Antarctic February-September extent correlation in 2016-2024 exceeded all 29 pre-2016 nine-year windows

We tested whether the Antarctic summer-winter extent association in 2016-2024 depends on recent extremes, linear trends, or comparison length. We used February and September means from Sea Ice Index Version 4 daily data for 1979-2024. The recent Pearson correlation was 0.873, compared with 0.111 in 1979-2015. The recent value exceeded all 29 earlier nine-year windows. The recent correlation was 0.840 after separate local linear trends were removed, and 0.843 without both 2023 and 2024. With year and previous September controlled, it was 0.804. A conditional three-year block interval for the raw correlation difference was 0.362 to 1.308. A February regression fitted before 2016 still had recent winter RMSE of 0.754 million square kilometres. These results extend an earlier campaign forecast study with a short-period robustness audit. They do not establish forecast skill, a physical cause, or a regime shift. Nine recent years, overlapping comparison windows, a selected period boundary, one extent product and block assumptions limit inference. Code, derived inputs, source hashes, all tests and a figure are included.

Antarctic sea ice · seasonal persistence · short records · window comparison · trend sensitivity