Antarctic Sea Ice’s 2026 Peak Was Third-Lowest—But Observers Measured Different Maxima

A new NASA Earth Observatory report, published October 12, puts Antarctica’s 2026 winter sea-ice maximum at 17.59 million square kilometers on September 14, the third-lowest annual peak in the satellite record. The figure comes from NASA and the US National Snow and Ice Data Center (NSIDC), whose October analysis also says Antarctic sea ice reached record-low daily extents after October 2.

That observation has a second story inside it: separate national observation services measured different peaks. Japan’s Arctic and Antarctic Sea Ice Information Center reports a September 11 maximum of 17.71 million square kilometers. The UK Met Office’s September briefing estimated 18.12 million square kilometers on September 11, provisionally. These institutions agree on the extraordinarily low rank, but not the peak’s exact area or date. The difference is a measurement question, not a reason to dismiss the underlying change.

Three observing systems, three figures

NASA and NSIDC identify the likely peak as September 14, when Antarctic sea ice spanned an estimated 17.59 million square kilometers, about 6.79 million square miles. NSIDC had originally described that date as the likely maximum because short-lived changes in the wind and ocean can drive renewed ice growth after a provisional peak. Its October review places the result against rapidly shrinking ice extent in early October.

Japan’s National Institute of Polar Research (NIPR), publishing an October 8 monthly assessment, identifies September 11 as the annual maximum at 17.71 million square kilometers. NIPR also puts the Antarctic *September monthly average* at approximately 17.39 million square kilometers, about 0.30 million below September 2025. That average ranks third lowest in its time series. Comparing its monthly figure directly against NASA’s one-day maximum would be a category error: an average and a peak measure different things.

The UK Met Office used a different sea-ice product in a September 21 briefing, provisionally putting the highest daily reading to that point at 18.12 million square kilometers on September 11. The researchers expressly allowed for a possible later peak, because the seasonal maximum can occur as late as early October. The UK, Japanese and American figures should therefore be attached to the precise observation methods and publication dates that produced them.

Different processing algorithms, satellite observations, concentration thresholds, grid resolution and daily sampling may contribute to differences in reported extent. That is an explanation of possible measurement mechanisms, not a claim that any one mechanism has been demonstrated to explain all three published values. A definitive reconciliation needs each institution’s original grid data and calculation method, not only the summarized headlines.

What “extent” counts—and what it does not

Scientists distinguish sea-ice extent from the area occupied by ice itself. For extent, satellite analysts typically count the entire area of grid cells where ice concentration meets a threshold, commonly 15 percent. A cell with partial ice can therefore contribute its full grid area to an extent estimate. Other products and map resolutions can draw the ice boundary differently.

NASA explains that passive-microwave sensors measure naturally emitted radiation from Earth’s surface. Ice and open water have different microwave signatures, and those signals can pass through clouds that often obscure optical imagery over the Southern Ocean. The measurements create a repeatable record extending to late 1978, supported over time by several satellite instruments. They are powerful precisely because their methods and limitations can be documented.

The season matters too. Antarctica’s sea ice usually reaches a maximum near the end of the Southern Hemisphere winter, around September, then retreats to a minimum around February. An annual maximum is a daily high, while the September average includes ice changes before and after that high. Neither is an ocean-ice volume measurement; sea ice thickness and concentration are additional questions.

NOAA’s September global climate summary, released in October, independently calls the region’s September mean ice extent its third smallest in a 48-year record, about 500,000 square miles below the selected average. This supports the picture of an unusually low September even though NOAA’s monthly statistic cannot arbitrate between competing daily peaks.

A volatile southern record

Antarctic ice history does not mirror the Arctic’s simple long-running decline. NASA reports that Antarctic sea ice increased by about one percent per decade between 1979 and 2014, amid substantial variability. The region reached winter record highs for three consecutive years, 2012 through 2014. That climb ended around 2015, followed by strikingly low extents from 2016 onward; 2023 marked record lows in both the annual minimum and maximum.

The 2026 peak belongs to this more recent cluster of low-ice winters. NSIDC says that after October 2, daily extents fell below previous records, especially in the Ross and Amundsen seas and the Indian Ocean sector. NIPR finds September ice below its comparison average through much of the Southern Ocean, apart from portions of the Weddell and Bellingshausen seas. The regional pattern matters because winds, ocean currents and ocean heat do not operate identically around the continent.

A third-lowest seasonal rank is an observation, not by itself a complete attribution study. Human-driven global warming is changing the climate system, but the magnitude of the human contribution to one Antarctic year and one regional ice pattern must be estimated using atmospheric, oceanic and ice dynamics. NASA explicitly cautions that researchers continue examining whether recent low extents indicate an enduring shift or unusually large variation.

The next comparison to demand

An easy but misleading headline would select whichever published number looks most dramatic and omit the others. A better public record retains the name of the measuring institution, the observed date, the physical variable, the reference period and the status of any provisional determination. Differences between otherwise rigorous datasets can illuminate where observations meet conventions and algorithms.

The jointly supported finding is narrower and more robust than any single figure: independent scientific institutions place Antarctica’s 2026 winter maximum among the three lowest measured since satellite records began, even as they report different dates and extents. NASA’s October 12 visualization adds historical context; the Japanese, British, NOAA and NSIDC products reveal the metrology underneath it.

The next useful research product would align those daily datasets on one calendar, inspect the ice-edge algorithms and propagate uncertainty before interpreting the differences. It would also track the fast-moving Antarctic retreat into the southern summer. Until those tasks are done, the responsible comparison preserves both the agreement and the measurement differences.