ENSO’s impact on linear and nonlinear predictability of Antarctic sea ice

While the influence of ENSO on Antarctic sea ice variability is well-known, its role in sea ice predictability, both linear and nonlinear, remains unexplored. This study utilizes deep learning models to quantify ENSO’s impact on Antarctic sea ice predictability. We find that ENSO events exert cross-timescale influences on sea ice’s subseasonal linear and nonlinear predictability. Within a 3-week lead time, ice persistence is the primary source of predictability. Beyond this period, ENSO becomes a key source of Antarctic sea ice predictability, with El Niño enhancing ice linear predictability more than La Niña. Specifically, El Niño improves ice linear predictability by 25.6%, 19.6%, and 30.4% in the A-B Sea, Ross Sea, and Indian Ocean, respectively, at an 8-week lead time. La Niña mainly enhances ice nonlinear predictability, particularly in the Ross Sea. We demonstrate that ENSO provides additional sources for Antarctic sea ice predictability primarily through generating more extensive ice anomalies. These insights deepen our understanding of sea ice predictability and are crucial for advancing forecasting models.