I wrote a post a couple years ago about the Arctic’s changing summer climate, here, so this post is partly a refresh, but I’ve also added a couple of new bits. Some of this is covered in annual reports such as NOAA’s Arctic Report Card and the various “State of the Climate” publications, so this is mostly just “my presentation style” rather than covering much new ground.

Temperatures

Summers are warming over most of the Arctic (Fig. 1), but the half century temperature changes over the Arctic Ocean are minimal. This actually is exactly what we expect: most of the summer warmth over the Arctic Ocean is going into melting snow on top of sea ice and sea ice itself, rather than warming the air. Across Arctic lands, virtually all of the Eurasian Arctic has warmed by 1°C or more in the past 50 years. In North America the warming is not as dramatic, and parts of Alaska and northwest Canada have seen only minimal change in the June through August average temperature.

Fig. 1 Change in June through August average temperature, 1976 to 2025 via Theil-Sen regression. Data from ERA5 courtesy of ECMWF/Copernicus.

The annual time series of summer average temperatures clearly illustrates the land vs. ice difference. Over land areas (including the Greenland Ice Sheet), average summer temperatures have warmed almost 2°C since the 1980s. In contrast, summer warming over the Arctic Ocean basin (defined as the seas north of the Bering Strait, Greenland and Svalbard), summer warming is only about half of the land warming.

Fig. 2 June through August average temperature, 1950 to 2025 for the Arctic land areas (red dots) and the Arctic Basin (blue dots). Time varying mean estimates via piece-wise linear regression. Data from ERA5 courtesy of ECMWF/Copernicus.

Sea surface temperatures

Looking specifically at sea surface temperature temperatures in areas that have minimal sea ice during most the summer, we find warming nearly everywhere. I’ve masked out in Fig. 3 areas where sea ice concentrations in recent years have still been high through well into August. Here, ocean water, whether exposed to the air or in contact with sea ice is very close to freezing (nominally -1.8°C)

Fig. 3 Change in average summer sea surface temperature via Theil-Sen regression. NOAA OISSTv2.1 courtesy of NOAA/Physical Science Laboratory.

Precipitation and snowfall

Precipitation and snowfall trends, like precipitation and snowfall itself, are quite variable over short distances. The most significant increases in average summer totals over the past 50 years are from Alaska’s North Slope northeastward into the Canadian Arctic Islands to northern Greenland. Trends are mostly weak in the Asian Arctic, with the most notable downward trend in central and eastern Siberia. Overall, summer precipitation has increased about four percent in the Arctic overall; across Arctic lands the increase is slightly less, at about three percent.

Fig. 4 Percent change in June through August total precipitation, 1976 to 2025 via Theil-Sen regression. Data from ERA5 courtesy of ECMWF/Copernicus.

In the Arctic, mainly north of 70°N and on the Greenland Ice Sheet (and high mountains farther south), some or even most of summer precipitation falls as snow (Fig 5).

Fig. 5 1991-2020 average percent of June through August precipitation that falls as snow. Data from ERA5 courtesy of ECMWF/Copernicus.

Summer snowfall on Arctic lands has decreased by nearly 30 percent in the past 50 years; the only significant land area with an increase is on the central and eastern Greenland Ice Sheet (Fig 6). Over the Arctic Ocean, there has been a modest increase in summer snow since the 1970s in some areas, mostly north of 75°N.

Fig. 6 Percent change in June through August total snow, 1976 to 2025 via Theil-Sen regression. Hatching masks areas where the 50-year median summer snowfall is less than 0.5cm liquid equivalent. Data from ERA5 courtesy of ECMWF/Copernicus.

Sea ice

The loss of late summer sea ice is one of the most iconic Arctic environmental changes. Figure 7 and 8 are two of my favorite ways to illustrate the changes. Figure 7 shows the typical sea ice extent on the day of the annual minimum extent (so far in the satellite era that’s always in September) in the first 11 years of the satellite record compared to the most recent 11 years. Using an 11-year median avoids any possibility of “cherry-picking” extreme years at either end. On the right side of the graphic (the most recent period), the retraction of the ice from northeast Asia and Alaska coasts is particularly dramatic.

Fig. 7 Median sea ice extent on the day of the annual minimum 1979-1989 (left) and 2015-2025 (right). All data from NSIDC.

Figure 8 shows the age of the sea ice (in years) at the end of the summer melt season in 1984 (left) and 2025 (right). The differences could hardly be more dramatic. In September 1984, about half of the sea ice had survived for four or more summers, and on the Pacific side, from the Laptev Sea eastward to the Beaufort Sea and the in central Arctic on Pacific side of the Pole, literally most of the ice was that old. Much of that old ice was lost in the 1990s and early 2000s through melt and movement out of the Arctic through Fram Strait (between Greenland and Iceland). Since about 2011 the oldest ice has been entirely confined to a narrow strip from northern Greenland to the northern Canadian Arctic Islands. Ice age is a proxy for ice thickness and the loss of so much thick ice is an important contributor to the warming Arctic, since thinner ice fractures more easily and allows for more heat to escape from the ocean water below the ice to the atmosphere above.

Fig. 8 Age of sea ice, in years, for the week of September 2-8m 1984 (left) and September 3-9, 2025. Graphics courtesy NSISC.

Technical details and underlying data sources
As usual in this newsletter, Arctic means “poleward of 60°N” unless otherwise specified except for sea ice, where “Arctic” means all northern hemisphere sea ice.
ERA5 monthly temperature, precipitation and snow data available through the Copernicus Climate Data Store, here.
Code by B. Brettschneider/NWS Alaska Region allowing rapid ERA5 regional analysis is invaluable for my work.
Sea surface temperatures from NOAA 0.25 degree daily Optimum Interpolation Sea Surface Temperature (OISST) version 2.1.
NSIDC Arctic Sea Ice information, graphics and data available here.
NSIDC sea ice age graphics are available here.

Reply

Avatar

or to participate

Keep Reading