Voluminous magmatism during the South Atlantic opening has been considered as a classical example for plume related continental breakup. We present a study of the crustal structure around Walvis Ridge, near the intersection with the African margin. Two wide-angle seismic profiles were acquired. One is oriented NNW–SSE, following the continent–ocean transition and crossing Walvis Ridge. A second amphibious profile runs NW–SE from the Angola Basin into continental Namibia. At the continent–ocean boundary (COB) the mafic crust beneath Walvis Ridge is up to 33km thick, with a pronounced high-velocity lower crustal body. Towards the south there is a smooth transition to 20–25km thick crust underlying the COB in the Walvis Basin, with a similar velocity structure, indicating a gabbroic lower crust with associated cumulates at the base. The northern boundary of Walvis Ridge towards the Angola Basin shows a sudden change to oceanic crust only 4–6km thick, coincident with the projection of the Florianopolis Fracture Zone, one of the most prominent tectonic features of the South Atlantic ocean basin. In the amphibious profile the COB is defined by a sharp transition from oceanic to rifted continental crust, with a magmatic overprint landward of the intersection of Walvis Ridge with the Namibian margin. The continental crust beneath the Congo Craton is 40km thick, shoaling to 35km further SE. The velocity models show that massive high-velocity gabbroic intrusives are restricted to a narrow zone directly underneath Walvis Ridge and the COB in the south. This distribution of rift-related magmatism is not easily reconciled with models of continental breakup following the establishment of a large, axially symmetric plume in the Earth's mantle. Rift-related lithospheric stretching and associated transform faulting play an overriding role in locating magmatism, dividing the margin in a magma-dominated southern and an essentially amagmatic northern segment.

The Namibian continental margin marks the starting point of the Tristan da Cunha hotspot trail, the Walvis Ridge. This section of the volcanic southwestern African margin is therefore ideal to study the interaction of hotspot volcanism and rifting, which occurred in the late Jurassic/early Cretaceous. Offshore magnetotelluric data image electromagnetically the landfall of Walvis Ridge. Two large-scale high resistivity anomalies in the 3-D resistivity model indicate old magmatic intrusions related to hot-spot volcanism and rifting. The large-scale resistivity anomalies correlate with seismically identified lower crustal high velocity anomalies attributed to magmatic underplating along 2-D offshore seismic profiles. One of the high resistivity anomalies (above 500Ωm) has three arms of approximately 100km width and 300km to 400km length at 120° angles in the lower crust. One of the arms stretches underneath Walvis Ridge. The shape is suggestive of crustal extension due to local uplift. It might indicate the location where the hot-spot impinged on the crust prior to rifting. A second, smaller anomaly of 50km width underneath the continent ocean boundary may be attributed to magma ascent during rifting. We attribute a low resistivity anomaly east of the continent ocean boundary and south of Walvis Ridge to the presence of a rift basin that formed prior to the rifting.

Geomar, Helmholtz Centre for Ocean Research, Wischhofstr. 1-3, 24148 Kiel, Germany 6 now at University of Leicester, University Road, Leicester LE1 7RH, United Kingdom 7 now at GEUS, Øster Voldgade 1, 1350 Copenhagen, Denmark 8 Institute of Geophysics, Polish Academy of Science, ul. Księcia Janusza 64, 01452 Warsaw, Poland 9 University of Barcelona, C. Marti I Franques, 08028 Barcelona, Spain 10 now at Forschungsanstalt für Wasserschall und Geophysik, WTD71, Klausdorfer Weg 2, D-24148 11 Kiel, Germany 12 Earthquake Research Institute, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-0032, 13 Japan 14 GFZ German Research Centre for Geosciences, Telegrafenberg 14473, Potsdam, Germany 15

Here we provide a short description of the methods used and present error statistics for the velocity models (Table DR1–Table DR4), data examples of recorded seismic sections (Fig. DR1), picked and traced travel times as well as the corresponding ray paths (Fig. DR2–Fig. DR5). Resolution (Fig. DR6, Fig. DR7) and DWS (derivated weight sum, Fig. DR10–Fig. DR11) plots demonstrate the model uncertainty. The different types of plots are due to different modelling approaches described in the methods section.

Upwelling hot mantle plumes are thought to disintegrate continental lithosphere and are considered to be drivers of active continental breakup. The formation of the Walvis Ridge during the opening of the South Atlantic is related to a putative plume-induced breakup. We investigated the crustal structure of the Walvis Ridge (southeast Atlantic Ocean) at its intersection with the continental margin and searched for anomalies related to the possible plume head. The overall structure we identify suggests that no broad plume head existed during opening of the South Atlantic and anomalous mantle melting occurred only locally. We therefore question the importance of a plume head as a driver of continental breakup and further speculate that the hotspot was present before the rifting, leaving a track of kimberlites in the African craton.

Fluid flow through marine sediments drives a wide range of processes, from gas hydrate formation and dissociation, to seafloor methane seepage including the development of chemosynthetic ecosystems, and ocean acidification. Here, we present new seismic data that reveal the 3D nature of focused fluid flow beneath two mound structures on the seafloor offshore Costa Rica. These mounds have formed as a result of ongoing seepage of methane-rich fluids. We show the spatial impact of advective heat flow on gas hydrate stability due to the channelled ascent of warm fluids towards the seafloor. The base of gas hydrate stability (BGHS) imaged in the seismic data constrains peak heat flow values to ∼60 mWm−2 and ∼70 mWm−2 beneath two separate seep sites known as Mound 11 and Mound 12, respectively. The initiation of pronounced fluid flow towards these structures was likely controlled by fault networks that acted as efficient pathways for warm fluids ascending from depth. Through the gas hydrate stability zone, fluid flow has been focused through vertical conduits that we suggest developed as migrating fluids generated their own secondary permeability by fracturing strata as they forced their way upwards towards the seafloor. We show that Mound 11 and Mound 12 (about 1 km apart on the seafloor) are sustained by independent fluid flow systems through the hydrate system, and that fluid flow rates across the BGHS are probably similar beneath both mounds. 2D seismic data suggest that these two flow systems might merge at approximately 1 km depth, i.e. much deeper than the BGHS. This study provides a new level of detail and understanding of how channelled, anomalously-high fluid flow towards the seafloor influences gas hydrate stability. Thus, gas hydrate systems have good potential for quantifying the upward flow of subduction system fluids to seafloor seep sites, since the fluids have to interact with and leave their mark on the hydrate system before reaching the seafloor.

The opening of the South Atlantic ocean basin was accompanied by voluminous magmatism on the conjugate continental margins of Africa and South America, including formation of the Parana and Entendeka large igneous provinces (LIP), the build-up of up to 100 km wide volcanic wedges characterized by seaward dipping reflector sequences (SDR), as well as the formation of paired hotspot tracks on the rifted African and South American plates, the Walvis Ridge and the Rio Grande Rise. The area is considered as type example for hotspot or plume-related continental break-up. However, SDR, and LIP features on land are concentrated south of the hotspot tracks. The segmentation of the margins offers a prime opportunity to study the magmatic signal in space and time, and investigate the interrelation with rift-related deformation. A globally significant question we address here is whether magmatism is the drives continental break-up, or whether even rifting accompanied by abundant magmatism is in response to crustal and lithospheric stretching governed by large scale plate kinematics. In 2010/11, an amphibious set of wide-angle seismic data was acquired around the landfall of Walvis Ridge at the Namibian passive continental margin. The experiments were designed to provide crustal velocity information and to investigate the structure of the upper mantle. In particular, we aimed at identifying deep fault zones and variations in Moho depth, constrain the velocity signature of SDR sequences, as well as the extent of magmatic addition to the lower crust near the continent-ocean transition. Sediment cover down to the igneous basement was additionally constrained by reflection seismic data. Here, we present tomographic analysis of the seismic data of one long NNW oriented profile parallel to the continental margin across Walvis Ridge, and a second amphibious profile from the Angola Basin across Walvis Ridge and into the continental interior, crossing the area of the Etendeka Plateau basalts. The most striking feature is the sharp transition in crustal structure and thickness across the northern boundary of Walvis Ridge. Thin oceanic crust (5-7 km) of the Angola Basin lies next to the 35 km thick igneous crustal root founding the highest elevated northern portions of Walvis Ridge. Both structures are separated by a very large transform fault zone. The velocity structure of Walvis Ridge lower crust is indicative of gabbro, and, in the lowest parts, of cumulate sequences. On the southern side of Walvis Ridge there is a smooth gradation into the adjacent 25-30 km thick crust underlying the ocean-continent boundary, with a velocity structure resembling that of Walvis Ridge The second profile shows a sharp transition from oceanic to rifted continental crust. The transition zone may be underlain by hydrated uppermost mantle. Below the Etendaka Plateau, an extensive high-velocity body, likely representing gabbros and their cumulates at the base of the crust, indicates magmatic underplating. We summarize by stating that rift-related lithospheric stretching and associated transform faulting play an overriding role in locating magmatism, dividing the margin in a magmatic-dominated segment to the south, and an amagmatic segment north of Walvis Ridge.

The Jan Mayen microcontinent was as a result of two major North Atlantic evolutionary cornerstones—the separation of Greenland from Norway (~54 Ma), accompanied by voluminous volcanic activity, and the jump of spreading from the Aegir to the Kolbeinsey ridge (~33 Ma), which resulted in the separation of the microcontinent itself from Eastern Greenland (~24 Ma). The resulting eastern and western sides of the Jan Mayen microcontinent are respectively volcanic and non-volcanic rifted margins. Until now the northern boundary of the microcontinent was not precisely known. In order to locate this boundary, two combined refraction and reflection seismic profiles were acquired in 2006: one trending S–N and consisting of two separate segments south and north of the island of Jan Mayen respectively, and the second one trending SW–NE east of the island. Crustal P-wave velocity models were derived and constrained using gravity data collected during the same expedition. North of the West Jan Mayen Fracture Zone (WJMFZ) the models show oceanic crust that thickens from west to east. This thickening is explained by an increase in volcanic activity expressed as a bathymetric high and most likely related to the proximity of the Mohn ridge. East of the island and south of the WJMFZ, oceanic Layers 2 and 3 have normal seismic velocities but above normal average crustal thickness (~11 km). The similarity of the crustal thickness and seismic velocities to those observed on the conjugate Møre margin confirm the volcanic origin of the eastern side of the microcontinent. Thick continental crust is observed in the southern parts of both profiles. The northern boundary of the microcontinent is a continuation of the northern lineament of the East Jan Mayen Fracture Zone. It is thus located farther north than previously assumed. The crust in the middle parts of both models, around Jan Mayen island, is more enigmatic as the data suggest two possible interpretations—Icelandic type of oceanic crust or thinned and heavily intruded continental crust. We prefer the first interpretation but the latter cannot be completely ruled out. We infer that the volcanism on Jan Mayen is related to the Icelandic plume.

The evolution of the South Atlantic in space and time is in general presented as a two plate system, where the separation of the two continent might be caused by the Tristan Hotspot. Evidence for massive subaerial and submarine volcanism is found on both margins and adjacebt basins. Namely, the Parana (Brasil) and Etendeka (Namibia) flood basalts onshore, and the Walvis Ridge and Rio Grande offshore are evidences for a long term volcanism related to the rift/drift of both continents. In science it is currently under debate, if hotspot volcanism is the driving force for the continental breakup or are e.g. the above mentioned features a by-product of the anyhow moving plates. Aeromagnetic investigations along East Antarctica show that from the beginning of the Gondwana breakup the South American and Africa plates moved in different directions and with different speeds. This most likely caused rifting along old zone of weakness between South America and southern Africa. The ridge in combination with the large igneous provinces (Etendeka and Parana) in South America and Namibia is today considered to be a classical model for hotspot driven continental break-up. For understanding the role of mantle plumes during continental breakup a large-scale geophysical experiment in 2011 on/offshore Namibia was conducted to investigate the crustal/upper mantle structure both under the Walvis Ridge and northern Namibia. Several seismic profiles in the strike and across the Walvis Ridge with in total more than 160 oceanbottom seismometers, magnetic telluric station were deployed. Onshore the number of instruments was more than 300 to map in detail the continent-ocean transition zone to enhance our understanding on how the crust was modified by the thermal anomaly. In addition, a larger number of seismological stations are operated both on- and offshore to investigate the structure of the upper mantle. Here, we reported the first results of this experiment and discuss some implications.

Expeditions 17/1 and 17/2 of R/V Maria S. Merian, carried out geophysical surveys between November 2010 and February 2011 in the area around Walvis Ridge, Southeast Atlantic Ocean. Among the data collected, reflection seismic and high-resolution bathymetric data offer some important preliminary insights into the tectonic evolution of the ridge and the adjoining lower continental slopes and ocean basins. The NE-SW trending Walvis Ridge has a trapezoid shape and is likely built up by thick sequences of plateau basalts, with top of basement inclined to the south. Sediments are almost absent on the NW side of the ridge, preserving a fascinating mountainscape formed early in the tectonic history. Sediment cover grows southward to thicknesses up to 2 s TWT, above a basement characterized by dipping reflector sequences inclined NW. There is progressive northward onlap of sedimentary strata, indicating that in the earlier part of its history parts of Walvis Ridge may have formed a land surface. This is further supported by denudational features, like steep cliffs up to 150 m high, and deeply incised valleys, defining paleo-drainages. Isolated, flat-topped guyots seaward of the ocean-continent boundary attest to a later history of wave abrasion and progressive subsidence of Walvis Ridge. The boundary zone towards the Angola Basin to the north is an important extensional or transtensional tectonic feature, conincident with the Florianopolis Fracture Zone on the Brazilian side of the Atlantic. The fault zone has recorded two important episodes of vertical differential movement, documented by thick half-graben fills and clastic sediment aprons on the Angola Basin crust. The early one is likely connected to continental break-up. The later one probably occurred in the Middle Tertiary, also affecting the SE side of Walvis Ridge. We conclude by stating that the record of sedimentary and geomorphologic processes around Walvis Ridge can contribute important information to decipher the tectonic history of the southeast Atlantic.

Oceanic island arcs are sites of high magma production and contribute to the formation of continental crust. Geophysical studies may provide information on the configuration and composition of island arc crust, however, to date only few seismic profiles exist across active island arcs, limiting our knowledge on the deep structure and processes related to the production of arc crust. We acquired active-source wide-angle seismic data crossing the central Lesser Antilles island arc north of Dominica where the oceanic Tiburon Ridge subducts obliquely beneath the forearc. A combined analysis of wide-angle seismics and pre-stack depth migrated reflection data images the complex structure of the backstop and its segmentation into two individual ridges, suggesting an intricate relation between subducted basement relief and forearc deformation. Tomographic imaging reveals three distinct layers composing the island arc crust. A three kilometer thick upper crust of volcanogenic sedimentary rocks and volcaniclastics is underlain by intermediate to felsic middle crust and plutonic lower crust. The island arc crust may comprise inherited elements of oceanic plateau material contributing to the observed crustal thickness. A high density ultramafic cumulates layer is not detected, which is an important observation for models of continental crust formation. The upper plate Moho is found at a depth of 24km below the sea floor. Upper mantle velocities are close to the global average. Our study provides important information on the composition of the island arc crust and its deep structure, ranging from intermediate to felsic and mafic conditions.

The region offshore Eastern Java represents one of the few places where the early stage of oceanic plateau subduction is occurring. We study the little investigated Roo Rise oceanic plateau on the Indian plate, subducting beneath Eurasia. The presence of the abnormal bathymetric features entering the trench has a strong effect on the evolution of the subduction system, and causes additional challenges on the assessment of geohazard risks. We present integrated results of a refraction/wide-angle reflection tomography, gravity modelling, and multichannel reflection seismic imaging using data acquired in 2006 south of Java near 113°E. The composite structural model reveals the previously unresolved deep geometry of the oceanic plateau and the subduction zone. The oceanic plateau crust is on average 15 km thick and covers an area of about 100 000 km2. Within our profile the Roo Rise crustal thickness ranges between 18 and 12 km. The upper oceanic crust shows high degree of fracturing, suggesting heavy faulting. The forearc crust has an average thickness of 14 km, with a sharp increase to 33 km towards Java, as revealed by gravity modelling. The complex geometry of the backstop suggests two possible models for the structural formation within this segment of the margin: either accumulation of the Roo Rise crustal fragments above the backstop or alternatively uplift of the backstop caused by basal accumulation of crustal fragments. The subducting plateau is affecting the stress field within the accretionary complex and the backstop edge, which favours the initiation of large, potentially tsunamogenic earthquakes such as the 1994 Mw = 7.8 tsunamogenic event.