Cryosat-2 Confirms Stunning Arctic Ice Loss

As ClimateProgress and others report, initial results are in from Cryosat-2. Since 2010, this European Space Agency satellite has surveyed polar ice to estimate its thickness, and by extension, its volume. It was a replacement for Cryosat-1, which was unfortunately destroyed in a launch failure. But the European Space Agency (ESA) considered this mission important enough to construct and deploy a replacement promptly, approving Cryosat-2 less than five months after the failure of Cryosat-1.

Initial results are now published in CryoSat-2 estimates of Arctic sea ice thickness and volume (Laxon et al., GRL, doi:10.1002/grl.50193). They compare estimates of Arctic sea ice volume during the Cryosat-2 period (2010 through 2012) to those during the ICESAT period (2003 through 2008), and to model results from PIOMAS, confirming that the stunning ice loss indicated by PIOMAS data is real. The abstract highlights this fact:

Between the ICESat and CryoSat-2 periods the autumn volume declined by 4291 km3 and the winter volume by 1479 km3. This exceeds the decline in ice volume in the central Arctic from the PIOMAS model of 2644 km3 in the autumn, but is less than the 2091 km3 in winter, between the two time periods.

It’s most interesting that the Cryosat-2 ice loss estimate during autumn is so much larger (by 1647 km^3, 62%) than the PIOMAS estimate, but the winter loss is quite a bit less (by 612 km^2, 71%). PIOMAS data already indicated that the annual cycle of ice volume has increased its amplitude, but the Cryosat-2 data suggest an even more dramatic change in the size of the annual cycle than that estimated by PIOMAS.

Cryosat-2 also indicates considerable ice volume decrease between winter 2011 and winter 2012, which they suggest may have been a contributing factor in the astoundingly low sea ice extent and volume observed at the 2012 September minimum. PIOMAS indicates no such winter decline from 2011 to 2012, although it does record the record low 2012 minimum. PIOMAS also suggests a decline between winter 2012 and this winter 2013. That could mean several things. It could presage another record minimum this coming September (I haven’t yet seen the Cryosat-2 estimates for winter 2013). Or it could mean that PIOMAS is simply “catching up to reality,” indicating the winter decline during 2013 rather than 2012 as shown by Cryosat-2. It could even mean that there is some undetected bias or uncertainty in the Cryosat-2 results, and that PIOMAS is more correct. And of course, it’s possible that the winter decline in 2012 isn’t really related to the 2012 record low minimum. Time will tell.

Cryosat-2 does indicate consistently higher ice volume than PIOMAS, typically around 20% during Autumn (Oct/Nov) and 10% during Winter (Feb/Mar).


Despite the bias between the two estimates, the decline is quite clear — and yes, properly described as “stunning.”

Another view of the sharp decline between Winter 2011 and 2012 levels in Cryosat-2 data, which is not seen in PIOMAS data, is this direct comparison of their estimates (Cryosat-2 data shown as triangles, PIOMAS estimates as circles):


Given the limited data available to calibrate ice thickness up to now, the Cryosat-2 results are a testament to the ingenuity of the PIOMAS team. Their model results gave decent estimates of ice volume, and their clearest result — the stunning loss of sea ice volume in the Arctic — is now confirmed. I expect the PIOMAS model soon will be revised by incorporating the results of Cryosat, and a newer, even more accurate, model estimate to be made available, one which has the advantage of covering more than three decades.

I also expect Arctic sea ice volume — and area, extent, and thickness — to continue to decline. Sea ice in the Arctic is warning us. Our climate, our planet, is changing and will continue to change. We’re not gonna like the result.


Click the graph!

As a reader pointed out, in the first graph the data labelled “Cryosat-2” are actually from ICESAT (the first data points) and from Cryosat-2 (the later data points).

35 responses to “Cryosat-2 Confirms Stunning Arctic Ice Loss

  1. This is a very nice post on an interesting article. It is worth noting that Seymour Laxon, the first author of the article, was tragically killed in a fall on New Year’s Day this year. He is a huge loss for the science community; this article is a fitting epitaph to him. He is much missed.

  2. Do Not Look At This Graph:

    [Response: Well howdy Dave! Why didn’t you respond at all to my post about your sea-level ideas? You seemed so eager to discuss the subject, you even made me promise not to censor your replies. But when the post appeared, you were tongue-tied. Not one word. Cat got your tongue. I guess you simply had no response. Too bad … that ship has sailed.

    Or how about the post about your crazy ideas concerning the effect of sea ice changes on albedo and insolation? You seemed so sure of yourself, until of course we actually analyzed actual data. The data which made your claims look ridiculous. But for some reason … what could that be? … again you had no response. Once again the cat got your tongue. Too bad, that ship has sailed too.

    But now you’re back! To grace us with … what? Steve Goddard?

    You should have heeded Mark Twain: It is better to remain silent and be thought a fool than to open one’s mouth and remove all doubt.]

    • The funny thing is, if “normal” is the average for 1979-2000, Goddard’s graph shows that current extent is well below normal, near but not quite 2 sigmas below normal.

      • reasonablemadness

        And it is even more laughable when you look at the *current* graph and see, what has happened the following days:

        The small uptick that Goddard wants to sell to his denier-friends as a “recovery”, “near normal” (even if nearly 2 sigma below average as you pointed out) or whatever far-from-reality-bullshit they might call it, is gone and the wiggle went the other way. So – to no surprise for everyone with a functional brain who follows arctic sea ice – the “recovery” did not happen and got canceled. The ice will not shoot through the roof and it is still thin and will decline further…

        But people like Goddard or Dave can’t grasp that. They are so trapped within their ideological belives that they can’t see the forest for the trees. It’s just sad…

    • Oops, I looked.

      I see a graph showing a startling decline in the ice: almost every day for the past four years has been significantly below average.

    • Dave, what are your predictions for volume, area, and extent for 2013 minimum?

      Also, you do realize that most of the Open Mind readers also check Neven’s sea ice graphs page regularly — yes? So, yawn. Hey, I’ll give you a free cherry: there’s actually a positive linear trend (10y) for a whole month starting at the end of March. Recovery! By 2200, the NH will be covered in ice . . . just in April (there will be no ice the rest of the year).

    • dikranmarsupial

      Perhaps dave burton could explain *why* we shouldn’t look at the graph, briefly summarise what lessons we can learn from it, and what it means for the future development of Arctic sea ice?

    • On the subject of Steve Goddard, is this as hilarious as I think it is…?
      Are we sure Steve isn’t a poe?

  3. Can’t wait to see Watt?-Me-Worry’s take.

  4. I thought he had disappeared while trying to arrest Sandy’s storm surge, because his party made sea level rise illegal…

  5. First chart might indicate IceSat not CS2 for 2006 data points. And yes, CryoSat 2 data should made available to improve all relevant models.

  6. Climate Ferret

    daveburton on February 20, 2013 at 6:56 pm
    Do Not Look At This Graph:
    Dave what message are you trying to communicate?
    1. That current ice levels are way below average.
    2. Or that you are crazy enough to believe small short variations are evidence for a forthcoming ice age?

    Where you one of the guys a few years back doing a naked dance around Steve Goddard’s ice blip? If so were you embarrassed?

    • it’s somewhat funny how similar people pointed to a similar late-spring upward blip in extent in 2012 as evidence that a ‘recovery’ was in sight, only to subsequently watch the worst melt year in recorded history.

      it’s like Groundhog Day with these guys.

  7. dikranmarsupial

    I suspect that my purely statistical prediction for NSIDC September sea ice extent of 4.1 ± 1.1 million square kilometers is likely to be optimistic as it doesn’t include any physics (such as the trend in volume). I think it is likely that the minimum will be higher than in 2012 simply due to “regression to the mean”, but the statistics strongly indicate that the decline is likely to continue.

    • I don’t ‘do’ statistics but if we go by volume rather than area or extent then there seems to be less variability around the trend and the decline is very steep, so every year except 2008 has set a new record since 2001. A new record minimum volume in 2013 would be a safer bet than record minimum area/extent.

      • dikranmarsupial

        My intuition suggests that my model is too optimistic in the estimation of the underlying trend (the volume is one of the reasons why I think this may be the case), but the extent is likely to be a higher than last year’s as last year was such an extreme event. So I think the extent is more likely to be between last years minimum and 4.1 million km^2. Fitting a statistical model to make predictions is an interesting activity, but the model isn’t reality.

    • Over at SkS I’ve been taking aim at this ‘regression to the mean’ concept, only now I see you’re a bit ‘Sk’ on this too. Tell you this, I’ve been thinking about my own predictions for 2013 since meeting you on this yesterday and I feel worried about the error bars to be included in my prediction. They would range from nil to what? Moreover, you could theoretically still hit pay dirt with your ‘regression to the mean’ re extent even if volume withers to a considerable new record low, which I do expect.

      If minimum extent drops to below 2 million square kilometers I will not be surprised at all (dare say I made that much clear, too).
      If minimum extent, on the other hand, achieves the values of 2008-2010 I’d be flabbergasted.


  8. Horatio Algeranon

    Looks like Microsoft’s Andy Lee Robinson basically reproduced (copied?) your graph simulation, Tamino.

    He made his quasi “3-D”, but it really does not add much, since yours actually looks 3-D anyway when it’s done.

    You were way ahead of the (Microsoft) curve.

    [Response: Any way to spread to word is a worthwhile effort.]

    • Well said Tamino, and I give you credit for inspiration.
      Horatio, I’m not going to let your comment go by unresponded to, the raw data is the same whoever graphs it, and here is a fairly full account how it arose. I was saving it for a blog one day, but now is as good time a time as any and stick my head above the parapet.
      I don’t think you fully appreciate the amount of effort and dedication involved, (and without any Microsoft software beyond notepad, though I used irfanview, putty for ssh, virtualdub and google chrome – everything else is on a Linux cluster).

      I’m an independent Linux system administrator and consultant, have designed server infrastructure for social networking sites and am semi-retired. I am beholden to nobody and nothing except the love of science, truth, beauty and honour. Something that all of us here share.

      Last year, I noticed a tweet by Richard Betts with Professor Michael Mann commenting on Tamino’s above piomas gif, wondering why the y-axis didn’t extend to zero, as it could possibly be seized on it as an attempt to mislead.
      I thought I could fix that easily there and then, so deconstructed the animated gif, edited each frame individually with a photoshop macro to extend the axis, reassembled and uploaded the result. Being helpful and being acknowledged is all the motivation I need.

      Scientists love dry graphs – purity conveying beautiful unobfuscated data intuitive to us, but only because we have years of training to develop the abstract and spacial abilities to interpret that data, and as those skills become natural and transparent, we become less conscious of them.

      The public, on the other hand, don’t in general have this ability to interpret graphs properly without some effort. It takes time for someone new to understand and quantify axes and build concepts required to create a contextual framework in which to assess, compare and digest the data, and as the average person now has less attention span and inclination to invest that time, the message and implications can be missed.
      (One effect of having fantastic tools like google, is that it can supplant our brains, making them lazier with a consequent tendency to atrophy with lack of exercise. All we need to know is how to find something, without actually having to remember it. There’s a study somewhere about this.)

      After modifying the 2D graph, I just wondered what it would look like in *proper* 3D. As I had plenty of time on my hands, new motivation and over 30 years of programming skills in many languages, I got the data and started to play with some ideas for fun and as an intellectual exercise, like some people do word puzzles.

      I soon realised it could make a great vehicle for communicating the magnitude of the Arctic collapse, so aimed to make something that looked professional, creative, intrinsically rewarding, entertaining and hopefully captivate a wider audience for long enough to be informed and remember. That became my new goal.
      (If there are more informed voters then there is a greater chance of eventually getting some educated politicians that take science seriously and start doing something, even if already too late).

      It took over a hundred hours of work, programming and experimenting, and was all produced using a text editor on Linux in a terminal window, using perl and php to unpack the data, control scene parameters and create scripts for the PovRay raytracer to render each frame.
      In order to produce the video in less time than it takes for a glacier to melt, I wrote a task scheduling system using perl and mysql to distribute and track the rendering tasks across 6 servers to use 20 processor cores in parallel, even including 3 cores from my remote webserver, After all that, and if I was happy with the result after the Nth time, I could assemble the images into a video using a command line tool called ffmpeg and upload to youtube. I then mentioned it on Neven’s Sea Ice blog for feedback, and the rest is history:

      I had no idea at the time, but this was to be very serendipitous. A couple of days later, BBC Newsnight found it, and included in their report. I saw it and promptly fell out of my chair – gobsmacked! It had reached a huge audience, and rattled a *lot* of cages with the effect that we are all rooting for – to be heard loudly and truthfully because mainstream media is still tiptoeing around of the herd of elephants in the room because of the fear of special interests and change, not realising that survival is not compulsory.

      Arctic ice melt ‘like adding 20 years of CO2 emissions’, Susan Watts, Newsnight Science editor, BBC News

      It soon spread and appeared on other sites:
      and on Yale forum by the revered Peter Sinclair
      A New Climate State: Arctic Sea Ice 2012

      Pondering the Path To an Open Polar Sea, Andrew Revkin
      A sobering take on Arctic sea ice (VIDEO)
      Arctic Has Lost Enough Ice to Cover Canada and Texas

      With hundreds of hours already invested in the code, it is now easy to automate with new data as it comes out, and could just attach it to a cron job and make one every month, but that may be excessive. I’ll make more over the coming years automatically closer to the next minimum.

      Meanwhile, while poring over the data last week, I found a concise and shocking way of displaying all the months’ trajectories together in Death Spiral form, bringing the end points in sharp focus, illustrating the geometric collapse of the minimum, and alluding to the 11th hour.
      It is going viral and has probably been seen by a million people by now – but still another 6,999,000,000 people to go…
      I’m sure you’ve already seen it by now, as well as the ice cube comparison over New York.
      So, I’m doing my bit in helping to communicate, and I have other animation projects in the works, to be revealed.

      Tamino, you do a great job in communicating the science, and we’re all in this together using whatever skills we have at our disposal.
      Keep it up!

      • “Being helpful and being acknowledged is all the motivation I need.”

        Well, I for one think this is all very, very helpful indeed. Thank you–many times over!

      • Horatio Algeranon


        Thanks for all the work.

        Sorry for minimizing it.

        Had no idea what was involved.

        Your rotating 3-D graph actually looks very much like some visualization software Horatio worked with over a decade ago and the assumption was made (mistakenly) that yours was actually simply a “canned” use of that kind of thing.

        Apologies for that.

      • Well, if the Microsoft quip is really applicable, then isn’t the correct phrase “embrace and extend”. :-)

  9. In the first graph the units on the yaxis are km^3, not thousand km^3.

  10. This thing from dave is obviously another LASPAB, “Look ! A squirrel playing a banjo” diversion – as it was brilliantly demonstrated in this post from Doctor Inferno :

  11. *(Sorry I realise this is off topic but the Rose article is closed for comments)

    OT: The 16 year old whacked mole just turned 17!

    Graeme Lloyd at The Australian decided that enough time has passed for the 16 year “no warming” claim to go to 17 years. No maths, no analysis – he just made it up. What’s more, he vaguely attributes it to The Met Office and Pachauri:
    “RAJENDRA Pachauri says a 17-year pause in global temperature rises would need to last “30 to 40 years at least” to break the long-term global warming trend.”

    The last time he opened his trap, his paper had to formally apologise to a sea level specialist, publicly retract its claim and write a follow up article. Here’s hoping it happens again before the mole gets a year older.

    It’s kinda fascinating. Like a biologist watching mitosis under a microscope, you can actually watch denier science evolve before your eyes.

    • (…oh and if you can’t read the article you can google the heading and click the search result link)

    • Certainly in the GISTemp data this is not true. Raw linear regressions from R lm routine of the annual means (Jan-Dec) are significant at the following years:

      Time period | Observed trend | probability
      1994 -> present .15deg/decade .0003
      1995 -> present .13deg/decade .002
      1996 -> present .13deg/decade .005
      1997 -> present .10deg/decade .02
      1998 -> present .09deg/decade .06 (NS)
      1999 -> present .13deg/decade .02
      2000 -> present .10deg/decade .07 (NS)
      2001 -> present .05deg/decade (NS)
      2002 -> present .03deg/decade (NS)
      2003 -> present .04deg/decade (NS)
      2004 -> present .05deg/decade (NS)
      2005 -> present .01deg/decade (NS)
      2006 -> present .08deg/decade (NS)
      2007 -> present .12deg/decade (NS)
      2008 -> present .29deg/decade (NS)
      2009,2010,2011-> present also all trend >0 (one just barely).

      Autocorrelation not taken into account.

      Anyway, it appears that global warming continued until 16 years ago, then magically disappeared 15 years ago, then magically again reappeared 14 years ago before magically disappearing again.

  12. Philippe Chantreau

    I fail to see what Dave’s point could possibly be. I guess in his deluded mind the fact that winter extent right now is not below all preceding winter extents on record is proof of something. Thought processes that crude are, well, words fail me. It would be nice if people like this could simply be ignored.

    • I imagine that in soccer, there are people who are proud of scoring an own goal, because it’s the only goal they score in their career.

      I’m thinking Dave Burton’s in that category.

    • “…the fact that winter extent right now is not below all preceding winter extents on record…”

      Right; it’s only the sea ice volume loss that has been monotonic.

      • I’m sure the NSDIC has good reason to define “covered” as 15% covered, but living at the southern end of the pack as I do, I REALLY can see the difference between the amount of ice out there when it is 15% covered and when it is 100% covered. Somehow, and especially as the pack thins I think this ought to become more of an issue, there needs to be some better notion of coverage developed. 15% of ice blocks is one thing. 15% spread out, thin icepans is another.

        [Response: There’s also sea ice area, for which daily data are available from Cryosphere Today.]

  13. Re. Comment…yeah, I know. I guess where I’m at is what measures get at the 2D aspects of ice packs (i.e., coverage w.r.t. air/surface interface)) best. The most oft quoted one now are the NSDIC 15% or more maps. I’m wondering if the changing nature of the pack–particularly thickness and age–may necessitate a rethinking of what measure best captures what coverage actually means. For example, thin ice pans often differ in transparency to my eye from old blocks of sea ice which are quite honeycombed (unlike glacial ice which is quite transparent).

    That said, I am totally ignorant, only observant that the nature of that 2D interface appears to be changing in my little area..

  14. John, I think that this is somewhat historical in that the notion of ice “extent”–which is what you are describing–made sense when most of the ice was within a relatively ‘solid’ pack with a fairly definite ‘ice edge.’ As that becomes less the case, the idea of extent loses some of its utility.

  15. I copied the icesat/cryosat/piomas comparison into Nevens new arctic sea ice forum (,8.msg890.html#msg890) – I hope You don’t mind, otherwise I delete it of course again.