Absurd January Warmth in Arctic Brings Record-Low Sea Ice Extent

This winter’s freezing season in the Arctic is falling short. The extent of Arctic sea ice this week is hovering near record-low values for early February, based on observations that extend back to the start of satellite monitoring in 1979. Data from the National Snow and Ice Data Center (NSIDC) shows that last month had the lowest overall Arctic sea ice extent of any January in the satellite record (Figure 1). As detailed in an NSIDC report on Thursday, the total extent of 13.53 million square kilometers (5.2 million square miles) was 1.04 million sq km below the 1981-2010 average and 90,000 sq km below the record from January 2011.



Figure 1. Departures from average in Arctic sea ice extent for January, 1979-2016. Image credit: NSIDC.


Figure 2. Sea ice extent for January 2016 (white), compared to the median January location of the ice edge for the period since 1979 (magenta line). The largest areas of open water where ice is usually present are in the Barents Sea, north of Scandinavia. Image credit: NSIDC.

Only a few weeks are left before the return of polar sunshine puts an end to the freeze-up that typically starts in September and peaks in late February or March. Last year’s maximum extent occurred quite early--on February 25--and it was the lowest in the satellite record, at 14.54 million square kilometers. This year appears to have a reasonable shot at breaking that record.

The not-so-frozen North
Hand in hand with the skimpy ice cover, temperatures across the Arctic have been extraordinarily warm for midwinter. Just before New Year’s, a slug of mild air pushed temperatures above freezing to within 200 miles of the North Pole. That warm pulse quickly dissipated, but it was followed by a series of intense North Atlantic cyclones that sent very mild air poleward, in tandem with a strongly negative Arctic Oscillation during the first three weeks of the month (see Figure 4).


Figure 3. Average air temperatures over the Arctic for January 2016 at the 925-mb level (about 2000 feet above the surface), expressed as departures from the long-term January average in degrees C. Image credit: NOAA Earth System Research Laboratory.


“January was absurdly warm in the Arctic,” said NSIDC director Mark Serreze. According to data from NOAA’s Earth Systems Research Laboratory, the average surface temperature in January between latitude 60°N and the North Pole was -18.2°C (-0.8°F), topping the previous record of -20.6°C (-5.1°F) set in January 2005. Just above the surface (925 mb), the average January temperature of -14.2°C (5.9°F) was well above the previous record of -16.5°C (0.7°F), also set in 2005. The fact that average readings at this level are warmer than at the surface reflects the strong inversion typical of the lower Arctic atmosphere, especially in winter, as cold air hugs the surface and milder air flows just above it.


Figure 4. This cross section through the polar vortex (between latitudes 65°N and 90°N) shows how the height of pressure surfaces (shown in millibars/hPa on the left-hand axis) varied over time from October 2015 to early February 2016. Red values show where a given pressure surface was unusually high in the atmosphere, corresponding to warmer-than-average temperatures. The temperatures have been normalized so that the right-hand legend shows standard deviations from the mean. Image credit: NOAA Climate Prediction Center.

Anyone for bathing in the Barents?
Some of the most visible ice-extent deficits right now are in the Barents Sea, north of Scandinavia. Warm waters carried by the Atlantic Meridional Overturning Current (AMOC) can push all the way northeast into the Barents, making it one of the most variable of the seas that fringe the Arctic. Data from Crysophere Today (University of Illinois at Urbana-Champaign) show that sea ice now covers less than half the area across the Barents that it did a year ago at this time. The difference in area--roughly 350,000 square kilometers--is bigger than the state of New Mexico. (Note that sea ice area is a somewhat different index than sea ice extent, as explained by NSIDC.)

Near the northwest corner of the Barents Sea, temperatures in Svalbard, Norway--at 78°N, the northernmost civilian community on Earth--have been far above average for the entire last month. From January 5 to February 3, the average in Svalbard was -4.7°C (23.5°F), which is a full 19°F above the norm--remarkable for a 30-day period. The coldest reading of the whole period, -11.9°C, was still above the average January high of -13.0°C! Temperatures pushed above freezing on four days, reaching 4.5°C on January 24 (still not a monthly record, though). The WU forecast shows Svalbard remaining unusually mild for at least the next week, with lows at or above the average highs.


Figure 5. The area covered by sea ice in the Barents Sea as of February 4 was around 250,000 square kilometers (right side of black trace), less than half of the value a year ago (left side of black trace). The red trace shows departures from average for the time of year. Image credit: Cryosphere Today/University of Illinois at Urbana-Champaign.

The outlook for 2016
Most of the attention around Arctic sea ice has focused on the dramatic losses in summertime, especially over the past decade. The minimum yearly extent set new record lows in 2007 and again in 2012. Both of these minimums were followed by a year or two of quasi-recovery, but then the long-term trend toward lower summer minimums resumed.

Could 2016 set another new record minimum? It’s far too soon to make any confident predictions. A record-low maximum could give the ice a head start on 2016 melting, but the ultimate outcome will depend mainly on weather patterns still to come, especially in early summer. Warm southerly winds and clear, sunny skies during June and July can make a huge difference in paving the way for a record minimum in September.


Figure 6. The years 2012 (green dashed line) and 2015 (blue line) show how the state of sea ice in late winter and spring doesn’t necessarily correspond to its minimum in September. Sea ice extent in March 2012 was near its long-term average in April, yet it was at a record low in September. Late winter ice was much less extensive in 2015, but that year’s minimum was still considerably higher than in 2012. The extent for 2016 is shown in the red line at far left. Image credit: NSIDC.


Those on the front lines of experimental sea ice prediction stress the role of weather conditions in modulating how any one season will fare. In a 2015 review paper for EOS, Julienne Stroeve (NSDIC) and colleagues from the Sea Ice Prediction Network took a look at what recent efforts have accomplished and what may be possible down the line. Since 2008, the network has issued monthly compilations of sea ice forecasts (available online) updated through the summer, created by more than a dozen participants using a variety of methods. The average of these predictions tends to do somewhat better than any one approach, but even that skill is still limited, especially more than a month or two in advance of the September minimum.

“Because the atmosphere is mostly unpredictable beyond 1 or 2 weeks, the sea ice forecasts initialized in late spring may not be able to accurately predict sea ice features that develop as a result of extreme summer atmospheric conditions,” said the EOS paper.

Idealized experiments suggest that the state of Arctic sea ice might be predictable with some skill as far as two years in advance. To reach this theoretical goal, we would need major advances in both Arctic observations and modeling. There’s another catch: climate change itself may make the task harder. “Interannual variability of summer sea ice extent will likely increase in coming decades,” noted Stroeve and colleagues, “and some scientists suggest that this might lead to a reduction in predictability.”

We’ll be back with a new post on Friday.

Bob Henson