History of Arctic Sea Ice, part 2

Since satellite observations began, the extent of sea ice in the arctic has declined dramatically; this year the summer minimum extent was the 3rd-lowest on record. There is also considerable observational evidence that arctic sea ice extent over the last few decades is much lower than it has been for at least a century prior to modern times.

Yet to put such dramatic change in context, one hopes to stretch even further back in time. A team of researchers led by Leonid Polyak of Ohio State University’s Byrd Polar Research Center has culled the available evidence from hundreds of studies of proxy data for sea ice extent. As reported at Science Daily, in a recent paper in Quaternary Science reviews they report their findings: that the present extent of arctic sea ice is at its lowest for at least several thousand years.

The array of proxy evidence is impressive, including plant and pollen remains (both marine and land), marine sediment cores, ice cores, ice-rafted debris, organic compounds (including specific biomarkers) in sediment, coastal conditions, driftwood remains, bones of marine mammals, and historical records, which while highly uneven, in some locations go back much farther than one might expect.

What story does all this evidence tell?

The period considered by Polyak et al. begins about 50 million years ago, just after the Paleocene-Eocene Thermal Maximum (PETM). At that time the arctic was much warmer than today, with summer temperatures estimated as high as 24 deg.C (75 deg.F!). Of course the Arctic was dramatically different, “This warm environment is consistent with forests of enormous Metasequoia that stood at the same time on shores of the Arctic Ocean – such as on Ellesmere Island across low-lying delta floodplains riddled with lakes and swamps.”

But such warm conditions didn’t persist, in fact the consistent trend over the last 50 million years has been cooling. This is testified by the steady progression of changes in land plants, especially in forests, from warmer to cooler to colder-tolerant forest species, and in northernmost areas eventually to none. As early as 47 million years ago there’s evidence of the onset of drifting sea ice and possibly even Arctic glaciers. Overall cooling continued throughout the remainder of the Eocene (up to 34 million years ago) as atmospheric CO2 concentration declined, culminating in an abrupt drop in CO2 and temperature at the Eocene-Oligocene boundary. This triggered the massive glaciation of Antarctica and possibly the first Greenland glaciers. By the early Miocene (about 23 to 16 million years ago), cool-temperature trees populated the high arctic, with much of the Canadian Arctic Archipelago hosting forests not unlike modern-day southern Maritime Canada and New England.

The cooling pattern continued as atmospheric CO2 declined, until by 13 or 14 million years ago sea ice may have persisted year-round, with debris rafted by sea ice over distances which would have taken at least a year to traverse. Shortly afterward pine forests begin to dominate in northern Alaska, and the Antarctic glaciation expands. Yet even in the early-to-middle Pliocene (about 5 to 3 million years ago) conditions were still much warmer than today, with persistent pine, birch, spruce, and larch along the arctic coast. Such flora are inconsistent with extensive perennial sea ice, as are foraminifers and ostracodes in sediment cores from that time.

In the late Pliocene (3 to 2.6 million years ago) arctic cooling wrought profound changes. The tree line retreated from the coast, and ice sheets of continental scale began to grow around the arctic boundary. Yet there are warm intervals throught the Pliocene and even past the beginning of the Pleistocene epoch (2.6 million years ago).

The Pleistocene epoch ushers in an era of cooler temperatures, and of the extreme climate swings commonly (but mistakenly) referred to as “ice ages,” driven by variations of earth’s orbit and axial tilt. The colder episodes are the glacial periods, with much colder temperature and massive ice formations, while the warmer periods between them are interglacials. During glacial periods ice dominates the arctic, in fact there’s evidence (but by no means certainty) that ice as thick as several hundred meters could have covered all or part of the Arctic Ocean. But during warm interglacials sea ice is much reduced; as recently as the last interglacial (about 100 thousand years ago) microfossils indicating seasonally open water are found in the central arctic, and most of the Arctic Ocean may have been ice-free in summer.

The present interglacial has been with us for about eleven and a half thousand years, and marks the commencement of the Holocene epoch. Its warmest period seems to have been the early holocene, around 8.5 to 6 thousand years ago. Orbital factors which warm the arctic peaked about 11 thousand years ago and have steadily declined since then, but warming in many areas was partly suppressed by lingering ice sheets. Yet the story is more complex than just warming and cooling, more or less ice, as many of the changes and their timing are not pan-Arctic but regional.

Along the northeast coast of Greenland, remnants of wave-generated beach ridges show that at summer minimum the sea ice could retreat up the northeast coast as far as 83 deg. N latitude, although ice persisted year-round along the northern stretch of coast. Still, that’s considerably less ice than at present; these days it remains icebound above 74 deg. N. After about 6 thousand years ago, the region of permanent sea ice creeps down the Greenland coast. About 5.5 thousand years ago ice shelves begin to form along Ellesmere Island, indicating that the ocean to the north had very high ice concentration, a condition which spread to the northeast Greenland coast by 3 thousand years ago.

But in other regions, the pattern was different. In the Canadian Arctic Archipelago warming peaked earlier, around 10 to 9 thousand years ago, opening up the northwest passage sufficiently for Bowhead whales to move freely between Atlantic and Pacific oceans. After about 8 thousand years ago the passageway appears to be blocked (or at least difficult), although during regional warmth around 5 thousand years ago it seems to have opened for Atlantic but not Pacific populations to reach the central Canadian Arctic Archipelago. Meanwhile, the North Atlantic and Nordic Seas show a very early Holocene minimum in ice cover, from about 11 to 10 thousand years ago, followed by increased ice, then another minimum around 6 thousand years ago. In one interesting case Polyak et al. note that close to the Alaska Chukchi coast more open-water conditions occurred at the same time as higher-ice conditions further north, illustrating the complexity and inhomogeneity of sea ice changes in the early Holocene.

Over the last few thousand years we can detect centennial variability in arctic climate. It seems to have been more subdued than climate variability in the northern hemisphere as a whole, although events such as the medieval warm period have left signs; a final peak in bowhead whale bones happened just before the year 1000 A.D., first in the Beaufort Sea and later in the eastern Canadian Arctic Archipelago. Biomarkers suggest reduced spring ice from about 800 to 1200 A.D., and it was at the end of this period that the Thule Inuit (hunters of bowhead among other prey) expanded their territory eastward, eventually reaching Greenland and Labrador. The subsequent decline in Bowhead abundance has since caused some hunting territories to be abandoned, and many Inuit to focus on other food sources.

Since the mid-19th century, not only has the arctic warmed more than 1 deg.C compared to the “little ice age,” it has reached its highest temperatures in at least the last two thousand years. Retreating glaciers have exposed material which hasn’t seen the light of day for millenia, while the ice shelves around Ellsemere Island declined by more than 90%, and continue to break up, after having been a stable feature of the arctic for over five thousand years. It is especially after the middle of the 20th century that the decline in sea ice becomes surprisingly rapid, and unlike so many other variations observed in the Holocene, the sea ice decline of the last several decades is not regional, but unquestionably arctic-wide. As Polyak et al. say, “On suborbital time scales, ice distributions varied in the Holocene, but no evidence exists for large, pan-Arctic fluctuations.” Not, at least, until the last several decades:

The current reduction in Arctic ice cover started in the late 19th century, consistent with the rapidly warming climate, and became very pronounced over the last three decades. This ice loss appears to be unmatched over at least the last few thousand years and unexplainable by any of the known natural variabilities.


35 responses to “History of Arctic Sea Ice, part 2

  1. I was quite interested this year in reports of a gray whale in the Atlantic. They have been thought extinct there for centuries. So the idea is that this one migrated through the Arctic from extant populations in the Pacific, facilitated by decreased sea ice. Using a whole bunch of assumptions one can look at genetic differentiation between populations and make estimates of how long they have been isolated. The estimates are poor (due to violations of assumptions) but they’re fun. With such estimates one could guess at the last time sea ice was low enough to allow a transarctic migration. Unfortunately Atlantic gray whale DNA is hard to come by (extinction sucks). Nice to know that there are plenty of other data to address this sort of question!

  2. David B. Benson

    To be more precise, during most of the
    ther ice cycled between stades (lotsa ice) and interstades (not so much ice) quite preiodically and in agreement with 40,000 cycling in obliquity:

    It has only been for the last 800,000 years or so that the more complex pattern of cycling between stades, interstades and interglacialials (even less ice) has arisen. The references and further reading in
    ought to clarify.

    • I don’t think there is any evidence that there weren’t stadial periods prior to 800 kyr ago – there is just little evidence that there were any. Interstadials are generally less than 10 kyr – and on a timescale of millions of years, that’s less than the blink of an eye.

      The clincher, of course, is that our longest ice core record end around 800 kyr ago. Coincidence? No. That’s as far back as we currently have high resolution climate data. Beyond that, climate is inferred from geology and proxies.

      It seems pretty likely that higher frequency components of orbital forcing would have caused interstadials during past ice ages, too.

      • David B. Benson

        There is good evidence for stadial/interstadial cycling from 2.6 mya until about 1–0.8 mya:

        After that the more complex 3 state transitions predominate.

  3. The current reduction in Arctic ice cover started in the late 19th century, consistent with the rapidly warming climate, and became very pronounced over the last three decades.

    Some argue that current warming is just a natural continuation of warming coming out of the LIA. No need to comment on that. Starting in the late 19th century to three decades ago, were there any scientific papers predicting a pronounced reduction in Arctic ice cover in mere decades due to “naturally” caused warming?

    • If there were, I imagine the more “sceptic” deniers would have found them by now – although we shouldn’t underestimate the capacity of deniers for failing to do the research.

      • I’ve been looking. I did find one that speculated the continued loss of glacier ice being experienced might result in a reduction of the WAIS. But no, nothing so far that says “hey, we’re coming out the LIA, so Mr. Natural says we’re going experience some “far out” warming.

    • You prompted me to dig a little. If you’re serious about this (and can get thru’ the paywall) I just came across this

      The Present Climatic Fluctuation
      1948 Hans W

      It doesn’t give what you ask for exactly but it’s a remarkable description of the mid 20th century warming period. It reads like something written by an alarmist in 2010 in the subjects it covers and the observations it makes. The difference being everything that is written as a disaster now is embraced as great news then. For example it talks about the improving Arctic ice conditions when it means less of the stuff to hinder daily life. A strangely different view of a warming world. It also contains the magnificent phrase “the climatic glory of the Stone Age”, who wouldn’t want to read about that?

      Anyway seriously it’s extremely interesting, I’d go as far as saying an eye openner, and contains heaps of references on the arctic in that period which might contain exactly what you don’t want to see. ;)

  4. Sounds to me like we passed a tipping point three decades ago.
    Methinks this climate carnival ride has left the station and we ain’t seen nothin’ yet…

    • TomG.

      More and more I find myself agreeing that we’ve already passed points of no return, although I still believe that the peri-millenium period was a window of opportunity for significant mitigation of the driving of further warming.

      Given the national and international heel-dragging that has been displayed over the last several years it is quite apparent that there will be no appreciable global action for at least another decade, by which time all efforts will only be to make the coming changes hurt as little as possible – a bit like jumping from a three storey window hurts less than jumping from a five storey window.

      I predict that in future the denialsts will, without blinking, change their mantra from “there is no global warming” to “so, we passed the warming tipping points decades ago and scientists couldn’t produce the evidence to elicit action. Fraud! Conspiracy!! Money-seeking tit-suckers!!! Murderers!!1!eleventy-one!”

  5. Darn it, I should have written this blog post. Great stuff, Tamino.

  6. This may be obvious to everyone else —

    History of Arctic (and Antarctic) Sea Ice, Part 1
    October 16, 2010

  7. I’d just like to thank tamino for such a pleasant and readable post. I really felt I had learned something about planet Earth.

  8. http://www.skepticalscience.com/Open_Mind_Archive_Index.html

    Skeptical Science decied your old posts were worth saving. Quality stands up in your older posts, a resource worth saving.

  9. “… at summer minimum the sea ice could retreat up the northeast coast as far as 83 deg. N latitude, although ice persisted year-round along the northern stretch of coast. Still, that’s considerably less ice than at present; these days it remains icebound above 74 deg. N.”
    Just a couple of random remarks:

    The northernmost point of Greenland is Cape Morris Jessup at 83.6 degrees north, so the northern stretch of coast referred to is only a couple of hundred kilometres along the north coast of Peary Land (the northernmost peninsula of Greenland). So in the early Holocene, Greenland was all but free from surrounding sea ice at minimum conditions.

    Perhaps “icebound above 74 degrees north” is normal (I couldn’t say), but in 2010, the September minimum saw ice-free conditions above 77.5 degrees north: http://rapidfire.sci.gsfc.nasa.gov/subsets/?subset=Arctic_r02c03.2010251.terra.1km
    For orientation of this satellite pic, north is at top right, the bear-shaped island in the centre is Shannon Island at 75.2 N, and the prominent peninsula halfway between Shannon Island and the top of the image is Danmark Havn at 76.8 N. There is open water at the coast for around 1 degree north of Danmark Havn, ie between 77.5 and 78 degrees north. That is about 40% of the way from recent “normal” in the direction of “warmest in last 6,000 + years”.

    (Only one year, highly variable, sensitive to fluctuating weather, very local etc etc…just remarking on it.)

  10. Thanks for the summary, Tamino. This is the logical follow-up after Klaufman et al. 2009.

    Off topic about statistical significance: this article in NewScientist says that 53.1% (of something more than 1.000 people) guessed something they couldn’t know. They say that result is statistically significant (because it’s more than 50%). I feel somewhat skeptical, because it’s so close to 50%! I tend to think that you would need to sample a lot more people to get a significant result so close to 50%; or, on the contrary, you would need a much higher percentage of guessing with just 1.000 people to conclude that it is significant. I would be grateful if you could let me know your first-glance opinion on the matter. Thanks!

  11. Ups, I’ve just realized that my off topic before cannot be answered because what matters is not the number of people but the number of experiments, which is not given in the NewScientist article, sorry. There’s a pre-print of the paper here. I think it was 100 sessions with 18 trials each. And I finish this off topic here, sorry again.

  12. Last really stupid question of 2010 (though I will not make a resolution that more may arrive in 2011!)

    During the glacial periods, i understand that the glaciers can be kilometres thick upon the land and that the oceans are thus drawn down 100’s of metres.

    My thought experiment/question – is there any chance that as he ocean levels drop, that the thick ice of the arctic would be strong enough that there could actually be open water under the ocean glaciers? You know, one really huge grand cave? If the ice could act as an insulator, could this modify the temp by the ocean surface enough to allow this to happen/

    Anywho, wishing all here, especially Mr T a wonderous and joy filled 2011!

    • I don’t understand the question.

      Remember, though, that large parts of Antarctica are below sea level, with only the ice rising above the ocean. Remember also that there is a considerable quantity of liquid water below the continental Antarctic ice.

    • Daniel Bailey

      Re: DeNihilist

      If you think of the sheer size of the Arctic you will realize that the idea of a huge “ice cave” up there, while entertaining, would be impossible. The heat released by the liquid ocean into the air cavity would eventually destabilize the ice above it, probably fairly quickly even for relatively small cavities (100 m across or less). The conceptual scale is simply too big.

      Also, the sea ice in the Arctic would be better referenced to as an “ice cap”; glaciers flow downhill (alpine mountain, Greenland and Antarctic). Sea ice in the Arctic floats on the surface (technically most is slightly underwater with just the “freeboard” rising above the surface).

      The Yooper

      • Yes–ice is not that great a structural material, and I don’t think a cantilever across the Arctic Basin would work, even with our most sophisticated alloys.

  13. Thanx for the replies gents. The crux of the thought is to do with the migration mystery to N.A. It appears that some digs in the recent past are finding sttlement dates that precede the thaw of the last glacial, yet most scientists are quite sure that an overland route through the glaciers would have been impossible.

    Besides, the image in my mind of kayaking under thousands of miles and hundreds of metres thick ice is quite compelling! :)

    • Daniel Bailey

      I’ve read some evidence of a sea route (along the edge of the lowered sea level ice-edge). As deep-time colonization of Australia and portions of Indonesia required sea travel across open ocean (even at glacial maximum), following the “kelp bed trail” up Kamchatka, Siberia, along the Aleutians & down the Pacific NW of NAmerica would have allowed for continued fishing the entire way.

      Way off topic, tho.

      The Yooper

  14. Any thoughts on how this graph could possibly be spun as “Arctic sea ice recovery”? It would be fun to see goddard tie himself in knots attempting to justify it… http://nsidc.org/images/arcticseaicenews/20110105_Figure3.png


  15. Tamino.

    “Overall cooling continued throughout the remainder of the Eocene (up to 34 million years ago) as atmospheric CO2 concentration declined, culminating in an abrupt drop in CO2 and temperature at the Eocene-Oligocene boundary.”


    “It is especially after the middle of the 20th century that the decline in sea ice becomes surprisingly rapid, and unlike so many other variations observed in the Holocene, the sea ice decline of the last several decades is not regional, but unquestionably arctic-wide.”

    I hope a follow-up post will clarify the recent disconnect with the usual previous Arctic ice behaviour (albedo change made by soot deposition).

    Best regards, Ray Dart.

    [Response: The real “disconnect” is the anthropogenic increase in greenhouse gases.]

  16. Suricat:

    I hope a follow-up post will clarify the recent disconnect with the usual previous Arctic ice behaviour (albedo change made by soot deposition).

    Which, if true, would also be an anthropogenic source, and why would this be better than the mainstream scientific explanation?

    Oh, yeah, we can then blame brown people in Asia!