In Which We Return to Arctic Sea Ice Decline and Find It Was As Bad As We Thought Seven Years Ago

For the last 3 years or so, I have rarely blogged about Arctic sea ice, because my model of how it worked seemed flawed and yet replacement models did not satisfy. 

Until 2013, measures of Arctic sea ice volume, trended downward, exponentially rather than linearly, at minimum (usually in early September).  Then, in 2013, 2014, and 2015, volume, area and extent measures seemed to rebound above 2012 almost to the levels of 2007, the first “alarm” year.  My model, which was of a cube of ice floating in water and slowly moving from one side of a glass to the other, with seasonal heat increasing yearly applied at the top, bottom, and sides, simply did not seem to reflect what was going on. 

And now comes 2016 (the year’s melting is effectively over), and it seems clear that the underlying trends remain, and even that a modified version of my model loosely fits what’s happening.  This has been an unprecedented Arctic sea ice melting year in many ways – and the strangest thing of all may be this glimpse of the familiar.

Nothing of It But Doth Change, Into Something Strange

The line is from early in Shakespeare’s Richard III, in which a character talks of his father’s drowning:  “Full fathom five my father lies/Of his bones are coral made/ … /Nothing of him but doth change,/Into something rich, and strange.”  And, indeed, the changes in this year’s Arctic sea ice saga deserve the title “sea-change.”  Here’s a list:

1. 1.       Lowest sea-ice maximum, by a significant amount, back in late March/April.
2. 2.       Lowest 12-month average.
3. 3.       Unprecedented amount of major storm activity in August.
4. 4.       Possibly greatest amount of melt during July for years when generally cloudy conditions hinder melting.
5. 5.       First sighting of a large “lake” of open water at the North Pole on Aug. 28, when two icebreakers parked next to an ice floe and one took a picture.  Santa wept.
6. 6.       First time since I have been monitoring Arctic sea ice melt when the Beaufort Sea (north of Canada and Alaska) was almost completely devoid of ice to within 5 degrees of the pole.

These events actually describe a coherent story, as I understand it.  It begins in early winter, when unusual ocean heat shows up at the edges of the ice pack, especially around Norway.  The ocean temperature (especially in the North Atlantic) has been slowly heating over time, but this time it seemed to cross a threshold:  parts of the Atlantic near Norway stayed ice free all the way through the year, leading to the lowest-ever Arctic-ocean maximum.

In May and early June, the sun was above the horizon but temperatures were still significantly below freezing in the Arctic, so relatively little melting got done. Then, when cloudy conditions descended and stayed late in June or thereabouts, the relative ocean heat counteracted some of the loss of melting energy from the sun.  But another factor emerged:  the thinness of the ice.  Over 2013, 2014, and 2015, despite the lack of summer melt, multi-year ice remained a small fraction of the whole.  So when a certain amount of melt occurred in July and August, water “punched through” in many places, so that melting was occurring not just on the top (air temperature) and bottom (ocean heat) but also the sides (ocean heat) of ice floes.  And this Swiss cheese effect was happening not just at the periphery, but all over the central Arctic – hence the open water at the Pole.

Then came the storms of August.  In previous years, at any time of the year, the cloudiness caused by storms counteracted their heat energy.  This year, the thin, broken ice was driven by waves that packed some of the storms’ energy, further melting them.  And, of course, one of the side effects was to drive sea ice completely out of the Beaufort Sea.

Implications For The Future

It is really hard to find good news in this year’s Arctic sea ice melting season.  Years 2013-2015 were years of false hope, in which although it seemed that although eventually Arctic sea ice must reach zero at minimum (defined as less than 1% of the Arctic Ocean covered with ice), we had reached a period of flat sea ice volume, which only a major disturbance such as abundant sunshine in July and August could tip into a new period of decline. 

However, the fact of 2016 volume decrease in such unpromising weather conditions has pretty much put paid to those hopes.  It is hard to see what can stop continued volume decreases, since neither clouds nor storms will apparently do so any longer.  One can argue that the recent el Nino artificially boosted ocean temperatures, although it is not clear how it could have such a strong relative effect; but there is no sign that ocean heat will return to 2013-2015 levels now that the el Nino is over.  

Instead, the best we can apparently hope for is a year or two of flatness at the present volume levels if such an el Nino effect exists, not a return to 2013-2015 levels.  My original off-the-cuff projection “if this [volume decreases 2000-2012] goes on” was for zero Arctic sea ice around 2018, and while I agree that the 2013-2015 “pause” makes 2018 very unlikely, 2016 also seems to make any “zero date” after 2030 less likely than zero Arctic sea ice at some point in the 2020s. 

A second conclusion I draw is that my old model, while far overestimating the effects of “bottom heating” pre-2016, now works much better in the “fragmented ice” state of today’s Arctic sea ice in July and August.  In this model, as ice volume approaches zero at minimum, volume flattens out, while extent decreases rapidly and area more rapidly still (unless the ice is compacted by storms, as occurred this year).  This effect will be unclear to some extent, as present measurement instruments can’t distinguish between “melt ponds” caused by the sun’s heat and actual open water.

Finally, the Arctic sea “ice plug” that slowed Greenland glacier melt by pushing back against glacier sea outlets continues to appear less and less powerful.  This year, almost the entire west coast of Greenland was clear of ice by early to mid June – a situation I cannot recall ever happening while I’ve been watching.  Since this speeds glacier flow and therefore melting at its terminus entering the sea, it appears that this decade, like the 1990s and 2000s, will show a doubling of Greenland snow/ice melt.  James Hansen’s model, which assumes this will continue for at least another couple of decades, projects 6-10 feet of sea rise by 2100.  And even this may be optimistic – as I hope to discuss in a follow-on post on 2016’s sudden CO2 rise.

Most sobering of all, in one sense we are already as near to zero Arctic sea ice as makes no difference.  Think of my model of an ice cube floating in water in a glass for a minute, and imagine that instead of a cube you see lots of thin splinters of ice.  You know that it will take very little for that ice to vanish, whereas if the same volume of ice were still concentrated in one cube it will take much more.  By what I hear of on-the-ground reports, much of the remaining ice in the Arctic right now is those thin chunks of ice floating in water. 

“Because I do not hope to turn again/Because I do not hope/ … May the judgment not be too heavy on us/ … Teach us to care and not to care.”  T.S. Eliot, Ash Wednesday