Low summer sea ice gave uncertain winter variance contrasts in 45 Arctic and 46 Antarctic seasons

We tested whether low summer sea ice extent was followed by more variable winter extent. We used NSIDC Sea Ice Index Version 4, with 45 Arctic September-to-next-March pairs for summer years 1979–2023 and 46 Antarctic February-to-September pairs for 1979–2024. We removed separate linear year trends and compared winter variance after the lowest third of summer residuals with variance after other summers. The Arctic variance ratio was 0.696, with a 95% three-year paired block interval from 0.184 to 1.786. The Antarctic ratio was 1.477, with an interval from 0.570 to 2.591. Both intervals include one. The Antarctic ratio was 0.513 when the record ended in 2015, and its median absolute deviation ratio was below one in the full record. These records do not establish a stable change in winter dispersion after low summer extent. We also reproduce a prior campaign endpoint result and label it as a check. Few low summers, changing climate states, retrospective groups, and extent-only data limit inference. The package preserves group evidence, resampling draws, code, source hashes, and an independent numerical audit.

sea ice · conditional variance · winter dispersion · Arctic · Antarctic · null result