Global average sea level is rising by approximately 3 millimetres per year. Given the huge economic and societal impacts of this change, accurate forecasts of sea level are urgently needed to inform policymakers considering mitigation and adaptation strategies. Melting of the ice sheets of Antarctica and Greenland currently contributes about one third of sea level rise. The future of this melting is highly uncertain, and the worst-case scenario involves a substantial ice-sheet contribution to dangerous sea-level rise. The largest contribution to sea level rise from ice sheets occurs when the ocean melts the base of ice shelves (floating extensions of the grounded ice sheet). The melt rate of ice in seawater is determined by the transfer of heat and salt from the ocean towards the ice. Observations reveal a turbulent boundary layer in the ocean beneath ice shelves, where vigorous mixing is driven by the flow of rising meltwater, large-scale circulation in the ocean, and tides. Mixing of heat and salt in the boundary layer influences the ice melt rate, but the physical processes involved are poorly understood and will not be resolved in climate models for the foreseeable future. The proposed project will improve our understanding of the ice shelf/ocean boundary layer and develop improved representations of ice-shelf melting for use in climate models. To achieve these aims we will use a suite of numerical models and the latest observations. We will start with direct numerical simulations (DNS) to model a small box of ocean next to an ice shelf (~1 cubic metre) at ultra-high resolution (~1 millimetre). This will provide insight into the turbulence near the ice and its interaction with melting. We will then use large-eddy simulations (LES) to study a larger volume (~1 square kilometre in area by 100 metres height) at high resolution (~10 centimetres - 1 metre). This will resolve the largest turbulent motions in the whole boundary layer. Both models will be validated using recent observations obtained from mooring sites at the George VI and Larsen C ice shelves (Nicholls, NE/H009205/1). The model results will in turn help interpret and understand the observations. We will use these numerical models to devise and calibrate parameterisations for ice melting and vertical mixing for use in ocean climate models. We will add candidate parameterisations to a one-dimensional (vertical) model that incorporates many popular ocean mixing schemes, and test them directly against the DNS and LES results. We will begin with existing parameterisations and modify them as needed to match the high resolution models. The successful parameterisations will be implemented in the UK ocean model (NEMO) and shared with climate modelling groups (including the Met Office) to improve predictions of sea-level rise.

In 2021, the first full inventory of emperor penguin colonies in Antarctica was generated using satellite imagery. It is therefore not surprising that their population vulnerability to changing climate is not yet well known, with colony movements only having been observed at a small number of sites. As Antarctica responds to warming climate and ocean conditions, sea ice is likely to decline presenting a potentially significant risk to the viability of emperor penguin colonies because they live on sea ice and rely on its stability for breeding and feeding. The extent of sea ice fell by over 2 million km2 compared to average around Antarctica between 2016 and 2018, and reductions of future sea ice loss suggest that the majority of colonies may become quasi-extinct by 2100 under current greenhouse gas emission scenarios. However, both historic and future colony responses are poorly known. For example, the models which predict future behaviour are based upon breeding factors measured at a single site and behavioural factors measured at only 9 sites over a short time period of 13 years. Thus there is a significant need to improve our understanding of past colony changes and how they link to changing sea ice habitat conditions so that we can better predict future colony vulnerability under a changing climate. Although sea ice loss (and thus emperor penguin habitat) is controlled on a large scale by warming climate and oceans, an additionally overlooked process which may be increasingly disrupting sea ice conditions is the calving of icebergs which can push, or cause the fracturing of, sea ice, leaving an embayment sea ice free. In response to the loss of sea ice, emperor penguins may move to another region where sea ice conditions are more stable, or if no such area is available, they have more recently been observed to climb onto the glaciers themselves. This is a dramatic response, but without it the colony may cease to exist. Such observations of movement are again limited to a few local studies, and the impact of calving-induced sea ice breakout events upon emperor penguin colonies has never been measured. Our aim is to understand the past, present and future vulnerability of emperor penguin colonies to changing glacier and sea ice conditions. We will use existing archives of freely-available satellite imagery to map past colony movements, sea ice and glacier calving conditions at each of the 61 newly identified emperor penguin colonies in Antarctica. This will allow us to establish how historic sea ice conditions have changed at each colony and will also allow us to understand the impact of specific glacier calving events over the last 30-40 years. Our work will allow us to determine whether colony ability to move onto glacier ice or to migrate to new sea ice areas is a common reaction to sea ice loss, or whether this is a new phenomena. Using this information, we will gain better understanding of colony vulnerability to sea ice changes. In areas where colonies currently appear at risk, we will use very high-resolution commercial satellite imagery to establish whether they remain viable as a breeding colony. This understanding will be used to control and enhance numerical models of penguin population dynamics and breeding success under future scenarios of sea ice and glacier calving conditions. In particular, as air temperatures warm or as glaciers calve at a particular frequency, we will test how colonies will respond. The outcome of this work is vitally important for our understanding of the species and its survival over the next century and it expected to form the foundation for a case to establish emperor penguins as a protected species in the face of climate change.

As our planet warms the ice cover shrinks, a process that transfers water from land to ocean and thereby raises sea level. The result, which could ultimately raise global sea level by 10s of metres, seems intuitively obvious. However, in the case of the Antarctic Ice Sheet, the processes at work are less than obvious. The atmosphere over the ice sheet is too cold to drive significant melting, so all the snow that falls in the interior is returned to the ocean as ice that only melts once it is afloat. The cold atmosphere creates cold surface waters, so most of the heat that melts the ice comes from deep within the ocean's interior. As it melts the floating ice from underneath, the thinning of the so-called ice shelves allows ice to flow off the land more rapidly, hence raising sea level. So, the underlying process is clear, but why should it drive a loss of ice from Antarctica as the climate warms? The waters that melt the ice are too deep in the ocean to feel atmospheric warming. However, as the atmosphere warms the circulation patterns change, influencing the winds that drive the ocean currents, and that delivers more of the deep warm water to the ice. Understanding how the processes work has been challenging. It is not immediately obvious why a change in the winds should deliver more, rather than less, warm water to the ice. Nevertheless, observation and modelling give us a consistent answer and our understanding of the processes grows as we focus our research on key unknowns. However, there is another puzzle that has received much less attention to date. More warm water leads to more rapid melting of the ice shelves, they thin and the flow of ice off the land accelerates. That acceleration of the flow delivers more ice to the ice shelves, and they should therefore start to grow, or at least thin less rapidly, unless the ocean heat delivery continues to grow. Until recently it was assumed that that is exactly what was happening, but as our record of ocean observations has lengthened, we have seen decadal cycles of warming and cooling. Why then should the ice shelves continue to thin? The answer must lie in the way in which the thinning of the ice shelves themselves affects the melt rate. Again, it is not immediately clear why the change in the ice should increase rather than decrease the melt. However, in this case observation of the key processes is exceptionally difficult because they take place beneath 100s or even 1000s of metres of ice. That is the challenge we will address with this project, by sending an autonomous submarine beneath the ice to make the critical measurements of the ocean, including the temperature of the water and the currents. Those direct observations of the ocean beneath the ice will allow us to verify that the ocean models we use to simulate the processes are correct, or to improve them if they are not. This will not be the first time such measurements have been made, but the new observations will differ in two important respects from the very few that have been made in the past. Some will be repeats of earlier measurements, so we will have observations from before and after a significant change in the extent of the ice shelf. Thus, we can directly answer the question of what change in the ocean circulation accompanied the change in shape of the ice cover. Other observations will target regions where the ice was grounded until recently. Because radar signals penetrate ice, but not seawater, we are able to map the topography only when the ice rests on the land and not when it is afloat. Thus, we paradoxically know the geometry of newly formed ocean cavities with much greater accuracy than we do the cavities that have been there since humans first explored the south polar regions. Our ability to understand the links between cavity geometry and ocean circulation is therefore enhanced in the newly opened cavities that are among the targets of our field campaign.

The surface ocean is home to billions of microscopic plant-like phytoplankton which produce organic matter in the surface ocean using sunlight and carbon dioxide. When they die, they sink and take this carbon into the deep ocean, where it is stored on timescales of hundreds to thousands of years. This storage helps to keep our climate the way it is today. This process of biological CO2 uptake and storage in the deep ocean is called the 'biological carbon pump' and, in order to understand how our climate will change in the near future, we need to understand what controls this process. Until fairly recently, the biological carbon pump was thought to work almost independently from the mixing processes that occur in the oceans, such as during storms, winter or by meandering ocean currents. However, recent work suggested that these physical processes may be very important for the biological carbon pump, providing a direct pathway for carbon to reach the deep ocean, and can contribute as much carbon to depth as the sinking of dead matter alone. Therefore, we urgently need to understand how the biological and physical processes interact to transport organic matter into the deep ocean. Two reasons explain this clear oversight: Physical and biological oceanographers often work independently, so that crossdisciplinary processes can get overlooked. In addition, the location where, and times when, these processes have the most dramatic effect on ocean carbon storage are hostile environments to work in, with very high waves and strong winds that make working from ships nearly impossible. ReBELS is an exciting programme that will bring together physical and biological oceanographers to closely work together on the biological carbon pump. To overcome the logistical challenges, ReBELS will take advantage of the recent developments in technology, using state-of-the-art marine autonomous robots that will be able to sample the ocean at times where we cannot do so with ships. Our study site will be the Labrador Sea in the Northwest Atlantic. There, organic carbon stays in the deep ocean much longer than anywhere else in the world (>1000 years). Moreover, the Labrador Sea has been identified as a very important location for the climate, as it is strongly affected by increasing temperatures and melting ice. Using autonomous technology, we will measure the biological carbon pump over the course of an entire year, and quantify carbon transport and carbon storage through the different biological and physical processes. To do so, we will measure the distribution of organic matter particles throughout the water column and determine whether they are sinking or being transported by ocean mixing. We will then extend our results to the entire Northwest Atlantic using proxies that can be determined on larger scales (for example from satellites). Finally, we will work with modellers to include these important processes when predicting climate in the future.

Large volcanic eruptions can have a major impact on climate, due to the emission of sulfur gases, which form small droplets (aerosols) that reflect incoming sunlight and cool the Earth's surface. When these aerosols form in the upper levels of the atmosphere (the stratosphere, 15-50 km altitude) they remain there for several years, resulting in pronounced global cooling. Indeed, this phenomenon has inspired controversial proposals to cool the planet to combat global warming through artificial stratospheric sulfur injections. However, despite its scientific and societal significance, understanding of volcanic impacts on climate is highly uncertain, due to the limited observational record of large explosive volcanism: only two eruptions, Pinatubo in 1991 and El Chichón in 1982, have impacted global climate within the satellite era. These eruptions are at least an order of magnitude smaller than the largest eruptions in the historical record, and so are not representative of the scope of how volcanoes can impact our climate. This makes it challenging to understand, and prepare for, the climatic and societal impact of large eruptions in the future. The limited observational record of volcanic sulfur emissions also creates a major issue for climate models, which need to know how much sulfur to add to their computerised stratospheres in order to mimic historical climate change events. To address these challenges, we are proposing a new way to reconstruct the amount of stratospheric sulfate from large eruptions over the last 2000 years, based on the record of volcanic sulfate found in polar ice cores. Although this approach is widely used, at present there are major uncertainties in how to convert the amount of sulfate found in ice cores into the original amount of sulfate that was in the stratosphere. This project will substantially improve this conversion - known as the "transfer function" - by using new ice cores, new measurement techniques, and new modelling approaches. First, we will make detailed comparisons of the amount of sulfate in the ice to measurements of the amount of sulfur that went into the stratosphere for eruptions during the last 150 years, a time period in which direct observations of the atmosphere (either by satellites or instruments that measure sunlight) exist. Compared to the last time this calibration was done, the number of available ice cores has grown from 11 to 90, allowing for much better spatial coverage and more representative data. We also have a new technique that measures sulfur isotopes to allow us to distinguish the climatically-important stratospheric sulfate from other sources of sulfate to the ice sheets, further improving the accuracy of the calibration. A new computer modelling approach will also be used to make sure that the transfer function is applicable to a broad range of different eruption characteristics (such as the size, season, and latitude of the eruption), and to help us characterise the transfer function's uncertainty. The insights from the ice core calibration and the modelling will be combined to generate a new record of stratospheric sulfate from volcanic eruptions over the last 2000 years. This record will be used widely in climate model simulations, including those used to inform the International Panel on Climate Change (IPCC). Indeed this work may lead to improvements in climate modelling, as if the amount of sulfate to be added to the models for historical eruptions is better known, we should be able to make better assessments of which models most accurately match the associated changes in climate. Looking forward, our work will also be valuable for policy makers and insurance companies interested in natural hazards, as it will allow them to better understand the frequency and potential impacts of the major eruptions that will occur in our future.