RealClimate logo

North Pole notes (continued)

Filed under: — gavin @ 22 August 2008

This is a continuation of the previous (and now unwieldy) post on the current Arctic situation. We’ll have a proper round up in a few weeks.

638 Responses to “North Pole notes (continued)”

  1. 301
    maikdev says:

    Re: 296 “Cryosphere today places 2008 almost tied with 2007 extent, the difference is so small, consider it a tie”.

    Now (Cryosphere Today): 3.199.000 km2 sea ice area
    2007 minimum (Cryosphere Today): 2.990.000 km2 sea ice area.
    209.000 km2 is the difference.

    Now (JAXA): 4.964.000 km2 sea ice extent
    2007 minimum (JAXA): 4.254.531 km2 sea ice extent
    710.000 km2 is the diference.

  2. 302
    Hank Roberts says:

    Remember not to draw conclusions from the pictures; they tell you how the pictures are made and where they get the data they use to make the pictures.

    For exampe from the main imagery page:
    * We now calculate daily maps of combined MODIS – AMSR-E data Click Here!
    o maps of MODIS RGB composites with AMSR-E ice contour lines
    o maps of snow grain size calculated from MODIS data with AMSR-E ice contour lines
    o maps of soot concentration calculated from MODIS data with AMSR-E ice contour lines
    Other daily updated products
    * Combined MODIS – AMSR-E sea ice maps
    * SSM/I sea ice maps

    You can look each of these up. Just as an example look at all the papers here about SSM/I from a simple Google search. If wossname had looked at these before counting pixels he might have used data instead:
    SSM/I and AMSR-E L1B from inside IUP

    The ASI sea ice concentration algorithm used here has been validated in several studies (Spreen et al. 2005, Spreen et al., 2008). However, no warranty is given for the data presented on these pages.

  3. 303
  4. 304
    Mark says:

    Crandles, #298.

    Since 2007 had quite a stunning minimum, the *retreat* this year could be a record and yet still have more ice extent than 2007.

    Precision is needed, my boy.

  5. 305
    Lauri says:

    RE 299
    Just visually assessing the Modis picture, it looks like one could sail around Greenland. I wonder how usual that is?
    See (of course even better with a higher resolution).
    It will be interesting to see Greenland ice mass reports once they get done.

  6. 306
    Timothy Chase says:

    Good news…

    Tamino has entered the running — a statician who (it goes without saying) has a great deal more mathematical acumen than myself. He is giving a decidedly different projection of what the next few weeks will bring in terms of sea ice extent — with this year’s minima being reached on 20th of September, much earlier than what I have projected, and with last year’s record remaining secure for the time being.

    Please see:

    More Less Ice
    September 1, 2008


    captcha fortune cookie: ol Knight

  7. 307
    Hank Roberts says:

    But note Tamino said right up front there that it’s a statistical curve, not based on field conditions:

    “Just for a little fun, I fit a 4th-order Fourier series to this year’s data and projected it into the future …”

    Wind direction, ocean currents, and temperatures thereof made last year unusual.
    Anyone know what the wind and water are doing this year? I haven’t seen much reported.
    Most ocean info comes from buoys and anchored instruments that have to be retrieved, with time lag, or from navy instruments that aren’t immediately reported.

    And have the petroleum companies started putting down instruments in the Arctic? Would anyone know?
    They’d certainly want to be collecting data in advance of drilling. How do they instrument areas, or do they?
    We do know last year’s melt left much less multiyear ice — making this year unusual.

  8. 308
    Hank Roberts says:


    The Global Ocean Observing System
    GOOS is a permanent global system for observations, modelling and analysis of marine and ocean variables to support operational ocean services worldwide. GOOS provides accurate descriptions of the present state of the oceans, including living resources; continuous forecasts of the future conditions of the sea for as far ahead as possible, and the basis for forecasts of climate change.


  9. 309
    Hank Roberts says:

    “… We now understand that an unusual weather pattern of warm winds and clear skies played a large role in 2007 melting, and we know that we started 2008 with an unusually large amount of new (first year) ice. An international group of researchers has, for the first time and starting from May, produced, shared, and compared monthly estimates of the 2008 minimum. Those groups, from 15 or more institutions, use recent, in some cases daily, satellite, ship and buoy data, climate models, weather models, and historical data in comparisons, correlations, extrapolations and estimations – you can follow their work in very interesting detail at the SEARCH Sea Ice Outlook page….”

  10. 310
    Hank Roberts says:

    OK, here’s the sort of thing I’ve been wondering about, including ice melt top and bottom info from buoys:

    These reports run several months behind the calendar because they call information from a large number of different sources and researchers and complile it. These pages have a lot of images, links and data.
    July’s report (the latest) includes the outlook for September, and so on.

    ——–excerpts follow————
    “The outlook for the pan-arctic sea ice extent in September 2008, based on July data, indicates a continuation of dramatic arctic sea ice loss. The July Sea Ice Outlook report is based on a synthesis of 22 individual projections utilizing a range of methods. Projections based on July data show no indication that a return to historical sea ice extent will occur this year.”

    “… Figure 3. Ice bottom and top melt reported by ice-mass-balance buoys.

    Figure 4 shows the present (13 August 2008) surface condition as evidenced by the web camera image from the NPEO Automated Drifting Station, the location of the ice-mass-balance buoy installation nearest Fram Strait. Unlike previous years at this time, and aside from right around the web camera buoy, we do not see many melt ponds. At this site melt pond coverage has been minimal all summer, arguably due to the limited snow cover in spring. This helps explain why the first-year, 1.9-m ice has only melted 0.2 m on the upper surface and provides one mechanism by which first year ice may survive the summer. …”

  11. 311
    Hank Roberts says:

    One last tidbit from the above July report. I’d been looking for Dr. Bitz’s work and it is part of this compilation.

    “… The following Figure from 10 August 2008 and the figures from Lars Kaleschke and Tom Agnew suggest, however, that there were large regions of low sea ice concentrations (black regions) within the boundary of sea ice extent at the end of July. While there may be some difficulty interpreting this figure due to summer melt ponding at the surface, the figure is certainly suggestive of further potential retreat in regions of thin first year sea ice. Such interpretations are important as some scenarios assume complete meltback of all first-year sea ice, while others limit meltback to ice classes thinner than a specific threshold (such as 1m). Total sea ice extent in early August declined at about twice the rate of any other time this summer (Bitz). …
    [ ]

    Text excerpt without figures from the above follows; long excerpt because this analysis discusses the statistics including autocorrelation/memory, which were mentioned recently in other threads.

    Sea Ice Outlook based on Statistics of Observed Ice Extent and
    Global Climate Model
    July 2008
    1. Name of contributor: Cecilia Bitz
    2. Estimate of the sea ice extent for the Arctic as a whole for the month of September 2008
    5.30 million square kilometers
    3. Principal method
    Statistical, based on observations and coupled climate model output.
    4. Short basis for prediction
    The 29 year observational record of September sea ice extent has zero autocorrelation at
    one-year lag and zero skew. The correlation with the extent in the prior July is significant,
    but the July 2008 extent lies very close to the long term trend. Therefore, my prediction for
    September 2008 is an extrapolation of the long term trend for September. These statistical
    relationships are in general agreement with much longer records that are available from the
    Community Climate System Model version 3, CCSM3.
    5. Longer basis for prediction
    With little deviation from the long term trend in July 2008 and no significant autocorrelation
    or skew from one September to the next in the observations (Fig. 1a), the conservative
    estimate for the future is on the trend line in September. An extrapolation of the trend line
    (Fig. 1b) to year 2008 gives 5.30 million square kilometers.
    The observational results were compared with a statistical analysis of an ensemble of
    20th and 21st century simulations and long control runs from CCSM3. With ensembles and
    multi-century control runs giving far more degrees of freedom, it is clear that CCSM3 does
    have a weak but significant autocorrelation in September ice extent from one year to the next.
    However, the autocorrelation is so weak that it did not compell me to modify my prediction
    based solely on the observations. In contrast, there is more considerable lagged correlation
    between thickness and extent, as expected owing to the much much greater memory in
    Figure 2 shows that years with September sea ice loss comparable to the 2007 observed
    loss are very rare.
    —- end excerpt—

    See links above for the real stuff, this is just a bit I grabbed as an excerpt to point to the real stuff.

  12. 312
    NeilT says:

    Hank, statistical curves are OK in their place, but the Ice area fell off a cliff in the last week and has dropped to almost the same as 2007. In fact when Cryosphere Today gets the graph to the end of the month, the current melt should be just below 2007 for that short spike.

    It’s going to be an interesting week or two, but the massive open water around Greenland and the Archipelago is unprecedented.

    Anyone know how much of the multi year ice which was in those areas survived? Did it move on or did it just melt?

  13. 313
    Chris says:

    #297 NeilT:

    You may be interested in the following comparisons. The first shows how in 1985, even by 2 weeks before this date, there was a huge amount more open water to the north of Greenland, the ice edge on the Atlantic side of the Arctic was already further north on average, the sea ice on the Atlantic side was lower in concentration, and there was already as much open water in the Laptev sea.

    The second shows how in 1993 the low (~60 per cent) concentration sea ice on the Pacific side of the Arctic reached a similar line to now i.e. from near Banks Island in NW Canada across to the Laptev Sea, with enough open water especially on the Siberian side to make today’s map seem a lot less dramatic.

    What these comparisons say to me is that both sides of the Arctic have seen significant melt before and recovered.

    “It is, today, massively lower than either 2005 or 2006”

    If that’s what you call massive, then 1993 was massively lower than 1992. What does that prove? [2005 vs 2008] [1992 vs 1993]

    “What you are missing is that 2008 is tracking 2004 SST and atmospheric temperatures with >1M Sq Km less ice. Woods and trees come to mind.”

    Or maybe one should assume that someone looking at lots of trees generally knows they are in a wood? Of course I’m aware that 2008 has ~1M Sq Km less ice, as I’ve already referred to in a previous post. My point was that there is a precedent for a rapid August melt followed by a much slower September melt – the fact that the extents are ~16% different doesn’t fundamentally change this.

    “Just how many firsts do we have to have before people understand that it is now out of control?”

    It would be “out of control” if buoys had shown continued dramatic thinning this year, the ice extent had “overtaken” 2005/6 in June rather than August allowing albedo changes to have a significant effect, and Arctic temperatures had shown significant rises.

    Instead we are in a situation where the seas and air are significantly colder than this time last year, with greater ice extent, area and even thickness, and thus a fair indication that Arctic re-freeze should set in earlier and more strongly.

    #301 CT latest area is 3.221, and the 2007 minimum was in fact 2.92. So 2008 is still ~300,000km2 ahead or ~10%

  14. 314
    Alan Millar says:

    I can’t understand all these hundreds and hundreds of posts about what is, after all, a regional weather event!

    It’s as though this is has to be emphasised as most other Global indicators of AGW are showing a negative trend and it is a good bet that this will continue for the forseeable future.

    Most modelers do not make a big issue of Artic sea ice melt as they clearly understand that the Arctic ice extent is controlled by many more factors than just Global temperatures.

    The globe is more than the NH after all. You have to factor in the SH at some stage! So why is there hardly any discussion of global polar ice extent?

    Let’s see, you cannot jump around over a few years data, when you are talking about the Earths climate.

    Arctic sea ice has increased this year Antarctic sea ice continues to show, a medium term, increase in extent.

    Global temperatures remain fairly stable with a slight cooling trend. Sea temperatures show no warming trend. There is no evidence of an accelerating sea hight increase.

    The models are not currently in agreement with actual Earth conditions over the last seven or eight years. Therefore, no need to panic, let’s wait awhile and be more certain before we leap off the cliff!


  15. 315
    LG Norton says:

    Re: #307 You cannot predict the future by doing statistical analysis on a curve that is dependent on a multitude of multivariate data.

    You predict the future by doing regression analysis on all the field conditions (solar loading, ice thickness, wind, air temperture, water temperture, ocean currents and probably a dozen others) variables.

    Then you work out problems between the variables, such as autocorrelation, multicollinearity and heterosedasticity. By then you have downed a 40oz of scotch.

    When that fails, you start fudging your data by using Principle Component Analysis or was that factor analysis, I forget which.

    Serriously, we don’t know all the factors that contribute to ice loss, and we don’t have enought data to differentatte what will happen in the next few weeks.

    Sit back and enjoy the ride.

  16. 316

    LG Norton leads me to agree. The melting for the next few weeks will be influenced mostly by local weather, and how know predictable is that? The other thing is that so many conditions this year are vastly different from the previous years’ that it is not even like comparing apples to oranges, it is more like comparing raspberries to bananas.

    However, Wayne’s comments #250 and #296 are very interesting because something sure is melting that ice.

    “squalid and” (Could ReCaptcha actually be an AI bot experimenting on all of us?)

  17. 317
    sidd says:

    Alan Millar wrote on 2nd September 2008 at 7:26 PM
    “Global temperatures remain fairly stable with a slight cooling trend. Sea temperatures show no warming trend. There is no evidence of an accelerating sea hight increase.

    The models are not currently in agreement with actual Earth conditions over the last seven or eight years. ”

    o dear me, i think every sentence above is incorrect. of course, i might be wrong.

  18. 318


    Then, again, you might not.

    Speaking of accelerating rates of sea-level rise, what is the latest on that? Anybody know?

    And, is the collapse of the landfast iceshelf off northeastern Greenland already affecting things further up the related glacier’s ice stream, or is that just my overactive imagination at work every time I look at the latest ENVISAT images?

    (Link below is good only for about the next 14 hours or so.)

    “Kickapoo France” LOLOL

  19. 319

    Re: #289


    How can “The SOI continues to lurch back towards La Nina” ?

    Has that ever occurred before?

  20. 320
    dhogaza says:

    It would be “out of control” if buoys had shown continued dramatic thinning this year, the ice extent had “overtaken” 2005/6 in June rather than August allowing albedo changes to have a significant effect, and Arctic temperatures had shown significant rises.

    Ah, the “La Niña will save us, as long as it happens every year and El Niño disappears” argument.

    Wanna lay odds on that?

  21. 321
    dhogaza says:

    The models are not currently in agreement with actual Earth conditions over the last seven or eight years.

    Ah, someone else who’s fallen for Lucia’s unpublished (and unpublishable), unstatistical, dear-lord-lets-toss-real-science-into-the-toilet bullshit.

    Which, of course, is winning her the Nobel soon, because, you know, overturning a ton of established science with tard-thinking is JUST WHAT the Nobel committee looks for (NOT!)

  22. 322

    #314: Weather? I think not, Polar ice maxima and minima extents is like a climate metric, more than weather. If you go back a little on RC, you will find literature explaining necessary temperature disparities between the Northern and Southern Hemispheres. Is not like CO2 will warm up the entire planet equally. By theory, the Northern Hemisphere will warm faster than the Southern, and lo and behold it does, and so the ice in antarctica behaves as the models predict.

    #301 Thanks for the numbers , Cryosphere Today Ice extent is likely within an undeclared margin of error, 5% difference between 07 and 08 is likely within that margin. Although:

    The graph shows identical extent (look at anomaly in red). Bremen and Danish maps disagree
    with Cryosphere, showing less ice. If you look at all the data it may be confusing. The melt in 08 was huge; there was an article around february 2008:

    Which has quoted an increase of average ice thickness by 10 to 20 cm with a gain of 2 million square kilometers. For those who say: “What’s the big deal?” it appears that Ice volume melt rivals 2007 with less sun, colder water and air…. Something to write about, as we say in Canada : Hey?

  23. 323
    Lauri says:

    Alan Millar wrote on 2nd September 2008 at 7:26 PM
    “Global temperatures remain fairly stable with a slight cooling trend.”

    As you use the quantitative word trend, let me correct you. A trend is computed based on a series of data points. For temperature, the slope of the trend depends on how many years back you include in your trend. So is there a cooling trend? To inform you, I computed the trend value for trends of different lengths back to the history starting with 2007 as the latest data point. The data are NASA data, global surface temperature anomaly.

    In the table below, the year refers to the beginning of trend years (e.g., 1995 0.036 means that the trend between 1995 and 2007 is 0.036 degrees Fahrenheit per year, i.e., 3.6 degrees for 100 years).

    1988 0.0199
    1989 0.0218
    1990 0.0216
    1991 0.0250
    1992 0.0285
    1993 0.0263
    1994 0.0226
    1995 0.0201
    1996 0.0214
    1997 0.0181
    1998 0.0178
    1999 0.0300
    2000 0.0250
    2001 0.0107
    2002 0.0043
    2003 0.0090
    2004 0.0160
    2005 -0.0250
    2006 0.0300
    2007 #DIV/0!

    Only from 2005 to 2007 there is a cooling trend. Do you really want to base your knowledge on this two-year period only and neglect all the others?

  24. 324
    Vincent van der Goes says:

    RE 314

    “The models are not currently in agreement with actual Earth conditions over the last seven or eight years.”

    This topic has been covered extensively on this site, see:

    Please have a look at it. It explains the point very well.

  25. 325
    Vincent van der Goes says:

    Also, Alan Millar:

    “The globe is more than the NH after all. You have to factor in the SH at some stage! So why is there hardly any discussion of global polar ice extent?”

    I agree that there is currently a lot of talk and speculation about the year-to-year NP ice extent (while the longer trend is more important). However, the focus on the north pole (rather than south) has a justification. I’ll try to explain as well as I can.

    Feedback mechanisms are strongest on the north pole. More warmth can relatively easy result in less ice, since it is all sea ice (unlike the south pole). When there is less ice, there is less reflection of sunlight, so the water will warm faster.

    As I understand it, this is why global warming has the most impact in this region, and why so many people talk about it. (a look at this picture can put things into perspective:

  26. 326

    Alan writes:

    Global temperatures remain fairly stable with a slight cooling trend.

    No, they do not:

  27. 327
    Nick Barnes says:

    NeilT @ 312: you can see from MODIS that the ice has moved away from the Greenland and Ellesmere coasts, rather than melting in place. Presumably this is due to wind.

    It’s still very disturbing, though, because now that this ice is not fast, and is out in the ocean, it will respond more readily to the transpolar drift pushing it out of the Fram Strait.

    Despite all the open water around Peary Land, Greenland is not circumnavigable at the moment because the Lincoln Sea is full of broken ice, and apparently so is the Nares Strait.

  28. 328
    Cobblyworlds says:

    #292, Chris.
    (I have been trying to post this for ages.)

    It is not the area/extent per-se that’s changed my mind, although to see it this close has amazed me. It’s mainly the rate of reduction this late in the season, the latitude to which that reached, and the loss of perennial ice implied by both this year’s melt and the transport out of the Fram Strait. From my understanding based on the papers I’ve read, perennial ice can be seen as a damper. That’s because it’s less likely to melt to open ocean than first year ice (due to thickness and salinity/structure) it damps the impact of weather on the melt. The recent events suggest to me that something has changed in the Arctic, last year was not a fluke. What we are seeing is the impact of ice volume reduction.

    Yes there are areas that haven’t thinned significantly, especially in the centre of the pack where a cold climate is provided by Greenland and the rest of the ice pack. But that story is not supported as you move away from the Canadian coast. (pdf here) – same as figure 4 of NSIDC 25 Aug 2008.

    As far as I can see we have a pretty typical year’s weather starting with a good ice-growing winter (cold, lesser snow thickness), yet producing a quite remarkable drop. All year I’ve watched developments and until recently have not seen anything to really convince me that the ice was not likely to recover showing 2007 to be the sort of outlier Bitz(ARCUS) points to in the model runs. 2 years of very low extent makes them even less probable outliers, more years would compound that improbability. To my amazement there was (is still?) the chance the so-far unbroken 1-year-autocorrelation* rule might be broken. I don’t think that’s likely, but I bet it will be broken within 5 years. *’the year after a record minimum extent year never produces a new record’.

    Graph 2 of Spreen/Kaleschke’s Sea Ice Outlook also shows what a distinct pair 2007/2008 are compared to previous years.

    The main reason I had doubts about us being in a SICI (Small Ice-Cap Instability) type transition is because GCMs don’t show a rapid transition to seasonally ice free state. Over on the previous part of North Pole notes I asked why, apart from the shock of 2007, did people think we were going through a Small Ice-Cap Instability (SICI) type threshold, here. I got no answer. But this recent late melt acceleration seems to me to be exactly what I would expect if we were. For me the key factor this year has been volume/thickness. We are not seeing unusual weather causing the recent unusual melt, not in the way unusually prolonged clear skies caused* 2007’s (*or at least enhanced the impact of storms).

    Have you ever seen Nghiem’s 2007 study of perennial ice extent? Nghiem 2007 “Rapid reduction of Arctic perennial sea ice.” 1.68Mb pdf, here. If not I recommend it, check out figure 3, March 2008 was down 1 million sqkm from March 2007. I keep going on about that paper but that’s only because I see it as a key observation.

    Hope this formats OK, Nth attempt to post.

    PS From the BBC, Major ice-shelf loss for Canada:

  29. 329
    Cobblyworlds says:

    #314 Alan Miller,

    I can’t understand all these hundreds and hundreds of posts about what is, after all, a regional weather event!

    Typically I’d agree re weather, however the Arctic has ‘memory’ in a way that most other regional environments don’t. That memory means what happens one year can have a substantial affect on what comes next, e.g. 2007’s crash preconditioning the ice for this year.

    It’s as though this is has to be emphasised as most other Global indicators of AGW are showing a negative trend and it is a good bet that this will continue for the forseeable future.

    Not the case on the appropriate (climatological) time scale. My reason for emphasis it’s a key issue is the potential it poses, see my post 254. In pursuing public opinion the denialist camp have manouevered themselves into irrelevance, so why would they factor at all? Public opinion is irrelevant to the ongoing process.

    The globe is more than the NH after all. You have to factor in the SH at some stage! So why is there hardly any discussion of global polar ice extent?

    Why would we consider the Antarctic ice state when looking at what’s happening in the Arctic? They are at opposite sides of the planet, and whilst the Arctic is a polar sea surrounded by land, the Antarctic is land surrounded by an ocean.

    Arctic sea ice has increased this year

    It’s still closer to 2007 than any previous year, and 2007 was substantially below previous years.

    I’ve lived in rough areas most of my life and have been in numerous “tight spots”, on many occasions had I waited for certainty I’d almost certainly not be typing this now: I’d be dead. This is the real world, sometimes one has to make the best judgment with limited and incomplete information.

    All that said, as far as I’m concerned you think what you want Alan, persuade people that it’s not happening if that suits you. But beware because that persausion will no more stop this process than King Canute’s regal status could stop the incoming tide.

  30. 330
    Chris says:

    #328 Thanks very much for the detailed post and the links.

    A few initial thoughts. Firstly, looking at the comparison between 2nd Sep 2007 and 2nd Sep 2008:

    From 165 degrees W through 180 degrees, 165E, 140E, to at least 125E, the 2008 ice extends to a lower latitude ( ~77N to ~80N compared with ~81N to ~84N in 2007) The extra ice is all first year ice since it was essentially open water at this time last year. It is therefore by definition much thicker (i.e. thicker than 0!) and if it survives the next few weeks it will be good evidence of how ice can recover, as well as adding to the total of multiyear ice. Also from 75E to 30E the ice is significantly more concentrated this year.
    On the other hand, there has been a retreat of the ice edge in the northern Beaufort from about ~74N in 2007 to ~76N now, and an area roughly the size of the new tongue at 165E to 180 degrees has thinned, which looks rather dramatic at first sight. However, note that the tongue has increased from ~10 per cent to ~90 per cent concentration, whereas the thinned area this year has merely gone from ~90 per cent to ~70 per cent concentration. Furthermore, there are currently signs of re-freeze in the area, and temperatures down to at least -5C are certainly helping on this front:

    “For me the key factor this year has been volume/thickness”
    I couldn’t agree more: the 2008 melt season started with massively less multiyear ice than 2007, and a legacy of massively thinner ice all across the Arctic. And yet it STILL has greater extent, area and thickness.

    “But that story is not supported as you move away from the Canadian coast. (pdf here) – same as figure 4 of NSIDC 25 Aug 2008.”
    I’ve already explained above why this story is absolutely supported – i.e. the point about ice being indisputably thicker than open water.
    But just to nail the point, take a closer look at the buoy data you refer to. I’ve already gone into a lot of detail on the buoys in previous posts. The only buoys which *appear* to support your point are the two with big yellow bars.
    The one (2007E) on the edge of the Beaufort was right on the ice edge, and has drifted into open water, so has obviously seen a lot of bottom melt. Its neighbours show a different story: 2008F at Lat: 76.832 N Long: 139.974 W has still failed to melt to less than 3m thick
    and 2007F at Lat: 72.609 N Long: 136.079 W appears to have shown no net melt at all since the start of the year despite substantial southward drift. i.e. still at ~3m thick
    As for Buoy 2006C in the central Arctic (i.e. the other one with a big yellow bar), the melt here has been half that in 2007, such that the thickness is identical to a year ago.

    This is a very exhausting process – I only want to establish the truth of what is going on, and yet I sense a massive resistance to points which I think should be obvious. Accepting there might be some truth to what I’m saying doesn’t mean that the strength of the AGW hypothesis is somehow called into question – I haven’t voiced an opinion on that. Whatever your hypothesis, I would say it is best if you examine things from all angles, and if the hypothesis is particularly strong then one shouldn’t have a problem with short term considerations which may not help to prove it as much as one might like.

  31. 331
    Ray Ladbury says:

    I would contend that to look at either 2007 or 2008 in isolation is to take them out of their proper context. In other words, a single record year is less important than the steady and seemingly inexorable decline in polar ice. A single record year could be weather–the decline is climate change. A month ago, we were wondering whether 2008 could nudge out 2005 for second place. It is now threatening 2007 for the gold. The interest in that race is due to the fact that people are wondering if we have indeed reached a tipping poing. While it is too early to tell, the rapid changes in the North when viewed in terms of the trends of ever decreasing summer sea ice, a melt season that extends later into the fall and globally rising temperatures make it difficult to argue that we’re in balance.

  32. 332
    Chris says:

    #331 I am perfectly aware of the longer term trend. I am also perfectly aware of the limitations of pre-supposing that trends will inexorably continue. That is why I have been trying to find out in as much detail as possible, and in all contexts, exactly what has happened in the Arctic in 2007 and 2008, and what the important drivers/factors have been. Seeing as there are already plenty of people asserting that we have reached or passed a “tipping point” on the basis of 2007 and 2008, I would say this is worth doing, wouldn’t you? Even if we are close to a “tipping point”, how can it not be worth trying to understand objectively what is actually happening?

  33. 333

    #330 Chris, lets think a little. 2008 March , ice was thicker by 10 to 20 cm. From the start, ice extent was greater than 2007 by about 2 million square kilometers. Wouldn’t there be greater albedo? If so
    from the onset, less melting was meant to be? So why is it nearly equal? Especially with all the weather favoring a lesser melt, more clouds, cooler surface and water temps and unfavorable winds…. Less old ice helps, but that does not mean something else is contributing to this melt…..

  34. 334

    “Less old ice helps, but that does not mean nothing else is contributing to this melt…..”

  35. 335
    kevin says:

    First, no one said that it is not worth “trying to objectively understand what is actually happening.” Your implication otherwise sounds like a cheap rhetorical trick.

    Second, you stated above regarding some first year ice that may survive the melt season that “if it survives the next few weeks it will be good evidence of how ice can recover, as well as adding to the total of multiyear ice.” This strikes me as true but trivial. Has anyone said that it is impossible for first-year ice to survive and become multi-year ice? Of course not. The issue is that first-year ice is less robust. My read on this summer’s melt is that, given the weather, it would not have been a spectacular melt season had there not been so much first-year ice. The fact that it is anywhere near a record minimum seems to be largely due to the fact that last year’s exceptional melt led to there being a lot of new, less robust ice out there. And now, since so much of it has indeed melted again, it will be the same situation next year. At some point in the near future, the summer weather in the arctic is going to be less favorable for ice than this summer was, and even less of the more robust ice will remain afterward. And so on. The only ways I see this not being a “tipping point” leading to smaller and smaller summer ice area/extent, i.e. the only reasons I have thought of that the trend wouldn’t inexorably continue, is if the arctic started cooling and/or the summer weather became calmer than it has been for long enough for the ice to rebuild significantly. Do you know of any reason to believe either of those things will happen?

  36. 336
    pat n says:

    Mean air temperatures at NWS climate stations in Alaska (Barrow, Fairbanks, Kotzebue and Nome) in August, 2008 were below 1971-2000 averages.

  37. 337

    #336 thanks Pat. Well that does not mean that the entire atmosphere cooled…. It would be great if NOAA came about with DWT stats, which is very significant as it incorporates the temperature of the entire atmosphere. The way I see it now, is that 2008 had a very interesting shift in warm air location, further above the surface, which means the surface record got colder, however the atmosphere is just as warm. How did this happen? I am not sure. but the ice was surely affected by an IR heat source not far above it. Between 1000 and 850 mb height. There are very few upper air soundings over the Arctic ocean, so you must look and see if GRIBB is showing something.

  38. 338
    Chris says:


    Let’s think a little bit more. When does the melt season really get going within the Arctic proper? Certainly not before late April as there’s still ice out into the Pacific.

    So what was the situation in late April/May 08 compared with a year before? Looking at the JAXA data, I make the difference in extent about 2 per cent (e.g. 2.12 per cent on 25th April and 1.88 per cent on 1st May if you want me to be more precise)



    So 2008 and 2007 started the melt season proper with very similar levels of ice extent. (I’m not really interested in the very thin ice floes that formed briefly out into the Pacific and Atlantic at the end of winter, thereby pushing up the maximum ice extent, before vanishing quickly around mid-April)

    Can you prove your statement that “2008 March , ice was thicker by 10 to 20 cm.” ? I would be extremely surprised if this was the case over the Arctic proper (i.e. away from the Pacific and Atlantic peripheries). Here’s an example of a central Arctic buoy that was in 3m thick ice in March 2007, but only 1.8m thick ice in march 2008, yet has seen only half as much melt this year:

    Meanwhile, I think it may have been you that speculated that the ice area as measured on Cryosphere today was essentially the same as a year ago? Well, that hasn’t happened yet. The minimum in 2007 was 2.92, and the lowest so far this year was 3.199 a couple of days ago i.e. 9.6 per cent above the 2007 minimum. In the last 2 days, ice area has gone up to 3.247, so that it is now back up to 11.2 per cent above the 2007 minimum (or approx 8 per cent more than this day last year)
    The exact figure for today’s extent is 4924219 which is 6.7 per cent or 7.5 per cent above this day last year depending on whether you factor in the leap year, and 15.7 per cent above last year’s minimum.
    (Incidentally, last year had 2 sub-minimums in extent with a difference of a mere 13,125 km2: 4267656 on 16th Sep and 4254531 on 24th Sep. To put this in perspective, the final correction made to today’s JAXA extent figure was 12,188 km2. Area minimum was on 16th Sep at 2.92 million km2 – google this and find a million hits to confirm it. So I wouldn’t assign too much significance to the precise date of 24th Sep last year re: extent)

    Just to spell a couple of things out, albedo would not have been significant before mid-April due to the weakness of the sun, and after mid-April it was not significant because extent was comparable to all previous years.

    Therefore, you are left with your mysterious belief in “something else” other than the variables which have actually been identified and measured. Would you care to elaborate, and perhaps provide some evidence? Personally I would say it’s quite simple: much of the ice, especially on the Siberian side of the Arctic ocean, was dramatically thinner, and therefore melted more quickly as the summer went on, probably helped by weather conditions in key areas.

  39. 339
    Andrew says:

    Re #336: Need to be careful about using monthly averages from a few weather stations as indicative of anything to do with climate. Especially true when the stations are all located near each other.

  40. 340
    Jim Galasyn says:

    Arctic becomes an island as ice melts
    By Auslan Cramb
    Last Updated: 4:01pm BST 31/08/2008

    The North Pole has become an island for the first time in human history as climate change has made it possible to circumnavigate the Arctic ice cap.

    The historic development was revealed by satellite images taken last week showing that both the north-west and north-east passages have been opened by melting ice.

    Prof Mark Serreze, a sea ice scientist at the National Snow and Ice Data Centre (NSIDC) in the US said the images suggested the Arctic may have entered a “death spiral” caused by global warming.

    Shipping companies are already planning to exploit the first simultaneous opening of the routes since the beginning of the last Ice Age 125,000 years ago. The Beluga Group in Germany says it will send the first ship through the north-east passage, around Russia, next year, cutting 4,000 miles off the voyage from Germany to Japan.

    Meanwhile, Stephen Harper, Canada’s Prime Minister, has announced that ships entering the north-west passage should first report to his government. The routes have previously opened at different times, with the western route opening last year, and the eastern route opening in 2005.

    The satellite images gathered by Nasa show that the north-west passage opened last weekend and the final blockage on the east side of the ice cap, an area of sea ice stretching to Siberia, dissolved a few days later.

    Last year the extent of sea ice in the Arctic reached a record low that could be surpassed in the next few weeks, with some scientists warning that the ice cap could soon vanish altogether during summer.

    Four weeks ago tourists had to be evacuated from a park on Baffin Island because of flooding caused by melting glaciers, and polar bears have been spotted off Alaska trying to swim hundreds of miles to the retreating ice cap.

    Measurements on August 26 showed an ice cap of just over two million square miles, confirming the second biggest ice cap melt since records began. New of the opening of the passages emerged as the British explorer and adventurer Lewis Gordon Pugh began a kayak expedition to the North Pole aimed at drawing attention to the dramatic impact of melting polar ice.

    “I want to bring home to world leaders, on this expedition, the reality of what is now happening here in the Arctic,” said the 38-year-old environmentalist in his blog.

    “The rate of change is clearly faster than nearly all the models predict, which has huge implications for climate change and how to tackle it.”

    Meanwhile Prof James Lovelock, of the University of Oxford, has claimed “planet-scale engineering of the climate” may have to be attempted to counter global warming.

  41. 341
    Chris says:


    “First, no one said that it is not worth “trying to objectively understand what is actually happening.” Your implication otherwise sounds like a cheap rhetorical trick.”

    I was responding to a post that dismissed a carefully researched and detailed post of mine with the argument “I would contend that to look at either 2007 or 2008 in isolation is to take them out of their proper context.” – perhaps this argument might sound like a cheap rhetorical trick to you as well? Seeing as I have no more taken 2007 or 2008 out of their proper context than have the other posters I have been debating with.

    You summarise my analysis of first year ice as “trivial”. It’s hard to argue with adjectives (as opposed to evidence) and I think the onus is on you to make your argument more precisely.
    Consider where open water was exposed in 2007
    and then consider what the situation is a year on:
    Why is this trivial? It seems clear that well over half of the open water exposed is now ice again. And this is before we talk about thickness, which you have ignored completely.

    “the only reasons I have thought of that the trend wouldn’t inexorably continue, is if the arctic started cooling and/or the summer weather became calmer than it has been for long enough for the ice to rebuild significantly. Do you know of any reason to believe either of those things will happen?”

    Here’s the Jun/Jul combined average from MSU for lower troposphere temperatures for the last 10 years in the “NoPol” region i.e. ~60N to ~82.5N:
    1998 0.64
    1999 0.46
    2000 0.27
    2001 0.68
    2002 0.79
    2003 0.85
    2004 0.53
    2005 1.04
    2006 0.83
    2007 1.47
    2008 0.52

    So I would turn the argument around and say, do you know of any reason to believe that 2005 and 2007 represent the likely temperatures of the next few years rather than 1998-2004, 2006 and 2008? Similarly, do you know of any reason to believe that summers should be stormier than normal in the next few years?

    Sorry if my tone seems rather aggressive, but I’m only responding to yours.

    I’ve made quite a few detailed points today, and I believe they are helpful and relevant. I’m going to try and leave it at that.

  42. 342
    Hugh says:

    #340 That’s funny, I thought James Lovelock said…

    Global heating would not have happened but for the rapid expansion in numbers and wealth of humanity. Had we heeded Malthus’s warning and kept the human population to less than one billion, we would not now be facing a torrid future. Whether or not we go for the recommendations for cutting back fossil fuel use discussed in Bali in 2007 or use geoengineering, the planet is likely, massively and cruelly, to cull us, in the same merciless way that we have eliminated so many species by changing their environment into one where survival is difficult.

    Before we start geoengineering we have to raise the following question: are we sufficiently talented to take on what might become the onerous permanent task of keeping the Earth in homeostasis? Consider what might happen if we start by using a stratospheric aerosol to ameliorate global heating; even if it succeeds, it would not be long before we face the additional problem of ocean acidification. This would need another medicine, and so on…

  43. 343
    Ray Ladbury says:

    Chris, We started out this year after a hard winter with a sea ice extent that was the highest in 5 years. We had a year with cooler temperatures than the last couple of years due to La Nina. And for most of the year, we stayed close to the profile of 2005. Then we just didn’t bottom out. Certainly, thinner sea ice played a role. However, it would appear that the dynamics have changed. It ain’t the same Arctic.

  44. 344

    #338 Chris, read the article:
    It was a significantly colder end of winter over the North American Arctic side of the Pole.
    Albedo is significant at every month, even a low sun has an impact. Increased ice surface automatically means more albedo, in April May June July, even in the places where you think ice is irrelevant: Bering strait etc.. Your single Buoy example is not very appropriate, it is even cherry picking.
    Take this one:

    Ice thickness has increased over the summer. The data is preliminary , by the way….

    Single ice stations are affected by so many factors, that its dangerous to generalize from it. However: you still fail to recognize summer weather conditions:
    cooler surface air, cooler water, greater cloud coverage and winds favoring scattering (greater albedo) of ice instead of compressing, all factors against a significant melt from happening. Yet it happened. I’ve stated that ice extent started at a much greater level than last year, that is well accepted pretty much everywhere, I look at graphs and Cryosphere today makes the melt extent almost , read the word, almost identical to last year. So I would say that, explaining a greater melt, from thinner ice makes sense, only if the melting conditions were similar.

  45. 345
    Peter Ellis says:

    Chris, from your posts you seem to have the impression that ice forms and melts entirely in situ. At least, that’s the only message I can take from your continued insistence on exactly which parts of the Arctic Oceas have ice this year but were open last year, and vice versa.

    That’s simply not the case: sea ice is mobile and follows the wind and water currents. The patch of ice you’re so animatedly pointing to at 165E isn’t necessarily new, recovering ice – it is much more likely to be an area of older ice that’s moved over a bit. That in turn helps explain the reduction in ice in the Beaufort sea, and so on.

    What we can say about this year’s ice coverage is that despite having started from a higher baseline, we have no more (or very little more) ice in terms of area than last year, but that the extent is slightly higher. That is, we have the same ice area, but more loosely packed. This is a vast amount of yearly melt, occurring despite the fact that conditions this year are much less favourable for melt than last year. Thus it appears that the thin first-year ice formed from last year has been largely unable to withstand even a cool summer. The fact that this year’s ice area is no greater than last year sets us up for the same thing to happen again next year – only will next summer be more favourable for melting?

    Even according to your own flawed analysis, you draw the wrong conclusions. What if the new ice area at 156E is indeed first-year ice which has withstood this summer? That means that if this year’s ice area is identical (or nearly so) to last year, then we must have lost an area of multi-year ice equal to the area of first-year ice you’re trumpeting as evidence of recovery. If so, then we’ve just replaced the last remaining bulwark of tough multi-year ice with ice that’s still very new. Scarcely a comforting thought!

  46. 346
    CobblyWorlds says:


    Note that there are differences between every year in terms of a regional pattern of melt, the correlation between September extent and the extent in months earlier than July is noted as being poor. There is very little inter-annual correlation. Personally I don’t think there’s much gain in examining such differences in great detail. I’ve followed Quikscat over the winter, and have spent more time than is healthy ;) going over the National Ice Centre Ice Charts. There is a lot of movement in the ice due to the Transpolar drift and Beaufort Gyre NSIDC. However this movement is area-dependent. For example around the pole and Ellesmere Island it is less, because the ice gets packed up by the transpolar drift pushing ice from Siberia up to the North coast of Greenland, whereas in the Beaufort Sea and on the East coast of Greenland it is significant. Furthermore the ice in the central ice pack won’t melt as much as in the periphery as it’s surrounded by ice and is kept cold. Given their locations and the progress of this year I don’t see a problem for my position with what is shown by any of the bouys you point out. (Which is not to say that I dismiss what you have noted.)

    Rather than get bogged down in the detail of the bouy data I’ll get straight to what seems to be the key thrust of your argument – this year as a prospect for recovery.

    In terms of the prospects for recovery it’s worth bearing in mind the time-constants of ice response. Bitz & Roe did a paper examining the loss of thick ice; “A Mechanism for the High Rate of Sea Ice Thinning in the Arctic Ocean” Journal of Climate 2003. They found that because thin ice can grow more quickly than thick ice there is a biasing towards loss of thick ice. For example it takes at least 6 years to make a given volume of 6 year old perennial ice, however because some volume is lost each summer in practice it takes more than that. Ice listed as 6 years old will contain some younger ice (new ice that freezes in fissures) and older ice (moved about by circulation). The old ice becomes thicker due to compaction, as the ice pack compresses ridges are formed, ‘sheets’ of ice overlap. In terms of thickness first year ice cannot get very much thicker than about 2 metres. That’s because additional thickness forms on the underside of the ice-pack, and for it to freeze there has to be a heat flux from the (relative) warmth of the water to the cold of the surface in the long Arctic night. The heat flux is inversely proportional to the thickness of the ice, so once the frazil ice has compacted and started to form ice-pack as we’d recognise it, growth is initially rapid but it then slows down. Furthermore snow will act as an insulator lesser snow cover will increase thickness, as happened last winter due to the late formation of ice in some areas (ARCUS May outlook report). So it’s a very quick process (1 year) to grow new first-year ice, but it takes much longer to replace the loss of old thick ice (many years).

    This is why I brought up the Nghiem 2007 paper in my post above. That shows what I see as the real tipping point in figure 3, the area/extent is to some degree a distraction. Steven Goddard has asserted over at “Watt’s up with that” pretty much what you assert; that the survival of ice above last year suggests prospects for a recovery. However Nghiem 2007 shows a persistent and intensifying year-on-year drop in perennial ice as measured by QuikScat. So in past years where extent/area has gone up between years (e.g. 2005 to 2006) there has still been a drop in perennial extent, thus suggesting that this is not a significant factor and cannot be read as a sign of the reversal of the loss trend. By the way I think it’s reasonable to suggest that an extent increase from 2007 will probably be similar to 2005-2006, as things stand at the moment, and after 2006 came 2007.

    Have you read Zhang’s team’s Sea Ice Outlook work using the PIOMAS model? Web page here. They took the weather of the years 2001 to 2007, as ensemble members 01 to 07 and used the weather to force the model from the initial condition of ice in March. What they found (top fig) was that only 2007’s weather caused a September minimum extent below 2007’s record, even then the drop was not as much as between 2006 and 2007. Their September minima for ensemble members 01 to 06 were all over 5 million kmsq, as things stand this year’s minima seems very likely to be below Zhang’s ensemble members 01 to 06, however in their paper they note tendencies towards regional overprojection, and their figures are September mean.

    Zhang’s study is interesting because it is relevant to Kevin’s post 335. Rather than go on more I simply say I second what Kevin says.

  47. 347
    Chris says:

    #345 If you’re determined to disagree with someone, you can always appear to find a flaw in their argument and generalise it. I can’t cover every aspect of the Arctic ice scene in every post, and just because I haven’t gone into detail on the shifting of ice doesn’t mean it obliterates my argument. I’ve dealt with the issue of the August incursion of anomalous heat from northern Siberia in previous posts, along with numerous other issues, which explain my “flawed analysis” as you call it in more detail. As for the “vast amount of yearly melt” you refer to, I would be interested (and extremely surprised) if anyone could provide evidence that within the Arctic ocean (i.e. away from the Pacific and Atlantic fringes which are not relevant since they have always melted every summer in recent history) there has been a volume loss approaching that of 2007.

    #344 I’ve explained my arguments – too much further debate would be needless repetition. The main thing I’ll say is that the weather conditions WERE more favourable to melt in the precise part of the Arctic that saw rapid melt of vulnerable first year ice in August i.e. the Siberian side. I followed the synoptic charts and weather data every day in August, and I watched the persistent warm southerlies eat up the ice in front of my eyes. I didn’t follow things so closely earlier in the summer so I can’t describe the causation behind the July Beaufort melt with any precision. But I do know that far north Canada had record heatwaves, and ice in the Beaufort had been broken up by winter storms.
    I haven’t cherry picked with the Buoys. I’ve been over the data from each individual buoy ad nauseam – see e.g. #330. The big picture is that 2007 saw record minimum average thickness by the end of summer of 1.3m – see Maslowski/NASA, hence 2008 started the melt season with dramatically thinned ice despite a cold winter (this simply is incontrovertible); hence if average thickness is >1.3m at the end of this summer then 2008 has seen much less volume loss over the Arctic ocean.

    I’ve consistently argued that an earlier and stronger re-freeze than last year is very plausible, despite the fact that last winter was relatively cold as people like to emphasise when talking about 2008 yearly ice loss (despite, as I keep having to repeat, the fact that most of the extra ice floes that were around briefly at the end of winter on the Pacific/Atlantic peripheries would have been of minimal thickness – who knows, a few cms in many areas?)
    No one has agreed with me, or even recognised that I might have a point. Well, let’s wait and see what happens. The fact that the (western) Chukchi sea is now 4C colder than this time last year, the NW Alaska coast is back to being colder than the long term average, and the southern Siberian sea is colder than at any time since 2001 makes me reasonably confident.

  48. 348
    Chris says:

    Sorry should have said “southern part of the East Siberian sea” at the end of my previous post. (N.b the southern part of the Beaufort is currently 1C warmer than last year, but this is surprisingly little considering how much longer it’s been open water)

  49. 349
    pat neuman says:

    Re: 339

    Temperature plots are available for public viewing at many climate stations in Alaska (back to about 1950), the Midwest (going back to the 1890s) and at Minneapolis MN back to 1820.

    Beginning dates of snowmelt runoff at river stations in the Upper Midwest from 1900 to 2008 can also be viewed, for the Red River at Grand Forks ND, the St. Louis River at Scanlan MN and the St. Croix River at St. Croix Falls WI.

    Plots of average dewpoints at a few climate stations are also available. Higher atmospheric humidity increases melt rates when air temperatures are above freezing.

    Plots at:

  50. 350
    Alan Millar says:

    “No one has agreed with me, or even recognised that I might have a point.”

    Chris I agree with you! It is obvious that there has been some recovery in the Arctic sea ice extent this year notwithstanding the poor intial conditions that you have amply demonstrated.

    Having said that I don’t see much relevance, from this whole protracted debate, to the AGW question.

    If the Arctic ice recovers to the long term average in the next few years does that falsify the AGW hypothesis?

    If the Arctic ice falls to new summer lows in the next few years does that prove the AGW hypothesis?

    The answer is absolutely no to both questions in my opinion. Why therefore, are people making so much fuss about it compared to other more relevant factors like sea temperatures, UAH satellte data etc?