← Back to the dashboard

How we track sea ice at both poles

What the dashboard shows, where the numbers come from, and what they can’t tell you.

The short answer

Sea ice is frozen seawater floating on the polar oceans. Its bright surface reflects most of the sunlight that hits it; open ocean absorbs most of it. So when ice shrinks, the ocean takes in more heat, which melts more ice.

The two poles are behaving differently. Arctic ice has been shrinking for more than four decades, in every season: its summer low has fallen by about 12% per decade, and its two smallest winter peaks on record came in 2025 and 2026. Antarctic ice held steady, and even grew slightly, until a record high in 2014; since 2016 it has dropped to a series of record lows. That is why the dashboard always shows the two poles separately: a single “global sea ice” number would hide both stories.

How we know

The numbers come from the Sea Ice Index, published by the US National Snow and Ice Data Center (NSIDC). Satellites measure the microwave energy given off by the surface: ice and open water give off different amounts, and microwaves pass through cloud and the dark of the polar night. That is why the record is daily and almost unbroken since late 1978.

Several instruments have carried the record: one on NASA’s Nimbus-7 satellite from 1978, then a series on US Defense Meteorological Satellite Program satellites. Since January 2025 the index has used AMSR2, an instrument on a Japanese satellite, because the older US instruments are past their design life and their data is ending. AMSR2 places the ice edge slightly further in, so recent values read a little smaller than the old instruments would have shown. NSIDC has published the full record in this form as Version 4 since August 2025.

“Extent” means the total area of ocean where at least 15% of the surface is covered by ice. Each day is compared with the 1981–2010 average for the same day of the year, so “9% below normal” always means below normal for this time of year. NSIDC uses 1981–2010 as its reference period, and we use NSIDC’s own averages. Other reports, such as the Arctic Report Card, compare with 1991–2020; because those years already had less ice, the same day looks closer to normal there.

Reading the dashboard

Below normal is not the same as a record. A day can be far below the seasonal average without setting a record for that date, because an even lower year already holds it. In the record calendar, each red square is a day this year that was lower than on the same date in every year since 1979.

Rank for the date. “4th lowest for the date” means that in the 48 years of the record only three had less ice on this calendar date. It is the quickest way to see whether a low reading is unusual even for recent years.

Winter peak and summer low. Each pole reaches its largest extent at the end of its winter (March in the Arctic, September around Antarctica) and its smallest at the end of its summer (September in the Arctic, February around Antarctica). We work them out the way NSIDC does: the highest or lowest five-day average of the season, ranked from lowest, with years within 40,000 km² of each other counted as tied. Our figures for 2026 match the ones NSIDC announced. Until the ice has clearly turned, the dashboard says “so far”.

How current each part is

What the dashboard can’t tell you

It measures area, not thickness. The Arctic can keep a similar extent while the ice gets thinner and younger, which it has been doing for decades: ice more than four years old has shrunk by 95% since 1985–2004. Thickness and volume need other satellites, such as CryoSat-2 and ICESat-2, and are not on this page.

The satellites also struggle in summer, when pools of meltwater on the ice can look like open water, and with very thin new ice. Day-to-day wiggles are normal; the trend over decades is what matters. From 1979 to 1987 the satellite measured only every other day, so for the rank of a date we fill a missing day from the days either side.

It cannot tell you what Antarctica will do next. Whether the lows since 2016 are a temporary swing or the start of a new, lower state is an open research question.

And sea ice does not raise sea level when it melts, because it is already floating. Rising seas come from warming water and melting land ice, such as Greenland’s and Antarctica’s ice sheets.

Further reading

Credits

The polar ice record exists because of the teams who have built and flown microwave instruments since Nimbus-7, NASA, the US Department of Defense and the Japan Aerospace Exploration Agency who kept them in orbit, and the scientists at NSIDC who calibrate and publish the record every day. This dashboard collects and presents their work.

← Back to the dashboard