The Vredefort Dome is located southwest of Johannesburg, South Africa, and represents the deeply eroded remnant of the central uplift of the world’s largest known impact structure, with an estimated diameter of ~300 km. The Vredefort impact structure is also the oldest known impact structure on Earth (~2.02 Ga). The Vredefort Dome comprises an ~40 km wide core of Archaean basement gneisses and an ~20 km wide collar of subvertical to overturned Late Archaean to Palaeoproterozoic supracrustal strata. This project presents the results of Landsat-TM and aerial photograph analysis, as well as field mapping of Witwatersrand Supergroup metasedimentary strata in the collar of the Vredefort Dome. The aim of this study was to investigate the structures (such as folds, faults, fractures), at all scales, and other deformation features (such as shatter cones and pseudotachylitic breccias) in the field area, and to establish geometric and temporal relationships between these features with regard to the impact cratering process. This study revealed a highly heterogeneous internal structure of the collar involving folds, faults, fractures and melt breccias that are interpreted as the product of shock deformation and central uplift formation during the Vredefort impact event. Broadly radially-oriented symmetric and asymmetric folds, with wavelengths from tens of metres to kilometres, and conjugate radial to oblique faults with strike-slip displacements of, typically, tens to hundreds of metres accommodated tangential shortening of the collar of the dome that decreased from ~17 %, at a radial distance from the dome centre of 21 km, to <5 % at a radial distance of 29 km. Ubiquitous shear fractures containing pseudotachylitic breccia, particularly in the metapelitic units, display variable local slip senses consistent with either tangential shortening or tangential extension; however, it is uncertain whether they formed at the same time as the larger faults during the rise of the central uplift or earlier, during the shock compression phase of cratering. Contrary to the findings about shatter cones of some earlier workers in the Vredefort structure, the Vredefort cone fractures do not show uniform apex orientations at any given outcrop, nor do small cones show a pattern consistent with the previously postulated “master cone” concept. The model of simple back-rotation of the strata to a horizontal pre-impact position also does not lead to a uniform centripetal-upward orientation of the cone apices. Striation patterns on the cone surfaces are variable, ranging from typically diverging, i.e., branching off the cone apex, to subparallel to parallel on almost flat surfaces. Striation angles on shatter cones do not increase with distance from the crater centre, as suggested previously. Instead, individual outcrops present a range of such striation angles, and a more irregular distribution of striation angle values with regard to the distance from the crater centre suggests localised controls involving the nature and shape of various heterogeneities in the target rock on this aspect of cone morphology. On the basis of the observations made during this study on small-scale structures in the collar of the Vredefort Dome, the relationship of shatter cones with curviplanar fractures (multipli-striated joint sets - MSJS) is confirmed. Pervasive, metre-scale tensile fractures crosscut shatter cones and appear to have formed after the closely-spaced MSJ-type fractures. The results of this study indicate that none of the existing models is able to explain all characteristics of shatter cones fully; therefore, a combination of aspects of the different models may currently be the best possible way to explain the formation and origin of shatter cones, and the formation of the related MSJ and their characteristic aspects (e.g., curviplanar shape, melt formation, etc.). The observed variety of shatter cone orientations, surface morphology and striation geometry in the dome concurs broadly with the results of some previous studies. The abundance of striated surfaces along closely-spaced sets of fractures (MSJ) observed in this study can be reconciled with reflection/scattering of a fast propagating wave at heterogeneities in the target rocks, as proposed by recent studies. This would mean that closely-spaced fractures and shatter cones were not formed during shock compression, as widely postulated in the past, but immediately after the passage of the shock wave, by the interference of the scattered elastic wave and the tensional hoop stress that develops behind the shock front. In addition to shatter cones, quartzite units show two other fracture types – a centimetre-spaced rhomboidal to orthogonal type that may be the product of shock-induced deformation and related to the formation of shatter cones, and later joints accomplishing tangential and radial extension. The occurrence of pseudotachylitic breccia within some of these later joints confirms the general impact timing of these features. Pseudotachylitic breccias in the collar rocks occur as up to several centimetre-wide veins with variable orientations to the bedding and as more voluminous pods and networks in zones of structural complexity, such as the hinges of large-scale folds and along large-scale faults, as well as locally, at lithological interfaces. In places, tension gash arrays along thin veins are observed indicating that movement occurred along these planes. Initial cooling calculations for pseudotachylitic breccias of different widths and compositions (metapelite or quartzite) suggest that thick veins (<10 cm) could have stayed molten over the entire duration of crater development (at least 10 minutes), making it possible for shock-induced melts to intrude dilational sites, such as fold hinges and extensional fractures, during the formation and subsequent collapse of the central uplift. Intrusion of such melts may also have lubricated movements along brittle and ductile structures. Thus, the presence of both shock- and friction-generated melts is likely in the collar of the Vredefort Dome. Based on the spatial and geometric relationship between the structures and other deformation features observed in the collar rocks of the Vredefort Dome, it is possible to establish a temporal sequence of deformation events. Shatter cones and related closely-spaced fractures were formed during the contact/compression phase of the cratering process. The formation of at least some shock-induced pseudotachylitic breccia also belongs into this phase. Large-scale folds and faults and friction-generated melts can be related to the initial formation of the central uplift and extensional joints to the subsequent collapse of the central uplift.

Les séismes intermédiaires (30-300 km) ont été largement documentés dans les plaques océaniques en subduction mais leur mécanisme physique reste énigmatique. Des séismes se produisent dans les plans de Wadati-Bénioff supérieur et inférieur. Ce dernier se situe dans le manteau lithosphérique plongeant, 15 à 40 km sous l’interface de subduction, et est considéré dû à la déshydratation de l’antigorite, serpentine de haute température.Pour tester cette hypothèse et comprendre quel mécanisme est en jeu dans le plan inférieur, des expérimentations (Griggs et D-DIA) et une étude de terrain (Balmuccia, Italie) ont été effectuées.Des péridotites artificielles ont été déshydratées pendant leur déformation dans des conditions typiques du manteau supérieur (1 à 3.5 GPa). Des émissions acoustiques sont enregistrées dans des échantillons comportant 5 %vol. d’antigorite. Les microfailles associées sont scellées par des pseudotachylites contenant des fluides, montrant que la déstabilisation de l’antigorite a déclenché une rupture dynamique et la fusion de l’olivine sur la surface de faille. Ces résultats mènent à l’établissement d’un model dans lequel un transfert de contrainte induit par déshydratation, et non par surpression de fluides, est le déclencheur de la fragilisation des roches du manteau.Parallèlement, une pseudotachylite de la péridotite de Balmuccia révèle l’enregistrement de l’histoire du glissement d’un séisme fossile de magnitude Mw > 6. La lubrification co-sismique est complète et transitoire, car le magma peut rapidement s’écouler dans les fentes en tension lors du passage du front de rupture, peut-être plus vite qu’il n’est produit. L’aspiration du magma mènerait à un refroidissement permettant le rétablissement de la résistance et l’arrêt du glissement.Cette pseudotachylite naturelle, un million de fois plus grande que son analogue expérimental, s’est formée dans les mêmes conditions de pression et de température. La grande similitude entre ces failles sur le terrain et au laboratoire indique un mécanisme similaire, et donc que les expériences montrent un mécanisme de rupture représentatif de ce qui se passe dans la nature. D’autre part, de l’H2O, trouvée fossilisée dans la pseudotachylite, était présente pendant la rupture sismique.Ce travail réconcilie des décennies d’études semblant contradictoires sur le lien entre séismes mantelliques et déshydratation de l’antigorite. À une certain échelle, une fraction d’antigorite de seulement 5 %vol. suffit à déclencher une sismicité, qui pourrait finalement être vue comme un indicateur du degré d’hydratation dans le manteau lithosphérique. / Intermediate-depth earthquakes (30-300 km) have been extensively documented within subducting oceanic slabs but their physical mechanisms remain enigmatic. Earthquakes occur both in the upper and lower Wadati-Benioff planes. The latter is located in the mantle of the subducted oceanic lithosphere, 15-40 km below the plate interface, and is thought to be due to the dehydration of antigorite, the high-temperature serpentine.To test this hypothesis and understand which mechanism is at play in the lower plane, both experiments (Griggs and D-DIA) and field work (Balmuccia, Italy) have been performed.Artificial peridotites were dehydrated during deformation at upper mantle conditions. Between 1 and 3.5 GPa, acoustic emissions are recorded in samples with only 5 vol.% antigorite. Associated microfaults are sealed by fluid-bearing pseudotachylytes, showing that antigorite destabilization triggered dynamic shear failure and frictional melting of olivine. These results lead to a model in which dehydration-induced stress transfer, rather than fluid overpressure, is the trigger of mantle rocks embrittlement.Simultaneously, a pseudotachylyte from the Balmuccia peridotite reveals the recorded sliding history of an ancient Mw > 6 earthquake. The co-seismic fault lubrication is complete and transient, as the melt could rapidly flow into tensile fractures generated by the rupture front pass through, possibly faster than it is produced. Melt suction within the fractures led to rapid cooling and may have promoted strength recovery and sliding arrest.This natural pseudotachylyte, one million times larger than the experimental ones, has formed at the same pressure and temperature. The high similarity between those experimental and natural faults indicates a similar mechanism at both scales, and thus that the experiments show a rupture mechanism representative of what happens in nature. Furthermore, H2O, found fossilized in the pseudotachylyte, was somehow present during the seismic rupture.This work reconciles decades of apparently contradictory studies on the possible link between mantle earthquakes and serpentine dehydration. At a certain scale, an antigorite fraction as low as 5 vol.% is sufficient to trigger seismicity, which could therefore ultimately be seen as an indicator for the degree of hydration in the lithospheric mantle.