This research proposal on sea ICE motion addresses a critical need for services to polar mariners, supporting safer arctic navigation and risk mitigation

IT leverages state-of-the-art remote sensing technologies, specifically satellite synthetic aperture radar (SAR) imagery, and advanced deep learning computer vision algorithms like convolutional neural networks (CNNS), transformers, and autoencoders. The study is pioneering a novel approach to comprehensively analyze sea ICE motion at multiple spatial resolutions. The research aims to provide detailed insights into spatial and temporal patterns and to develop scalable analysis techniques. The proposed work will support operational ICE services in NOAA (US national ICE center, and alaska sea ICE program) and validate forecasting models, including those used by noaas office of response and restoration (or & R) in the event of an oil spill in ICE covered water. Anticipated outcomes include impactful research contributions to environmental conservation, navigation safety, and the advancement of the scientific community, including conference presentations and the publication of a research paper. Additionally, a live web system will be developed to facilitate the seamless execution of the generated deep learning models, allowing users to apply these cutting-edge systems to their specific datasets. This will enhance accessibility and foster broader utilization of the research tools and outcomes, positioning the developed models as a foundational framework for future studies in polar research and climate science.