Seasonal Arctic sea ice forecasting with probabilistic deep learning

Anthropogenic warming has caused unprecedented year-round reduction in Arctic sea ice extent. This paper introduces IceNet, a probabilistic deep learning sea ice forecasting system trained on climate simulations and observational data to forecast sea ice concentration maps for the next six months. IceNet outperforms a state-of-the-art dynamical model in seasonal forecasts, particularly for extreme sea ice events, advancing the accuracy and range of sea ice forecasts and contributing to conservation tools mitigating risks associated with rapid sea ice loss.

Tom R. Andersson, J. Scott Hosking, María Pérez-Ortiz, Brooks Paige, Andrew Elliott, Chris Russell, Stephen Law, Daniel C. Jones, Jeremy Wilkinson, Tony Phillips, James Byrne, Steffen Tietsche, Beena Balan Sarojini, Eduardo Blanchard-Wrigglesworth, Yevgeny Aksenov, Rod Downie, Emily Shuckburgh