Nonlinear sensitivity of glacier mass balance to future climate change unveiled by deep learning

Glaciers and ice caps are experiencing strong mass losses worldwide. This study performs the first-ever glacier evolution projections based on deep learning, modeling 21st-century glacier evolution in the French Alps. By the end of the century, a glacier volume loss between 75% and 88% is predicted. Deep learning captures a nonlinear response of glaciers to air temperature and precipitation, improving the representation of extreme mass balance rates compared to linear statistical and temperature-index models. Temperature-index models, often used in large-scale studies, are shown to be over-sensitive to future warming, potentially biasing projections under low-emission scenarios and for flatter glaciers and ice caps, leading to long-term biases in sea-level rise and water resources projections.

Jordi Bolibar, Antoine Rabatel, Isabelle Gouttevin, Harry Zekollari, Clovis Galiez