Crazy Cryosphere: Record Low Sea Ice, An Overheated Arctic, and a Snowbound Eurasia

There are weather and climate records, and then there are truly exceptional events that leave all others in the dust. Such has been the case across Earth’s high latitudes during this last quarter of 2016, on track to be the planet’s warmest year on record. Sea ice extent and area have both plummeted to record lows for this time of year in both the Arctic and Antarctic. Such dramatic losses rarely occur at the same time, which means that the global total of sea ice coverage is phenomenally low for this time of year. The weirdness extends to midlatitudes: North America as well as the Arctic have been bathed in unusual mildness over the last several weeks, while Eurasia deals with a vast zone of above-average snowfall and below-average temperatures. Let’s look at each of these to see what’s up and where they may (or may not) be related.


Figure 1. Global sea ice area, including both Arctic and Antarctic. Sea ice extent is typically larger than sea ice area because it includes all data cells with at least 15 percent ice coverage (see NSIDC definitions). Global sea ice extent is experiencing a similar departure from average as global sea ice area. Experts usually analyze Arctic and Antarctic sea ice separately rather than together (see discussion below). Image credit: Wipneus, using data from National Snow and Ice Data Center. (NSIDC was not involved in producing this image.)


Figure 2. The normalized value of global sea ice area as of November 17, 2016, was so far from any other total in the 37-year record that it represented a departure of about 8 standard deviations below the average! Image credit: Wipneus, using data from National Snow and Ice Data Center. (NSIDC was not involved in producing this image.)


Figure 3. Departures from the 1981-2010 average for sea ice extent, in millions of square kilometers, across the Arctic (blue) and Antarctic (green) in the year 2016 through November 17. The departures from average were almost equally large by mid-November, leading to a total global sea ice extent of more than 4.2 million sq km below average. Image credit: Zachary Labe, based on data from the National Snow and Ice Data Center. (NSIDC was not involved in producing this image.)

Strange times at both poles: coincidence or connection?
Round-the-clock darkness usually forces a rapid growth in sea ice across the Arctic by November, but that process has been much slower than normal over the past month or so. There is now far less mid-November sea ice in the Arctic than in any other year since satellite records began in 1979. For the five-day average ending November 17, the difference in Arctic sea ice extent between this year and the next-lowest year (2012) was 582,000 square kilometers, an area about a third larger than California. It’s an especially dramatic example of the long-term decline in sea ice across the Arctic that’s been evident for upwards of 20 years.

Experts agree that the laggard sea ice this month around Antarctica is a separate matter from the Arctic, because sea ice in the northern and southern polar regions is produced by two markedly different circulation regimes and geographies. “At NSIDC, we generally frown on the practice of looking at the global sea ice extent,” said Mark Serreze, director of the National Snow and Ice Data Center, “the reason being that ice in the two hemispheres tends to behave rather differently; while Arctic extent shows clear downward trends in all months, the pattern for the Antarctic has been much more complex.” Serreze and several other ice experts I contacted agreed that there was no obvious explanation for why sea ice extent would suddenly dip in unison in both the Arctic and Antarctic when the two processes are typically so uncorrelated. Previous record-warm years didn’t behave this way. Could some previously dormant or absent connection be popping up just now? If so, it’s not an obvious one. NSIDC’s Ted Scambos: “I’d say that to link the two poles with a single causality chain at the seasonal/annual level is probably about a decade of research in the future.”

Unlike the Arctic, sea ice extent around Antarctica has actually shown a slight increasing trend over the last couple of decades. This might seem odd in a global climate that’s warming, but there are several plausible explanations, as we discussed in an October 26 post. Just two years ago, in September 2014, Antarctic sea ice extent hit the highest values observed at any time of the year since monitoring began in 1979. We’re now seeing the lowest values on record for mid-November, and the margin between this year and all other years has been increasing. For the five-day average ending November 17, the difference in Antarctic sea ice extent between this year and the next-lowest year (1986) was an enormous 1.12 million square kilometers.

An Arctic that’s having trouble cooling down
Temperatures north of 80°N smashed records for warmth throughout the winter of 2015-16. Now they’re on an even more torrid pace. In mid-November, temperatures across the high Arctic spiked to readings more typical of September, about 40°F above average for this time of year (see Figure 3 in our November 17 post). “Continued persistence of this pattern may significantly affect sea ice thickness into 2017,” tweeted Zach Labe (@Zlabe, University of California, Irvine) on Monday.

It’s difficult to measure sea ice thickness and volume in a continuous way, but the University of Washington’s PIOMAS model, which estimates sea ice volume using the available data, dove into record-low territory this month, just weeks after a rapid refreeze took place early in the autumn. “Whatever the respective roles of natural variability and [anthropogenic global warming], these wild swings do not inspire confidence in a semi-stable system,” noted Neven Acropolis in an early-November update on the Arctic Sea Ice Blog.


Figure 4. The huge contrasts between a far-warmer-than-average Arctic and a much-colder-than-average North Asia are projected to continue for the period November 18 - 22, 2016, as forecast by the GFS model on Thursday, November 17. Shown are anomalies (departures from average) in degrees Fahrenheit (top of legend) and Celsius (bottom of legend). Image credit: ClimateReanalyzer.com, University of Maine.

The atmospheric circulation this autumn has favored southerly flow from the eastern Pacific into the Atlantic, which has pushed recurrent bouts of unusually warm air across North America and the North Atlantic into the Arctic. One focal point of the warmth has been the Kara and Barents Sea, north of Scandinavia and western Russia, where sea ice has seen little or no expansion for November thus far. Temperatures in Longyearben, Norway--Earth’s northernmost permanent settlement, at latitude 78°N--have varied between about 7°C and -4°C (45°F and 25°F) since October 22. The average high and low for November 15 are about 12°F and 0°F.


Figure 5. Temperatures in Svalbard, Norway (in degrees C) from October 2015 to October 2016, including daily highs and lows (spiky line) and a running average (smoothed line). Readings above freezing (0°C) are shown in red, with readings below freezing in blue. Only a couple of days in the entire past year have stayed below average (black curve), and only four days this autumn through November 17 have failed to get above freezing. Image credit: Norwegian Meteorological Service.


Figure 6. Daily air temperature (highs and lows averaged) at Vize Island, Russia, in the northern Kara Sea (latitude 79°N) have cooled very little since August. The temperature on November 15 was close to freezing (32°F), compared to an average for the date of around -1°F. The island is experiencing some of the most extreme coastal erosion on Earth, as permafrost melts and stronger winds and waves reach the area. Image credit: Richard James, World Climate Service, via Brian Brettschneider, @Climatologist49.


Unusual cold and snow in Eurasia
The only place in the middle and high latitudes of the Northern Hemisphere that’s been consistently cold and snowy this autumn is Eurasia (see Figure 4 above). It’s as if the hemisphere’s entire allotment of chilly, snowy weather has been rounded up and consigned to one area, albeit a big one. For this, we can credit or blame what’s called a “wave one” pattern, where the upper-level circulation around the North Pole is dominated by a single elongated loop, shunted in this case toward the Eurasian side. Although the cold in Eurasia hasn’t been enough to balance the warmth elsewhere, it’s been quite dramatic on its own terms. On November 9, Stockholm, Sweden, experienced its heaviest one-day November snowfall (39 cm or 15.4 inches) in records going back to 1904. Across Siberia, October produced what appears to be the greatest snow extent for the time of year since 1998, and some areas got record totals for so early in the season, according to a weather.com report. In central China’s Hubei province, hundreds of homes were damaged and thousands of power poles were brought down by heavy snow during the second week of November, according to Xinhua.

By comparison, the snows over North America have been paltry indeed. On November 15, only 0.2% of the entire contiguous U.S. was covered by snow, the lowest coverage for mid-November in at least 14 years.


Figure 7. A woman walks through a record-setting snowfall on November 9, 2016, in Sundbyberg, near Stockholm, Sweden. Image credit: Jonathan Nackstrand/AFP/Getty Images.

Autumn snow cover in the Northern Hemisphere: On the increase?
Overall, it appears there has been an increase in snow cover across the Northern Hemisphere during autumn in recent years. This tendency has been largely overshadowed by the much more distinct and dramatic loss of snow cover during the Northern Hemisphere spring, and there are some good reasons. For one, because there is so much more sunlight in spring versus autumn, the loss of spring snowcover would have a much bigger effect on Earth’s radiative balance than a corresponding gain in autumn snowcover. Still, the apparent autumn trend is worth noting, especially since it may be playing a role in winter weather across North America and Eurasia.


Figure 8. Monthly anomalies (departures from average) in snow cover extent across the Northern Hemisphere, in millions of square kilometers, for the months of (top to bottom) October, November, April, and May. Image credit: Rutgers Snow Lab.


The images at right paint the general picture. Based on NOAA satellite observations, snow cover has increased slightly across the Northern Hemisphere (NH) in autumn and decreased sharply during spring. For the period 1967 - 2012, the spring trends were statistically significant (March through June) whereas the autumn trends weren’t, according to the most recent assessment report (2013) from the Intergovernmental Panel on Climate Change. Since that report came out, autumn snow extent has continued to run high across the Northern Hemisphere. The value for October 2016 was the third highest in the 51-year database, and the four prior Octobers were all well above the 1981-2010 average.

“The flakes have sure been flying in recent Octobers,” said David Robinson, state climatologist for New Jersey and leader of the Rutgers Snow Lab project. However, Robinson and other snow researchers I consulted still aren’t quite ready to classify the autumn increase as a significant trend. “My initial reaction is one of patience before being confident that there is a trend to identify,” said Robinson. “How many years that might take is the question.”

Complicating the task, ironically, is the improvement in satellite-based snow sensing technology over the years since 1967. In a 2013 paper titled “Is Eurasian October snow cover extent increasing?” (Environmental Research Letters), two scientists from Environment Canada, Ross Brown and Chris Derksen, presented evidence that the NOAA dataset was inconsistent with other sources of snow cover data for the period from 1982 to 2005, apparently due to improvements that allowed more early-season snow cover to be detected. After adjusting for this, the NOAA dataset showed a declining trend in Eurasian October snow cover through 2011, consistent with other datasets. Subsequent work along these lines has been carried out by Paul Kushner and Lawrence Mudryk (University of Toronto). “My conclusion from this is that the trends aren’t clear over the 1980s to early 2010s. I have not investigated the last few years,” Kushner told me.

A number of studies have pointed to the increasingly large expanse of open Arctic waters during autumn as a potential factor in producing heavier and/or more extensive snow cover. In particular, the open sections of the Karents and Bara seas provide a convenient source of moisture for early-season snowfall in adjacent Siberia. This is one of the key variables used to predict winter conditions across North America in the seasonal forecasting technique developed by Judah Cohen (Atmospheric and Environmental Research) and colleagues, as we discussed in an October 20 post. Importantly, Cohen’s technique doesn’t hinge on the presence of an increasingly snowy Eurasia, as Brown and Derksen point out: “The conclusion that October [snow cover extent] has not experienced significant increases over Eurasia in recent years does not undermine the arguments presented in Cohen et al (2012) linking Arctic moistening, Eurasian snow anomalies and extratropical winter cooling. This process depends on snow cover anomalies (not trends) and the physical processes involved in generating a strong surface cooling anomaly (albedo and surface temperature) depend on the areal extent and the depth of snow cover.”

In his weekly forecast update published on November 14, Cohen maintained that the above-average extent of snow this past October over Siberia should favor eventual disruption of the polar vortex and an enhanced chance of cold intrusions across Eurasia, possibly extending into eastern North America as we head into December and beyond.

We’ll be keeping an eye this weekend on slow-to-develop Invest 90L in the southwest Caribbean and will post updates as needed. See our post from earlier this morning for more details. Next week we’ll be covering the new GOES-R satellite, which at last check was scheduled to be launched at 5:42 pm EST Saturday. Sky and Telescope has details on how to watch the launch in person or online.

Have a great weekend, everyone!

Bob Henson