L’objectif de cette thèse est de comprendre (1) si l’exhumation des roches mantelliques aux dorsales océaniques est compatible avec l’enregistrement des inversions de polarité du champ magnétique terrestre, (2) quels sont les processus associés aux anomalies magnétiques observées à l’aplomb des transitions océan-continent et (3) quelles sont les conséquences de ces processus sur les mécanismes de l’océanisation. Afin de travailler avec des données de résolution maximum, une méthode de calibration et d’interprétation des données magnétiques acquises proches du fond est développée dans le premier chapitre. Dans le deuxième chapitre, cette méthode est appliquée aux données magnétiques acquises à l’aplomb des domaines de manteau exhumé de la dorsale Sud-Ouest Indienne (SWIR). Il est ensuite montré que l’exhumation des roches mantelliques est associée à un signal magnétique de faible amplitude et de faible continuité spatiale rendant impossible l’identification des anomalies magnétiques d’accrétion océanique. De fait, il est proposé que contrairement aux basaltes, les roches mantelliques exhumées de la SWIR ne portent pas une aimantation rémanente suffisamment stable pour fossiliser la direction du champ magnétique terrestre. Dans le dernier chapitre, il est démontré que l’anomalie « J », antérieurement interprétée comme la première isochrone associée à la partie distale des marges Ibérie/Terre-Neuve, est en fait identifiée à l’aplomb d’une structure crustale particulière caractérisée par un haut topographique et un épaississement crustal. Par conséquent, il est proposé qu’un évènement magmatique majeur est à l’origine de la création de cette structure crustale et que cet événement est le déclencheur de la mise en place de la première dorsale océanique. / The aim of this work is to constrain (1) whether exhumation of mantle rocks at mid oceanic ridges is compatible with the record of polarity reversals of the Earth magnetic field, (2) what is the origin and the processes responsible for the magnetic anomalies observed at magma-poor rifted margins and (3) what are the consequences of these processes on continental breakup. In a first part, in order to work with high-resolution data, we develop a method for calibration and interpretation of deep-tow three component magnetic data. In a second part, we apply these methods to the data acquired above the large exhumed mantle domains of the ultraslow-spreading Southwest Indian Ridge. We show that, in these areas, neither the magnetic properties of the dredge samples nor the deep-tow magnetic data are consistent with the seafloor-spreading magnetic pattern commonly observed at mid oceanic ridges. We further suggest that in contrast to mid oceanic ridges basalts the exhumed serpentinized mantle rocks do not carry a sufficiently stable remanant magnetization to produce marine magnetic anomalies. In the last part, we show that the “J” anomaly, previously interpreted as the first seafloor-spreading anomaly of the Iberia and Newfoundland passive margins, is associated with locally high topography and thickened crust. We propose that this peculiar crustal structure results from voluminous magma both erupted at the surface and added beneath the exhumed mantle domain. We therefore propose that the J anomaly did not form during seafloor spreading but instead represents a pulse of magmatism that have triggered continental breakup.

The Gulf of Mexico Basin Opening project (GUMBO) is a study of the lithological composition and structural evolution of the Gulf of Mexico (GoM) that uses Ocean Bottom Seismometer (OBS) data from four transects in the Northern GoM. I examine 39 OBS shot records in the easternmost transect for shear wave arrivals and pick shear wave travel times from the 11 usable records. I then carry out a tomographic inversion of seismic refraction travel times. I use the resulting shear-wave velocity model in conjunction with a previously constructed P-wave model to examine the relationship between Vp and Vs. I compare velocities in the sediment and basement with empirical velocities from previous studies for the purpose of constraining lithological composition below the transect and make an interpretation of the structural evolution of the eastern GoM. The seismic velocities for crust landward of the Florida Escarpment are consistent with normal continental crust. Seaward of the Escarpment, velocities in the upper oceanic crust are anomalously high (Vp = 6.5 – 7 km/sec; Vs = 4.0 – 4.6 km/sec). A possible explanation for this observation is that GoM basalt formation consisted of basaltic sheet flows, forming oceanic crust that does not contain the vesicularity and lower seismic velocities found in typical pillow basalts. Increased magnesium and iron content could also account for these high velocities. Seismic refraction and reflection data provide a means of investigating the nature of the Moho in the northeastern GoM. I use a finite difference method to generate synthetic record sections for data from eight instruments that are part of the two easternmost GUMBO seismic lines (lines 3 & 4). I then vary the thickness of the Moho in these synthetic models and compare the results with the original receiver gather to examine the effects this variability has on amplitudes. The data from the instruments chosen for these two lines are representative of continental and transitional crust. The finite difference models indicate that the Moho beneath GUMBO 3 is ~1500 m thick based on the onset and amplitudes of PmP arrivals. All five instruments display consistent results. The instruments along GUMBO 4 suggest a Moho almost twice as thick as GUMBO 3 on the landward end of the transect that grades into a Moho of similar thickness (1750 m) in the deep water GoM. The three instruments used to model the Moho in this area show that the Moho ranges from ~1750 to 3500 m in thickness. The sharper boundary beneath continental crust in GUMBO Line 3 supports other evidence that suggests magmatic underplating and volcanism in the northern GoM during the mid-Jurassic. The thicker Moho seen on the landward end of GUMBO Line 4 that is overlain by continental crust was likely unaffected by GoM rifting. Therefore, the Moho beneath the Florida Platform might be as old as the Suwannee Terrane, and complex Moho structure is not uncommon for ancient continental crust. / text

Les parties distales des marges passives pauvres en magma représentent la transition complexe entre les domaines continentaux et océaniques. Ces zones encore peu étudiées sont pourtant des endroits clefs pour comprendre les processus impliqués durant les premiers stades de l’accrétion océanique, et plus particulièrement ceux du magmatisme et du métasomatisme. Durant ces premiers stades, ces deux processus sont gouvernés par l’exhumation mantellique. L’interaction entre les liquides magmatiques, les roches du manteau et les fluides marins vont affecter le régime thermique de la marge. De par le monde, seulement deux Transitions Océan-Continent (TOC) ont pu bénéficier d’investigations scientifiques poussées et constituent naturellement les deux sites d’études de cette thèse, à savoir, les marges actuelles conjuguées d’Ibérie-Terre Neuve du sud de l’Atlantique Nord ainsi que les marges fossiles de la Platta et de Tasna, fragments de TOCs de la Téthys Alpine Jurassique. En combinant les études de terrain ainsi que les investigations minéralogiques, pétrologiques et géochimiques, nous avons pu contraindre trois processus clefs se déroulant dans les TOCs. (i) La percolation de liquide magmatique imprégnant le manteau sous-continental hérité dans les marges Ibérie-Terre Neuve permet une refertilisation de ces marges distales. (ii) La transition géochimique visible entre les basaltes des TOCs et les basaltes de dorsales océaniques peut s’appréhender par la fusion partielle du manteau sous-continental refertilisé. (iii) Le rôle des fluides hydrothermaux, ayant des températures comprises entre 60°C et 190°C, joue un rôle sur le métasomatisme de la lithosphère en produisant une intense serpentinisation et rodingitisation, respectivement du manteau sous-continental en exhumation et des dykes basaltiques. Ces températures étant cohérentes avec une exhumation mantellique au niveau du plancher océanique. / Distal parts of magma-poor rifted margins represent a complex transition between continental and oceanic domains. These areas remain poorly understood while being a key-place to unravel magmatic and metasomatic processes involved during the first stages of oceanization. At this time, these processes are enhanced by mantle exhumation, and the interaction between melts, mantle rocks and fluids affect the thermal regime of the margin. So far, only two Ocean-Continent Transitions (OCT) have been particularly investigated, namely the present-day Iberia Newfoundland conjugate margins and the fossil analog Platta-Tasna nappes, remnants of the Jurassic Alpine-Tethys OCTs. Studies presented in this Ph.D. thesis have been focused on these two margins. Here, by combining field-works, petrological, mineralogical and geochemical investigations, we have unraveled in OCTs three key-points: (i) The deep porous-flow melt percolation impregnating the long-lived inherited subcontinental mantle in Iberia-Newfoundland margins allow the refertilization of these distal domains; (ii) The geochemical transition depicted from OCT-basalts towards MOR-basalts can be explained by the partial melting of the refertilized subcontinental mantle; (iii) The role of active hydrothermal fluids, on both the exhumed mantle and basalt dikes, lead to the serpentinisation and the rodingitization respectively, at temperature ranging between 60°C and 190°C. These temperatures being consistent with the ongoing mantle exhumation towards near-seafloor conditions.