Significant contribution of internal variability to recent Barents–Kara sea ice loss in winter

Abstract

The Arctic has experienced a rapid sea ice loss in the Barents and Kara Seas in winter over the satellite era. Such sea ice loss has been modulated by anthropogenic forcing and internal variability, but a precise estimate of their relative contribution remains unclear. Here, using large ensemble simulations and machine-learning techniques, we successfully reproduce wintertime Barents-Kara sea ice trends as the joint impact of anthropogenic and internal variability components. Over the whole satellite period, anthropogenic forcing contributes about 70% to the Barents-Kara sea ice loss, with internal variability contributing to the remainder. However, internal variability is more important in explaining varying sea ice trends over shorter periods (~20 years), including the accelerated sea ice loss up to 2017, consistent with an extreme, unforced atmospheric circulation dipole trend in the Euro-Atlantic sector. Overall, this study highlights that internal variability plays a more important role in shaping recent winter Arctic sea ice loss than previously thought, and has implications for Arctic sea ice projections and the use of machine learning methods in future attribution studies.