S U M M A R YDeep penetration multichannel reflection and Ocean Bottom Seismometer wide-angle seismic data from the Congo-Angola margin were collected in 2000 during the ZaïAngo cruise. These data help constrain the deep structure of the continental margin, the geometry of the pre-salt sediment layers and the geometry of the Aptian salt layer. Dating the deposition of the salt relative to the chronology of the margin formation is an issue of fundamental importance for reconstructing the evolution of the margin and for the understanding of the crustal thinning processes. The data show that the crust thins abruptly, from a 30-40 km thickness to less than 10 km, over a lateral distance of less than 50 km. The transitional domain is a 180-km-wide basin. The pre-salt sediment layering within this basin is parallel to the base of the salt and hardly affected by tectonic deformation. In addition, the presence of a continuous salt cover, from the continental platform down to the presumed oceanic boundary, provides indications on the conditions of salt deposition that constrain the geometry of the margin at that time. These crucial observations imply shallow deposition environments during the rifting and suggest that vertical motions prevailed-compared to horizontal motions-during the formation of the basin.

SUMMARY The deep structure of the West African continental margin between 5°S and 8°S was investigated using vertical reflection and wide‐angle reflection/refraction techniques, during the ZaïAngo project, a joint programme conducted in 2000 April by Ifremer and TotalFinaElf. To penetrate below the salt layer, a non‐conventional, low‐frequency seismic source was used in the ‘single‐bubble’ mode, together with ocean bottom instruments (hydrophones and seismometers) and a 4.5 km long streamer that recorded multichannel seismic reflection (MCS). The data show that the continental crust thins abruptly over a lateral distance of less than 50 km, from 30 km thick below the continental platform (based on gravity data), to less than 4 km thick below the Lower Congo Basin that formed prior to the Aptian salt deposition. This subsalt sedimentary basin (180 km wide, 4 km thick, with velocities varying from 4.7 km s−1 to 5.8 km s−1 at the bottom) is located between the foot of the continental slope and the oceanic domain. It is underlain by crust of an intermediary or transitional type, between continental crust and what can be recognized as oceanic crust. In the transitional zone, a crustal upper layer is present below the pre‐salt sedimentary basin, 3 to 7 km thick, with velocities increasing from 5.8 km s−1 at the top to 6.8 km s−1 at the bottom of the layer. This layer appears to thin regularly, from 6–7 km thick below the depocentre of the pre‐salt sedimentary basin to 3–4 km thick below the western termination of the basin. Below this upper crustal layer, an anomalous velocity layer (7.2 to 7.8 km s−1), is documented, below the eastern side of the basin, where the crustal thinning is at a maximum. The origin of this layer is unknown. Several arguments, like rifting duration (between 15 Ma and 30 Ma) or the absence of seaward‐dipping reflectors, precludes the hypothesis of underplated mantle material, but other hypotheses (such as serpentinized material or high‐grade metamorphic crustal rocks or a mixture of mafic and ultramafic crustal rocks) are plausible. Near the ocean termination of the basin, the transitional zone is bounded to the west by a basement ridge that is clearly documented on two profiles (‘7+11’ and 14) having a dense ocean bottom seismometer/hydrophone (OBS/OBH) spacing. On these profiles, an anomalous velocity layer is present in the westernmost part of the transitional zone (below the basement ridge) and in the oceanic domain. This layer, absent on profile 3, may be related either to oceanization and slow seafloor spreading processes or to a consequence of the rifting process.

Gas hydrates and gas bubbles were collected during the MARNAUT cruise (May-June 2007) in the Sea of Marmara along the North Anatolian Fault system, Turkey. Gas hydrates were sampled in the western part of the Sea of Marmara (on the Western High), and three gas-bubble samples were recovered on the Western High, the Central High (center part of the Sea of Marmara) and in the Çinarcik Basin (eastern part of the Sea of Marmara). Methane is the major component of hydrates (66.1%), but heavier gases such as C 2 , C 3 , and i-C 4 are also present in relatively high concentration. The methane contained within gas hydrate is clearly thermogenic as evidenced by a low C 1 /C 2 + C 3 ratio of 3.3, and carbon and hydrogen isotopic data (δ 13 C CH4 of − 44.1‰ PDB and δD CH4 of − 219‰ SMOW). A similar signature is found for the associated gas bubbles (C 1 /C 2 + C 3 ratio of 24.4, δ 13 C CH4 of − 44.4‰ PDB) which have the same composition as natural gas fromK. Marmara-af field. Gas bubbles from Central High show also a thermogenic origin as evidenced by a C 1 /C 2 + C 3 ratio of 137, and carbon and hydrogen isotopic data (δ 13 C CH4 of − 44.4‰ PDB and δD CH4 of − 210‰ SMOW), whereas those from the Çinarcik Basin have a primarily microbial origin (C 1 /C 2 + C 3 ratio of 16,600, δ 13 C CH4 of − 64.1‰ PDB). UV-Raman spectroscopy reveals structure II for gas hydrates, with CH 4 trapped in the small (5 12) and large (5 12 6 4) cages, and with C 2 H 6 , C 3 H 8 and i-C 4 H 10 trapped in the large cages. Hydrate composition is in good agreement with equilibrium calculations, which confirm the genetic link between the gas hydrate and gas bubbles at Western High and the K.Marmara-af offshore gas field located north of the Western High. We calculate the characteristics of the hydrate stability zone at Western High and in the Çinarcik Basin using the CSM-GEM computer program. The base of the structure II hydrate stability field is at about 100 m depth below the seafloor at the Western High site, whereas in the Çinarcik Basin, P-T conditions at the seafloor correspond to the uppermost range for structure I hydrate formation from microbial gas.

Volcanic eruptions are foundational events that shape the Earth's surface and provide a window into deep Earth processes. How the primary asthenospheric melts form, pond and ascend through the lithosphere is, however, still poorly understood. We document an on-going magmatic event offshore Mayotte Island (North Mozambique channel), associated with large surface displacements, very low frequency earthquakes and exceptionally deep (25-50 km) seismicity swarms. We present data from the May 2019 MAYOBS1 cruise, which reveal that this event gave birth to a 820m tall, ~ 5 km 3 deepsea volcanic edifice. This is the largest active submarine eruption ever documented. The data indicate that deep magma reservoirs were rapidly drained through dykes that intruded the entire lithosphere and that pre-existing subvertical faults in the mantle were reactivated beneath an ancient caldera structure.

Understanding of the evolution of fluid-fault interactions during earthquake cycles is a challenge that acoustic gas emission studies can contribute. A survey of the Sea of Marmara using a shipborne, multibeam echo sounder, with water column records, provided an accurate spatial distribution of offshore seeps. Gas emissions are spatially controlled by a combination of factors, including fault and fracture networks in connection to the Main Marmara Fault system and inherited faults, the nature and thickness of sediments (e.g., occurrence of impermeable or gas-bearing sediments and landslides), and the connectivity between the seafloor and gas sources, particularly in relation to the Eocene Thrace Basin. The relationship between seepage and fault activity is not linear, as active faults do not necessarily conduct gas, and scarps corresponding to deactivated fault strands may continue to channel fluids. Within sedimentary basins, gas is not expelled at the seafloor unless faulting, deformation, or erosional processes affect the sediments. On topographic highs, gas flares occur along the main fault scarps but are also associated with sediment deformation. The occurrence of gas emissions appears to be correlated with the distribution of microseismicity. The relative absence of earthquake-induced ground shaking along parts of the Istanbul-Silivri and Princes Islands segments is likely the primary factor responsible for the comparative lack of gas emissions along these fault segments. The spatiotemporal distribution of gas seeps may thus provide a complementary way to constrain earthquake geohazards by focusing the study on some key fault segments, e.g., the northern part of the locked Princes Islands segment.