Arctic Climate Review: Winter 2025-26 Forecast & February Update

by Archynetys News Desk

The Arctic in February and for the 2025-26 winter season was quite different climate-wise than winter 2024-25, which was overall mild and wet. This winter, temperatures were closer to the 1991-2020 baseline average and Arctic lands were actually slightly cooler than the 30-year average. Precipitation totals were overall slightly below average, but of course with large regional variations.

February around the Arctic was notably colder than the 1991-2020 baseline in many areas in ERA5 reanalysis. Northwest Territories, Canada had the second coldest February since 1994 and the Nordic Arctic the coldest February since 2011. For the Arctic overall (land and seas poleward of 60°N), this was the coldest February since 2009.

Only a third of the Arctic had a February average temperature above the 1991-2020 baseline average, and only 23 percent of Arctic lands were above the 30-year average (and more than half of Arctic lands were colder than the 1951-1980 baseline). That’s the lowest fraction of the Arctic to be warmer than the era-specific normal in February since 1979.

Fig. 1 February 2026 average temperature departure from the 1991-2020 baseline average. Data from ERA5 courtesy of ECMWF/Copernicus.

Winter 2025-26 temperatures

Winter 2025-2026 was the 14th warmest since 1950. Overall, about 13 percent of the Arctic had a top-three warmest winter (since 1950), while less than 0.5 percent of the Arctic had a top-three coldest winter.

Regions notably colder than average included Alaska and the western and central Canadian Arctic as well as the Eurasian Arctic from Scandinavia eastward through central Siberia. The eastern Canadian Arctic, western Greenland and especially the Russia far eastern Arctic had a mild winter. More than 60 percent of Arctic lands were colder than the 30-year average, the most since 2012-13.

For Alaska, this was second coldest winter in the past 25 years (2019-20 colder), while the Yukon Territory had the coldest winter since 2008-09 and the Northwest Territories the second coldest since 1993-94 (2021-22 slightly colder).

Comparatively warm areas included Qikiqtaaluk (Baffin Island) portion of Nunavut, Canada, where this was fourth mildest winter since 1950. In the Russian far northeast, the Magadan Oblast (north of the sea of Okhotsk) had by far the warmest winter since (at least) 1950 and Chukotka (easternmost Russia, west of the Bering Strait) had the third mildest winter since 1950.

Fig. 2 Winter (December-February) 2025-26 average temperature departure from the 1991-2020 baseline average. Data from ERA5 courtesy of ECMWF/Copernicus.

Overall, the Arctic-wide winter average temperature was about 0.5°C above the 1991-2020 baseline and very close to the winters 2021-22 and 2022-23 (Fig 3). Although unremarkable for the past decade, this winter was still milder than any prior to 2007.

Fig. 3 Winter (December through February average surface temperature 60° to 90°N (lands and seas) each season 1949-50 to 2025-26. Time varying average estimated via piece-wise linear regression. Data from ERA5 courtesy of ECMWF/Copernicus.

Precipitation in February (Fig. 4) was below the 1991-2020 baseline average over much of the Arctic Ocean and from eastern Greenland to the Nordic Arctic, while most of the North American Arctic had above average totals. Canada’s Yukon Territory had the third highest February total since 1950, though it was not nearly as wet as February 2022. Overall, precipitation for the Arctic was almost 10 percent below the 1991-2020 baseline average, notable primarily because most Februarys since 2015 have had significantly above average precipitation.

Fig. 4 February 2026 total precipitation categories based on percentiles of the 1991-2020 reference period. Data from ERA5 courtesy of ECMWF/Copernicus.

Winter 2025-26 precipitation

Precipitation for the full winter season (Fig. 5) generally followed pattern of February, with above normal across much of the North American Arctic but well below normal totals from central Greenland eastward through the Nordic Arctic and into western Siberia. Overall, precipitation was about 5 percent below the 1991-2020 average, though in recent years, winters 2012-13 and 2020-21 were drier.

Fig. 5 December 2025 through February 2026 total precipitation categories based on percentiles of the 1991-2020 reference period. Data from ERA5 courtesy of ECMWF/Copernicus.

Regionally, Iceland had the second driest winter season since 1950 and Svalbard the driest since 2004. In contrast, Alaska’s North Slope had the second highest winter precipitation (only 1968 higher), Canada’s Northwest Territories the third highest and Yukon Territory fifth highest.

Large scale patterns in the atmosphere and the differences from normal (Fig. 6) help explain the overall pattern of temperatures and precipitation around the Arctic this winter. Aloft (Fig. 6 left), high pressure dominated over the Greenland to Svalbard region and especially the Bering Sea. This correlates especially well with areas that had below normal precipitation At the surface (Fig. 6 right), higher pressure than average correlates with the below normal temperatures across mainland Alaska and northwest Canada as well as the Nordic Arctic.

Fig. 6 December 2025 through February 2026 left: average 500 hPa heights (contours) and departures from 1991-2020 baseline average (colors) with arrows showing the geostrophic wind direction; right: average average sea level pressure (contours) and departures from 1991-2020 baseline average (colors). Data from the NCAR/NCEP R1 reanalysis data courtesy NOAA/ESRL.
Technical Details

Table of Contents

As usual in this newsletter, Arctic means “poleward of 60°N” unless otherwise specified.
ERA5 monthly temperatures and precipitation data is available here.
Code by B. Brettschneider/NWS Alaska Region allowing rapid ERA5 regional analysis is invaluable for my work.
ERA5 monthly temperature and precipitation data available through the Copernicus Climate Data Store, here.
The NCEP/NCAR reanalysis is available here and remains a useful near-real time tool for large scale, free-atmosphere evaluation. However, this data set schedule being terminated.

Related Posts

Leave a Comment