Arctic Atmosphere-ice Variability and Marine Accessibility: A Data-driven Approach

Navigating the Arctic Ocean presents daunting challenges due to its darkness, remoteness, and harsh conditions, but has nevertheless prompted human incursions spanning millennia. The ability to traverse the Arctic Ocean has become of increasing importance over time, supporting people living in the Arctic and the strategic development of commercial opportunities. Navigability – the capacity to support safe passage for vessels – is critical for all Arctic operations, from global shipping to local fisheries, from resource extraction to military deployment, from tourism to traditional lifeways. Key to the assessment of this maritime access is understanding and predicting the frequency and intensity of extreme weather in the Arctic. This project will use computational approaches to develop future scenarios for the conditions that lead to shipboard ice accretion and dangerous sea ice convergence. The results will have far-reaching impacts including assessing the utility of these scenarios for Arctic operational planning on decadal time horizons, developing approaches to emerging high resolution climate information that can be generalized to other applications, and training of early career scientists in these challenging computational techniques. To address the research goals, first, numerical weather prediction at daily resolutions will be used to develop a model of the probability along maritime shipping routes of (i) icing conditions leading to adverse operational impacts and (ii) sea ice convergence leading to dangerous ridging conditions. Second, the project team will develop and deploy machine learning techniques to enhance the spatial resolution of sea ice projections in geographically constrained segments of maritime routes, such as the heavily trafficked Arctic straits and emerging Canadian Archipelago routes. Third, these perspectives will be incorporated into future multi-model climate projections to develop more accurate scenarios of marine accessibility with regard to climate futures, jurisdictions, and routes. These generalizable insights will be developed on the use of machine learning to enhance the spatial resolution of climate model output variables for a range of applications. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.