Arctic Sea Ice Primed for Phenomenal Melt Season

The Arctic Ocean’s coating of sea ice—now remarkably thin and sparse after a record-warm winter—could plummet by late summer to the lowest extent in 38-plus years of observations. Weather conditions over the next few weeks will determine just how much melting ultimately occurs. However, the ice is so depleted that even a melt season from here on that’s average by recent standards could leave the ice at a record-low extent.

Thus far, the 21st century has seen two grand dips in Arctic sea ice extent, in 2007 and in 2012. In both cases, the ice saw a modest but incomplete recovery, and the 38-year observation period can be easily split into “before” and “after”: each year after 2007 has seen a lower minimum than each year beforehand.

Figure 1. Arctic sea ice extent in 2007 (dark blue), 2012 (dashed dark green) and 2017 (light blue). The heavy grey line is the 1981 – 2010 median ice extent. Regions lying below the light grey shaded area are below the lowest 10% of measured values. This year has featured a lower extent than 2007 and 2012 thus far, but other years—2010, 2015 and 2016 (not shown)--have featured even less extent in early June. Image credit: NSIDC.

The sorry state of Arctic sea ice right now isn’t readily apparent if you look only at extent (the amount of ocean covered by at least 15% concentration of ice). On Wednesday, the National Snow and Ice Data Center reported that Arctic sea ice extent in May was the fourth lowest on record. That brought an end to the string of record-low monthly values set or tied in January, February, March, and April. Colder-than-usual weather at high latitudes helped keep the ice extent from shrinking in May as much as it typically does.

Sea ice is more than skin-deep, though. The total 3-D volume of Arctic sea ice is the least on record for this time of year in estimates from PIOMAS (the University of Washington’s Pan-Arctic Ice Ocean Modeling and Assimilation System). Since the ice’s horizontal extent isn’t quite at a record low, this implies that the average thickness is exceptionally small right now, as suggested by PIOMAS.

The fear among Arctic observers is that early-summer conditions could take advantage of that thin ice and produce a dramatically accelerated melt at some point in the next few weeks, perhaps leading to a “cliff” (a striking drop that shows up clearly on a graph of sea ice extent).

Figure 2. Arctic sea ice volume has reached record lows in every month of 2017 thus far, as shown by the blue triangles on these month-by-month plots of data from the PIOMAS system. Each of the 12 plots shows volume for that calendar month for the period 1979 through either 2016 or 2017. Image credit: Larry Hamilton.

How to melt an oceanful of ice

If you were devising a recipe to maximize a summer’s worth of Arctic sea ice loss, you’d probably start by placing a strong version of the predominant surface high pressure system, the Beaufort High, over the Beaufort Sea and Canadian Basin of the Arctic during June and July. The Beaufort High has shown intensification during recent years, and it tends to hinder cloudiness, allowing round-the-clock sunshine to weaken the ice. Melt ponds atop the ice surface are key at this time of year, as their dark color allows them to absorb more solar heat than the lighter snow and ice around them.

If the Beaufort high intensifies in tandem with low surface pressure over high latitudes of Eurasia, it can lead to a setup dubbed the Arctic rapid change pattern, or dipole anomaly, which exerts a double whammy on sea ice: pulling warm air north to attack sea ice on the Pacific side of the Arctic, and pushing sea ice on the Atlantic side toward lower latitudes.

Later in the summer, your recipe might include some big Arctic cyclones to generate high winds that fragment the weakened ice, thus allowing even more of it to melt.

Figure 3. NASA researchers investigate melt ponds atop Arctic ice on July 12, 2011, as part of the ICESCAPE field project. Image credit: Kathryn Hansen/NASA, via Wikimedia Commons.

Unfortunately, June has started out with melt-favorable Arctic weather, and the long-range ECMWF and GFS outlooks call for more of the same, with the potential for a dipole pattern rearing its head. In a June 4 post at the Arctic Sea Ice Blog (ASIB), Neven Acropolis viewed the weather outlook with dismay: “I can’t imagine a worse forecast at this time of year. The past couple of melting seasons were marked by relatively weak preconditioning and thus a lack of melting momentum during July and August. Last year the Arctic clearly dodged a bullet because of it, but this year we may not be so lucky, given the record warm winter, record low sea ice volume and now open skies to cause plenty of melt ponding on that massive expanse of first-year ice.”

Jim Pettit (who goes by the handle Neapolitan at WU) shared similar concerns in an email on Wednesday: “With the current record-low volume coupled with the not-record-low extent, we can assume that the bulk of the ice is thinner than usual. Satellite images seem to bear this out; it appears that much of the ice has already cracked up, and that which is still together is nevertheless frangible and fractured, and not likely to last until mid-September. The next 100 days or so are going to be very interesting indeed.”

Figure 4. The amount of loss of total sea ice volume one might expect through the summer of 2017 by starting with the May 31 value (as estimated by the PIOMAS system) and extrapolating the melt curves that were actually observed in each year from 2003 through 2016, to get a sense of how much or how little melt could be reasonably expected from this point on. Because volume is running so low right now, even an “average” June-to-September volume loss for the period 2003-2016 (dashed red line) would produce a minimum volume lower than the record value observed in 2012. Image credit: Jim Pettit, with data from Polar Science Center/APL/University of Washington.

Zach Labe (@ZLabe), a doctoral student researching Arctic sea ice at the University of California, Irvine, said in an email: “This is a challenging/transition time of year before the dominant weather pattern has set up across the Arctic Ocean, so I think it is a wait-and-see time in regards to watching how sea ice extent unfolds. If we do see a sustained dipole pattern, I would expect unusually steep drops in extent this summer.”

Of course, the $64 million question is what the atmosphere’s next move will be. “What will this summer’s weather bring? No idea,” NSIDC director Mark Serreze acknowledged in an email. He added: “The melt season has started off in a deep hole. The ice decline in May was overall rather unspectacular, but the rate of loss has really picked up in the past few days.” Ice in the Chukchi Sea northwest of Alaska has been melting especially quickly, he noted. “Yes, it was a warm winter in that area, but I suspect this has a lot to do with the oceanic heat flow into the Chukchi Sea through the Bering Strait.”

Will weather save the day or sound the alarm?

As vulnerable as the Arctic’s sea ice appears now, it’s entirely possible that it will get through summer without experiencing a record-low extent. Think of a tinder-dry forest in June: it’s primed for potential wildfire if the summer is hot and rainless, yet a cool, wet summer would reduce the risk considerably. Likewise, the weather this summer could help protect the Arctic ice that remains—or it could push the ice into a state that’s unprecedented in modern times.

At least one model-based technique offers some hope for a less-than-record-smashing melt. Steven Cavallo and Nick Szapiro (University of Oklahoma) and William Skamarock (National Center for Atmospheric Research) have combined several disparate models into a new approach to sea ice prediction. Their technique draws on a 50-member ensemble of long-term ocean-atmosphere-ice simulations from the Community Earth System Model (CESM). After identifying the ensemble member that most closely matches the state of conditions on June 1, the team then uses the high-resolution Model for Prediction Across Scales (MPAS) to refine the ensemble member’s lower-resolution solution for the summer. MPAS does not necessarily need to predict exactly where and when a given surface feature will be; the key feature in terms of sea ice is MPAS's potential to capture the overall nature of where and when storminess will be most prevalent across the Arctic in a given summer. “This is very experimental, but the methods are physically based,” said Cavallo.

In a “hindcast” test of summer 2012, again starting with June 1, the CESM-MPAS duo successfully called for a parade of ice-grinding Arctic storms in late summer, which helped produce a record-low extent. For 2017, the just-completed CESM-MPAS outlook is calling for an ice-depleting pattern dominated by surface ridging in June and July and more frequent surface cyclones in August. The outlook calls for a minimum extent of around 4 million sq. km. which would be very low but still above the 2012 record of 3.62 million sq. km.

The CESM-MPAS forecast is being submitted to the Sea Ice Prediction Network for inclusion in its first batch of Sea Ice Outlook contributions from various sources. These June outlooks, which typically number several dozen, should be released around the end of the month. Each outlook gives a prediction for the average Arctic sea ice extent in September, which is typically the month with the lowest extent. This year, for the first time, the network will also compile outlooks for the Antarctic, where sea ice typically peaks in September.

As for sea ice volume, the go-to source for year-round data is PIOMAS, the Pan-Arctic Ice Ocean Modeling and Assimilation System based at the University of Washington’s Polar Science Center. PIOMAS incorporates observations of sea ice thickness gathered by satellite—which are complicated by a number of factors, such as snow cover and less-than-year-round coverage—into a modeling system that calculates anomalies in sea ice volume (departures from the 1979-2016 average) on a daily and monthly basis. Starting this month, the PIOMAS team is planning to provide mid-month summaries during the summer melt season on top of its standard end-of-month summaries.

Figure 5. Predictions of September 2016 sea ice extent for the Arctic compiled through the Sea Ice Outlook, categorized by type of forecast method (colors) and month issued (June, July, and August, from left to right within each color block). The actual observed sea ice extent for September 2016, as reported by NSIDC, is shown by the horizontal gray bar. The outer edge of each plot shows the highest and lowest forecast value within that forecast type. Image credit: SEARCH/SIO.

Next steps on the polar-midlatitude connection

Meanwhile, polar scientists will be gathering in Colorado next week for an Aspen Global Change Institute meeting focused on polar amplification, the accentuation of global warming at high latitudes. Even though this amplification is clearly seen in observations as well as depicted by climate models, there are still questions about the mechanisms at work.

High on the agenda in Aspen will be a set of coordinated modeling experiments now taking shape: the Polar Amplification-Model Intercomparison Project (PA-MIP). “There has been great progress on designing these experiments,” said Xiangdong Zhang (University of Alaska Fairbanks). One hoped-for benefit is to reduce uncertainties involving the impact of Arctic warming on midlatitude weather, as we discussed in a post earlier this year.