- Walls of sea ice on the margins of the German Baltic: February 2026 surprised us with an unusually high amount of ice on the Baltic
- The Arctic: Researchers discover a previously unknown recrystallization process in which the surface of sea ice appears to transform into snow
- The Antarctic: A surprisingly large amount of sea ice survived the summer, as Stefanie Arndt reports from on board the research icebreaker Polarstern
The Baltic: Up to 80-centimetre-thick ice in Finnish waters
For those living near the Baltic Sea, the winter 2025/2026 will go down in history as a particularly ice-rich one. Just as in January, in February the northern and eastern regions of the inland sea were characterised by an unusually high amount of sea ice (Figure 1). In Bothnian Bay, the northernmost part of the Gulf of Bothnia, experts from the Finnish Ice Service observed 30- to 80-centimetre-thick sea ice anchored to the shore (fast ice). In contrast, in the Gulf of Finland the fast ice was up to 50 centimetres thick. In both regions, thicknesses on the open sea ranged from 10 to 40 centimetres.
In terms of the bodden along Germany’s Baltic coastline, at the end of the month the Ice Service of the Federal Maritime and Hydrographic Agency (BSH) reported both dense accumulations of ice and areas of open water. A particular eye-catcher: several-metre-high walls of ice on the shoreline, formed by the wind and waves. These ice masses even drew visitors from other parts of the country to the Baltic coast and were surely captured in thousands of photographs – including those taken by Sea Ice Portal reader Frieder Söling (Figures 2 & 3).
Figure 1: Mean sea-ice concentration in the Baltic in February 2025 and February 2026. The comparison underscores the fact that a remarkably large amount of sea ice formed on the Baltic in the winter 2025/2026.
Figure 2: On the east coast of the isle of Rügen, the ice piled up to four metres high at the end of the month. Photo: Frieder Söling
Figure 3: However, at the same time there were patches of open water in the Greifswald Bodden, as can be seen here. Photo: Frieder Söling
The Arctic: No end of declining sea ice in sight
The sea-ice development in the Arctic in February held no surprises. In keeping with the time of year, the sea-ice extent grew by 500,000 square kilometres in the first three weeks of the month, reaching a preliminary winter maximum of 14.36 million square kilometres on 20 February (Figure 4). “After that date, the sea-ice extent dropped significantly. But this trend could reverse itself in March, which is why we can’t consider this the absolute winter maximum at this point in time,” explains Dr Klaus Grosfeld, a climate researcher at the Alfred Wegener Institute and co-initiator of the Sea Ice Portal.
The monthly mean sea-ice extent was 14.19 million square kilometres, putting it in fourth place in the February time series (Figure 5). “The monthly mean figure confirms the long-term decline in Arctic sea ice. Especially in the Barents Sea, Sea of Okhotsk and Davis Strait to the west of Greenland, there is currently less sea ice than in the reference period 2003 – 2014,” says Klaus Grosfeld (Figure 6).
Satellites detected somewhat more pack ice than the monthly mean in the Bering Sea and Fram Strait. “But these gains aren’t enough to compensate for the losses in other regions,” the sea-ice expert adds.
Figure 4: Development of sea-ice extent in comparison. Though the curve for 2026 (sky blue line) rose in the first half of February, it never exceeded the lower edge of the turquoise band, which indicates the span of minima and maxima in the period 1981 – 2010. By comparison: in the winter 2024/2025 (dark blue line) there was far less sea ice; in the winter 2021/2022 (dark red line) there was more.
Figure 5: The monthly mean sea-ice extent in the Arctic for February 2026 was only slightly above the trend line for the month, confirming the long-term decline in Arctic sea ice.
Figure 6: Difference in the mean position of the ice margin in February 2026, compared to the long-term mean for the years 2003 – 2014. Regions marked in blue had more Arctic sea ice in the second month of 2026 than in the reference period; those marked in red had less.
When sea ice turns into snow
The snow cover on sea ice has a major influence on how much energy from the atmosphere passes through the ice to reach the ocean, and vice versa. Accordingly, sea-ice physicists also investigate the properties of this snow. Following up on the MOSAiC expedition (2019 – 2020), an international team of researchers have now identified a previously unknown form of interaction between the sea ice and snow cover. They have successfully shown that, over time, the structure of the uppermost part of the Arctic sea ice becomes snow-like, causing the sea ice itself to grow thinner (Figures 7 & 8).
“Given the substantial temperature difference between the frigid snow and the ocean-warmed sea ice, part of the ice’s surface becomes gaseous without previously melting. It is transformed directly into water vapour and rises through the cold snow-cover layer, where it refreezes into something like rime ice,” explains Dr Hanno Meyer, an expert on the geochemical properties of snow and ice at the Alfred Wegener Institute, Helmholtz Centre for Polar and Marine Research.
“In the course of this recrystallization process, the ice’s structure and chemical properties change, which is what helped us make this discovery. The former sea ice then actually looks like snow and is much more porous and therefore permeable than sea ice or compacted snow cover,” adds his AWI colleague Dr Martin Werner, who also analysed the ice and snow samples from the Arctic (Figure 9).
During the MOSAiC expedition, the snow from the sea ice in the samples gathered was up to 6.3 centimetres thick, as the experts relate in a recently released article. “This finding very much surprised us, as until then, we had assumed that our snow-depth measurements reflected the actual amount of snowfall. Instead, the snow cover consists of the fallen snow and sea ice that has been transformed. And that means our snow and mass-balance readings aren’t as clear and precise as previously assumed,” says Martin Werner.
The researchers assume that the snow and sea ice in the Antarctic undergo the same process under comparable conditions. As such, their new findings are relevant – especially for remote sea-ice sensing – in both regions.
As the experts report in their article:
Snow that comes from sea ice influences the snow depth and density measured, has its own unique surface and grain size, and could potentially also influence the distribution of brine in the snow cover. Understanding these connections is essential to gaining a better grasp of backscatter signals in remote-sensing applications involving altimetry.
Figure 7: The snow samples and data painstakingly gathered during the MOSAiC expedition to the Central Arctic (2019/2020) offered a point of departure for the researchers’ new study. Photo: Alfred Wegener Institute / Esther Horvath
Figure 8: Conceptual illustration of the newly discovered transformation of sea ice into a snow-like crystal structure. The column-like images are from a micro-computer tomography (CT) scan and show the initial crystal structure of the sea ice samples (day 0) and the post-recrystallization structure 21 days later. The diagram reflects data from isotope-based experiments, which allowed the experts to confirm the ice’s transformation. Graphic: Amy R. Macfarlane et al. (2026)
Figure 9: AWI researchers Dr Martin Werner (left) and Dr Hanno Meyer (right) are specialists for the geochemical properties of ice and snow. They contributed to the new study as co-authors, as did the snow experts Prof Stefanie Arndt and Daniela Krampe. Photos: Alfred Wegener Institute / Kerstin Rolfes & Jan Pauls
The Antarctic: Return to the sea-ice extents of the past
This summer, the sea-ice extent in the Antarctic declined significantly less than in the previous five summers. It reached its minimum on 26 February, at 2.8 million square kilometres, putting it at 31st place on the list of record-low sea-ice minima in the Antarctic. 1st place: the minimum sea-ice extent from the summer 2022/2023, at 2.1 million square kilometres (Figure 10).
“In comparison to the extremely low sea-ice extents in recent years, this summer the Antarctic sea-ice extent is once again closer to the climatological mean,” says AWI researcher and Sea Ice Portal expert Dr Renate Treffeisen. “In the first, second and third week of February, the annual curve essentially matched the long-term mean for the period 1981 – 2010. It was only in the fourth week that it dipped slightly below it. In other words, in February 2026 there was markedly more sea ice in the Antarctic than we’ve seen in the past several summers.”
The sea-ice distribution is also interesting. Especially in the northwest Weddell Sea, off the coast of East Antarctica, and off the coast of Adélie Land, pack ice levels were unusually high in February 2026. The satellite data shows fewer floes than in the reference period 1981 – 2010 in the Ross Sea, eastern Weddell Sea, and off the coast of Wilkes Land (Figure 11).
Figure 10: Development of Antarctic sea-ice extent in comparison; in February, the curve for 2026 (sky blue line) largely matched the long-term mean (turquoise line) for the period 1981 – 2010.
Figure 11: Difference in the mean position of the ice margin in February 2026, compared to the long-term mean for the years 2003 – 2014 (left), to the previous year (centre) and to the record-low year 2023 (right). Regions marked in blue had more Antarctic sea ice in the second month of 2026 than in the reference period; those marked in red had less.
In the region with the most pronounced sea-ice growth, the research icebreaker Polarstern is currently fighting her way through the pack ice. One of the 50 scientists on board: sea-ice physicist Prof Stefanie Arndt. Together with the two expedition leaders Dr Ilka Peeken and Prof Christian Haas, she sent us the following brief report.
Report from on board: RV Polarstern currently in the northwest Weddell Sea
The SWOS expedition has now been at sea for more than three weeks, including roughly two-and-a-half weeks in sea ice – bound for the northwest Weddell Sea (Figures 12, 13 & 16). The sea-ice extent in the target area currently reflects practically ‘normal’ summer conditions, following several summers with greatly reduced extents. The comparatively solid ice cover naturally posed certain limitations for us; it made it impossible for us to penetrate the ice much farther than 67 degrees south. Nevertheless, reaching this region marks an important success, as very little field data on it is currently available.
Figure 12: The research icebreaker Polarstern ploughing through up to 2-metre-thick pack ice in the northwest Weddell Sea. Photo: Alfred Wegener Institute / Christian Haas
Figure 13: Position of the research icebreaker Polarstern on 3 March 2026. Screenshot from the AWI expedition blog
Despite the dense ice conditions, we have already taken extensive readings on the snow and ice properties at twelve ice stations – from on board the ship, and with the aid of our helicopter. These efforts were supplemented by numerous EM-Bird flights, which provide macro-scale information on ice-thickness distribution and allow us to see our individual readings in a regional context. They also make the pronounced spatial variability of ice conditions especially clear: toward the Antarctic Peninsula, on the continental slope, thick and compact ice with heavy snow cover is dominant, whereas ice conditions farther to the east and south are generally looser and more readily passable.
It’s also striking that at the majority of the stations already visited, the snow cover is unusually light, and in some cases barely there at all. The detailed analysis of these observations over the months to come will pave the way for more precise insights into the underlying processes and seasonal influences.
A special highlight of the journey so far was the return to a floe where samples had been taken and sensors deployed the previous year as part of the HAFOS expedition (Figures 14 & 15). However, after a year of drifting, not much was left of the once three-kilometre-long floe: only fragments measuring a few hundred metres long.
Nevertheless, the ice was sufficiently stable for fieldwork. As such, we were able to repeat the measurements and samples gathered last year. This new round of sampling will contribute an extremely valuable dataset on the changing ice and snow properties during the floe’s drift through the Weddell Sea, which continued for more than a year.
Figure 14: AWI sea-ice physicist Stefanie Arndt couldn’t be happier: in this picture from 19 February 2026, we see her standing next to a sea-ice buoy that she and her team had deployed on a large floe in the southeast Weddell Sea a year earlier. In the interim, both the floe and the buoy have drifted through the Weddell Gyre – a large, circular current in the Weddell Sea that transports pack ice to the northwest. Photo: Alfred Wegener Institute / Stefanie Arndt
Figure 15: Northward, ho! Over the past twelve months, sea-ice buoy 2025S141 followed this course through the Weddell Sea. You can find the temperature, atmospheric pressure, and snow-depth data it recorded here: https://data.meereisportal.de/relaunch/buoy.php?lang=de&active-tab1=method&active-tab2=buoy&singlemap&buoyname=2025S141#galerie_buoy Map: Meereisportal.de
Figure 16: Following rains in the expedition area, the snow on the sea ice melted. Now the dark ocean below can be seen in areas where the sea ice is sufficiently thin. Photo: Alfred Wegener Institute / Christian Haas
Contact
Prof. Dr. Stefanie Arndt (AWI)
Questions?
Write us an email or use the contact form.
Author
Sina Löschke (Science Writer)
Graphics