The oldest, thickest sea ice in the 'last ice area' of the Arctic - a region thought to be most resilient to climate warming - unexpectedly broke up twice in the past year. Our current theories assume that the end-of-summer ice-covered area will steadily retreat into the Central Arctic Basin as global warming accelerates over coming decades. However, the dynamic break-up events witnessed in 2018 challenge this prevailing view. Here we hypothesise that a weaker, increasingly mobile Central Arctic ice pack is now susceptible to dynamic episodes of fragmentation which can precondition the ice for rapid summer melt. This mechanism of dynamic seasonal preconditioning is unaccounted for in global climate models, so our best current projections are overlooking the possibility for rapid disintegration of the Arctic's last ice area. Our team has demonstrated that seasonal preconditioning is already responsible for the neighbouring Beaufort Sea becoming ice-free twice in the past five years. Even ten years ago this region contained thick perennial sea ice, mirroring the Central Arctic Ocean, but it has now transitioned to a marginal Arctic sea. Could the processes responsible for the decline of the Beaufort Sea ice pack start to manifest themselves in the Central Arctic? Currently, a shortfall in satellite observations of the Arctic pack ice in summer prevents us from testing our hypothesis. We desperately require pan-Arctic observations of ice melting rates, but so far satellite observations of sea ice thickness are only available during winter months. Our project will therefore deliver the first measurements of Arctic sea ice thickness during summer months, from twin satellites: ESA's Cryosat-2 & NASA's ICESat-2. We have designed a new classification algorithm for separating ice and ocean radar altimeter echoes, regardless of surface melting state, providing the breakthrough required to fill the existing summer observation 'gap'. Exploiting the recent launch of multiple SAR missions for polar reconnaissance, our project will integrate information on ice-pack ablation, motion and deformation to generate a unique year-round sea ice volume budget in the High Arctic. This record will inform high-resolution ice dynamics simulations, performed with a suite of state-of-the-art sea ice models from stand alone (CICE), ocean-sea ice (NEMO/CICE), to fully coupled regional high resolution (RASM), and global coarser resolution (HadGEM) models, all now equipped with the anisotropic (EAP) sea ice rheology developed by our team. Using the regional and stand-alone models we will analyse the role of mechanics in this keystone region north of Greenland to scrutinise the coupling and preconditioning of winter breakup events - such as those witnessed in 2018 - to summer melting rates. Using the coupled models, we will quantify the likelihood of the Arctic's last ice area breaking up much sooner than expected due to oceanic and atmospheric feedbacks and how this will affect the flushing of ice and freshwater into the North Atlantic.

Until recently, awareness of the importance of winter carbon dioxide emissions from arctic soils was highly limited, resulting from incorrect assumptions that emissions from frozen soils beneath snow were insignificant compared to other sources. Consequently, carbon dioxide emissions during arctic winter months are frequently omitted from global carbon cycling budgets and our capacity to measure atmosphere-snow-soil processes controlling carbon dioxide emission and simulate them in climate models are under-developed. This limits our ability to make future climate projections, especially in arctic tundra and forested regions, which characterise about 27% of the Earth's land surface and are warming more than twice as fast as the global average since the late twentieth century. Carbon dioxide, a gas which causes the Earth's atmosphere to trap heat causing the planet to warm, is emitted by microbes decomposing organic material in soil. Decomposition can occur when the soil is frozen, but rates of carbon dioxide emission decrease as soil temperatures decrease, down to -20 degrees Celsius when carbon dioxide emissions become negligible. Winter snow cover has an important impact on arctic soil temperatures, acting like a duvet covering a bed. A thick duvet with lots of air trapped between the feathers provides insulation. Air trapped between the snow crystals within a snowpack acts in a similar manner, limiting the loss of heat from soils warmed in the summer to the cold atmosphere during long arctic winters. As the ground is often snow covered for at least half of the year in Arctic regions, it is vital that we understand processes that control the impact of snow cover on soil temperatures and carbon dioxide emissions, and accurately represent these processes in climate models. Here we ask, how sensitive are measured carbon dioxide concentrations within arctic snowpacks to the variability of snowpack physical properties (e.g. size of the snow crystals)? Can more realistic simulations of snowpack density and thermal conductivity in climate models reduce the underprediction in carbon dioxide emissions from arctic snowpacks? And, how may future changes in winter soil temperatures and snow cover affect future carbon dioxide emissions? In order to answer these questions, we will create a new field measurement database of arctic meteorology, soil and snow properties, and carbon dioxide concentrations. We will use this database to develop more realistic representations of processes controlling winter carbon dioxide emissions in climate models, which will lead to confident model projections of future winter carbon dioxide emissions from the wider Arctic region. By combining field and laboratory measurements with climate modelling, this partnership between Canadian, Finnish and UK scientists will increase our predictive understanding of Arctic environmental change resulting from, and contributing to, our warming planet.