By This Hour Science Desk
Arctic sea ice reached its annual low point on Sept. 12, covering an estimated 4.6 million square kilometers, or 1.78 million square miles, NASA and the National Snow and Ice Data Center said. The figure places 2026 in a tie with 2008, 2010 and 2025 for the 10th-lowest annual Arctic minimum in the continuous satellite record.
The ranking does not make 2026 a new record low. It nevertheless adds another year to a striking feature of the modern record: NASA and NSIDC say the 20 lowest annual Arctic minimum extents observed since continuous satellite measurements began in late 1978 have all occurred from 2007 through 2026. That sequence is the central context for reading a single season whose final extent was comparatively stable against the most extreme years, but still low by the standards of the earlier satellite era.
September is normally the point at which Arctic sea ice reaches its smallest seasonal footprint. Ice expands during the dark, cold autumn and winter, then retreats through spring and summer as temperatures rise. The Sept. 12 date marks the conclusion of that annual melt season as assessed by the agencies, rather than an indication that Arctic ice has disappeared or that conditions are identical across the polar ocean.
A minimum that sits within a long run of low years
NASA and NSIDC placed the 2026 minimum at 1.78 million square miles. In the agencies’ accounting, that matches the extent reported at the annual minimum in three earlier years: 2008, 2010 and 2025. The resulting four-way tie is ranked 10th-lowest in the satellite record.
Such a ranking provides a narrow but useful comparison. It locates this year against other late-summer minima measured over nearly five decades, but it does not by itself describe every aspect of the ice cover. The supplied account reports an estimate of extent and a position in the historical ordering; it does not provide separate measurements of ice thickness, ice age, regional distribution, or changes within individual Arctic seas. Those distinctions matter for a fuller physical assessment, but they are not established in the material available for this report.
Nor does the reported tie erase the broader pattern identified by NASA and NSIDC. The agencies say every one of the 20 lowest annual minimum extents has fallen in the period beginning in 2007. In practical terms, the recent record has repeatedly produced late-summer ice coverage below anything observed during the first decades of continuous satellite monitoring. The 2026 result belongs to that group even though several years registered lower minima.
Walt Meier, a senior research scientist at NSIDC, characterized recent Arctic extent as having levelled off while remaining low compared with the earlier portion of the record. That distinction is important. A relatively steady run of September extents is not the same as a return to conditions seen before the modern stretch of low minima. The reported plateau describes recent behavior in one measure at one seasonal point; the comparison to the earlier record provides the longer perspective.
Weather can shift one summer’s outcome
The agencies caution against treating any individual melt season as a simple straight-line expression of the longer record. Weather conditions can substantially alter how much ice melts during a given Arctic summer. That year-to-year variability helps explain why annual rankings can move even within a period in which the overall set of recent minima remains historically low.
Linette Boisvert, a sea-ice scientist at NASA’s Goddard Space Flight Center, attributed part of the relative stability in recent September extents to greater cloud cover over the past decade. In the account provided by NASA, that cloudier pattern has limited the additional solar radiation that would otherwise accelerate melt. The point is not that cloud cover makes the Arctic immune to summer melting. Rather, it is offered as one reason that the September extent figures have not continued to decline uniformly from year to year.
This seasonal complexity sets limits on what can be inferred from the 2026 number alone. A minimum of 4.6 million square kilometers establishes the agencies’ estimate for this year’s low point and its ranking in their record. It does not, on its own, identify the precise combination of weather conditions behind the outcome, quantify cloudiness in 2026, or settle how future summers will unfold. The source material instead frames the figure as consistent with patterns already visible across the satellite-era record.
The annual cycle itself also matters. Arctic ice does not stay at its September minimum: the approach of autumn and winter brings darkness and colder conditions that support seasonal growth. For that reason, the annual minimum is a defined point in a repeating cycle, and comparisons in the NASA and NSIDC account are specifically comparisons among annual minimum extents. A reader should not treat the reported September figure as an all-year measure of Arctic ice cover.
Nearly five decades of observations underpin the comparison
The historical comparison rests on a satellite record that began with observations from NASA’s Nimbus-7 mission in October 1978. NASA and the National Oceanic and Atmospheric Administration subsequently used a succession of satellite instruments to continue the sea-ice record. The account identifies instruments on Defense Meteorological Satellite Program spacecraft from 1987 and NASA’s Aqua satellite from 2002 to 2011 as parts of that continuity.
Scientists now use data from the Advanced Microwave Scanning Radiometer 2, or AMSR2, which launched in 2012 aboard the Japan Aerospace Exploration Agency’s GCOM-W satellite, to continue the record cited for the 2026 minimum. The use of successive instruments is relevant because a multi-decade ranking depends on maintaining observations across changes in spacecraft and missions. NASA and NSIDC describe the record as continuous from late 1978, giving the 2026 placement its stated historical frame.
The length of that record also explains the care needed with the word “minimum.” It refers to the lowest extent in a particular calendar year within a monitored series, not to an absolute statement about all Arctic ice at all times. The agencies’ finding that 2026 tied for 10th-lowest is thus a ranked result within the satellite-era annual-minimum dataset. It is a robustly specific description of the reported measurement, while leaving other questions outside its scope.
NASA illustrated the season with daily imagery showing the retreat of ice around the pole between March 15 and Sept. 12. The visual chronology corresponds to the seasonal pattern described by the agencies: loss through the warmer months followed by the endpoint used to identify the annual low. It does not change the numerical finding, but it emphasizes that the reported minimum was reached after months of seasonal retreat rather than in isolation.
Antarctic conditions offer a different, unfinished seasonal picture
At the opposite pole, Antarctic sea ice was approaching its annual maximum when the Arctic reached its minimum, NASA said. Antarctic ice typically peaks in late September or early October. The agency said it and NSIDC would announce the 2026 Antarctic maximum after determining that seasonal growth had ended.
NASA reported that Antarctic extent fell by about 300,000 square kilometers, or roughly 116,000 square miles, during six days in August before growth resumed. Meier linked the episode to weather that pushed and compacted thin, mobile ice near the outer edge of the cover. He described events of that kind as commonplace but said the size of this particular fluctuation was unusual.
The agencies draw a clear distinction between the poles. Antarctic sea ice changes more sharply from year to year than Arctic sea ice, in part because Antarctica’s ice is not enclosed by surrounding land and can respond more freely to wind and weather. NASA also said recent Antarctic annual maxima have generally been lower, but Boisvert and Meier were cautious about describing that observation as a long-term trend because variability is large. That caution is specific to the Antarctic discussion and should not be used to blur the separate Arctic result.
The report on the Arctic minimum has not been independently corroborated for this article. It is based on the supplied NASA account, which attributes the measurement and historical ranking to NASA and NSIDC. The available material contains no competing estimate, but it also does not supply the underlying data analysis, regional breakdowns, or an independent assessment of the agencies’ conclusion.
For now, the confirmed claim within the supplied record is limited and clear: NASA and NSIDC say Arctic sea ice bottomed out on Sept. 12 at 4.6 million square kilometers, tied for 10th-lowest in their satellite-era series. As the Arctic moves into its growth season, that figure will remain the agencies’ stated benchmark for the 2026 summer melt season, while the separate Antarctic maximum is still pending confirmation.
For further context on this subject, see Researchers Reportedly Used Claude in Effort That Reached OpenAI Account and GitHub Data.
Reporting notes
What is confirmed: The agencies ranked 2026 in a four-way tie for 10th-lowest, alongside 2008, 2010 and 2025.
Why this matters: The result extends a run in which the 20 lowest Arctic annual minima in the satellite record have occurred since 2007.
What remains unclear: The supplied material does not provide underlying data, regional detail, thickness measurements, or independent confirmation. This report is based on one source and has not been independently corroborated.
