The El Niño-Southern Oscillation (ENSO), the most prominent interannual climate variability signal, has been widely studied for its teleconnections with Antarctic sea ice variability. However, its influence on the predictability of Antarctic sea ice remains poorly understood, hindering the development of accurate sea ice prediction models.

A recent study led by Prof. Li Xiaofeng from the Institute of Oceanology of the Chinese Academy of Sciences (IOCAS) has shed new light on this issue by examining how ENSO affects both linear and nonlinear predictability of Antarctic sea ice across varying lead times. The findings were published in npj Climate and Atmospheric Science.

Antarctic sea ice is a critical component of the global climate system, regulating inter-ocean water exchange, facilitating poleward heat transport, and influencing the global thermohaline circulation. Over the past decade, increased variability in Antarctic sea ice extent has underscored the need for improved predictability research.

Using a deep learning model, the sea ice prediction network (SIPNet), and its linear counterpart, the research team investigated the effects of El Niño and La Niña events on Antarctic sea ice predictability. The study revealed that ENSO exerts cross-timescale influences on subseasonal predictability.

Within a three-week lead time, sea ice predictability is primarily driven by the persistence of sea ice anomalies, with minimal ENSO influence. However, as the lead time extends beyond four weeks, the role of sea ice persistence diminishes, and ENSO’s influence becomes increasingly dominant.

The study also found that El Niño events have a stronger overall impact on sea ice predictability compared to La Niña events. Specifically, El Niño enhances linear predictability, improving predictability in the Amundsen–Bellingshausen Sea, Ross Sea, and Indian Ocean sector by 25.6%, 19.6%, and 30.4%, respectively, at an 8-week lead time. In contrast, La Niña primarily enhances nonlinear predictability, particularly in the Ross Sea. However, both ENSO phases significantly reduce predictability in the western Pacific sector.

ENSO influences Antarctic climate through atmospheric teleconnections, generating larger and more persistent sea ice anomalies that provide additional predictability signals. “SIPNet effectively captures and interprets these signals, advancing our understanding of Antarctic sea ice predictability and offering scientific guidance for refining sea ice prediction models,” explained Prof. Wang Yunhe, the study’s first author.

These findings provide critical insights into the mechanisms driving Antarctic sea ice predictability and offer a foundation for enhancing polar climate forecasting capabilities.