrounding rocks and decollement layers have exactly the same mineral assemblage, i.e. smectite/illite, chlorite, kaolinite, quartz and feldspars. The proportions vary from one sample to another but the couple smectite/illite is always majority.The geochemical study of the solid fraction indicates that the samples are rich in Si02, low in Fe2O3, MgO, MnO and alkaline, with a few exceptions. The major element composition is organized between a pole and a pole clay alkaline carbonate. The REE patterns are similar and characteristics of clays They also have a low degree of fractionation, due to the formation of carbonates. So these are the same minerals that control the chemistry of samples.The liquid phase geochemical study shows that water involved in the MVs is mainly of marine origin, that water/rock reactions are similar, including weathering of illite in smectite, allowed us to obtain an estimate of the water equilibrium temperature between 60 and 110°C, implying an origin depth of 2-3 km.The geophysical study indicates that directly above the MVs and the two hot springs studied, the continental crust has substantially the same thickness and the depth of the subducting oceanic crust is around 20 km. At this depth, the melting of the peridotite is not possible and the resulting fusion of continental crust, responsible of arc volcanism, neither. The geothermal gradient measured at TePuia is influenced by another phenomenon, but the modeling of the lithosphere does not allow us to find which one.These studies show geochemical, petrographical and mineralogical characteristics common to mud volcanoes in New Zealand. Fluids involved in these structures therefore come from the same parent bed, covered with the same sedimentary package. The geophysical study gives us no information about it but nevertheless allows us to establish with certainty that the thermal regime is the same from Northern Hawke's Bay to the south of the ECB; TePuia region is a special case, perhaps influenced by the Matakaoa volcanic complex.

Les zones de subduction impliquent le recyclage de la lithosphère océanique dans le manteau convectif pour former une croûte continentale par migration ascendante de la partie fondue du slab. Ce processus à long terme, où la plaque supérieure accumule des contraintes qui sont instantanément libérées sous la forme de séismes, se répète sur un processus cyclique : le cycle sismique de subduction (SEC). Néanmoins, le glissement rapide sur les failles n'est pas le seul processus contribuant à la déformation pendant le SEC. Des processus de glissement asismique sont observés sous la forme de creeping constant et de glissement lent transitoire. Exemple du dernière est le phénomène d'afterslip et les slow slip events (SSE) dont la durée est beaucoup plus longue que les séismes "normaux". De la même importance est l'effet du stress induit par un sésime dans le manteau. Parce qu'il se comporte comme un corps viscoélastique, le manteau détend ce stress au cours d'un processus de déformation lent de longue durée et à grande échelle. Tous ces phénomènes ont un impact sur la déformation de la plaque supérieure et doivent être pris en compte dans l'analyse du budget de glissement sur le SEC afin d'obtenir l'estimation réelle du risque sismique d'une région. En partant de cette prémisse, la principale motivation de ce travail est de contribuer à la détermination de l'état actuel de couplage de l'interface sismique de la zone nord de la subduction chilienne. La majeure partie de cette étude a été menée dans deux directions : l'analyse et l'interprétation des déformations transitoires à court terme (jours, semaines) et à long terme (années) détectées par les stations GPS permanentes déployées au nord du Chili. Premièrement, en recherchant de façon exhaustive les signaux transitoires à court terme dans des séries géodésiques, nous cherchons à contribuer au débat actuel sur la question de savoir si des glissements lents se produisent dans la zone de subduction chilienne ou non. Jusqu'à présent, un seul cas de glissement lent associé à la phase de nucléation d'Iquique 2014 a été rapporté. Nos résultats, obtenus après filtrage et analyse approfondie des séries temporelles consacrées à la réduction de leur bruit inhérent, indiquent l'existence de seulement 3 événements susceptibles de correspondre à de petits épisodes de glissement sismique. Ils se produisent entre 2009 et 2011, avec des durées de quelques semaines et des amplitudes qui ne dépassent pas 4 mm. De plus, l'occurrence de l'épisode de glissement sismique plus important simultané à l'activité du séisme d'Iquique 2014 est confirmée, et l'analyse est poussée plus loin à l'aide d'enregistrements par inclinomètres à longue base. Grâce à cette analyse affinée et à la sensibilité de cet instrument, 4 événements de glissement lent dist incts peuvent être identifiés dans la plus grande région de glissement lent révélée par les données GPS, survenant au cours des 3 mois précédant le choc principal avec des magnitudes comprises entre Mw 5,8 et 6,2. Enfin, une analyse à long terme a été effectuée afin d'identifier les changements mineurs mais significatifs des tendances sur de longues périodes. Les données disponibles indiquent une diminution à long terme de la déformation de la plaque supérieure après le séisme de profondeur moyenne de Tarapaca 2005 (Mw 7.7). Nous testons l'effet viscoélastique de l'asthénosphère sur la déformation de surface déclenchée par une rupture profonde (~ 100 km). Pour cela, nous construisons un modèle 3D viscoélastique réaliste d'éléments finis viscoélastiques et avons essayé les rhéologies viscoélastiques de Maxwell et Burgers. Nous trouvons une très bonne correspondance entre le signal postsismique observé associé à cet événement et la déformation postsismique modélisée en utilisant une rhéologie de Burgers avec une viscosité long terme de 1.9e+18 Pa s, contestant l'hypothèse d'une diminution du couplage interséismique conduisant finalement à le mégathrust d'Iquique. / Subduction zones involve the recycling of oceanic lithosphere in the convective mantle to form continental crust by upward migration of the melted part of the slab. In this long-term process, the upper plate accumulate stresses that are instantaneously released in the form of earthquakes, repeating over a cyclic process: the subduction earthquake cycle (SEC). Nevertheless, fast slip on faults is not the only process contributing to the deformation during the SEC. For instance, the low frequency process of aseismic slip is observed in the form of steady fault creeping and transient slow slip. Examples of the latest are the phenomenon of afterslip, and slow slip earthquakes (SSEs) with much longer durations than ''normal'' earthquakes. Of the same importance is the effect of stress induced by an earthquake in the mantle. Because it behaves like a viscoelastic body, the mantle will relax this stress during a slow, long-lasting (up to decades) and large-scale (up to thousands of km) deformation process. All these phenomena impact the upper plate deformation and have to be considered in slip budget analysis over the SEC in order to obtain the real estimation of the seismic hazard of a region. Following this premise, the principal motivation of this work is to contr ibute to the determination of the actual locking state of the north seismic gap of the Chilean subduction zone. The main part of this study was conducted along two directions: analyzing and interpreting both short-term (days, weeks) and long-term (several years) transient deformations detected by permanent GPS stations deployed in north Chile. First, with an exhaustive search for short-term transient signals in 15-years long geodetic time series, we aim at contributing to the current debate of whether slow slip events do occur in the Chilean subduction zone or not. Up to now, only a single and debated case of slow slip associated to the nucleation phase of Iquique 2014 has been reported. Our results, obtained after thorough filtering and analysis of the times series devoted to reduce their inherent noise, indicate the existence of only 3 events that are likely to correspond to small episodes of aseismic slip. They occur between 2009 and 2011, with durations of few weeks and amplitudes that do not exceed 4 mm. Additionally, the occurrence of the larger aseismic slip episode simultaneous to the foreshock activity of Iquique 2014 earthquake is confirmed, and the analysis is pushed further with the help of long-base tiltmeter records. Thanks to this refined analysis and to the sensitivity of this instrument, 4 distinct slow slip events can be identified in the larger region of slow slip revealed by GPS data, occurring during the 3 months previous to the mainshock with magnitudes ranging between Mw 5.8 and 6.2. Finally, a longterm analysis of the same cGPS time series was conducted in order to identify small but significant changes of trends over long durations. Available data indicate a long-term decrease of the upper plate deformation after the intermediate depth earthquake of Tarapaca 2005 (Mw 7.7). We test the viscoelastic effect of the asthenosphere on surface deformation triggered by deep failure (~ 100 km depth). For this, we build a realistic 3D viscoelastic finite element model and tried Maxwell and Burgers viscoelastic rheologies. We find a very good correspondence between the observed postseismic signal associated to this event and the modeled postseismic deformation using a Burgers rheology with a long-term viscosity of 1.9e+18 Pa s, challenging the hypothesis of a decrease of interseismic coupling eventually leading to the megathrust failure of Iquique.

La partie superficielle des zones de subduction (0-10 km de profondeur) a longtemps été considérée comme asismique. Cependant la découverte de séismes très basses fréquences dans cette zone, ainsi que la propagation très superficielle de la rupture cosismique lors du séisme de Tohoku-Oki (Japon) remettent en question cette hypothèse jusqu’alors largement admise. L’une des raisons pour lesquelles le potentiel sismogénique de cette zone est mal contraint réside dans le fait que les processus qui y règnent sont complexes, mêlant déformation et diagenèse (principalement la transformation des argiles de type transition smectite vers illite), et ainsi difficilement reproductibles en laboratoire. Au cours de cette thèse, des expérimentations en presse triaxiale sur échantillons smectitiques (représentatifs des matériaux accrétés dans les prismes d’accrétion) et illitiques ont été réalisées sous différentes conditions de pression de confinement (de 50 à 200 MPa) et de température (20°C et 300°C). Ces expériences ont été menées afin d’identifier les modes de déformation de ces échantillons, et de déterminer les effets couplés de cette déformation et de la diagenèse sur la rhéologie de ces roches et notamment leur potentiel à générer des instabilités de glissement. Dans ces expériences, le style de déformation est à chaque fois similaire, avec, tout d’abord, une localisation progressive de la déformation le long d’une zone de cisaillement, puis formation d’une fracture. Malgré cette constance dans le style de déformation, le comportement rhéologique des échantillons, lui, est drastiquement opposé entre les expériences à 20 et à 300°C, avec un comportement exclusivement durcissant à 20°C (i.e. stable) quel que soit la minéralogie, et un comportement qui évolue systématiquement vers du stick-slip (i.e. instable) à 300°C (pour les échantillons smectitiques). Ces résultats montrent que la réactivité chimique des smectites au cours de la diagenèse (activée dans les expériences à 300°C) conditionne la formation d’instabilités de glissement. Nous proposons que la réactivité chimique des smectites dans les zones de subduction pourrait promouvoir la propagation de la rupture cosismique vers la surface. / The shallow portion of subduction zones (0-10 km depth) has long been considered as unable to store and release seismic energy. However, the detection of very-low frequency earthquakes in this zone, as well as the propagation of the coseismic rupture to the trench during the Tohoku-Oki earthquake, question this hypothesis. The difficulty to assess the seismogenic potential of this shallow portion lies principally in the complexity of the processes that occur in this zone, combining deformation and diagenesis (especially the smectite-to-illite transition), and hence not easily reproducible in laboratory. In order to analyse the mechanical properties of the shallow portion of subduction zones, triaxial tests have been performed with smectitic and illitic samples, under confining pressure between 50 and 200 MPa and at temperature of 20 and 300°C. The aim of these experiments was to identify the deformation modes of such sedimentary material and to determine the effects of deformation and diagenesis on rheology of these materials and on the rock potential to exhibit instable failure. In the experiments, deformation operates under the same pattern with a progressive localisation from shear band to fracturation. Even if the deformation style does not differ much between experiments, the rheology of the samples tested at 20°C and at 300°C contrasts drastically. Indeed, while the samples (smectitic and illitic) tested at 20°C show exclusively a strengthening behaviour (i.e. stable), the smectitic samples tested at 300°C exhibit a rheology that systematically evolves from strengthening to stick-slip behaviour (i.e. unstable). These results indicate that the chemical reactivity of smectite under diagenetic conditions (diagenesis is activated in the experiments conducted at 300°C) constitutes a weakening mechanism promoting unstable sliding. Finally, we propose that, at the subduction zone scale, the chemical metastability of smectite could promote the propagation of the coseismic rupture to the very shallow portion of accretionary prisms.

Non renseigné / Most mega-earthquakes (i.e. earthquakes with Mw ≥ 8.5) occur along subduction mega-thrusts, the interfaces between the subducting - and the overriding plates in convergent margins. These events may have catastrophic impact on human societies due to their destructive potential. For this reason being able to predict the timing and size of these earthquakes became one goal of the international scientific community. The subduction seismic cycle is influenced by many different parameters. The interplay between these parameters governing the frequency and size of megathrust earthquakes still remains unclear, mainly due to the short (i.e. limited to the last century) seismic record.The seismogenic part of the subduction thrust fault spans between depths of 11±4 and ± 51 km (Heuret et al. 2011). In this zone a combination of temperature, pressure and rocks characteristics creates conditions favourable for seismic behaviour. Whether a specific area in the subduction thrust fault has the ability to trigger mega-earthquakes can be expressed using the degree of seismic coupling, i.e. the amount of slip that occurs with respect to the total amount of plate convergence (e.g. Scholz 1998; Scholz & Campos 2012). When a fault is fully coupled, all of the fault slip occurs during earthquakes instead of also during aseismic behaviour (e.g. slow slip events). The internal structure of the interplate fault zone mainly determines whether an area within a subduction zone behaves seismic or aseismic (Wang & Bilek 2011). This is influenced by the topography of the plate interface (e.g. subducting seamounts; Wang & Bilek 2014), but also subducted sediments and fluids in the subduction channel may play an important role.The main goal of this project is to understand which parameters affect the behaviour of mega-earthquake ruptures. This will be done by comparing natural data (e.g. seafloor roughness, sediment thickness and fluid content in the subduction channel) to rupture characteristics of major recent earthquakes. With this analysis also more knowledge can be gained on the triggering of slow earthquakes instead of mega-earthquakes. These are slow slip events with lower frequencies and longer durations than ‘regular’ earthquakes (Saffer & Wallace 2015).The database of natural data, implemented by the long-term scientific joint venture between the Univ. Montpellier and the LET (Roma Tre) will be used for the analysis. Ongoing work is done on determining a method for estimating the seafloor roughness, i.e. the distribution of high, low and smooth areas (by Michel Peyret in collaboration with Serge Lallemand, Univ. Montpellier). Also data is available on the trench sediment thickness around the world (Heuret et al. 2011). In the frame of this project, information on the roughness of the seafloor will be added to the database. In addition the rupture characteristics of major recent earthquakes will be collected. By performing a multiparametric statistical analysis of the database, a conceptual model will be realized, exploring the possible link between all the different parameters. The aim is to validate this model in the lab using scaled 3D analogue models. This will be done both at the LET and at Univ. Montpellier by using a broad range of geometries and contact materials with different rheologies (e.g. gelatin, foam rubber and a new analogue material; Caniven et al. 2015; Corbi et al. 2013). This jointed experimental approach with both the Univ. Montpellier and the LET involved creates a rich environment where differences and similarities of the two different approaches can be used to validate the results.