The present-day seabed in the eastern English Channel is an erosional landscape dissected by a complicated network of palaeovalleys. The largest of these palaeovalleys has been interpreted as the product of catastrophic flooding though the Straits of Dover during the Mid Quaternary. Whilst the eastern English Channel is a valuable source of aggregates, little attention has been paid to the sedimentary record preserved on the continental shelf in terms of its ability to document landscape change throughout the Quaternary. This thesis aims to establish the first stratigraphic model of deposits preserved on the continental shelf in the eastern English Channel and chronometrically constrain the timing of deposition using Optical Stimulated Luminescence (OSL) dating. The model will be used as a framework to reconstruct landscape change and address questions regarding the preservation potential of sediments, and the imprint sedimentary processes have on the landscape, in continental shelf settings over glacial-interglacial cycles. The stratigraphic model was constructed through the integration of high resolution mutlibeam bathymetry, shallow sub-surface 2D seismic profiles, lithological information from vibrocores, and chronometric data obtained through OSL. A variety of drowned landscapes including terrestrial (fluvial and colluvial), coastal and shallow marine, were identified. These landscapes document palaeoenvironmental change on the continental shelf from MIS 6 to MIS 1. The fluvial landscape in the English Channel is dominated by multiple phases of lateral and vertical erosion, primarily in response to changes in sea level, but also as a result of reorganisation of drainage basins and variable discharges due to fluctuating ice margins. Exposure of the continental shelf during cold periods is documented in the form of remnant periglacial deposits and extensive palaeosols. The most volumetrically significant sediments preserved on the continental shelf were deposited in shallow marine and coastal settings. These sediments are typically restricted to palaeovalleys where accommodation created during relative sea-level rise enabled deposition. Elsewhere, sediments are preserved as relict coastal landforms, in particular, as part of an exceptionally rare drowned barrier complex at Hastings Bank. Over multiple sea-level cycles, sediments are recycled by fluvial and marine processes, with the most recent phase of deposition having the greatest preservation potential. Erosional processes have the greatest persistence in the landscape record. However, they create a composite record and distinguishing between different events without a correlative sediment package is problematic. The results presented in this thesis highlight the timing and nature of ‘normal’ sedimentary regimes in a continental shelf setting over multiple glacial-interglacial cycles. Further, they reveal evidence for erosion and deposition by fluvial processes in the Northern Palaeovalley during the last glacial period, thus contradicting an existing hypothesis that states the palaeovalley formed through catastrophic flooding.

This thesis presents one of the most comprehensive studies on submarine canyons yet. It integrates data on the geology, geochemistry, sedimentology and oceanography of the Nazaré, Setúbal and Cascais canyons, west Iberian margin, in order to constrain the processes and controls of past and present sedimentation in this area. The results indicate that, during the glacial stages, turbidity currents are the dominant process of sediment transport, erosion and deposition in these canyons. Turbidity currents are mostly in the form of small-volume, high-frequency events that are generated by fluvial and hydrodynamic processes, and the flows remain mainly within the upper canyon. A smaller proportion of turbidity currents are large-volume, low-frequency, seismictriggered events that flush through the entire canyons. Catastrophic mass wasting is most prominent in Setúbal Canyon due to its closer proximity to the region’s active fault zone along the southern Iberian margin. During the Holocene, sedimentary activity in Setúbal and Cascais canyons continued in the form of frequent, canyon-flushing turbidity currents that ceased abruptly ca 6.4 ka. This interval corresponds to both a regional aridification event that affected the Mediterranean and northern African regions, and to a decline in rising sea level. A combination of the two events is interpreted as being the direct cause of the cessation in sedimentary activity on the continental margin. In Nazaré Canyon, sedimentary activity during the last ~1000 years has been dominated by a mid-canyon resuspension depocentre that is controlled by small-volume, low-energy, hydrodynamic-generated turbidity currents. This depocentre provides a unique high-resolution record of recent sedimentation in a deep-sea setting, preserving a potential forest fire debris signal that is linked to a change in climate during the Little Ice Age. The conclusions from this work are that sedimentation in the west Iberian canyons is controlled by the complex interplay between several variables, the most important ones being the source and supply of sediment, the hydrodynamic conditions on the shelf and slope, and the canyon morphology. The sedimentary activity in the river-fed Setúbal and Cascais canyons is found to be affected more by regional climatic changes than by eustatic sea-level changes, and in Nazaré Canyon by the supply of sediment along the shelf.

The formation of new ocean lithosphere at mid-ocean ridges is a fundamental component of the plate tectonic cycle, and through hydrothermal interactions with seawater is a major control on the composition of the oceans, ocean crust, and upper mantle. Two complementary approaches are used to investigate the thermal implications of endmember theoretical models that describe the accretion of the lower oceanic crust at fast spreading rates. The first approach uses the record of hydrothermal alteration of the ocean crust, including Sr and O-isotopes, to investigate and quantify the role of hydrothermal circulation during the accretion of the ocean crust. The second method uses diffusion based geospeedometry techniques to determine cooling rates in the ocean crust. Samples from two locations of ocean crust formed at fast spreading rates at the East Pacific Rise are used in these investigations, ODP Hole 1256D and Hess Deep. Hole 1256D provides the first intact sampling of a complete section of upper oceanic crust formed at a fast spreading rate and recovered the first in situ sampling of the dike/-gabbro boundary. Hess Deep is a tectonic window where the westward propagation of the Cocos-Naza Ridge has rifted ocean crust formed at the EPR and exposed the lower ocean crust at the seafloor. The whole rock profile for Hole 1256D reveals Sr isotopes in the volcanic sequence to be only slightly shifted from primary MORB values (0.70284-0.703814 compared to 1256 MORB of 0.70283). In contrast, Sr isotopes in the sheeted dike complex (0.70294-0.70536) are strongly elevated towards hydrothermal fluid compositions (0.70505-0.70525). Rocks of the plutonic complex are characterised by elevated Sr ratios along igneous contacts (up to 0.70524) but only limited increases in Sr isotopes relative to MORB in the centres of the gabbro bodies (0.70290-0.70396). The complementary oxygen isotope profile records the downwards transition from low temperature to high temperature hydrothermal alteration but contains small scale variation associated with changes in secondary mineral abundances and local fluid/rock ratios. Both the detailed Sr and O isotope profiles document the importance of dike margins and other igneous contacts as focussed pathways for fluid flow through the crust. The time-integrated fluid flux required to cause the observed Sr isotope profile through the sheeted dike complex is 2.0 - 2.6 x 106 kg/m2 and is consistent with fluid fluxes calculated for other crustal locations (e.g, Hole 504B, Pito Deep, Hess Deep). The heat flux required to sustain this fluid flux is equivalent to half of the latent heat released during the crystallisation of the lower ocean crust. At Hole 1256D the removal of heat by hydrothermal fluids was effcient and demonstrates that the fluid flux in the sheeted dikes must have removed some portion of the heat flux out of the lower ocean crust. In order to remove all of the latent heat of crystallisation from the lower crust, there must be significant hydrothermal circulation in the lower ocean crust.

Rifted continental margins and extensional basins, provide lasting records of the processes that occur during continental break-up and initial spreading. The eastern Black Sea (EBS) basin provides a good setting to study the development of continental margins, because of ongoing sedimentation during its development and the close proximity of its conjugate margins. Here, I present an analysis of a new wide-angle seismic dataset that reveals the structure of the deep sediments, crust and upper mantle within the EBS basin. These data provide a unique look at the formation of extensional basins, as the dataset includes a prole that is orientated parallel to the rift axis. This prole places new constraints on the variation in magmatism that accompanied continental rifting and the lateral extent over which these variations occur. The wide-angle data show 8-9km of sediment in the centre of the basin, and reveal a wide-spread low-velocity zone (LVZ) within the deep sediments. The depth of this LVZ coincides with the organic-rich mud layer identied as the Maikop, and indicates overpressure within this formation. From the seismic velocity model, excess pore pressures of 60 - 70 MPa above hydrostatic were estimated within the Maikop. The wide-angle data also reveal highly thinned continental crust (7km thick) in the western EBS, and crust interpreted as thick oceanic crust (13km thick) in the eastern EBS, implying a transition from magma-starved to magmatically robust rifting. Lateral variations in mantle temperature and composition can account for a gradual increase in magmatism, but the wide-angle data reveal that this transition is abrupt (30km) and coincides with one of a series of basement scarps. These results impact the interpretations of other rift zones, where a variation in the syn-rift magmatism is observed, but the nature of the transition is not known.

Constraining the magmatic systems of active volcanoes is important for comprehending the mechanisms that drive magma supply. Particularly important are the geometry and characteristics of magma storage regions in the upper and middle crust, which can in uence the style and time-scales of eruptions. Seismic tomography can provide detailed resolution of the physical characteristics of complex three-dimensional structures in the subsurface, and has been successfully employed at several active volcanoes to constrain the structure of their magmatic systems. The island of Montserrat, Lesser Antilles, has been the subject of an land-sea active-source seismic experiment to constrain its upper crustal structure, with particular focus on the magma reservoir feeding the current eruption. A two-dimensional inversion of seismic travel-times from a subset of the data collected, including wide-angle refractions and reflections, constrains the upper crust to a depth of 5-6 km, along a south-east to north-west section. The two-dimensional model delineates the high-velocity cores of the volcanic edifices that make up the island, the slower surrounding volcaniclastic deposits and pelagic sediments, and the underlying upper crust. Analysis of field recordings and synthetic waveforms calculated with a viscoelastic finite-difference method indicates that the high-velocity cores correspond to a geologically heterogeneous region which causes anomalous scattering of the seismic wave field. A three-dimensional seismic velocity model from first-arrival travel-time tomography delineates the high-velocity cores in greater detail and reveals a low-velocity volume at 4 to 8 km depth beneath the active Soufriere Hills volcano, inferred to correspond to the active upper-crustal magma chamber. The magma chamber is further constrained with the help of numerical models of incremental magma chamber growth, which suggests that a magma chamber of about 18 km3 formed at 5.5 to at least 7.5 km depth by incremental intrusion of sills over a few thousand years and is therefore larger and deeper than previously inferred. The new tomographic results are integrated with previous geological, petrological and geodetic constraints to provide one of the most complete models of the magmatic system of an active volcano from the magma source in the mantle wedge to the volcanic edifice at the surface. The volume estimate and chamber location are critical parameters required for models of eruption dynamics, which in turn are key to forecasting the likelihood and characteristics of future eruptions and the mapping of hazard.

Submarine landslides at open continental slopes are the largest mass movements on Earth and can cause damaging tsunamis. To be able to predict where and when such large landslides may occur in the future requires fundamental understanding of the mechanisms that cause them. Due to the inaccessibility of these features this understanding is based on poorly tested hypotheses. Recent studies have proposed that more landslides occur during periods of sea level rise and lowstand, or during periods of rapid sedimentation. These hypotheses are tested by comparing a comprehensive global data set of ages for large submarine landslides to global mean sea level and local sedimentation rates. The data set does not show statistically significant patterns, trends or clusters in landslide abundance, which suggests that the link between sea level and landslide frequency is too weak to be detected using the available global data base. The analysis also shows no evidence for an immediate influence of rapid sedimentation on slope stability, as failures tend to occur several thousand years after periods of increased sedimentation rates. Large submarine landslides occur on remarkably low slope gradients (<2�), which makes them difficult to explain. A widely used explanation for failure of such low angle slopes is high excess pore pressure due to rapid sedimentation and/or focused pore fluid flow to the toe of the slope. If these hypotheses are universal, and therefore also hold for continental margins with comparatively low rates of sediment deposition (where numerous large landslides are observed), is tested in this thesis. Fully coupled 2D stress-fluid flow finite element models are created that simulate the excess pore pressure and drainage response of a continental slope to the deposition of new sediment. Homogeneous models with a wide range of physical-mechanical properties as well as models with an aquifer are loaded by low rates of sediment deposition. All models turn out stable and resulting excess pore pressures are too low to significantly decrease effective stress anywhere along the slope. Hence, factors other than sediment deposition must be fundamental for initiating slope failure, at least in locations with slow sedimentation rates. The results obtained in this thesis not only indicate that failure mechanisms that have previously been considered important may not be universal. They also emphasise the large uncertainties in our current understanding of the occurrence, timing and frequency of large submarine landslides at open continental slopes.

A significant proportion of the world’s organic carbon is trapped in submarine methane hydrates. When ocean bottom waters warm, these hydrates may be destabilised, releasing gaseous methane into the surrounding sediments and potentially into the overlying water column and atmosphere. Increased atmospheric methane contributes to further warming as methane is a potent greenhouse gas. Release of methane from hydrate may have accompanied some paleoclimate warming events, but observations of hydrate destabilization due to current global warming remain unconfirmed. The discovery of more than 250 seafloor methane bubble plumes close to the limit of the gas hydrate stability zone offshore Western Svalbard has recently been linked to increases in bottom water temperature in this region over the past 30 years. To assess the source and fate of this methane, this thesis presents a geochemical study of hydrate, sediments, seawater, and gas in the vicinity of the seafloor methane seepage. Analyses of the gas molecular and isotopic compositions reveals that hydrate-bound gas, free gas in shallow sediments, and gas bubbles entering the water column at seafloor seep sites all have the same source. The gas is thermogenic gas produced offshore that has migrated laterally to the continental slope and shelf region. Transport-reaction modelling of pore water chemistry shows that active anaerobic oxidation of methane in sediments is an effective barrier to release of methane into ocean bottom waters. However, small fractures and faults allow ~90% of the methane that enters near-surface sediments to bubble into the water column at localized seafloor seeps. Analyses of the methane distribution in the water column indicate that the methane in the bubbles rapidly dissolves in seawater, and is transported northwards at depth in the West Svalbard Current. As a result, there is limited vertical exchange of methane between deep and surface waters. Surface waters are nevertheless supersaturated due to isopycnal mixing with methane-rich waters from the shallow shelf onto the upper slope. Measurements of methane mixing ratios in air indicate that the sea to air methane flux offshore Western Svalbard does not make a significant contribution to the local atmospheric methane budget. Sedimentary records of the δ13C-CH4 signature of benthic foraminifera provide evidence for intermittent methane seepage at the current limit of hydrate stability (~400 m water depth) over the last ~20,000 years. Although this is likely due to changes in hydrate stability as a result of changes in bottom water temperature, we find no evidence for this in the current data set.

No description available.

Rifting is the process that leads to the formation of oceans. Rifting is the break up of continents, leading to the formation of new oceanic floor between the two continental plates. Although the concept of continental rifting is accepted within the scientific community, it is still debated what controls the volume and composition of igneous material generated at these constructive plate boundaries. Here I present the results of dynamic modelling of rifted margins. I have explored the consequences of margin and mantle structure on the melt generated during continental extension and breakup. The central aim is to understand how melting affects the rifting of continents, especially in the North Atlantic. In order to understand the enigmatic melt production observed around the North Atlantic various tools are developed for interpreting the model output. These are predictions of primary major element composition of the melt, rare-earth element composition of the melt, predictions of the crystallised mid-oceanic ridge basalt composition and the seismic velocity of the lower crust. The thickness of the lithosphere has a very large impact on the subsequent rifting style. Extension of a 125 km thick thermally and rheologically defined lithosphere that has no prior thinning produces little melt during breakup. The Southeast Greenland margin rifted above a pre-thinned lithosphere and at initial fast half spreading rates. Further- more, to generate the thickness, chemistry and seismic velocities observed off this margin, rifting was coincident with the arrival of a 50 km thick, 200 ◦C thermal anomaly. This thermal anomaly is not a plume, rather an exhaustible thermal layer that has drained along the sub-lithospheric topography from a distal plume. The melts generated are high in MgO, and depleted in TiO. They are depleted in rare-earth elements. This would lead to high seismic velocities within the underplate being, as observed off Southeast Greenland.

The evolution of extensional processes at continental rift zones provides important constraints on the underlying lithospheric deformation mechanisms, level of seismic hazard and location of likely hydrocarbon traps. The Corinth rift in central Greece is one of the few examples that has experienced a short extensional history (< 5 Myr), has a relatively well–known pre–rift structure, is experiencing pure extension, and is located in a fluctuating marine–lacustrine setting producing characteristic cyclical stratigraphy. Traditionally, the rift has been described as an asymmetric half–graben controlled by N–dipping faults on the southern margin. This view has been challenged by increasing seismic data from the off-shore part of the rift which show it is more complex, analogous to more developed rifts like the East African rift and Red Sea. High resolution and deep penetration seismic reflection data across the entire offshore rift zone are combined with onshore geomorphological data to constrain: the architecture of major rift–bounding faults; basin structure; spatial and temporal evolution of depocentres; total extension across the rift; and slip rates of major faults from stratigraphic analysis and dislocation modelling of long term deformation. Stratigraphy within the offshore Corinth rift is composed of a non reflective older unit (oldest syn–rift sediments are ca. 1–2 Ma) and a well stratified younger unit separated by a ca. 0.4 Ma unconformity. Net basement depth is greatest in the present centre of the rift zone (2.7–3 km) and decreases to the east and west (1.5–1.6 km). The 0.4 Ma unconformity surface records an important change in rift geometry. Pre. 0.4 Ma, sediment deposition occurred in 20–50 km long isolated basins, controlled by both N and S–dipping faults. Post 0.4 Ma, sediment deposition and basement subsidence has been enhanced in areas between these originally isolated basins creating a single 80 km long central depocentre. Since 0.4 Ma activity has became focused on mostly N–dipping faults. However, in the west, N tilting stratigraphy and basement indicate S–dipping faults are locally structurally dominant. Late Quaternary averaged major fault slip rates are 3–6 mm/yr on the N-dipping south margin faults, >1.8 mm/yr on S–dipping offshore faults, and 1–3 mm/yr on faults in the eastern rift. Total extension over rift history (Late Pliocene to present) has been greatest in the west (8 km), with extension distributed over many faults (most now inactive) spaced at 5 km intervals. To the east total extension is reduced (5–6 km) and is distributed over fewer faults spaced at 15–35 km intervals. There are large differences in rift character along the rift axis and throughout rift history. The highest geodetic rates over the last 10–100 years are in the western part of the rift and do not correspond to the area of greatest offshore basement depth. This suggests a recent change in the locus of strain focusing, potentially analogous to the change that occurred in rift geometry ca. 0.4 Ma.