The fundamental goal of Earth Science is to reconstruct geologic history in order to understand how past events have shaped the evolution of planet Earth. Rock deposits from the Earths ancient oceans hold clues to the past events. The Earth's history periodically records evidence that the ocean's abruptly become depleted in oxygen (anoxic) for short periods of time (0.5 to 1 million years). It is believed oceanic anoxic events are linked to lapses in key oceanic current circulations, climate perturbations, and/or intense magmatic activity that have resulted in the deposition of black carbon-rich rocks. The causes of these global oceanic anoxic events remain a hotly debated topic and resolution of the problem will have profound implications for our understanding of why oceanic anoxic events are common just to the Cretaceous. One specific oceanic anoxic event we propose to study occurred 93.5 million years ago at the boundary of the Cenomanian-Turonian geologic time stages. Geoscientists who study ancient oceans rely on chemical signals in oceanic sediment to reconstruct oceanic processes and environments, for example isotopes of carbon, oxygen and strontium. Isotopes of osmium determined from oceanic sediments can also be used, whereby they monitor changes in the input and flux of continental weathering and magmatic activity in to the oceans. Our current understanding suggests that the Earth 93.5 million years ago had a similar plate tectonic structure to today's Earth for which we have well established ocean current circulation paths and a single osmium isotope composition that suggests that osmium is predominantly derived from detritus from the Earth's crust. However, just prior to the Cenomanian-Turonian boundary oceanic anoxic event a recent study suggests that osmium in the oceans is sourced from intense magmatic activity, which may have triggered a global oceanic anoxic event. However, osmium isotope values from two sections we have studied compared to this recent study show that these time-correlative values are disparate, suggesting that the osmium isotope budget in the ocean was drastically different. Thus very unlike today's ocean. As a result this raises an important question: can osmium isotope values from one location, which researchers have relied on thus far, be used to interpret global processes? Detailed osmium isotope analysis across two global sections will aid in defining a comprehensive understanding of the osmium isotope homogeneity in the ocean during the Cenomanian-Turonian boundary oceanic anoxic event. This will permit an improved understanding of individual oceanic areas (e.g., inputs into the ocean and weathering rates) and global ocean currents. As a direct result the research will test the hypothesis of whether magamtic activity triggered the cause of global oceanic anoxia. Especially significant is that the outcome of the proposed study will benefit researchers who are investigating the ocean to understand the Earths evolution.

Methane is a strong greenhouse gas produced in sediments directly from microorganisms, or from prolonged heating of fossil carbon. Release of methane to the ocean and atmosphere is known to have shifted global climate and nutrient cycles in the past, in particular when released in massive quantities and over short time periods from frozen subsurface reservoirs known as marine gas hydrates. With global warming, unstable sources such as gas hydrates pose a potentially large threat to the marine environments, climate, and thus society. It is therefore pivotal to recognize mechanisms of increased methane emission in the past. Equally important is understanding its sources and pathways from the subsurface, identification of processes which cycle released methane, and quantification of subsequent ocean-atmosphere and biogeochemical feedbacks. This research project sets out an ambitious organic geochemical programme to explore the process of aerobic microbial oxidation of methane in oceans over the last 1 million years of the late Quaternary and assess its effects on climate. The target process has received little attention to date but possibly plays a more important role in carbon cycling and oxygen availability in the ocean than commonly considered. As a new tool in past climate research we propose to use molecular compounds (specific biohopanoids, BHPs) generated exclusively by bacteria which feed on methane (methane-oxidising bacteria). We recently analysed BHPs in the sedimentary record (down to 100 m depth equivalent to about 1 million years) from a giant deep sea sediment fan in front of the Congo river in tropical Africa. These new data push direct evidence for the process of massive methane release and its aerobic microbial oxidation far back into the geological past; previous studies suggesting a similar mechanism were limited to the last 45 ka. The new Congo fan data are encouraging and provide strong support for previously unrecognised methane emission events and aerobic turnover in the eastern tropical Atlantic. Building on that, we will focus our research on two contrasting sediment records, (1) the Congo and Amazon Fans which had oxygenated water conditions throughout the study period and (2) the Mediterranean. The latter is well known for its pronounced and rapid changes in oxygenation throughout the past few million years which alternated between oxygenated and oxygen-fee (anoxic), and frequently even toxic (sulfidic/euxinic) conditions leading to the formation of sediments exceptionally rich in organic carbon, commonly termed sapropels. We here propose a multidisciplinary approach, utilising BHP markers with other geochemical and isotopic evidence and modelling. This integrated approach will allow us to fully evaluate the existence of previously unrecognised methane emission events and subsequent methane oxidation in the Quaternary ocean. Combined with coupled atmosphere-land-ocean modelling the data will allow us to address the relevance of the target process in oxic settings and others which are approaching oxygen-free conditions. This will then provide the first quantitative estimates on greenhouse gas volumes emitted by the emission events and explore climate effects.

The crust that underlies the world's oceans forms as a result of seafloor spreading - a process that sees the rigid oceanic plates pulled apart at fast (>100 mm/yr), intermediate (100-55 mm/yr) or slow (55-20 mm/yr) rates. As plates separate the mantle beneath rises to fill the gap and as it does so it melts due to the lower pressure. This molten rock, or magma, solidifies to form the ~6-8 km thick oceanic crust, comprising a layer of erupted and rapidly cooled magma (basalt) at the top and a layer of slowly cooled magma (gabbro) beneath. Over the last decade, observations have shown that the crust created where oceanic plates are pulled apart at slower rates, does not form by such a simple process of symmetrical, magmatic construction as our current models predict, but instead the magmatic construction is interspersed with periods of apparent magma-starvation. During these amagmatic phases plate separation is accommodated by large-offset faults along which rocks from the lower crust and the upper mantle beneath are brought to the surface. These regions of exhumed lower crust and upper mantle rocks are called oceanic core complexes (OCCs). About 25% of the Earth's mid-ocean ridges spread at very slow rates of less than 20 mm/yr. However, most of these ultraslow ridges are located in remote areas that have poor weather or ice cover that impedes their investigation. Consequently, how the crust forms and ages at these slowest spreading centres, which current models predict should be predominantly magma-starved and cold, remains poorly understood. Recent seabed imaging and sampling studies of the ultraslow Mid-Cayman Spreading Centre (MCSC) in the Caribbean, have observed the deepest and hottest black smoker hydrothermal vents on Earth, and regions of exhumed lower crust and upper mantle juxtaposed against volcanically erupted rocks of the "normal" upper oceanic crust. Here we will establish the crustal context of these contrasting observations that challenge the predictions of traditional models, and we will determine the time and space interplay between magmatic construction and amagmatic extension and the controls on, and relationship between, faulting and hydrothermal activity. As part of a British, German and American partnership, we will use sub-seabed seismic imaging to study the structure and lithology of the crust at the Mt Dent OCC on the MCSC and determine the relationship between this and the adjacent volcanic domain that also hosts hydrothermal vents. We will also investigate how the crust changes as it cools and ages as it spreads away from the ridge axis. Using the pattern of local earthquakes we will map sub-seabed fault geometries and whether or not these faults are connected at depth. As the southern tip of the MCSC also abuts against the continental crust of the Caribbean plate across the Swan Island Transform Zone, this also provides a unique opportunity to determine not only how the mantle rises up and melts beneath the ridge and how this melt is distributed along-ridge, but also if this process is impeded by the cooling affect of adjacent thick, cold continental lithosphere. To achieve our goals we will deploy ocean-bottom seismographs (OBSs) onto the seabed to determine the variation in velocity associated with, and the interfaces between the different rock types deep into the crust and upper mantle using man-made seismic signals. We will also use the OBSs to record the signals that occur naturally when faults move. We will measure the gravity field to determine crustal density as a test of our seismic models, and to image deeper into the mantle to depths beyond which our seismic signals will penetrate. Finally, we will measure reversals in the magnetic field to reveal seafloor spreading rate and crustal age and, jointly with the seismic data, determine how frequently phases of amagmatic extension have occurred from the current time to at least 20 million years ago.

Submarine landslides are a common geological feature in the deep ocean. They occur in a range of environments, from steep volcanic island flanks to areas of gently sloping sediment-covered open slope. Some submarine landslides disintegrate during their passage downslope and transform into sediment-gravity flows that can transport huge volumes of sediment for hundreds of kilometres over relatively flat ocean floor (<1o). Submarine landslides and sediment-gravity flows are the dominant process for global sediment transport from the continental shelf to the deep ocean, and are a major threat to an increasing worldwide network of seafloor infrastructure, e.g. oil/gas pipelines and telecommunications cables. Submarine landslides can also generate catastrophic tsunamis. For example, the giant Storegga Slide offshore Norway about 8200 years ago produced a tsunami that devastated coastal communities from Norway to Scotland. In addition, the deposits of sand-rich flows form many of the World's largest oil and gas reservoirs, while mud-rich flows may sequester globally significant volumes of organic carbon in the deep ocean. Improving our understanding of landslide and sediment-gravity flow hazards requires field data from past events; these data provide insights into important parameters such as volume and recurrence interval. These data also help us to model landslide-generated tsunamis and assess the associated risks. In this study we propose to generate the first ever field dataset tracing a large-scale submarine landslide and its associated sediment-gravity flow from source-to-sink. We will focus on the Moroccan Turbidite System offshore NW Africa, where the World's largest sediment-gravity flows were able to transport >100 km3 of sediment across distances up to 2000 km. The volume, source area and timing of several geologically recent (last 200,000 years) flows has been identified, using a dataset of >200 shallow sediment cores collected from across the entire depositional area over the last 30 years. Previous work has shown that most of these flows originated from (as yet unmapped) landslides in and around upper Agadir Canyon, which is one of the largest canyons in the World at 450 km long, up to 30 km wide and 1250 m deep. Most of upper Agadir Canyon above 4000 m water depth is unexplored, so we plan to map and sample landslides in this area using geophysical tools and sediment corers. The new results will allow us to undertake a novel 'mass balance' analysis, where we can quantify 1) the volume of material evacuated during the initial landslide, 2) the rate and extent of disintegration of failed material, 3) the volume of material removed by the resulting flow, and 4) the volume of eroded seafloor sediment incorporated in the flow. This unique quantitative field dataset will allow us to tackle three important science questions: 1) How quickly do large submarine landslides disintegrate into long run-out sediment flows, and how is this process influenced by shape of the slope? 2) How efficiently do landslides remove failed material, i.e. what proportion of landslide debris is deposited on the slope and how much transforms into a flow that is transported distally? 3) How much sediment is incorporated into the flow through seafloor erosion, and where does most of this erosion take place? The results will be vital for ongoing landslide-tsunami and sediment gravity-flow modeling being undertaken by NOC and others in the NERC community, and will improve assessment of associated global geohazards.

Although global climate is expected to warm over the next century in response to increasing levels of greenhouse gases, regional changes over the next decade or so are likely to be dominated by unforced natural variability of the climate system. Some of this natural variability is potentially predictable months or even years in advance because it is related to relatively slow processes, especially those in the ocean (El Niño, fluctuations in the thermohaline circulation, and large-scale anomalies of ocean heat content). Another potential source of long-term variability comes from the well-known 11-year solar variations in the Sun's output but the mechanisms for how the signal reaches the surface is not well understood and, because of this, it is not well represented in climate models used for decadal predictions. The percentage variation in total solar irradiance over an 11-yr cycle is very small, but the variation in the UV is much larger and this can impact stratospheric temperatures and ozone production. Various mechanisms have been proposed to explain how this stratospheric solar signal may extend its influence to the surface, including amplifying mechanisms through atmospheric circulation changes. Analysis of observations show that the surface response to solar variability is regional. There has been controversy surrounding the observed signal over Europe, but recent analysis of long-term observational records over Europe have confirmed a strong statistically significant signal at lags of 3-4 years. Climate models, including the Hadley Centre HadGEM model, are able to capture an upper stratospheric response to changing UV, but do not reproduce the observed signal in the lower stratosphere nor the observed lagged signal over Europe. Sensitivity tests with the HadGEM model with an exceptionally large UV change was able to reproduce the lagged nature of the signal, thus showing some promise for its ability to reproduce the signal, but the surface response amplitude is still much too weak, suggesting that there is room for significant improvements which should lead to improved decadal forecasts. This prime aim of the proposal is to improve the representation of mechanisms of solar influence on the Earth's surface in the HadGEM climate model so that forecasts using the Met Office DePreSys operational decadal forecast model can be improved. The project will employ both the HadGEM model and a simpler model for extensive testing of mechanisms, with a particular focus on improving the representation of those mechanisms that transfer the solar signal to the surface via stratospheric heating anomalies and a surface amplifying mechanism that involves atmosphere-ocean coupling in the North Atlantic. The resulting improvements to the HadGEM model will be tested by comparing results from re-forecasts (hindcasts) of selected years, with particular attention to improvements in the Atlantic / European sector. The project will be performed by researchers at Oxford University who will carry out the HadGEM investigations and Imperial College who will perform the mechanistic model investigations. Extensive support will be provided by Project Partners at the Met Office, who will be closely involved in the interpretation of the HadGEM experiments and implementation / testing of improvements in the DePreSys forecast system and a Project Partner at Kiel University who will advise on solar spectral forcing and contribute to interpretation of results in the context of other major climate models.