Melting Arctic ice is dumping vast quantities of freshwater into the ocean, but scientists cannot reliably trace where that water came from—river, glacier, or sea-ice—because the data needed to distinguish these sources is scattered, incomplete, and hard to interpret. This matters because the source of freshwater determines which nutrients it carries, and those nutrients fuel the marine food webs that sustain fisheries from the Arctic down to the tropics. Without a clear picture of where the freshwater originates, it is impossible to predict how Arctic change will ripple through global ocean ecosystems. The AISIT project will solve this by building the first standardised, machine-readable database of oxygen isotope (δ¹⁸O) measurements, temperature, and salinity data from Arctic waters. Over six months, the team will collate existing data from the BIOPOLE programme and international partners, then work with AI specialists to design a benchmark that lets machine-learning models extract patterns from the sparse, messy dataset. If successful, the database will allow researchers to map freshwater sources across the Arctic for the first time. This is primarily fundamental science—it will not directly change a supply chain or a weather forecast tomorrow—but it will give climate modellers and marine ecologists the tool they need to understand how a changing Arctic reshapes the planet’s oceans.
View original technical description
Context - Warming in the Arctic is leading to a significant increase in glacial meltwater discharge, large alterations in major Arctic river outputs and unprecedented levels of sea-ice retreat. This introduces large amounts of extra freshwater into the Arctic ecosystem with numerous ecological impacts, both locally and globally. These impacts will vary depending on the source of this freshwater since that influences the amount, and mix, of nutrients introduced into the ocean. Arctic nutrients fuel the biological growth essential for healthy marine ecosystems, not only in the Arctic itself but in temperate and tropical regions. Therefore, tracing these sources of freshwater from land to ocean is a priority if the global consequences of Arctic change are to be understood. The different sources of freshwater inputs, from rivers, glaciers, and sea-ice, can be tracked through tracers such as the stable oxygen isotope d18O, as well as certain trace metals and Rare Earth Elements. Freshwater tracer measurements are particularly insightful when combined with accompanying temperature and salinity measurements. Challenge - The collection of freshwater tracer samples, plus accompanying temperature and salinity data, can be easily accommodated within polar field campaigns. Nevertheless, the many challenges of these campaigns, plus the required laboratory analyses, has resulted in sample coverage gaps. Furthermore, scientific programmes are often short-term and geographically focussed, leading to sample results being disparate and uncollated. Finally, ocean circulation and mixing means that there are numerous factors that complicate the interpretation of freshwater tracer data that require powerful analytical approaches. Together, these issues have resulted in the lack of a comprehensive overview of freshwater sources into and out of the Arctic Ocean. Nevertheless, such an understanding can be achieved with a thorough collation of the many sources of freshwater tracer data already available, combined with a growing suite of machine learning methods that can help interpret this data. Aims and Objectives – AISIT is a 6-month project that will enhance data accessibility and use of Arctic freshwater tracer data. The core objective is to develop a standardised and machine-readable database of freshwater tracer data combined with temperature and salinity. It will initially focus on d18O data collected from the NERC cross-centre National Capability programme BIOPOLE, which considers the complexities associated with the land-to-sea transport of nutrient-laden freshwater in polar regions. Through its own fieldwork, and in collaboration with international partners, BIOPOLE has generated, and has access to, a good amount of suitable data for this project. A follow-on objective is to collaborate with other programmes to make the database even more comprehensive in terms of coverage and incorporating other types of freshwater tracers. Throughout the database collation phase, regular interaction with the AI community will be maintained to ensure the database is fit for purpose and that an appropriate AI benchmark is established for wider community engagement. Potential applications and benefits - AISIT will further our capability to understand the wider implications of Arctic change to the Earth system as well as to the marine economy of the UK. Through our interactive approach in co-designing an AI ready database and accompanying AI-benchmark, AISIT will catalyse new downstream research opportunities to exploit sparse data more effectively and improve scientific insights using multi-modal AI to accelerate scientific discovery. Furthermore, AISIT will help further establish the UK as being a leading nation in impactful Arctic science.
Plain English summaries and category classifications on this site are generated by AI and may not perfectly reflect the original research.
Is something wrong? Let us know