Sources, Sinks and Snow (S3): winter carbon emissions from future Arctic landscapes

Arctic regions are warming around three times faster than other parts of the Earth, causing vast tracts of these sensitive landscapes to rapidly change. The Arctic contains large quantities of frozen carbon in the ground, known as permafrost (representing 50% of global soil organic carbon), which is now thawing and becoming available to the atmosphere. Warmer Arctic air temperatures, and thawing permafrost, are increasing the amount of carbon being emitted into the atmosphere leading to accelerated rates of global climate change. To assess the importance of these carbon emissions and to model how much carbon will be emitted in future decades, we must monitor how processes controlling carbon emissions are changing and refine computer models to incorporate these processes. Arctic winter seasons can extend over nine months of the year and are warming at a faster rate than other seasons. Our understanding of ecosystem processes controlling carbon emissions during long winter periods are, however, severely limited by our lack of field measurements across different Arctic landscape types such as tundra and boreal forest. Furthermore, current computer models poorly simulate winter snowpack and energy exchanges between the air and land, causing them to poorly estimate winter carbon emissions from the ground. The S3 project will address these challenges. We will conduct extensive field measurement campaigns and harness novel low-cost instrumentation, to measure carbon fluxes through snowpacks across diverse and rapidly changing Arctic landscapes. Field measurements will be used to evaluate and improve the way we represent ecosystem processes, which control winter carbon emissions from soil to the atmosphere in climate models. Combining our results with others measured elsewhere in the Arctic, we will then create and test a new, computationally fast statistical model of ecosystem processes, which reproduces the functions controlling carbon emissions in a full climate model. This statistical model, known as an emulator, can be used to isolate and attribute how much each process contributes to the overall uncertainty in simulations of carbon emissions. The attribution of uncertainty in each process cannot currently be achieved by just repeatedly running a climate model; instead, an emulator will allow us to understand which processes are most important for accurate future projections of winter carbon emissions. This will help determine what causes the Arctic land surface to act as an annual carbon emitter (source) or carbon absorber (sink), and assess how this will change in the future across different regions of the Arctic. Policymakers and scientists will benefit from improvements made to an open-source community climate model and the attribution of uncertainty to individual ecosystem processes. This will help improve confidence in projections of future carbon emissions at policy-relevant timescales. Accurate projections are critical for forthcoming decades, as land subsidence, increased shrub growth, forest disturbance, and hydrological change are all profound environmental consequences expected because of permafrost thaw in Arctic regions. These consequences will directly affect local communities in Arctic regions. Improved projections of future change will inform community mitigation and adaptation land management strategies, as well as evidencing the urgent need for national and international carbon emission policy-making.