- Markedly low levels of winter sea ice in the Barents Sea and Bering Sea; Arctic is losing substantial amounts of older sea ice through Fram Strait
- In Svalbard, meteorologists observe rapid temperature spikes, freezing rain, and cold days in March
- In the Antarctic, winter has returned: new sea ice is forming, though currently less than expected
Arctic Ocean: A record low in numbers
According to our satellite data, on 21 March of this year the sea-ice cover on the Arctic Ocean reached its maximum winter extent; the winter maximum was 14.45 million square kilometres, the lowest total winter extent in the Arctic since the beginning of satellite observations in 1979 (Table 1). In terms of the monthly mean value, the March sea-ice extent was 14.21 million square kilometres, matching the previous lowest monthly mean value, from 2017 (Figure 2).
The US-based observation service NSIDC had, on the basis of its own satellite data, reported a winter maximum of 14.33 million square kilometres on 22 March, but also referred to it as a new record-low sea-ice extent for the Arctic. The NSIDC figure differs slightly from the Sea Ice Portal’s, as the two institutes apply different methods and algorithms to analyse sea-ice satellite data.
In turn, at the beginning of the week, Norddeutscher Rundfunk (NDR) reported a third set of data, which estimated the Arctic sea-ice extent at 12.8 million square kilometres on Sunday, 30 March 2025 – that is, 1.65 to 1.53 million square kilometres less than the Sea Ice Portal and NSIDC numbers respectively. The base data used in the NDR article were measurements of the actual sea-ice extent, which are always substantially lower than the sea-ice extent data used by the Sea Ice Portal and the NSIDC. You can find detailed explanations and an infographic on the two parameters in our glossary under the term “sea-ice cover”.
Table 1: Overview of the maximum winter sea-ice extent in the Arctic since 2015, together with the corresponding mean monthly values for March. Table: meereisportal.de
Figure 1: Development of mean sea-ice extent in the Arctic for the month of March. The light blue line underscores the long-term decline in sea ice.
Sea-ice losses in several regions of the Arctic Ocean
“Conclusively identifying the causes of the lower sea-ice extent this past winter is no mean feat, as the differences between individual years, unlike for the summer, are relatively small,” says AWI sea-ice physicist Dr Thomas Krumpen. Nevertheless, the satellite data does support some initial theories: there were fewer ice-covered areas than the long-term mean in e.g. a region north of Svalbard, and in the northeast part of the Barents Sea. “In both regions, warm Atlantic water comes into contact with the sea ice, slowing the formation of new ice,” Thomas Krumpen explains. The satellites also detected below-average sea ice levels in the Bering Sea (Figure 2).
The low sea-ice extent can be partly explained by pronounced ice dynamics. Satellite data from between October 2024 and March 2025 shows that, particularly in the Laptev Sea, Kara and Barents Seas, powerful offshore winds drove the sea ice from the coast and toward the Central Arctic (Figure 3). “In colder regions like the Kara and Barents Seas, this led to above-average ice formation, whereas we observed a northward shift in the ice margin in the warmer Barents Sea,” the AWI expert relates (Figure 4).
Due to the lack of sea-ice cover in the Barents Sea and Bering Sea, the surface water was also comparatively warm there in March: the monthly mean sea-surface temperature was 2 to 3 degrees above the long-term mean for the reference period 1971 – 2000 (Figure 4).
“If we look at the regions of the Arctic Ocean individually, none of them experienced record-breaking sea-ice losses. But combined, they led to the lowest maximum total extent for Arctic sea ice since the beginning of satellite observations in 1979,” says Thomas Krumpen.
Figure 2: Difference in the mean position of the ice margin in March 2025, compared to the long-term mean for the years 2003 – 2014. Regions marked in blue had more Arctic sea ice than the reference period; those marked in red had less.
Figure 3: Drift-speed anomalies and mean sea-ice drift on the Arctic Ocean for the period October 2024 – March 2025 in comparison to the long-term mean for 2010 – 2025. Figure: Thomas Krumpen/Alfred Wegener Institute
Figure 4: Mean monthly sea-surface temperature anomalies in the Arctic in March 2025. In those regions marked in dark orange and red shades, the sea-surface temperature deviated from the long-term mean for 1971 – 2000 by more than 1.5 degrees Celsius.
Besides ice drift, air temperatures most likely played an important part. According to Thomas Krumpen: “From October to March, air temperatures in broad expanses of the Arctic were nearly constantly 5 to 6 °C above the long-term mean. These conditions most likely limited ice growth and could potentially also explain the low ice thicknesses that we’re currently seeing in the CryoSat-2 satellite data.”
Observational data from AWI sea-ice buoy 2024I15 shows just how great the temperature changes in the Central Arctic were last winter. At the beginning of the year, the buoy drifted near the North Pole, where it recorded air temperatures between minus 2 and minus 10 degrees Celsius in the first half of February (Figures 5 and 6). “At that time of year, the air temperatures are normally around minus 30 degrees Celsius.”
Figure 5: The map shows the sea-ice cover and the position of AWI sea-ice mass-balance buoy 2024I15 on the last day of March 2025. Map: meereisportal.de
Figure 6: Air temperature readings taken by AWI-sea-ice mass-balance buoy 2024I15 from September 2024 to the end of March 2025. Temperature spikes from two warm-air incursions in February 2025 are clearly recognisable. Figure: meereisportal.de
Figure 7: Air temperature anomalies in March 2025 compared to the long-term March mean for 1971 – 2000. Based on daily mean temperatures measured at an altitude of ca. 760 metres. As the map shows, in March too, it was unusually warm in the western part of the Arctic Ocean.
Winter rain and slippery footing in Svalbard
The meteorological observatory at the German-French Arctic research station AWIPEV in Ny-Ålesund, Svalbard also reported rapid air-temperature rises in the course of the winter. On several occasions, southerly winds transported warm and moist air masses over the Norwegian Sea and Greenland Sea to Svalbard. In fact, the air temperature was above freezing for several consecutive days twice in February (Figure 8). “At these times, it stopped snowing and started raining. And, because temperatures subsequently dropped again, the ground around the station was soon no longer covered with snow, but with patches of ice, making it very slippery,” recalls Dr Marion Maturilli, an AWI researcher and head of the meteorological observatory at AWIPEV Station (Photo 1).
Conversely, in March the AWIPEV team repeatedly reported extremely low temperatures. “This marked drop in daily mean temperatures was something we’d also seen in previous years. The prevailing northerly wind at this time of year brings cold air to Svalbard; it’s why March has now become the coldest month of the year,” Marion Maturilli explains. All in all, the temperatures in the first three months of 2025 matched the climatic conditions of the preceding winter.
Figure 8: This temperature curve from the German-French Arctic research station AWIPEV in Svalbard exhibits major temperature fluctuations, with warm-air incursions in February 2025 and substantial cooling in mid-March. Graphic: Marion Maturilli/Alfred Wegener Institute
Photo 1: The ground near the “Blue House”, part of AWIP research station on Svalbard, was very slippery in February. Wherever rain fell on the snow and subsequently froze, patches of ice formed. Photo: Sofie Tiedeck/Alfred Wegener Institute
A sneak preview of the sea-ice development through summer
Given the sea-ice development in the first three months of the year, AWI experts can now make the first cautious predictions concerning how the sea-ice situation in the Arctic could change by the end of summer. “Some aspects indicate a low summer sea-ice extent. For instance, over the past few months we’ve observed an unusually high amount of older and thicker sea ice leaving the Arctic through Fram Strait – most likely due to atypical drift constellations in the past few years,” says Thomas Krumpen. And wherever older sea ice is lacking, the resilience of the remaining sea ice during the summer months is significantly reduced (Figure 9).
Figure 9: This diagram shows the age of the sea ice that leaves the Arctic via Fram Strait and ultimately melts. Over the past few months, the exiting sea ice was unusually old compared to the figures for the past 20 years. Nevertheless, the mean ice age can no longer match the figures seen in the 1980s and 1990s. Back then, the percentage of multiyear ice was much higher – a clear sign of the progressing transformation of the Arctic sea-ice system. Figure: Thomas Krumpen/Alfred Wegener Institute
Near Fram Strait and in the Central Arctic, still more old sea ice is waiting to be exported. “But in this case, the old sea ice doesn’t come from the ‘Last Ice Areas’ north of Greenland and Canada as is usually the case, but partly from Siberia’s Laptev Sea, where it amazingly survived the past two summers,” says Thomas Krumpen. With a bit of luck, this multiyear ice will drift into the research area of the Polarstern expeditions planned for this summer, where AWI experts can then examine it – a chance for the 40-plus-year-old research icebreaker to show what she can do.
The Antarctic: Slowly but surely, winter is coming
In the Antarctic, winter has arrived. Since the summer minimum on 22 February 2025 (with a sea-ice extent of 2.16 million square kilometres), the sea-ice cover on the Southern Ocean has been rebounding (Figure 10). However, as the growth in March was only at an average level, the monthly mean sea-ice extent – at 3.24 million square kilometres – was only slightly above the previous record from March 2022 (3.10 million square kilometres) and the figure for March 2023 (3.12 million square kilometres) (Figure 11).
“The minor fluctuations in sea-ice extent that we’re currently seeing are the product of local weather conditions. Only the next few months will tell whether the sea-ice cover can make up for the losses of the past two years in the course of the winter,” says Dr Klaus Grosfeld, an AWI physicist and expert from the Sea Ice Portal.
Figure 10: In the second week of March, the daily sea-ice extent curve for the Antarctic was below the Sigma span (turquoise band). However, so much new ice formed in the second half of March that the curve rebounded to the lower edge of the Sigma span.
Figure 11: Development of mean sea-ice extent in the Antarctic for the month of March. The light blue line underscores the long-term decline in sea ice. In March 2025, the sea-ice extent was far below the trend line for a fourth consecutive year.
In March 2025, there was substantially less sea ice than in the previous year in e.g. the Ross Sea, Amundsen Sea, eastern Weddell Sea, and off the coast of East Antarctica (e.g. off the Queen Mary Coast). Conversely, satellites detected more pack ice in the western Weddell Sea and at the northwest margin of the Ross Sea. Once again, sea-ice cover was particularly low in the Bellingshausen Sea. In this important overwintering region for the Antarctic krill, the sea-ice cover now melts to a far greater extent in summer than in the years 2003 – 2014, as a comparison of the ice margin’s current position with the long-term mean shows (Figure 12).
Another striking aspect in March 2025: the comparatively high sea-surface temperatures north of the 65th parallel south in the Indian Ocean (Figure 13). However, whether these higher temperatures could have influenced the sea-ice development just off the coast of East Antarctica is a question that can’t be answered using the satellite data alone. Air temperatures were also far above average in broad expanses of the Antarctic. The map of mean air temperature anomalies shows air masses up to 4 degrees Celsius higher, particularly over West Antarctica, the South Pole region, and the Weddell Sea (Figure 14).
Figure 12: Difference in the mean position of the ice margin in March 2025, compared to the long-term mean for the years 2003 – 2014. Regions marked in blue had more Antarctic sea ice than the reference period; those marked in red had less.
Figure 13: Mean monthly sea-surface temperature anomalies in the Antarctic in March 2025 in comparison to the long-term March mean for 1971 – 2000.
Figure 14: Air temperature anomalies in March 2025 compared to the long-term March mean for 1971 – 2000. Based on daily mean temperatures measured at an altitude of ca. 760 metres.
Graphics
