Banner
    Arctic Ice November 2010
    By Patrick Lockerby | October 26th 2010 06:04 PM | 93 comments | Print | E-mail | Track Comments
    About Patrick

    Retired engineer, 60+ years young. Computer builder and programmer. Linguist specialising in language acquisition and computational linguistics....

    View Patrick's Profile
    Arctic Ice November 2010

    Return to previous Arctic conditions is unlikely

    Record temperatures across Canadian Arctic and Greenland, a reduced summer sea ice cover, record snow cover decreases and links to some Northern Hemisphere weather support this conclusion

    Arctic Report Card 2010
    http://www.arctic.noaa.gov/reportcard/
    In the Arctic Report Card 2010, the term 'previous conditions' refers to the 20th century.  My own studies go back to the settlement of Iceland in 857 CE. In that context, return to previously stable Arctic conditions isn't just unlikely, it is extremely improbable.

    As I write these words, the ice extent shown by NSIDC is only a little above 2007 levels for this date, and growth in extent is quite slow for the season:

    NSIDC graph 28 October 2010, 50% reduced size.
    source: http://nsidc.org/arcticseaicenews/

    Arctic sea and air temperatures remain unusually high for the month.  Of particular interest are the sea areas around Greenland and the Canadian archipelago.  Multi-year ice continues to be exported through Nares Strait into a relatively ice-free Baffin bay.  There is also older ice in the passages between the Canadian islands, around which new ice is forming.

    First-year ice is not necessarily thin ice.  It can form into thick floes by over-riding ( slabbing ) and by forming pressure ridges.  However, first-year ice remains salty, and as such is much weaker than older ice and more susceptible to melting.

    Indications are that the melting of first-year ice in 2011 will cause a rapid reversion across the Arctic to the relatively small and mobile older floes seen this Summer.  A further, and perhaps final, loss of multi-year ice can be expected.

    Unless the Arctic sees unusually low temperatures before April 2011 - an unlikely event - it is highly likely, I suggest,  that the central Arctic ocean will be virtually ice free by the end of Summer 2011.


    An aside on climate science

    Climate science is not so much a branch of science as the whole tree, roots and all.  Findings from disciplines as diverse as linguistics, anthropology, psychology and nuclear physics can all add to our knowledge of climate history.  Looking at the Arctic and Northern peoples, their histories and legends and their archaeological legacy we can determine that the events of the 21st century in the Arctic lie well outside the range of natural climate variability from 875 to 1999.

    There are three Cs in science: continuity, cohesion, and confirmation.  Scientific knowledge has a continuous history: facts don't just spring up from nowhere - they have antecedents.  The parts of scientific knowledge fit together in a highly logical fashion: there is a cohesion of data and methodology which transcends the artificial boundaries of scientific disciplines.  Science requires that observations are confirmed by independent observers: we have a wealth of independent observations about the Arctic covering many centuries.

    The whole point of science is to investigate the unknown as far as is humanly possible. There will always be gaps in our knowledge, but the history of practical application of science to the solution of problems shows that problems can be solved without perfect knowledge.  To give but one example: the steamer San Fransisco was disabled in the Atlantic in December 1853.  Captain Jefferson Maury1, working in his study from theoretical principles of oceanography was able to predict the drift track of the disabled ship with such accuracy that the abandoned ship was found.
    there is much about which we are still in ignorance. But the investigations of late years—especially those conducted under the superintendence of Captain Maury of the American Navy, and Doctors Carpenter and Thompson of England—have shown that our atmosphere and our ocean act in accordance with a systematic arrangement, many facts regarding which have been discovered, and turned, in some cases, to practical account.
    R.M. Ballantyne,
    The Ocean and its Wonders, 1874.
    http://www.gutenberg.org/ebooks/21754

    Observations about a general warming in the Arctic regions have been made and confirmed since interest in the Arctic was re-awakened in the 1500s.  The theory of Arctic warming is backed up by solid evidence: evidence which is continuous, cohesive, and confirmed.  If a team could get together and produce maps based only on confirmed observations in written records, I am very confident that we could reconstruct Arctic ice extent back to about 1500 with an acceptable degree of accuracy.  Any oddness in reports can be filtered out using the tools of linguistics and psychology.  Even sea monsters may be found to have a basis in fact.


    Two colossal sea monsters

    The
    Carta marina created by Olaus Magnus, circa 1539, carries descriptions of some of the things featured.  These two creatures - which look to me like an artist's rendition from a vague description of a walrus and narwhal - are described by Olaus Magnus as:
    Two colossal sea monsters, one with dreadful teeth, the other with horrible horns and burning gaze - the circumference of its eye is 16 to 20 feet.
    Perhaps the 'circumference of its eye' is a misunderstanding of the 'radius of action' of a walrus from a hole in the ice, or the size of the hole it can punch.  A walrus will commonly attack a hunter by submerging and then surfacing through the ice where the hunter would have been if he wasn't smarter than the walrus2.

    Note also the driftwood off the coast of Greenland.  That is a very interesting detail for such an early map.  Driftwood was an important source of timber and firewood for the Viking settlers in Iceland and Greenland.  Today, it is an important source of proxy data regarding the historic drift patterns of sea ice3.

    I hope to write more about the Carta Marina in another article.  The point made about it here is simply that good information can be extracted from superficially dubious sources.


    Reports of Arctic warming

    Investigations have revealed that in the last twenty years the Arctic Ocean has grown considerably warmer.  According to nansen's observations of 1895, the temperature of the surface layer of cold water was about -1oC.  The depth of this water layer was 656 feet.  At a greater depth the temperature was above zero.
    The words above are from a 1939 book, The Conquest of the Arctic by Louis Segal.

    The following extract mentions "recent climatic improvement", "climatic amelioration", and glacial mass loss "starting at some places 200 years ago and at others within the last 50 to 60 years".  It was published in 1956.

    With some exceptions, especially in Alaska, glaciers in arctic and subarctic regions have been shrinking for many decades and mostly at accelerating rates (Thorarinsson, 1940). In areas bordering the North Atlantic this shrinkage, starting at some places 200 years ago and at others within the
    last 50 to 60 years
    , has been nearly universal and locally catastrophic. Many Alaskan glaciers have behaved in similar fashion, but others are now at their most advanced position in centuries (Field, 1932; 1937, pp. 69-76; Cooper, 1942).  Most North Atlantic glaciers are still receding, but a number of Alaskan glaciers are advancing, have advanced recently, or have been in equilibrium for several years (Baird and Field, 1951, pp. 121-127; 1954).

    The modern glacier recession in North Atlantic areas (Ahhlman, 1946, p. 24; Eythorsson, I949b) has been attributed to the "recent climatic improvement" involving a rise in winter, spring, and autumn temperatures, and to less marked degree in summer temperatures (Liljequist, 1949).  Climatic amelioration is further suggested by the decrease in average thickness of the arctic ice pack from 144 inches to 86 inches between 1893-95 and 1937-40 (Ahlmann, 1946. p. 23). The navigation season to Svalbard has become longer, birds and fish are now found farther north, frozen ground is deteriorating in many areas, and vegetative growth has increased (Hustich, 1949).  This climatic improvement is attributed to increased northward circulation of warm air, brought about by changes in atmospheric pressure gradient (Pertensen, 1949), and to an accompanying secondary effect produced by the Gulf Stream System (Helland-Hansen, 1949).  The North Atlantic low, and perhaps also the North Pacific low, have moved farther north to produce these conditions.  This is a grossly oversimplified statement, but, regardless, a rise in temperature more than a change in precipitation has been the major result, and temperature is the factor most strongly affecting the behavior of these glaciers (Ahlmann,  l940a, p. 100).

    This climatic amelioration culminated in late 1930s or early 1940s (Kincer, 1946, p. 342; Hustich, 1949, p. 103) and was followed by more continental conditions (Ahlmann, I953, pp. 22-27).  Most glaciers of the North Atlantic region as yet display no marked reaction to this latest change, but the erratic and diverse behaviour of Alaskan glaciers may be an early response to the new conditions. Whether this is going to be a relatively minor perterbation in the history of general shrinkage or heralds a major reversal in behaviour is for the future to tell. Unfortunately, means of evaluating glacier fluctuations quantitatively are still inadequate. Data on total regime and over-all shrinkage of glaciers are more significant than fluctuations of the margins, the type of information usually available. The still scanty knowledge of the influence of various meteorological factors on accumulation and wastage makes evaluation of climatic fluctuations through glacier changes difficult and uncertain.

    Glaciers in the Arctic   /   Sharp, R.P.
    Arctic, v. 9, no. 1 and 2, 1956, p. 78-117, ill., maps
    Contribution - California Institute of Technology. Division of Geological Sciences, no. 714
    ASTIS record 9783
    Index:
    http://pubs.aina.ucalgary.ca/arctic/
    This article:
    Arctic9-1&2-78.pdf

    The Arctic cannot warm in isolation.  If the Arctic is warming, then so is the whole planet, albeit more slowly.  The Arctic is indeed warming, and a general historic warming trend is traceable through various proxies back to at least the age of the Vikings.  The little ice age was but a mild hiccup in a long-term warming trend.  The rate of that warming has accelerated markedly since the start of the industrial era and again post 1950s.  If global warming is a hoax, then it is a very long-running hoax perpetrated by explorers and sea-captains since at least the middle of the 17th century, for reasons which I am entirely unable to fathom.

    -------------------------

    [1] - Captain Jefferson Maury was grandfather of Matthew_Fontaine_Maury, 'father of modern oceanography'.

    [2] - There is an excellent description of the contest between a hunter and a walrus in chapter 30 of Elisha Kent Kane's Arctic Explorations in Search of John Franklin, 1876.


    [3] - Changes in Driftwood Delivery to the Canadian Arctic Archipelago: The Hypothesis of Postglacial Oscillations of the Transpolar Drift,
    Arthur S. Dyke, John England, Erk Reimnitz, Hélène Jetté
    Arctic, Vol 50 - 1.
    http://arctic.synergiesprairies.ca/arctic/index.php/arctic/issue/view/64

    Comments

    Patrick,

    Great to see you publishing again...especially stuff about Arctic history that I didn't even know that I didn't even know. I hope you are feeling and doing better. Things look pretty ominous for the rest of us though.

    I also see that you have come around to my way of thinking on the amount of ice at the end of summer being nearly "0" in 2012, instead at the end of next summer, 2011. It's probably an insignificant academic difference whether the ice is gone next year or the following because the weather patterns in especially the Northern Hemisphere will change and worsen drastically when the ice has melted to that point... at least that is the general consensus of what I have read here or elsewhere. The flooding problems that happened in Pakistan this year will probably look minimal compared to the new weather regime.

    Thanks again for another very interesting article.

    Hi Patrick. Thanks for all your efforts, and this 'historic' post! I wonder, do you have any insights into the data release plan for ESA's CryoSat-2? I think most Arctic watchers agree that sea ice volume information will be critical to accurately forecasting the first sea ice free season. Cheers, mate!

    logicman
    Vaugh A., Lodger: thank you both for your comments.  I enjoy the historic research - the more so when I get kind comments like yours.  If all goes well I hope to write something about the Vikings in the Arctic.  Teaser:  English seafarers may have had a hand in the disappearance of the Vikings from Greenland.
    Interesting post as always. I've enjoyed your postings and musings on Arctic ice and climate - they have been very informative, and generally right on the mark. I have one quibble unfortunately with a small part of this article - you say in the Arctic "a general historic warming trend is traceable through various proxies back to at least the age of the Vikings. The little ice age was but a mild hiccup in a long-term warming trend." But so far as I am aware, the long-term trend for the Arctic, until the past 150 years or so, was cooling, not warming. See for exaple Kaufman et al (2009) or Polyak et al (2010). Otherwise I really enjoy your articles, and keep the good information flowing!

    Patrick:

    I second the notes of Vaughn and Dodger.

    As for your bold prediction:
    "the central Arctic ocean will be virtually ice free by the end of Summer 2011."

    I do not think there are enough forcings to do this and it will be many more years and perhaps a decade or more before this can occur. Oscillations are more likely than a steady annual linear decline, and I would expect the trend to slow as we near the bottom as it is still pretty cold up there in September. Predicting arctic sea ice is risky as the 2010 experience showed, but below are the predictions I am willing to make for the next three months.

    PLAYING WITH NUMBERS: GUESSES OF ARCTIC ICE EXTENT THROUGH JANUARY 31

    We are entering the relatively uneventful period of arctic ice extent of November and December, during which accurate guesses appear possible based on the JAXA data for 2002 through 2010. The JAXA data show that the average November and December increases are roughly the same at 1.9 million km2 , with January showing a reduced average increase of 1.1 million km2. The variation in annual extent diminishes at this time of year, particularly in December, resulting in a “choke point” since the arctic ice extent is influenced as much by geography as it is by conditions.

    The data for November 30 and change from October 31 is shown below:

    30-Nov
    Extent Change
    2002 11,044,688 2,244,844
    2003 10,639,688 1,910,313 10-28 data
    2004 10,814,375 2,079,219
    2005 10,556,406 1,676,406
    2006 9,740,938 1,040,938
    2007 10,252,500 2,249,219
    2008 10,714,375 1,822,031
    2009 10,398,281 2,086,718

    Avg. 10,520,156 1,888,711

    The October 31, 2010 extent was 8,038,906 km2. While the first 15 days of October averaged almost 100,000 km2 increases per day, this slowed for the last half of October to under 70,000 km2 per day. The current state is comparable to the October 31, 2007 ice extent of 8,003,281 km2. (UPDATE: as of November 2nd 2010 is
    88,906 km2 behind the same day in 2007.)

    Based on the numbers in the JAXA data set, November 30 should be in the 10.1 to 10.3 million km2 range given the low level of 8.0 million km2 for October 31 and the average increase for last 3 years of 2.1 million km2. The largest November increase in the 8 year period of the data was 2.25 million in 2007. While November of 2008 reached 10.7 million km2, the highest value since 2004, the 2008 November number was “helped” by the high October 31 values of 8.9 million km2, therefore it is unlikely that November of 2010 will reach the 2008 level.

    December 31 should be in 12.2 to 12.4 million km2 range since all 2004 through 2009 values fall in this range. The values for 2002 and 2003 were approximately 12.7 million km2. Both of these years were helped by high values at November 30 of 11.0 and 10.6 million km2 respectively.

    31-Dec
    Extent Change
    2002 12,720,625 1,675,937
    2003 12,655,938 2,016,250
    2004 12,265,000 1,450,625
    2005 12,282,188 1,725,782
    2006 12,290,000 2,549,062
    2007 12,369,688 2,117,188
    2008 12,230,156 1,515,781
    2009 12,255,469 1,857,188

    Avg. 12,383,633 1,863,477

    January 31 should be in the 13.1 to 13.7 million km2 range as all values since 2004 fall in this range. The wide range of possible numbers for January diminishes the value this guess. The value for 2003 was 14.0 million km2; however, it was “helped” by the high December 31 2002 figure of 12.7 million km2, the highest December amount for this period.

    31-Jan
    Extent Change
    2003 14,022,188 1,301,563
    2004 13,503,438 847,500
    2005 13,250,938 985,938
    2006 13,053,281 771,093
    2007 13,442,031 1,152,031
    2008 13,685,156 1,315,468
    2009 13,670,625 1,440,469
    2010 13,313,125 1,057,656

    Avg. 13,492,598 1,108,965

    Predicting ice extent 3 months in advance is not very accurate and further guesses of the arctic ice extent after January is not prudent or meaningful as the range of values is too large. Predicting one month in advance in the summer of 2010 was difficult.

    For the antarctic, the value at November 2nd was .106 million km2 above the average, which represents less than a 1% difference from the 1979 to 2000 average average ice extent. This difference is a significantly smaller than the percentage difference experienced in the arctic. The University of Bremmen chart does not appear to show much annual variability from the average for the next month. Based on the chart, the November 30 antarctic ice extent should be at the average for October 31.

    Additional correction - Iceland was settled around or after 871 CE, not 857. There is no verified archaeological evidence of human activity beneath the Landnam tephra marker of the same age, dated by its presence in ice cores, and a great deal immediately afterwards, including the tephra existing in cut turves, indicating the turves were cut within a few years of 871 CE. There are anecdotal suggestions of earlier Celtic hermits reaching the south coast, but none confirmed, and if you want to get an Icelandic archaeologist's knickers in a twist, the easiest way is to suggest that the Celts got there early!

    Interesting breakdown of the numbers there Will, very useful as a reference for the coming months.

    logicman
    Andy C: you are perfectly correct.  My mistake! There are many studies of the landnam tephra, Greenland ice cores and archaeological remains pointing clearly to a settlement date of 871 CE plus or minus 2 years.

    My mistake was to cite 871 CE as a settlement date.  A few people landing, building temporary shelters and then leaving doesn't qualify as a settlement.  It's a bit like Columbus, Cabot et al versus the Pilgrim Fathers.  According to the Hauksbok Landama translated by Vilhjamur Stefansson, 1943: the explorers Gardar, Naddodd and Floki all landed on Iceland before it was settled.  It was Floki who, from a mountain, saw ice floating in the sea and so named the place Iceland.  Note to historical revisionist-propagandists: Iceland was most specifically not so named because it was at any time in human history mainly or totally covered with ice.

    Judging from investigations by other scholars and from papers on driftwood dating, the date 857 CE represents an earliest plausible date for any Viking exploratory landing on Iceland.  Although it is plausible that any sailor lost and adrift in the region might end up by accident on Iceland, there is no archeological or other physical evidence to support claims that the Irish, the Welsh or any other claimant nation was the first to land on Iceland.

    Declaration of interest: I should point out that in supporting claims that the Vikings were first to do x, y or z I may be biased: my family name can be traced back to Viking roots.  :-)
    Thanks Patrick, I really appreciate your updates. Great stuff.

    John Hasenkam.

    Patrick,
    Realy good to have you back. The perspective you provide is amazing, shows we have been changing the climate longer than generally realised.

    Dear Tony,
    I think Patrick shows that the climate has been changing longer than generally realised.
    You don't have to feel guilty, you were not a viking in 871 or whenever.

    Another fascinating post, Patrick - your historical researches have given a great long-term perspective to situations where we are used to short datasets, 30 years of CT, 8 years of IJIS.

    Will - I think there is a danger in the numbers game - the assumption that this year is only a bit different from the last few years. I don't think that that is necessarily the case.

    For all that this year's lowest extent was low, it was nowhere near as low as it might have been if the weather had been different. In 2007, huge amounts of ice were exported from the basin through Nares and Fram Straits. That didn't happen this year - if it had, we would surely have seen a new record low extent. Instead large quantities of ice were pushed up against the northern fringe of the Canadian Archipelago, and stopped dead for months. However the melt opened the NorthWest Passage early, and over July and August the fast, longer term ice that plugs the channels in the archipelago gradually gave way, allowing some export through these channels. Since September, ice has been slow to reform (lots of "grease ice" still, and Coronation Gulf and Victoria Strait area still open water!) and of course, much of what is now covering the archipelago is weak first year ice. One presumes that the NWP will open even earlier next year as a result of the weaker ice that is there now.

    Now, if those channels to the north of M'Clure Strait open next year earlier than they did this year, then even unfavourable weather pushing ice into the archipelago will still act to reduce extent. When these channels are open, the amount of ice that can be exported from the cental basin doubles. I see that as being quite different to years prior to 2010. In particular, it favours Patrick's vision of "the central Arctic ocean will be virtually ice free by the end of Summer 2011".

    I interpret the "central Arctic ocean" not as the Pole but as the centre of the basin - a point roughly equidistant from Greenland, Point Barrow and the Siberian coast, centred on around 85 N, 150 E. Ice displaced through the Canadian islands makes way for ice on the Russian side of the pole to be transported elsewhere. I would suggest that the E Siberian and Laptev Seas will show even stronger negative anomalies in 2011 than this year. In addition, this year a large patch of very poor, loose ice formed in this area - was that more or less "random" or are there particular geographic effects at work (an upwelling of warm water, perhaps?) that favour a weak spot in this area?

    Now I don't believe that we are looking at an "ice free" Arctic next year. If the record is broken, it might go into the 3's, but not the 2's. But I do think you will see a larger ice free region in the central Arctic Ocean, north of Siberia reaching almost to the pole. Depending on the operation of the Beaufort Gyre, that could be between Wrangel Island, the New Siberian Islands and the pole, or further west, between NSI, Severnaya Zemlya and the Pole. By "large", I'm thinking in the million sq km range. Basically, I think that at minimum, there will be very little ice on the Russian side of the 0/180 meridian.

    just my 2 cents

    FrankD:

    Thanks for the detailed analysis of the possibilities for arctic sea ice decline.

    I am just playing with the numbers and have not tried to analyze the specifics the way you have or the way the models do. However, the numbers you are predicting appear too low even if a non-linear curve is used to describe the observed trend of September ice extent data (which gets to a zero arctic ice extent faster than a straight line trend), it does not appear a virtually ice free arctic will occur before 2020. Please see the graph at http://www.skepticalscience.com/news.php?n=443 (see the graph in the comments for a zero axis). Based on the graph, there will still be 2 million km2 at the September minimum in 2020.

    While ice volume has declined (awaiting CROSAT-2 info for confirmation), this alone is not sufficient to generate an annual minimum. A prediction for an ice free arctic needs to take into consideration that while the trend will continue downward, there will be some years reporting increases from the prior year and some years reporting decreases. It is possible that 2011 will be more like 2008 and will have a higher ice extent than 2010. If the 2007 weather conditions which were highly favorable to ice reduction do not exist in the summer of 2011 then 2011 will not beat out 2007.

    Based on the observed pattern, at least four or five years out of the next ten years will show increases over the level of the preceding year. (For 2001 through 2010, five out of the last ten years showed an increase.) If you look at the graph, the fact that 2008 and 2009 showed modest increases after the precipitous decline of 2007 is not that surprising. Even with four or even five years with annual increases, 2020 will be below the 2010 minimum as the size of the declines will outstrip the size of the increases.

    I like the references in your analysis to specific areas of the Arctic and the sophistication of your analysis compared to my guess. While "looking at the numbers" would never have predicted the huge 2007 decline , I doubt it could be predicted using the analysis of observed conditions at March 31, 2007. I am wary of predictions of such a big decline for 2011 at such an early point in time.

    In 2010, the late start to the melt season at levels near the historical average for the end of March was followed by the rapid decline that resulted in a new low point for June. Even with the June low, 2010 did not set a minimum record for September sea ice extent. From this recent history, it appears weather conditions have a bigger impact than the starting point of the ice.

    Currently the extent is tracking near 2007 levels, which puts 2011 on the same path as 2008. Looking back to 2006, November 30 was very low, at 9.74 million km2, but December saw a 2.55 million increase that brought the December 31, 2006 figure to 12.29 million, which is close to other years. March 31 of 2007 was .2 million km2 higher than March of 2006 and just 85,000 km2 below 2005, and yet the minimum for 2007 was 1.1 million km2 below the previous record established in 2005. It took an extraordinary amount of melt in 2007 to establish that record. Even if March of 2011 starts at the same point as 2007, it will take weather conditions that are very favorable to a decline for June through August to get a new minimum record.

    Perhaps the starting point of the melt in March of 2011 will be well below the comparable level for 2007, and 2011 will match 2007's melt, otherwise, I do not think we will get below 4 million km2 in 2011. Even if 2011 goes below the 4.0 million level, it does guarantee that the 2012 minimum will not return to levels above 4.0 million km2, which is still a very low level compared to the 1979 to 2000 average minimum.

    The forces acting on the arctic ice extent are complex and include air temperature, sea surface temperature, clouds, ocean currents transporting heat into the Arctic basin, the air pressure index (Arctic Oscillation), ice transport out of the Arctic through Fram and Nares Strait and weather patterns. You will need all of these items to line up in order to get the severity of decline predicted in the above comment.

    Time will tell.

    FrankD:

    You may find this link interesting which predicts that as the ice extent declines, the rate of decline will level off.

    They made a June 1 statistical estimate of the 2010 September average extent of 4.78 million km^2 based on what they call "Grumbine and Wu statistical ensemble" which they indicated was their "best guess". (Even they do not want to guess the actual minimum.)

    http://moregrumbinescience.blogspot.com/2010/06/when-will-arctic-ice-be-...

    This source is predicting ice free in 2035 plus or minus 7 years.

    logicman
    Friends: thank you all for your contributions.

    Something we don't have is a month-by-month account of ice volume over a historically long period.  If we had such a dataset it would certainly show just how rapid the modern ( circa 50 years ) decline in ice volume is.

    From the earliest exploration of Iceland through the rise and fall of the sealing and whaling industries and the Grand Banks fishing industry we have records showing that the average end-of-summer ice edge has slowly retreated in the Baffin Bay, Greenland Sea, Bering Sea and Kara Sea areas.

    For the past few days, Arctic sea ice extent has been close to or under the November 2007 levels, making it the lowest extent ever for early November.  Any growth in ice extent during the next 5 months will be as a result of fragmentation and dispersal of existing ice or the creation of new ice.  At the start of the melt season in April the ice formed in winter will accordingly be 5 months old or less.  Older ice continues to be exported through Nares Strait.  If the combination of factors which led to the 2010 volume loss is repeated in 2011 then I suggest that an ice-free central Arctic is highly likely by September 2011.

    I now hand the discussion back to my readers.
    Hi Will,

    My inclination is to work the numbers too, but it was in so doing that I realised that reality wasn't corresponding to my numbers. In a better statistician, that would have provoked a more rigourous analysis. Because I'm lazy and only semi-numerate, I started looking at it qualitatively, and asked, what am I missing. Off the back of some particularly sage comments from Patrick, and from some of Neven's residents, I realised that playing with numbers like this meant I was missing a vital point - volume.

    If you take your links views of 2020 to 2040, and combine them with the PIOMAS model of volume, you find a strange phenomenon. By 2020 (eyeballing Skeptical science, Grumbine is working probabilities, not extents), extent will be in the range of 2 million sq kms. But in ten years (probably sooner) the PIOMAS volume anomaly will almost certainly have hit the floor at -13,500 km^3. Since the September average is 13,500 km^3, obviously when that happens in September, ice volume will be zero.

    So what... we are projecting a situation where we have 2 million sq kms of zero-thickness ice?

    The problem that I have realised - largely from reading the Chatterbox - is that ice is losing thickness faster than it is losing extent, and ultimately it will reach zero in the vertical dimension before either of the horizontal ones. So any analysis resting on area or extent is ultimately flawed. It is quite apparent to me that one average, the ice will melt through from the bottom much faster than it will melt in from the edges. And it is equally apparent that while a lot of good brains have been working extent, they are wasting brainpower on the wrong metric.

    What does volume look like? We only have models for now, but well trained ones: http://psc.apl.washington.edu/ArcticSeaiceVolume/images/BPIOMASIceVolume...

    The supposed mini-recovery of 2008 and 2009 wasn't. Volume in 2008 was barely higher than 2007, while 2009 was a bit lower, and 2010 catastrophically lower again. It is presently recovering, but how far it will improve over winter will be interesting. But think about it - if extent is increasing and volume is not, whats happening? Obviously, the ice is getting thinner and more spread out, and more vulnerable to sudden collapse.

    BTW - the monthly average against which those volume anomalies are measured is here: http://psc.apl.washington.edu/ArcticSeaiceVolume/images/PIOMAS_daily_mea...
    September average, 13,500 km^3. Once the volume anomaly drops to that level in September you have no more ice, whatever the best linear regression analysis of extent or area tells you. 2010 was unusual in that the max anomaly was in June. Recently, the September anomaly has been at or near the low point. The PIOMAS site applies a linear trend, but clearly - as with the skeptical science graph you linked to - a second order polynomial is a better fit. And extrapolating that curve (I don't have their data, so Eyeball Mk I), you get to -13,500 in around 2015.

    In June, Wieslaw Maslowski published some analysis of volume data, and concluded that a virtually ice free Arctic is likely by 2016, +/- three years. I doubt the data that has come in since has led to any increase in optimism. Climate Progress wrote it up here: http://climateprogress.org/2010/06/06/arctic-death-spiral-maslowski-ice-...
    and mention: "September Ice Volume was lowest in 2009 at 5,800 km^3 or 67% below its 1979 maximum."

    This year, the September volume was about 4,000, after a couple of months of favourable weather.

    As to Grumbine and Wu, well in June, I made a statistical estimate of 4.8 too (posted at Neven's blog). So... you know....whatever.... ;-)

    Frank,
    Thank you for the statistical analysis. I have also noticed the disconnect in the linear models portraying Arctic Ice extent, cover, and volume. I too have been too lazy(also distracted by a job...high school teacher... and other interests etc.) to do enough analysis to show when the lines for these 3 conditions reach "0" at the same time. I also know that trying to make a linear function accurately fit a non-linear function is difficult at best and usually skewed and thereby flawed. I also know that there are few...if any...linear functions that accurately describe natural events, yet we are attempting to show a linear function with many of the models that predict future events(I wonder if it is an artifact of the difficulty computers have/or had with handling calculus?)
    e^x, or a variation thereof, is a formula that is often used to describe natural events( e is approximately 2.71828183). The difficulty is to figure out "x". In each of the 3 described conditions, extent, cover, and volume, "x" will be a different value. What is needed are values of "x" where the graphs of each of the conditions all reach "0" at the same time. This would also be a check and balance of the different predictions and because of the check and balance would likely be be much more accurate than looking at different linear assumptions of which some have to be inaccurate simply because of their divergence. Again, I too lazy to do all of the necessary calculations, but I would certainly like to see them happen.

    "The forces acting on the arctic ice extent are complex and include air temperature, sea surface temperature, clouds, ocean currents transporting heat into the Arctic basin, the air pressure index (Arctic Oscillation), ice transport out of the Arctic through Fram and Nares Strait and weather patterns. You will need all of these items to line up in order to get the severity of decline predicted in the above comment. "

    Will, I wanted to treat this seperately, because it suggests you've missed or misunderstood my comment (although I presume you are referring to Patricks original write up when you talk about "the above comment". You say that several things have to "line up". Many of these depend on one or more of the others in complex ways, but I grouped them roughly for convenience:
    * ocean currents transporting heat into the Arctic basin, sea surface temperature - not the same, but similar for my purposes. Excess heat is being transported into the arctic and there is little sign of that slowing down. While it might vary a little each season, it is probably the most dependable of the things that need to line up.

    * clouds, air temperature, - granted, highly variable. There is some indication that increased temps cause increased cloud, a negative feedback. But that only slows the process, it doesn't stop it - most HEAT is being brought in by ocean currents, and some reduction in direct solar heating isn't going to help much, especially since it depends on higher air temps anyway. Critical for individual seasons, but fairly unimportant in the long run.

    * the air pressure index (Arctic Oscillation), ice transport out of the Arctic through Fram and Nares Strait and weather patterns.- This is where i think you missed my point. With the AO in one mode, ice exist through Fram and Nares. With the AO in the other mode, historically, it has piled up against the archipelago and not been exported. But with the channels in the archipelago opening up, it now *can* be exported from the Arctic basin. It may take a bit longer to melt than if it went out through Fram or Nares, but its still gone. Once again, that is a new paradigm - previously, you were correct, the AO had to be right to achieve low extents. Now, in EITHER mode, multi year ice is exported from the Basin. Historically, it seems true that ice extent falls sharply when the AO moves from negative to positive (as far as I can see it is the change in state, not the existence of a postive AO, per se), so there is a timing factor here, but that is all.

    So of the three (really) things that have to line up, we can count on one, one only effects timing, and the other is no longer a constraint, or at most, also a timing issue. Most of the things that line up will determine exactly which year becomes the first September zero. But they don't change the underpinning state. I think you are wrong in saying taht all these things need to line up to produce an ice free central arctic (see above for what I mean by that - Patrick may or may not want to be more specific about his meaning). Rather, I think they all need to line up to stop it from happening.

    Frank D:

    I appreciate the analysis of the factors and the fact that the volume is dramaticly lower, but it still appears that a shift in wind directions could result in a larger minimum ice extent in 2011 than 2009. This is not a denial that the trend is for a decrease in arctic ice extent or a denial of AGW, it is only a caution that predicting a new minimum of below 4 million km2 in 2011 has a very low probability of occurring.

    I am not suggesting that we will see a return to ice extent levels that existed before 2000. The rapid decline in recent years should not extrapolated in a linear fashion to predict the significant decline suggested by Patrick or the estimates you are making for 2011. It is possible that 2011 will see a higher arctic minimum than 2010 and the probability that this could occur is higher than a minimum below 4 million km2. This would not be proof that AGW does not exist or that the impact of AGW has stopped. It only means that there is a great deal of inter-annual variability to ice extent and predicting the next year's minimum number is more a matter of luck than science.

    In order to get the kind of decrease you and Patrick are predicting you will need wind patterns that are favorable to such a decrease, not just warm ocean currents or higher than "normal" temperatures,. Without the wind pushing the ice out of the Arctic, you will not see a new minimum in 2011.

    The artcle below found that wind direction has a bigger impact on ice extent than other factors and also cautioned against using ice extent as an indicator of the ice's climatic "state of health." In this regard, the analysis of ice volume that you have provided may be a better indicator of climate impact than sea ice extent.

    http://www.sciencedaily.com/releases/2010/04/100427111449.htm

    "The Arctic climate paradox

    A few years ago, US researchers discovered what they termed the "Arctic climate paradox." Since 1980, the researchers had been observing a decrease in ice cover. They explained this through a slow process of climate change combined with fluctuations in patterns of atmospheric pressure and air currents over the Arctic. It was believed that the positive phase of the Arctic Oscillation (AO) was a major cause of the receding ice cover.

    The AO is normally influenced by three pressure systems located over the Azores, Iceland and the Northern Pacific Ocean. Since 2000 the AO has been in a negative phase. As a result, researchers predicted that the pace of reduction in the ice cover would slow down.

    Instead it accelerated.

    Unknown factor

    "The US researchers argued that the ice was responding to something else, another factor that nobody had considered," explains Asgeir Sorteberg, Associate Professor at the Geophysical Institute at the University of Bergen. He has been investigating this phenomenon along with his colleagues in the project entitled the Norwegian Component of the Ecosystem Studies of Sub-Arctic Seas (NESSAS).

    When the Norwegian researchers began their work, they noticed in particular a dramatic change in the weather pattern in the Arctic beginning about the year 2000. The change corresponded to the point in time when the reduction of ice cover in the Arctic began to accelerate.

    The researchers began to analyze the circulation patterns over the Arctic.

    "We found that these patterns can explain in large part why the ice cover decreased so much more rapidly after 2000. Wind patterns depend on the position of major high-pressure and low-pressure systems. We discovered that months with very little ice cover and high temperatures corresponded with crucial variations in the wind patterns," explains Mr Sorteberg.

    "Up until 2000, the Arctic Oscillation (AO) had the greatest impact on the winter ice cover in the Arctic. But the change around 2000 meant that more of the weather and wind over the Arctic after that year was determined by high-pressure and low-pressure systems in northern Russia. In other words, the AO, which was usually so crucial, played a much less important role."

    Ice is pushed away

    "We have now managed to document what has occurred in connection with this change," says Mr Sorteberg.

    The changed wind direction pushes large ice masses away from the Arctic and down along the eastern coast of Greenland. At the same time, less ice forms when the winds over the Arctic are determined by the pressure systems in northern Russia rather than those over the North Atlantic and the Pacific Ocean, as is normally the case.

    The conclusion from this research is that we should be cautious about using the extent of the ice cover as an indicator of the ice's climatic "state of health."

    The extent of the ice cover is highly dependent on the wind direction, and short-term changes in the ice cover give very little indication of whether climate change is occurring in the Arctic.

    "The dramatic changes in the extent of Arctic sea ice in recent years have mainly been caused by atmospheric circulation patterns that have tended to reduce ice cover, combined with a slow process of climate change. Variations in the circulation patterns are part of the natural fluctuations in the weather. In certain periods these fluctuations will reinforce human-made changes, while at other times they will mask them," says Mr Sorteberg.

    Climate change leads to thinner ice

    Mr Sorteberg believes we should be cautious about interpreting the dramatic decrease in Arctic ice cover in the past decade as an indication that the Arctic will be ice free in 10 to 20 years.

    However, he emphasizes that he and his colleagues do not reject the assertion that climate change is affecting Arctic ice cover or that the IPCC is wrong when it states that the Arctic may be nearly ice free in summer towards the end of this century.

    "There is no doubt that the Arctic sea ice has become thinner in recent years. The thickness of the sea ice is a much better indicator than the extent of the ice cover if we want to study how climate change may affect the ice in the Arctic," says Mr Sorteberg."

    It seems like we are well below "normal" ice extent for the last 8 years.

    Espen:

    I like your focus on a longer time period than a single year as the appropriate manner in which to assess climate impact. Rather than make a claim for a specific year generating a new minimum extent figure or predicting the "disappearence" of arctic ice by a particular year, we should be looking at the general trend of multi-year decline in average ice extent.

    Wacrump:

    I think I was my focus was more than just 1 year, when I compare this year with all years since 2002, only 2006 got a little less ice extend than 2010 this time of the year :
    http://www.ijis.iarc.uaf.edu/seaice/extent/AMSRE_Sea_Ice_Extent_L.png
    But anyway, to me it looks like we can see an ice free arctic summer sea very soon, and sooner than many experts expect, and I am not an expert maybe that makes a difference?

    Espen:

    The decline in September ice extent in the period 2002 to 2010 based on the data from the IJIS (using September 9 of 2002 and September 18, 2010 is under 850,000 km2. The year 2002 may be a bad starting point since it was a decline from 2001; however, the 2002 level does not appear to have been significantly different from the lows which occurred in 1995 and 1990 based on the NSIDC chart at:

    http://nsidc.org/arcticseaicenews/2010/100410.html

    The September 2010 minimum of 4.8 million km2 was the third lowest in the satellite record (based on September 18 according to IJIS). By comparison, the 2007 minimum was 4.25 million km2. It took a decrease in excess of 1.5 million km2 from the 2006 September minimum of 5.78 million km2 to reach the 2007 minimum. The previous year minimum for 2005 set a record for a minimum low and was 5.32 million km2, which is only .5 million km2 higher than 2010.

    Predictions of an ice free arctic summer very soon would require that we consistently have absolute declines comparable to 2007 for several years in a row. The data during the satellite period does not include any period which shows that such year after year declines of this magnitude are possible. Declines comparable to the 2007 crash are not gonig to occur as the rate of decline should decrease as the area of sea ice extent decreases. The prediction of an ice free Arctic summer relies too heavily on the anomalous decline in 2007 and does not fit the data we have for the last thirty years.

    The minimum in 2010 represented a decline from 2009, but was almost .6 million km2 higher than the 2007 minimum, even though ice volume in 2010 was lower than 2007. While the Arctic is headed for an ice free or nearly ice free summer during this century, the pattern and data do not suport that this will occur as soon as Patrick's or Frank D's or the prediction you are making.

    In making such a prediction please include a probability or confidence factor. and a specific time period. My non-expert view is that the likelyhood that Arctic sea ice extent will go below 3 million km2 in the next 3 years is very low since there is no historical precedent for such an absolute three year decline and there is no precedent for a percentage decline of this magnitude. Based on the satellite record, there appears to be a more than 50% chance that one of the next three years will be higher than 2007.

    While I appreciate the information you have provided in other posts and your, Frank D's, and certinly Patrick's level of knowledge and analysis concerning the Arctic sea ice, I am reluctant to join in your predictions. Perhaps you guys will be correct, but I just do not see it.

    Based on the conditions that existed at the end of 2007 and knowing that ice volume was going to decline after 2007, would the analysis that you three are applying have predicted that 2010 would be .6 milion km2 higher than 2007?

    I do not think the ice volume anomaly decrease of 11 million km3 relied upon by Frank D will be mantained. The model is already showing a sharp increase in the volume as it returns to the level indicated in the downward trend line. While there is a high probability that a new low will be set in the next 5 years, I am reticent about predicting an ice free arctic or an arctic with less than 1 million km2 at the minimum in the next ten years as this appears to be well outside what the historical data and other sources are predicting.

    Wacrump:

    I understand your arguments and they be may be right, but my gut feelings says something else, because something very wrong is happening, and it is not only up there in the north but also down here where I live, I see it everyday!

    Patrick, FrankD and Espen:

    Sorry to keep coming back, but when I look at the regions of ice coverage I do not see where the future September ice reduction will come from as many of the areas outside the central Arctic Basin were nearly depleted at the September 2010 minimum. There no longer appears to be an opportunity for additional significant September declines in the areas outside the Arctic Basin to sustain a linear downward trend line for the September minimum.

    The areas outside the Arctic Basin, which were the easiest areas to deplete, have almost nothing left to give up and will not be a source of significant future Arctic sea ice loss at the September minimum. The Arctic Basin core will not be as easy to wipe out as these lower latitude regions were. Therefore, the September minimum declining trend line should slow and not continue on a linear basis.

    The central Arctic Basin in the area anamoly chart on Cryosphere today, had a lot of ice left at the September minimum, (looks like 2.5 million km2), and this will not disappear as fast as the ice in areas outside the basin. This area was only .75 million km2 below the 1979 to 2008 average. It should take more than a few years to make this area ice free at the September minimum as it has only shown a 25% or so reduction from the average.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

    Other areas which have suffered significantly higher percentage declines were nearly wiped out at the September 2010 minimum. It looks like most of the ice loss at the September 2010 minimum besides the Arctic Basin came from the Siberian side of the Arctic. The Chukchi Sea melted out completely in September 2010 and therefore can not contribute any additional loss to a future September minimum. The East Siberian Sea was under .1 million km2 in September 2010, the Laptev Sea was under .05 million km2, the Kara Sea had less ice than the Laptev Sea, and the Barents Sea melted out completely, so these areas do not have much left to give up to generate a smaller September minimum than the minimum which occurred in 2010.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.10....
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.9.html
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.8.html
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.7.html
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.6.html

    On the Canadian side, the Beaufort Sea nearly melted out in September 2010, so there is not much left to contribute to a lower September minimum in future years.
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.11....

    The Greenland Sea, which did not show any reduction from the average September extent, only had .1 million km2 of sea ice at the September 2010 minimum, so there is not much left to disappear there. Much of this ice may be ice transported from other areas of the Arctic and there may not be significant additional declines.
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.5.html

    Likewise, the Canadian Archipelago contributed a .1 million km2 deficit to the September 2010 anamoly, but it only had an additional .12 million km2 left to melt out at the 2010 September minimum. This ice is trapped between various land masses and is not subject to the same heat transfer from ocean currents as other regions and may not melt out completely. http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.12....

    Other areas are historically ice free at the September minimum (these show zero September anamoly and zero September area) such as the Bering Sea, St. Lawrence, Baffin Sea, Hudson Bay, and the Sea of Okhotsk, and therefore they do not contribute anything to the loss of ice at the September minimum.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.2.html
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.3.html
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.4.html
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.13....
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.14....

    On this basis, future declines will not be able to generate the same percentage declines that were generated in 2007 or prior years as the easy to melt ice is nearly gone. This may explain, in part, why the ice loss has stalled after the steep decline in 2007. Future years may exhibit a diminishing decline rate in the September minimum compared to the pre 2008 rate.

    The current anamoly deficit of just over 1.05 million km2 appears to be due to a slower than normal freeze up in Hudson Bay and Baffin Bay which did not make any significant contribution to the September minimum anamoly since they were nearly ice free at the September 2010 minimum.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.13....
    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.4.html

    I would expect that Hudson bay would show a substantial recovery in extent by the March 2011 maximum but I am not sure how Baffin Bay will react as this area recieves transported ice from the Nares Strait.

    Let me know what you think of this regional ice analysis method and whether you think it is useful in predicting future declines in the September minimum.

    Colorado Bob
    Patrick -
    Good to see you posting stuff. Been doing a few things as well -



    Scientific American published this article,  Major, Worldwide Damage to Corals Seen This Year. In it , SA quotes C. Mark Eakin, who coordinates the National Oceanic and Atmospheric Administration's Coral Reef Watch. "2010 has been a major, major year of coral bleaching in all of the oceans around the world."
    This disaster unfolded over months the article doesn't give a blow by
    blow account of the thing. I knew there would be reef by reef reports
    hidden away on the web. so far the list has 15 countries :

    http://coloradobob1.newsvine.com/_news/2010/11/12/5457136-global-reports-of-the-2010-coral-bleaching-event-

    Colorado Bob

    The Hindu Kush Before and After the Great Pakistani Floods


    I saw this image from the Swat Valley in August, and I was stunned . Not for what is in the foreground, I was in the Big Thompson Flood in 1976. I know what happens when it rains like hell in the mountains. What struck me are those mountains in the background. That is the Hindu Kush . These two pictures were shot in the last week of July (left), and from the same spot 3 weeks later. Look at all that missing snow.

    At the time, it occurred to me that the Modis Terra satellite would show this record event even better . Three months later I took the time to go look , and do some more "before and after" images. Again I am stunned.

    http://coloradobob1.newsvine.com/_news/2010/11/10/5439030-the-hindu-kush-before-and-after-the-great-pakistani-floods
    Colorado Bob
    I'm about to get the set of passes from July/Aug., 2009 to compare with the set from 2010 .


    Really interesting spot there Colorado Bob. Did you manage to locate comparable images for 2009 or other years? It would be intriguing to see if there is any remotely comparable loss of snow in previous years in MODIS data, though it'll be a substantial job to find out. From your imagery, there are likely to be some very unhappy Himalayan glaciers in that region (IPCC-related irony duly noted!)... [as an aside, I'd guess the long-term mass-balance impacts will be slight, unless the melting can be measured in many tens of metres. Rain can melt ice pretty fast, but I don't know how quickly that kind of remarkable rain could eat into ice]. In the short-term, presumably the effect on regional water resources within the next 6-12 months may be substantial?

    logicman
    Colorado Bob:  you have been busy!  Good work!

    The global coral bleaching is very worrying, but I expect it will be a long time before we get more than a 'so what?' from the usual suspects.

    Your corals link was broken, so I repeat it here:
    http://coloradobob1.newsvine.com/_news/2010/11/12/5457136-global-reports...
    Colorado Bob
    Thanks fixed the link.
    Looks like we are having a relatively heavy melting "downstairs" (antarctica)

    Best regards Espen

    Around 10 years ago I became interested about arctic sea ice melting and have been reading everything I can find on it ever since. Early on, as I became familiar with the subject, it occurred to me that as the ice deteriorated it would become much more susceptible to wind and waves. This would tend to bring it into contact with more sea water and also increase the tendency for ice to be transported southward through the straits. Both of these tendencies would accelerate the deterioration of the ice. I have talked to dozens of knowledgeable people and have not gotten anywhere with this hypothesis. Reading RealClimate and other blogs got me nowhere. Only the postings of Patrick, and other commenters to his posts, have given me any insight on this subject. Thank you !

    ( I think I am correct)

    logicman
    Mike: many thanks for your generous thoughts.

    ... it occurred to me that as the ice deteriorated it would become much more susceptible to wind and waves. This would tend to bring it into contact with more sea water and also increase the tendency for ice to be transported southward through the straits.

    Spot on! 

    Areas of ice in contact with warm water melt faster than areas which are in similarly warm air or are being warmed by the sun.  Every time a piece of ice breaks in two it presents two new surfaces to the sea.  This is a binary sequence - 2, 4, 8, 16 etc. new edges -  which, if continued long enough, will present a greater new edge area to the sea than the original bottom area and edge area combined.  If you add in the effect of waves washing over the ice you get a situation in which - if the sea is warm enough and violent enough - ice cannot be expected to persist for long.
    Thanks again. I guess that what I was thinking was that crushed ice melts faster in a glass of water than one large cube. Add some stirring (wave action) and it melts even faster. Some things are intuitive.

    Frank D:

    I thought about incorporating the volume analysis, but I did not see much of a correlation in the volume declines and the extent changes over the last three years, although intuitively there should be a correlation. In the three years since the low of 2007 there have been significant volume declines, but ice extent has not set a new September low.

    The volume decline analysis certainly suggests an ice free artic, but there is no reason to believe that the future volume declines will be linear. With CRYOST-2 we will have data to watch the changes in volume.

    http://www.esa.int/esaEO/SEMQ7KIRPGG_index_0.html

    Since volume can not become negative, the rate of volume decline will flatten out. The analysis by Wieslaw Maslowski using a linear trend for volume decline is a bit of a cherry pick since the line is drawn only for the period after 1996 rather than the entire period of data presented in the chart. It is an interesting exercise, but ice extent still appears to be tracking recent year numbers rather than showig a significant decline that the ice volume model would indicate.

    The November 30 , 2010 ice extent will be higher than the 2006 level of 9.7 million km2 even though the ice is significantly thinner than in 2006.

    Based on the November 7, 2010 PIOMAS chart, the ice volume anomaly is showing signs of a pick-up, although it is still well below the trend line.

    http://psc.apl.washington.edu/ArcticSeaiceVolume/images/BPIOMASIceVolume...

    Have you seen any studies that correlate the volume chnages with the change in ice extent? I have no explanation for why the volume decline is not showing up in the extent numbers.

    Colorado Bob
    Nov. Observed Events -

    Rain from Anchorage to Barrow an ‘extraordinary event’

    Read more: http://www.adn.com/ 2010/ 11/ 22/ 1568263/ rain-from-an...

    Hudson’s Bay -
    Higher-than-normal temperatures have prevented ice from forming in the region, putting it three to four weeks behind schedule, according to the Canadian Ice Service, a division of Environment Canada.

    Read more: http://www.cbc.ca/ canada/ story/ 2010/ 11/ 17/ hudson-b...

    Record high temperatures return to Russia
    The weather in Russia will remain abnormally warm for at least the next five days, said the head of the Russian Hydrometeorological Center, adding that the record temperatures have also been recorded in Siberia and a number of other Russian regions.
    http://en.rian.ru/ russia/ 20101111/ 161293803.html

    Temp . 2 days ago at Nuuk Greenland -

    15.8C (60.88F)

    http://www.dmi.dk/dmi/vejrarkiv-gl?region=7&year=2010&month=11

    Bob, amazingly warm temperatures in Greenland and Russia. Right now it is abnormally cold here: Ridgefield, Washington, 25 miles north of Portland, Oregon. Current temperature November 24, 2010:
    -6C or 20F
    Forecast low tonight:
    -9C or 16F
    Coldest weather for here this early in the year in about 25 years.
    Very unusual strange weather.

    Colorado Bob

    GISS Surface Temperature Analysis - Nov. 2010 -

    http://data.giss.nasa.gov/cgi-bin/gistemp/do_nmap.py?year_last=2010&month_last=11&sat=4&sst=1&type=anoms&mean_gen=11&year1=2010&year2=2010&base1=1951&base2=1980&radius=1200&pol=pol

    Hudson's Bay didn't start icing until Nov. 23 - 24

    http://arctic.atmos.uiuc.edu/CT/animate.arctic.color.0.html

    Late freeze of sea ice threatens Hudson Bay bears


    Date: Wednesday Dec. 8, 2010 1:47 PM ET




    Scientists say polar bear moms and their cubs near Churchill in
    northern Manitoba are suffering the worst effects of a late freeze-up of
    sea ice on Hudson Bay.



    The bears are just now setting out for the sea ice they use as a
    hunting platform for seals, said University of Alberta researcher Andrew
    Derocher.



    That's weeks later than usual -- and comes on top of an early spring
    thaw that drove the bears off their hunting ground nearly a month sooner
    than usual.



    "This year's been pretty challenging on the population," said
    Derocher from Inuvik, N.W.T. "They were early off the ice and now
    they're late getting on.



    "Some of these bears have had a very long on-land period. A lot of the bears are just running out of steam."

    http://www.ctv.ca/CTVNews/Canada/20101208/polar-bears-101208/

    Hansen -

    The extreme warmth in Northeast Canada is undoubtedly related to the
    fact that Hudson Bay was practically ice free. In the past, including
    the GISS base period 1951-1980, Hudson Bay was largely ice-covered in
    November. The contrast of temperatures at coastal stations in years with
    and without sea ice cover on the neighboring water body is useful for
    illustrating the dramatic effect of sea ice on surface air temperature.
    Sea ice insulates the atmosphere from ocean water warmth, allowing
    surface air to achieve temperatures much lower than that of the ocean.
    It is for this reason that some of the largest positive temperature
    anomalies on the planet occur in the Arctic Ocean as sea ice area has
    decreased in recent years.

    2010 — Global Temperature and Europe's Frigid Air

    http://data.giss.nasa.gov/gistemp/2010november/
    Colorado Bob
    The Ice Island will be on the move early this spring.  The west Greenland coast was still reporting readings in the 50's yesterday.
    I know it is late December now but I think it is worth to report that the ice extend for this time is record low according to the data from http://www.ijis.iarc.uaf.edu/en/home/seaice_extent.htm

    Merry Christmas and a Happy New Year to all of U
    Espen

    Espen:

    What do you make of the decline in ice extent from the December 15th 11,411,406 km2 to the December 20 value of
    11,328,594 km2 (Per IJIS-JAXA)?

    It looks as if the decline in total Arctic ice extent is due to the recent decrease in Hudson Bay ice area of .1 million km2. (Per chart at Cryosphere Today http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.13.... )

    The Arctic sea ice area anomaly at Cryosphere today reported a negative 1.36 million km2 anomaly for December 21. The slower than average freeze-up of Hudson Bay is contributing approximately .5 million km2 to the negative anomaly and Baffin Bay is contributing aproximately .3 million km2 to the anomaly. The Hudson Bay anomaly should return to the average when the maximum freeze up occurs in March of 2011, but the Baffin Bay anomaly may persist.

    The central Arctic Basin is tracking the 1979 to 2008 mean of 4.25 million km2, which appears to be the maximum extent possible for this region. Thttp://arctic.atmos.uiuc.edu/cryosphere/IMAGES/seaice.recent.antarctic.pnghe central Arctic Basin will not be ice free at the September minimum in 2011. Other regions in the Arctic may be within .2 millon km2 of the zero extent - zero volume level in September of 2011 if the melt in these areas is similar to the 2010 pattern, but there will be more than 2.0 million km2 in the Central Arctic Basin in 2011 (an area greater than Alaska) at the September minimum. The 2011 minimum for the Arctic Basin may be as high as the 2009 and 2010 minimum of 2.5 million km2. It will take a 2007 summer wind pattern to cause the Central Arctic Basin ice to go significantly below 2.5 million km2. See figure 8 in the link below:

    http://www.arcus.org/search/seaiceoutlook/2010/pre-release

    I simply believe is due to the extreme weather conditions (not climate) in the area around Greenland and Hudson Bay, temperatures well above normal, that is resulting in very cold conditions in Europe, and that is normal too when this happens

    And on top of this, or rather in the bottom of this, I expect very low ice extend in the Antarctic. So a combined ice extend will be probably be very low this winter season.

    According to JAXA: The latest value : 11,506,719 km2 (December 23, 2010) record low for date. Hudson Bay is well under half the ice coverage expected. Some of the other areas are down but none as significantly as Hudson Bay.

    Anyone know when the ice is considered part of the ice extent. Frazzle, greasy nilas or later? (Yes I looked them up)
    Did we just have a halt in the freeze or a compaction event? Clearly the thinner the ice the more likely a compaction event will cover a greater area.

    Craig Dillon
    Thanks for the update. It is important to understand the graph, to understand its implications. One way of understanding climate vs weather is that climate is just weather that repeats itself. So, warm weahter that repeats itself gives a warm climate, etc. Well, if you look at the graph, notice the 2 std deviation bar. Notice that both 2007 and 2010 are well outside of it, more like 4 std deviations. Now, 3 std. dev. is about 99%. If you look up 2008 and 2009, you will find that they, too, are outside 3 std deviation from the norm. THEREFORE, what we are witnessing is a NEW climate state. A break from the past. 4 years in a row of weather outside 3 std dev. from norm is too much to be just a momentary weather phenomenon. This is a climate change. Another way of looking at it, is that now the Beaufort, Kara, Leptev, East Siberian, Barents, Chuchki seas have all become virtually ice free for each of the last 4 years. Only the Arctic Basin has remained mostly iced over each summer. But, even that is covered with mostly first year ice which is only about 4 to 6 ft thick. How many more years will that last? So, even though the Arctic Ocean is not ice free in summer yet, much of it does become ice free each summer. And with those seas becoming ice free, the resulting loss of albedo, is letting much more solar heat to be absorbed by those seas. [Those seas are shallow, and cover huge reserves of methane clathrate. Will the annual absorption of solar heat during the long summer days warm those seas to release that methane? Just a thought.]
    Colorado Bob

    The Arctic sea ice animation thru Dec 23, notice the retreat of the ice in Hudson’s Bay beginning on the 15th .

    http://arctic.atmos.uiuc.edu/CT/animate.arctic.color.0.htmlhttp://arctic.atmos.uiuc.edu/ CT/ animate.arctic.color.0.html

    Colorado Bob
    Here's the Weather underground page for Greenland, remember it's dark there, on it's way to the dead of winter.

    Narsarsuaq, Greenland

    Currently 41F -

    http://www.wunderground.com/global/stations/04270.html

    The high Sat is forecast 59F in the dark on the first day of 2011.

    Colorado Bob

    Data of Sea Ice Extent

    These Greenland numbers show up here rather well :



    The red line clearly indicates we're not in Kansas any more.
    PATRICK:

    In your original post you stated:

    "Unless the Arctic sees unusually low temperatures before April 2011 - an unlikely event - it is highly likely, I suggest, that the central Arctic ocean will be virtually ice free by the end of Summer 2011."

    While I appreciate your level of knowledge, your suggestion is not supported by an analysis of the Arctic ice in the Central Arctic Basin (called the Arctic Basin region by Cryosphere Today). What is going to cause the huge expanse of ice in this region to disapear in 2011? While temporary open areas can develop within this region, the region as a whole will contain a significant ice area at the September minimum.

    While many regions in the Arctic outside of the Central Arctic Basin are virtualy ice free at the September minimum, the Central Arctic Basin still reports a substantial amount of ice.

    According to Cryosphere Today run by the University of Illinois, the 2010 sea ice area minimum was 3,072,000 km2 which was only 2.6% above the record sea ice area loss of 2007 which stood at 2,992,000 km2 (a bare 80,000 km2 or 2.6% above 2007 sea ice area minimum). Almost all of this ice at the 2010 minimumwas in the Central Arctic Basin where there was aproximately 2.5 million km2 of ice area (this was the same as the 2009 level and was 100,000 km2 higher than 2008 and 400,000 km2 higher than the 2007 record minimum). The 2010 level was only .75 million km2 below the 30 year average.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

    While a new record minimum is likely before 2016, an ice free Arctic is not likely by this date. Predictions based on Arctic wide trend analysis (including the PIOMAS volume trend) are flawed as they ignore the trend for the Central Arctic Basin. The September area (volume data is not available) trend line for this region does not show the rapid decline that the Arctic as a whole has shown. See figure 8 at:

    http://www.arcus.org/search/seaiceoutlook/2010/pre-release

    While the December 31, 2010 Arctic sea ice extent will set a record low number for this date per the JAXA data and may be below 12.0 million km2 (previous record low figure for December 31 was just over 12.2 million km2 which was the low end of a prediction for a December 31, 2010 extent that I posted at the beginning of November), this record low is due to the slow freeze up in Hudson Bay, Baffin Bay, and other regions outside of the Arctic basin. For the northern most regions of the Arctic, the, Cryosphere Today charts indicate that these regions are already at or close to their winter maximum level. The Central Arctic Basin is already at its winter maximum area of 4.25 million km2. This ice wil not vanish in 2011.

    Perhaps the Central Arctic Basin region will show an accelerating rate of decline in September due to the loss of ice in regions surrounding the Central Arctic Basin and added heat from ocean currents and declining albedo, but so far the ice area data produced by the University of Illinois has not shown an increase in melt rate sufficient to clear the Central Arctic Basin of ice by 2016 much less by 2011.

    I believe we will see an ice free arctic sea sooner than most believe, due to climate changes, but it could be next year or the year after depending on the weather conditions, like we now are seeing a slow freeze of the west coast of Greenland and the Hudson Bay, I also believe we will see a record low ice extend around Antarctica this season, and I am happy to inform everyone that commercial use of the Arctic Sea as a route to and from Asia is not feasible according to an "expert" report made for the EU.

    Colorado Bob
    Here's the heat wave that has affected southern and western Greenland for the last 60 days. Narsarsuaq, set 17 new high daily temperature records there since Nov. 19th. It's 56 degrees there now.

    Narsarsuaq, Greenland at 9:50 PM WGT / 12-29-2010

    http://coloradobob1.newsvine.com/_news/2010/12/29/5735005-narsarsuaq-greenland-at-950-pm-wgt-12-29-2010?last=1293691012&threadId=3016017&sp=0&pc=25#last_1


    Colorado Bob
    "The thing that I found absolutely dramatic is that we're up there in November and we're standing on a beach, on an arctic beach, and all we're seeing is waves," he said. "There was not a hint of ice."

    http://www.canada.com/technology/Close%20encounter%20with%20free%20bears...
    wacrump,

    With all due respect, your treatment of the central basin as a distinct entity is not correct. The Basin is protected from melt by the ice in the surrounding seas. Over the last few years those seas have reached a point where they are ice free (which was not historically the case for several), and the heat can attack the central basin properly. I don't have long term data for the central basin, but I would be surprised if it had dropped significantly below 3.5 million sq km coverage at any time prior to 2006 - though I would welcome correction on that point.

    To take an analogy, suppose I have a stack of 20 coins. Every evening I add a coin to the top. Every morning I remove two from the top. So the trend for the whole stack is -1 coin per day. Obviously, at that rate, the stack will be gone in about 20 days. But you are measuring the trend in the bottom five coins, observing (correctly) that there is no daily change in that region, and assuming that that situation will continue more-or-less indefinitely. In fact, once that bottom five section is breached, the whole lot will be gone in 5 days.

    You say that PIOMAS volume analysis is wrong because it ignores the trend in the central basin. I would argue that it is rightfor that very reason. There is no reason to assume that the central basin is somehow protected from melting once an adjoining sea has melted out. And it is much more relevant that the most rigorous analysis of extent or area, because those are derived numbers - they simply reflect the distribution of the meaningful number, the amount of ice (mass or volume).

    As noted earlier in the discussion, "ice-free" is a bit of a moveable feast, because no one really means absolutely zero sea ice in the Arctic. But extending PIOMAS modelled data for September gives an ice-free (actually zero) Arctic for that month in 2016 +/- 3 years (95% confidence).

    I concur that the central basin will last a little while, but your numbers suggest multi-decadal timeframes. I think that is way off, and (again, with respect), you are wasting your efforts analysing the wrong numbers.

    regards, and a happy new year to all.

    I have not been able to get radar images of the central arctic basin, but the ice on the edge of this region in northern Greenland appears thin with a number of small open areas.

    Image of Kenedy Strait including the Petermann Glacier:

    http://ocean.dmi.dk/arctic/images/MODIS/Kennedy/201101040008.ASAR.jpg

    Image of Lincoln Sea and the opening to the Nares strait::

    http://ocean.dmi.dk/arctic/images/MODIS/Lincoln/201101021627.ASAR.jpg

    Image of Eastern portion of Lincoln Sea and North Greenland:

    http://ocean.dmi.dk/arctic/images/MODIS/MorrisJessup/201101032046.ASAR.jpg

    Images of North Eastern Greenland:

    http://ocean.dmi.dk/arctic/images/MODIS/Nord/201101032046.ASAR.jpg

    http://ocean.dmi.dk/arctic/images/MODIS/NEW/201101032046.ASAR.jpg

    Kane Basin ice bridge has not formed and upper eastern portion of Baffin Bay shows broken areas of ice:

    http://ocean.dmi.dk/arctic/images/MODIS/Kane/201101021627.ASAR.jpg

    http://ocean.dmi.dk/arctic/images/MODIS/Qaanaaq/201101020115.ASAR.jpg

    FrankD:

    You and I will continue to disagree, but:

    Re: "I would be surprised if it had dropped significantly below 3.5 million sq km coverage at any time prior to 2006"

    Both sources indicated in the original post show levels below 3.5 million km2 for years prior to 2006.

    Per the graph in the link below prepared by Adriene Tivy it appears all years are below 2.9 million km2 at the minimum.

    See figure 8 at:

    http://www.arcus.org/search/seaiceoutlook/2010/pre-release

    The current graph for the Arctic Basin by Cryosphere today indicates that the 1979 to 2008 average ice area for the Arctic Basin in September was 3.25 milion km2 (2.5 million observed for 2010 plus .75 million for the September anomaly), so some of these years must be below 3.5 million km2.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

    These two sources must be measuring slightly different areas or methods when describing the central Arctic Basin, but both sources indicate numbers lower than the 3.5 million km2 at the minimum for years before 2006. If these sources are wrong, then my observations are not correct, as these are the sources I have relied upon for my "analysis".

    My main point is that the central Arctic Basin will not be "ice free" in 2011. I assure you that the 2011 minimum for the Arctic Basin will not be under 1.5 million km2 (as measured by Cryosphere Today - for reference, the record wind driven minimum in 2007 was 2.1 million km2 and the 2009 and 2010 minimums were at 2.5 million km2). The central Arctic Basin appears to have reached its historical maximum of 4.25 million km2 beginning in mid November, which leaves plenty of time for thickening of the ice. Based on the Cryosphere today chart, in 2010 the central Arctic Basin region was able to maintain 4.0 million km2 of ice until the beginning of July even with the rapid rate of melt shown in other regions of the Arctic. This level appears to be the same as the 1979 to 2008 average. After July 1 the level of ice in the region plummeted by 1.5 mllion km2 to go .75 million km2 below the average. I would expect a similar pattern to prevail in 2011. Perhaps the central Arctic will dip to 2007 levels in 2011, but this would still leave it with plenty of ice.

    While I respect the various predictions of an "ice free" Arctic at some point in the near future, I am challenging the most rapid projection by Dr. Maslowski that it will be as soon as 2016 +/- 3 years (95% confidence). Based on the science literature on this point, I am in good company, because Maslowski does not represent the consensus view of the scientist who embrace AGW and most models show a later ice free date. There are models in the science community that would agree with a multi decade time scale for an ice free Arctic,so this is not some fringe view. I do not have the skill to put a model together to show when the Arctic will show consistently produce "ice free" periods at the September minimum. My central theme is that Arctic wide trends overstate when the central Arctic will be "ice free" and that the "ice free" analysis need only be done for the central Arctic Basin region since the other regions are already at an "ice free" point at the September minimum.

    Your coin analogy has the same flaw as treating the Arctic as one region. Instead of one pile of coins, set up 10 different piles of 20 coins, each of which has a different, but steady rate of coin takeaway. The stack of coins with the slowest takeaway rate will be the last pile to disappear. That is all I am saying, you must look at the rate for the slowest pile, not the average rate for all the piles.

    I respectfully disagree with the statement

    "There is no reason to assume that the central basin is somehow protected from melting once an adjoining sea has melted out."

    The central Arctic Basin is different due to the depth of the ocean compared to surrouding regions, the Arctic above 80 degrees north receives less solar heat input (due to sun angle) than Arctic regions below 80 degrees north, and the average annual temperature for the Arctic above 80 degrees is lower than regions between 66 and 80 degrees north, The water temperature for 80 degrees north is also likely to be lower than that for regions below 80 degrees north, but there does not appear to be any data on this. The ice in regions surrounding the central Arctic basin have pretty much melted out since 2007, and yet, the central Arctic Basin has shown ncreases in ice area, so it must somehow be protected from melting even with this level of melt in surrounding regions. I submit that the current observed extreme low ice levels in Hudson Bay and Baffin Bay do not have a measurable impact on the central Arctic Basin.

    Since the severe minimum of 2007, ice in regions surrounding the central Arctic have nearly disappeared at the minimum. While we only have three years of data since 2007 (which I agree are too few for a trend analysis) it appears that regions outside the central basin are struggling to match their 2007 level (some are up some are down, but the average is unchanged) while the central Arctic has shown an increase of 400,000 km2 since 2007. This increase of ice in the central Arctic Basin is why a new minimum has not occurred since 2007.

    I am willing to concede that the melt rate in the central Arctic will accelerate compared to its historical observed trend due to the loss of ice from surrounding regions, but it will not match the historical rate for the Arctic as a whole, which has seen some regions already ago to a zero ice level at the September minimum (and I do not mean some muddled 20% ice free level, I mean zero).

    Before I can embrace the position that the central Arctic basin (as measured by Cryosphere Today) will be "ice free" at the minimum by 2019 (for this purpose, I will concede that under 400,000 km2 at the minimum is sufficient for the central Arctic Basin to be declared "ice free"), I would at least like to see a couple of years go below 1.5 million km2. Currently, the minimum for this region is at 2.5 million km2 and has not been below 2.1 million km2. It is this observation data that suggests there is a flaw in treating the Arctic wide trend line as being representative of the future melt in the central Arctic Basin.

    These may be the wrong numbers as you say, but they still represent ice, even if it is thin ice.. My guess for this region in 2011 is that it will be above 2.0 million km2 and that between now and 2016 it is likely to stay above 1.5 million km2..

    What is your prediction for the minimum ice area for the central Arctic Basin region for this period (as determined by Cryosphere Today) ?

    Yes we will continue to disagree - I hope that is as okay for you as it is for me. This has been a lot more interesting as a result. I also don't think you represent a "fringe" position, but one which the majority who have ventured an opinion would agree with. My own - "Cassandra-ish", and based largely on a model for which there is little cross-validation (PIOMAS) - is definitely the fringe. Patricks post, and my support for it, is something of a paradigm shift. Every "paradigm shift" starts out as "fringe", but not every "fringe" becomes "paradigm shift" (thankfully).

    However, I would point out that whatever numbers are wafted about, the question of whether area is fundamental or derived remains. More on that later. (BTW -"melt" below is shorthand for all those factors that cause ice loss).

    While my 3.5 should have been 3.0 (misread the CT graph when I added the anomaly back in), the point I was trying and failing to make, still stands. You say that the coin analogy suffers from the same flaw as treating the Arctic as a whole - but that was my whole point, you can't just arbitrarily break up the data and get useful trends out of it. You may be right that Hudson Bay doesn't directly affect the central basin, but most of the other seas do. In any case HB area is irrelevant to September trends, because it is always zero by then anyway. The seas that surround the central basin render it relatively immune from area loss along any boundary as long as those surrounding seas have ice in them. The basin can't lose extent along the boundary with the Beaufort, for example, as long as the Beaufort has a reasonable amount of ice in it (probably in the 10-25% range), because inflowing warm water will melt ice in the Beaufort first, and because any open water that does appear in the central basin will tend to be replaced by ice drifting in from adjacent areas. Once those seas are ice free, inflowing warm water + the ice/water albedo change allow the central basin to melt. The centre can lose volume before area, by "spreading". To use your multiple piles extension of the coin analogy - you can only start taking coins from the central pile once an adjacent pile has been removed, and the rate at which you can take them from the central pile is limited by how many adjacent piles are gone.

    To take the reductio ad absurdum - forget about the whole basin, let's just consider the ~1 million sq km area bounded by Greenland, the Canadian Archipelago and the Pole. Within our observing period, this area would barely have dropped at all (apart from some trivial noise, and the polynya at the head of Nares Strait). Any variation in the Arctic Basin will mostly have occurred in the surrounding areas outside this triangle. Can we deduce from that zero trend that it will not melt in future? Hardly.

    While I agree with your statements of how the central basin differs from the surrounding regions, I disagree that "the central Arctic Basin has shown increases in ice area, so it must somehow be protected from melting even with this level of melt in surrounding regions." Your "somehow" I see purely as a timing issue. You seem to see the annual changes in the central basin as happening simultaneously with those in the other seas. I see it more sequentially. It's not all or one, but to me, it melts from the edge to the core, and freezes from the core to the edge. So you say "the basin is full by November", and I say "of course. And...?" The only protection offered (that I can see) is the heat taken up by melting out the surrounding basins. As long as there is ice between the centre and the warmer seas further south, that will melt before the centre does. Before ~2006, It took most of the melt season just to clear the surrounding seas. More recently they have melted out earlier, and I believe 2011 will continue that process. (Of course, seasonal variability impacts on the detail - if the Beaufort melts out early while other seas are still fairly full, the centre can be attacked on the boundary with the Beaufort and lose area while total area remains high). In a sense, what is key here is time: the melt season is getting longer (2010 started 3-4 weeks earlier and finished 2-3 weeks later in several adjoining seas). That gives warm water, sunshine and warm air more time to work on the central basin.

    But perhaps you can suggest some physical mechanism that intervenes to protect the centre, even in the absence of surrounding ice. The large polynya / diffuse area near the pole in 2010 suggests otherwise.

    You cite an increase in minimal area of 400,000 (15%) as a good thing. Set against a decline in modelled volume of ~25%, to me it just means that the ice is less compact, more mobile (easily exported), more diffuse (easily melted). This feeds the non linear response - if the ice holds its area but keeps losing volume, eventually it will simply collapse.

    For concrete predictions see my first post on this thread: Total sea ice extent for 2011 between 3 and 4 million (IJIS Extent). That probably translates to between 2 and 3 million on CT Area - essentially, little ice one the Russian side, still plenty between the pole and the Canadian islands/Greenland. The Canadian islands will also remain at a decent number - 2010 saw a huge loss of ice but this was replaced by ice from the central basin. I presume that process will continue.

    So we more or less agree on next year. We might even agree on 2012 but after that...? The point I have tried to make throughout is that even if area holds relatively steady for now, if volume continues to drop, that area figure must eventually crash. Tipping points, and all that.

    These numbers are predicated on PIOMAS being an accurate model, which is something of an unknown.

    I am also keenly aware that I do not have a good handle on the detail of all the factors involved here. Even setting aside advection, temperature / cloud feedbacks, ice / water albedo feedbacks etc, there is the question of area versus volume. Both area and volume are relevant metrics, but I don't know which is key. Area is a fundamental variable in some processes, but in others, is simply a derived reflection of the volume and irrelevant to the process. We have processes where area is irrelevant but volume matters (heat absorbed by melting - mass really, but a 1:1 relationship to volume), where area matters but volume is irrelevant (heat absorbed /reflected from insolation), and some where both are relevant (rate of melt from warmer water is a function of volume / surface area).

    Ultimately - "man proposes, nature disposes". We'll see.

    But let me ask you a question - throughout this discussion, you have responded to my remarks about volume with remarks about areal trends. So setting aside statistics, do you consider volume to be relevant in deetermining this stuff, or is area the only factor?

    Based on your post, I decided to take a closer look at the validity of my position that the central Arctic Basin can be treated as a separate and distinct region in assessing iArctic sea ce trends.

    The various regions of the Arctic that Cryosphere Today uses in its ice reports are not as arbitrary as the suggestion of only looking at the core 1 million km2 to determine ice trends, which as you astutely pointed out (and I agree) would be an invalid manner of establishing a region to test sea ice trends. In particular, I looked at the following seas that surround the Arctic Basin with a view of determining if there is a sufficient basis to treat these regions seperately from the Arctic Basin:

    Beaufort Sea
    Chukchi Sea
    East Siberian Sea
    Laptev Sea
    Kara Sea
    Barents Sea

    Based on the geographic descriptions for the above seas I am comfortable with separating them from the central Arctic Basin in looking at sea ice trends(I know that is not surprising given my bias) . Each of these regions is based on geographic descriptions that were made without reference to ice conditions. These regions are bounded by land in many cases and are fed by fresh water inflow from various rivers. They appear to be rather shallow compared to the central basin and are subject to dirfferent currents. The sea water in these regions appears to have different salinity and temperture profiles than the central Arctic.

    For example, I found this short piece on the Laptev Sea:

    "The Laptev Sea is a major source of arctic sea ice. With an average outflow of 483,000 km2 per year over the period 1979–1995, it contributes more sea ice than the Barents Sea, Kara Sea, East Siberian Sea and Chukchi Sea combined. Over this period, the annual outflow fluctuated between 251,000 km2 in 1984–85 and 732,000 km2 in 1988–89. The sea exports significant amounts of sea ice in all months but July, August and September.[7]

    Ice formation starts in September on the north and October on the south. It results in a large continuous sheet of ice, with the thickness up to 2 meters (6 ft 7 in) in the south-eastern part of the sea as well as near the coast.[7] The coastal sheet ends at the water depth of 20–25 m which occurs at several hundred kilometers from the shore, thus this coastal ice covers some 30% of the sea area. Ice is drifting north to this coastal band,[3] and several polynyas are formed by the warm south winds around there. They have various names, such as the Great Siberian Polynya, and can stretch over many hundreds kilometers.[3] The ice sheet starts melting in late May-early June, creating fragmented ice agglomerates on the north-west and south-east and often revealing remains of the mammoths. The ice formation varies from year to year, with the sea either clear or completely covered with ice."

    I agree that the melting of the ice in the seas bordering the central Arctic does have an impact on the manner in which the central Arctic will melt in the future, but I am reasonably comfortable that a sufficient non-arbritrary basis exists for looking at this region seperately from the surrounding sea in determining ice trends.

    Please feel to disagree, but I think there is a basis for treating the central arctic basin separately from these regions.

    Will,

    There's quite a bit to work through, so just some quick responses to points raised in individual posts.

    I'll concede the point that there is a basis for treating the various seas as distinct geographical entities. All boundaries except Chukchi/Beaufort are marked by islands that no doubt create somewhat different conditions in each sea. Equally the southern edge for all is land, so there is no doubt about the validity of drawing a border there. But what is the basis for demarking the northern edge of these seas (and seperation from the Arctic Basin? I venture to suggest that it was the northernmost point readily accessible by explorers before reaching the "permanent" ice cap. And since we are considering a declining ice cap, perhaps those definitions are now rather arbitrary...?

    Barents, Kara, and Greenland Seas do have reasonable northern boundaries. Laptev, E Siberian, Chukchi and Beaufort I am less convinced about.

    Sorry to keep bombarding you, but this is helping me think through the analysis in depth (sorry for the pun).

    I have been looking for sources which would show changes in ice volume in the central Arctic Basin.

    One source I have found is an image based on ICESAT data from 2004 through 2008 at the following link:

    http://www.nasa.gov/images/content/365869main_earth2-20090707-full.jpg

    Additional information through 2010 that includes "ice concentration" is at the following link:

    http://nsidc.org/sotc/sea_ice.html

    Generally, ice thickness at the North Pole has never been particularly thick and it agenerally appears to be close to the thickness of first year ice of approximately 1.7 meters and is at most 2.5 meters. The ice thickness at the winter maximum at the North Pole does not appear appreciably thicker than ice in the seas surrounding the Arctic Basin which become "ice free" at the September minimum. (Please let me know if you have a different view or sources that indicate a different view of north pole ice thickness.) This may explain why the submarine in open waters at the north pole pictures that the denialists are so found of are so prevalent. The real trick would have been to see the subs surface just north of Canada or Greenland in the heart of the thickest ice of 5 meters or more (might not happen without risking damage to the sub).. I would guess that this thin ice at the North Pole would be particularly susceptible to disappearing during the melt season so there must be some mechanism to replenish this ice as it melts.

    The thickest ice appears along the north coast of Greenland and the Canadian Archipeligo. This multi-year ice appears to require a land mass that it can "pile up" on due to wind and ocean surface current conditions. With the exception of the north coast of Greenland and the Canadian Archipeligo there is no place for this to occur in the central Arctic basin and the ice in this portion of the basin where the ice does not "pile-up" is not that thick (except perhaps at lines in the ice where pressure ridges form). The piling up of ice on the north coast of Greenland and the Canadian Archipeligo may make the realization of an "ice free" summer Arctic difficult. This may partially explain why the loss of ice volume in these two portions of the central Arctic Basin has had such a small impact on ice extent. in the Arctic Basin. The complete loss of ice in the regions surrounding the central Arctic Basin has also contributed heavily to the volume loss with no impact on central Arctic Basin ice extent, which leads me to believe that there is some manner in which the central Arctic ic protected from loss of ice extent that does not operate in the regions surrounding the central Arctic Basin. Assuming the ice in the central Arctic Basin has dramatically thinned, it has not resulted in significant open water areas at the ice minimum compared to the regions surrounding the Arctic Basin with similar ice thickness.

    So what is it that "protects" ice extent in the central Arctic Basin you ask?

    The melting view you expressed above of a static ice sheet melting from the outside edge toward the center, while intuitve, may not accurately represent what actually occurs. The Arctic ice does not stay in one place and just melt. It is in constant motion. Generally, it appears that winds and surface currents push the ice from the sea regions surrounding the Arctic Basin into the Arctic Basin. This transport of ice into the central Arctic Basin may provide a "fresh" supply of ice to replace the portion of the winter maximum cover that disappears during the melt season. Thus, even if the winter maximum ice thickness in the central Arctic Basin has declined significantly, it will not necessarily result in open water areas at the summer minimum as ice that escapes the Beaufort Gyre and the seas near Russia replaces the ice that melted or was otherwise been transported out of the Arctic Basin.

    This is the mechanism which "protects" the central Arctic Basin ice extent from declining as rapidly as ice in other regions of the Arctic that do not receive a "fresh" supply of ice from other regions during the melt season. Based on bouy movements, it appears that the ice cover in the central Arctic Basin in September consists of ice that came from other parts of the Arctic and not ice that was at that location at the winter maximum. This suggests that the central Arctic Basin will not become "ice free" at the summer minimum due to the "melting" of its winter maximum ice cover, but rather will become "ice free" when the winter maximum ice cover in regions surrounding the central Arctic Basin diminishes to the point that it does not survive the melt season trip into the central arctic Basin..

    Below are links that show the Arctic ice minimum of September of 2007 and 2010. In order for the prediction of an "ice free" Arctic to occur these significant sized white areas will have to turn blue. At this point I am not convinced that this will happen in the time frame suggested by Dr. Maslowski.

    http://nsidc.org/images/arcticseaicenews/20100927_Figure1.png

    At the 2007 minimum, there was an ice extension that went to Siberia and divided the open water into two areas.

    http://nsidc.org/news/press/2007_seaiceminimum/images/20071017_extent.png

    Re: ice thickness at the Pole. It is true that ice is thicker on the Canadian/Greenland side than at the pole itself - a point that recurs in my descriptions above of where I think ice will or will not be lost. Ice at the pole itself is variable. For example, when USS Nautilus became the first submarine to transit the North Pole, it was not able to surface due to thick ice. When USS Skate followed the next year, they were able to surface (and scattered Sir Hubert Wilkins ashes there). But I suggest you look at the MODIS satellite mosaics for late August / early September 2010, and observe the state of the ice near the north pole in the direction of the Bering Strait. Right there is where you can see my problem with extent as a measure. There is the better part of 1 million square kilometres of "brash" - a slurry of ground up ice and water that is treated the same way as a five metre thick slab butting up against Greenland. Clearly the two are not equivalent, and should not be treated the same way - remember that anything greater than 15% concentration counts (30% for some gruops), so the pack can be very diffuse and still rack up extent.

    "it has not resulted in significant open water areas"- In 2010 there were very significant areas of open water, however, they were mixed on a relatively fine scale with poor quality ice (as above). It still counted as extent because it was above the 15% threshold, but it was open water with ice floes, not pack ice. The Healy blog illustrated this clearly, for example, making good speed though water with thin floes that was marked on IJIS / CT etc as decent conentration ice. It is my personal opinion that nearly 1 million sq km's of "sea ice extent" was little more than the sort of crushed ice you'd find in a good margarita.

    Your link to 2004-2008 maps was interesting, but I urge you to seek 2009 and 2010 data (I'd post links but my "Holidays = continuous YouTubing" teen has maxed our bandwidth, and I'm a bit limited to dig them out for now). You can find some here at the Chatterbox though. A large proportion of that multi-year ice has been dragged into the Beaufort Sea, and melted there over the last couple of years. It has resisted melting for a long time - you can see it as a finger pointing into the Beaufort on the 2010 NSDIC map you linked to. It helped keep a high extent in 2010. Had that ice remained on the coast of the archipelago, it would have caused a lower extent, but a higher volume (which in my opinion would have been a healthier result).

    I agree that I was oversimplifying with my depiction of a static sheet. It is dynamic, but if last season is anything to go by, transport is becoming less predictable as conditions change (a point I made early in the discussion). For the point I was trying to convey, transport simply confused the issues, which was why I simplified. But your description of transport into the basis surprises, and I'm off to read some more.

    Sorry, that last sentence should read "transport into the basin".

    Will,

    Coincidentally, I was reading RealClimate today and came across these maps, illustrating the change in old thick ice over the last two years:
    6 Jan 2009: http://topaz.nersc.no/topazVisual/matlab_static_image.php?action=NA_ARC_...

    6 Jan 2011: http://topaz.nersc.no/topazVisual/matlab_static_image.php?action=NA_ARC_...

    There are a number of interesting points of comparison (eg Chukchi Sea), but the great big one is the absence of thick (>2 metre ice) along the north shore of the archipelago / Greenland.

    FrankD:

    A problem with using a trend line analysis for the volume data is that it creates a result that is not believable. The following trend line analysis that was done by Neven shows that the Arctic will be ice free in December of 2023 and will be ice free from July through December. See link below:

    http://img543.imageshack.us/img543/2145/piomasmonthlyvolumes.png

    If the volume trend line analysis does not produce a believable result for July, August, October and December, why should I believe that the trend line analysis for September is valid?

    I suspect that the rate of future declines in volume are not going to accelerate in the manner predicted by the trend lines that Neven has created. While I am impressed with his graphing skills and knowledge of Arctic ice conditions, I do not believe this graph provides an accurate prediction of future trends. I suspect that it is far more likely that the rate of decline in volume will slow (in a fashion similar to the concept of diminshing returns). Too many regions have reached a zero volume level at the September minimum. These regions have nothing to contribute to future volume declines. The thick multi-year ice north of Greenland and north of the Canadian Archipelago is substantially diminished and may not have as much to contribute to future volume declines as it has in the past.

    Another reason these trend lines are not reliable is that they predict that north of Greenland and north of the Canadian Archipelago will be ice free by 2016 and that is not believable. These two areas are likely to have thinner ice by 2016, but there is no reason to believe they will not be able to maintain at least a layer of first year ice in 2016 given the manner in which the winds and currents drive the ice from other areas of the Arctic toward these two areas.

    Will, if you reread the thread at Nevens where you found that image, you will note that Neven didn't produce it. But someone posting under the name FrankD. That would be me.
    Further if you re-read the thread you will note:
    1. That I performed the exercise because the quadratic trend gives a clearly better fit that the linear trend on the original PIOMAS graph. It exists more as thought experiment than a geniuine prediction, as it was hedged with comments about why it was inaccurate.
    2. I explicitly stated that I thought that the curve was s-shaped, not the quadratic I'd used, but had posted the results because I hope others might do a better job of it. In fact the only other commentator who did has come up with even more "alarming" numbers.
    3. The discussion talks a lot about the tail of the s-curve (your "law of diminishing returns"). But here's the rub. Assuming the model is relatively accurate, September volume loss can only sustain 3 more years at the rate of the last couple of years before it is gone. I believe other factors will intervene to extend that a little, but not much.

    But I yet again we seem to come to the same point - you talk about "regions that can't contribute any more loss". I do not understand what you mean in a physical sense. If a region is ice free then the physical effect is to allow more heat to enter deeper into the Arctic and melt more of your central region where ice does still exist. This is why I have repeated talked about the acceleration of melt in the central basin. All this means to me is that ever growing amounts of heat are working on ever smaller volumes of ice for longer. I know what the result of that is, but I'm honestly not sure what you think will happen.

    I agree that what is left of the thicker ice on the Canadian side will resist for longer, and that is not represented well in the graph I've done. But at September 2010, Ice volume was 4000 cubic kilometres, a decline of 1500 cubic kilometres in one year, and less than 25% of the long term average September volume. As I say, its only going to take three more years at that rate to melt the remainder. The only thing that can really stop that are:
    1 - changed transport conditions - as above, I think that opening the NW Passage makes that worse, not better. You say "there is no reason to believe they will not maintain...", I think there is every reason to so believe.
    2 -reduced heat entering the Arctic, which seems unlikely - most likely changes, including the net effects of changing radiation / cloud / albedo feedbacks, look worse not better.
    3 - the addition of extra ice from glaciers in Greenland and Canada, which migh preserve the numbers a little, but is hardly a good sign.

    If its any consolation, when I ran the numbers, I didn't believe them either. However what you or I believe doesn't really matter. If you can advance a reason why the curves are wrong (beyond the agreed addition of a small s-curve tail), I'd honestly love to hear it because I found the results quite depressing and would gladly be proved wrong. But Argument from Incredulity is a textbook logical fallacy, and is to be avoided. Your reasoning elsewhere has been generally sound, but this isn't.

    I am impressed by your graphing skills!

    I have gone back to Neven's site and I see where you invoke the s-shaped curve analysis. Sorry I missed that this was your graph, there are a lot of good comments to go through on his web site and I will go back and read them.

    I have seen a similar s shaped curve suggested by Robert Grumbine on his site and I have posted a request that he perform his analysis using only data for the central Arctic basin, but I expect that will go nowhere.

    My main point continues to be that I think there are weaknesses in applying the trend analysis based on the Arctic as a whole. I have asked that the trend analysis be done only with respect to the central Arctic Basin because it is clear such a trend analysis will produce a later "ice free" date than the Arctic as a whole analysis. While I have aggressively pushed looking only at the central Arctic basin in our discussions, my view is that the "ice free" date can not be accurately predicted using trend line analysis. Additionally, I am concerned that a failure of these aggressive "ice free" predictions will be seized upon by the misinformation groups as "proof" that all AGW predictions are invalid.

    Experience has taught me that applying trend line analysis to complex processes that are not well understood provides a false sense of predictive power. There are multiple processes that are causing the disappearance of the Arctic ice and the interaction of these processes are too complex to think that they will play out in some neat fashion that can be predicted by trend line analysis. Based on the reading I have done, it appears that there has been a dramatic shift in air pressure patterns in the Arctic with something called the Arctic dipole anomaly and some others that I can't remember just now and there are some indications of shifts in the amount of heat brought into the Arctic from the Atlantic. It does not appear to me that the causes of these phenomenon are sufficiently understood that an accurate prediction can be made of how they will play out in the future.

    The recent crash in ice volume is more dependent on these phenomenon than it is on any simple model predicting higher temperatures due to increased CO2 forcing. Believing that these newly witnessed conditions will continue to play out as they have in the last few years may be correct, but it is also possible that these pressure differences could return to levels that do not promote such extreme ice melting conditions.

    I think we all go to the trend line analysis because it is so much easier than looking at the messy complexity of the various factors that are causing the loss of the ice. The IPCC4 models appear to have substantially under estimated Arctic ice loss and I think Dr. Maslowski's trend line analysis is over estimating ice loss.

    I found the drop in ice volume in the PIOMAS model rather shocking and it made me wonder if there was a problem with the model, even though it tested well against the ICESAT data. Even accepting the PIOMAS model, it appears the ice volume anomaly is very volatile as the December 31, 2010 PIOMAS model is showing a sharp pick-up of 3000 km3 in the volume anomaly.

    http://psc.apl.washington.edu/ArcticSeaiceVolume/images/BPIOMASIceVolume...

    More years, more research, more data and Cryosat-2 will no doubt provide a better answer to the future of Arctic ice than I can.

    Just to clarify one point, when I say that the already zero regions will not contribute to future declines in volume, I mean that the region itself has no more ice to lose and therefore that the region will not show any future decline in ice volume. Your point that an ice free area can contribute to future declines in other regions by is well taken.

    I will try to get to some of the other questions you have asked.

    I thank you for taking time out of your busy schedule to respond to my musings I will look at your other postings here and elsewhere to advance my understanding of the Arctic ice.

    Hi again Will,

    Did everyone else leave the pub? And where is our friendly host?

    I just did those curves on Excel, so they're nothing special. If I ever find time, I'm gonna get me a copy of R, and do some real stats!

    I agree with you comment about possible limitations of PIOMAS. In this regard I feel I can watch this, and even take a turn with the dice, with a certain degree of detatchment. No prediction of mine will change anyone's mind about AGW in general, if I'm right no one will notice and if I'm wrong, no one will care (least of all me). I find the subject interefascinating, but I am strictly the interestedd amateur, so I have nothing riding on it.

    Again, kicking this around with you has turned over many interesting stones and given me much to think about before we get into the next melt season.

    You are of course correct that what is needed for accurate prediction is not a simple numerical model but an authoratative physics model. I don't know about you, but before I get one of them, I need a BSc, an MSc, a PhD and some funding to book some hours on a supercomputer. I'm pretty sure that even your timeline will see an ice-free Arctic before that happens! :^P

    See below for more on PIOMAS anomalies.

    FrankD:

    Thank you for your taking the time to respond. I hope to pick up some pointers from this discussion to sharpen my limited understanding of Arctic Sea Ice.

    I readily concede that volume is a more important indicator of the condition and quality of the Arctic Sea ice than area or extent and it is obvious that all three of these indicators will reach a zero level at the same time.

    I agree with your 1 million km2 example, as you could always cherry pick a sufficiently small area to show no change even from one season to the next, however, the region I have selected is far from inconsequential at 4.25 million km2 and does shrink to 2 to 2.5 million km2 with the seasons. This region is reacting to the warming of the Arctic as the 2010 anomaly was .75 million km2 below the average. My view is that it is not reacting as fast as the Arctic as a whole.

    I am just being cautious as I would prefer to see the central arctic basin region drop below 1.5 million km2 for a couple of years due to something besides the wind direction before making a 2020 "ice free" prediction. That would indicate that the volume analysis for the Arctic as a whole method by Dr. Maslowski has more merit than I have been giving it.

    I have been directing the discussion to area and extent when determining whether the arctic is "ice free" as this is the easiest measure for me to "see" when looking at the various satellite images and there appear to be more data sets for two dimensions than there are for three dimensions. Volume is harder for me to "see". An Arctic that has a thin veneer of ice might be considered "ice free" using the volume analysis , but I might still be able to "see" an ice covered white area in the NSIDC simulated image rather than a blue area used for "water" or the color scheme used in the MODIS images or the Cryosphere Today image. For a lay person like myself, the Arctic is not ice free if the image shows white stuff. If the images were done on the basis of volume and ice depth of less than say, three inches , was treated as water then the image might appear "ice free". How thick does the ice have to be before the satellite measures it as ice instead of water?

    I am aware that the satellite measurements sometimes denote areas as white even if they only contain 15% ice, which is why I have used area based information rather than extent. I have read that there are issues with sensor function where melt ponds appear on the surface of the ice. I have also noticed in looking at the 2 dimension oriented data sets that there are several different ways the ice is "interpreted". It is very confusing, so I try to indicate which data set I am looking at when discussing the ice.

    Based on the PIOMAS volume model and observations of area and extent, it appears to me that the volume is decreasing faster than area and extent in the central Arctic Basin, although I have not seen any volume data for the basin that would allow me to back up this claim. My expectation is that the decline in volume for the Arctic as a whole will slow because the oldest and thickest ice is gone and the "easiest to melt" ice is gone and I expect the decline in 2-dimensional ice in the central arctic basin will accelerate for all the reasons you mention. I also expect that the Arctic Basin will show greater inter annual variability in area and extent as the ice conditions become more sensitive to winds that push the ice away from Asia toward Canada and Greenland, but I do not see that this will result in consistently "ice free" conditions for the central Arctic before 2020.

    Some of the research suggests that the arctic will shrink in stages rather than a steady decline with some interim new equilibrium points temporarily established on the way to an ice free Arctic. The Arctic and Antarctic will become "ice free" at some point, since they have been ice free in the past when global temperatures rose to the level expected from anthropogenic forcings. It is just a matter of timing.

    One small aside, I do not think the increase in 400,000 km2 was a good thing, I am only noting that it ocurred. A "good thing" would be a return to the volume and extent levels that existed in 1970, but that is not going to happen.

    Will,

    Back at you - we disagree, but your comments have helped me question my own thinking and improve my knowledge, so I have greatly enjoyed them (even if I've argued.. err, firmly(?).. with a lot of them).

    Your last installment contains much that we can agree on. In fact there are no satellite series for volume as yet. The first satellite capable of delivering that data was launched only last year. PIOMAS and PIPS are models, using satellite images with other data collection methods. The constant rider on any discussion of volume is "if its accurate".

    The issue of what counts is vexed. I think you are correct that melt ponds reduce the number wrongly, but see my earlier reply tonight for some observations about how it can be over-estimated as well. My opinion is that as volume has declined, there are more areas of loose uncompacted ice of indifferent thickness. This transition reduces volume in "real time", but this loose ice spreads out an extent is preserved artificially high. Area is also preserved, but to a lesser degree. This is why I think volume is crashing while extent is showing a much slower trend. That imbalance cannot continue forever. Either new feedbacks will come into play to preserve the remaining volume, or, more likely IMO, extent will eventually crash.

    I see in our discussions something of the catastrophist -v- gradualist division that has been a feature of earth sciences for two centuries. Several points you make suggest changes occuring in gradual transitions. I think the transitional period we seem to be in is quite untable, and we are more likely to see a rapid switching between states (of course rapid is a relative term depending on whether you are talking about Arctic Sea Ice or the Greenland Icecap). The term "tipping point" has become a bit passé lately, but it still works for me.

    I was today rather pointedly corrected on the fact that Dr Maslowski's forecast is explicitly for the central Arctic basin, not for the Arctic as a whole. For the purposes of our discussion, its probably much of a muchness, but it does indicate that he too was focussed on trends in the central area. The declining volume from 1995 onwards is sepecifically in that central region, and with that in mind, you may find it rewarding to review his graph again - in particular, I found the annual fluctuation in volume (again, in this confined region) quite interesting.

    wrt to your remark about volume decreasing faster than extent, I completely agree, and you can use Maslowski's data with the area / extent information you obviously already have to explore that. I'm not sure if his definition of the central basin corresponds exactly with the various 2d datasets.

    Best
    FD.

    FrankD:

    You stated that there are no satellite series for volume. Was ICESAT capable of measuring ice thickness when it was operating?

    http://www.nasa.gov/images/content/365869main_earth2-20090707-full.jpg

    The PIOMAS volume model has shown significant fluctuations recently. In September of 2010, the negative anomaly was 9,400 km3. After September, the anomaly reached a negative anomaly of 11,000 km3, but it "recovered" to a negative anomaly of 8,000 km3 by December 31, 2010. Such a large fluctuation in a short periiod of time appears unusual.

    http://psc.apl.washington.edu/ArcticSeaiceVolume/images/BPIOMASIceVolume...

    Due to the season, ice volume is increasing but I did not expect the volume anomaly to go up that fast. CRYOSAT-2 may be able to resolve the accuracy of PIOMAS and explain why this occurred.

    The negative volume anomaly is not showing up in the Ice extent/area data as generally the Arctic basin regions are at the maximum with the exception of the Barents Sea, which has more ice extent this year than the last four years, and according to Cryosphere Today's anomaly chart, the Barents Sea region is almost at the average.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.6.html

    Hudson Bay and Baffin Bay are reporting significant negative area anomalies, but I do not think this is driving the volume anomaly calculated by PIOMAS as these areas consist of first year ice that is likely to be thinner than the first year ice in the Arctic basin regions as it freezes at a later date than the Arctic basin regions..

    My comment about the Barents Sea is based on some charts I have came across from the NSIDC (have you seen these before?) that compare the current period to the same period in the previous 4 years on a region by region basis at:

    http://nsidc.org/data/masie/masie_plots.html

    According to the PIOMAS linear trend for volume, the rate of decline is 3,500 km3 per decade and

    "The model mean seasonal cycle of sea ice volume ranges from 28,600 km^3 in April to 13,400 km^3 in September. The blue line represents the trend calculated from January 1 1979 to the most recent date indicated on the figure. Monthly average Arctic Ice Volume for Sept 2010 was 4,000 km^3, the lowest over the 1979-2010 period, 78% below the 1979 maximum and 9,400 km^3 or 70% below its mean for the 1979-2009 period."

    The Polar Science Center that issues the volume graph has a model that indicates that the Arctic does not become ice-free in late summer until around 2050.

    Below is the September 2015 prediction that shows ice in the central Arctic Basin:

    http://psc.apl.washington.edu/BEST/PSW2007/ICEPREDICTIONS/heff2015.gif

    They even show some ice at 2025:

    http://psc.apl.washington.edu/BEST/PSW2007/ICEPREDICTIONS/heff2025.gif

    It is nearly "ice free" by 2045.

    http://psc.apl.washington.edu/BEST/PSW2007/ICEPREDICTIONS/heff2045.gif

    Do you know why Dr. Maslowski, who is using data generated by this group, came up with such a different estimate of an ice free date?

    The NSIDC posted the following chart concerning the age of Arctic ice in October of 2010:

    http://nsidc.org/images/arcticseaicenews/20101004_Figure6.jpg

    While the more than 2 year-old ice appears to be at a record low proportion of the total, the chart shows that there is a significant base of 1 to 2 year ice that may result in a higher proportion of more than 2 year-old ice in 2011. Do you think that the average ice age in 2011 will be greater than 2010 and therefore the negative volume anomaly per PIOMAS will return to the level indicated by the blue trend line?

    The NSIDC made the following comments related to ice volume in early October:

    "[In 2010} the wind patterns associated with the negative phase of the Arctic Oscillation transported a great deal of multiyear ice from the coast of the Canadian Arctic into the Beaufort and Chukchi seas. Scientists speculated that much of this ice, some five years or older, would survive the summer melt period. Instead, it mostly melted away. At the end of the summer 2010, under 15% of the ice remaining the Arctic was more than two years old, compared to 50 to 60% during the 1980s. There is virtually none of the oldest (at least five years old) ice remaining in the Arctic (less than 60,000 square kilometers [23,000 square miles] compared to 2 million square kilometers [722,000 square miles] during the 1980s).

    Whether younger multiyear ice (two or three years old) in the Arctic Ocean will continue to age and thicken depends on two things: first, how much of that ice stays in the Arctic instead of exiting into the North Atlantic through Fram Strait; and second, whether the ice survives its transit across the Beaufort and Chukchi Seas or instead melts away."

    The melting of the multiyear ice from the coast of the Canadian Arctic may be what drove the negative jump in the volume anomaly.

    In early December the NSIDC updated this description saying:

    "Recent research from scientists at the NASA Jet Propulsion Laboratory shows that summer melt of old, thick ice in the Beaufort Sea has contributed substantially to the overall loss of thick multiyear ice in the Arctic. Using data from the QuikSCAT satellite, researchers Ron Kwok and Glenn Cunningham found that the Beaufort Sea lost 947,000 square kilometers (366,000 square miles) of multiyear ice during the summers of 1993 to 2009.

    The study also showed that multiyear ice loss increased in the last few years. From 2005 to 2008, the Beaufort Sea lost 490,000 square kilometers (189,000 square miles) of multiyear ice, 32% of the total loss of multiyear ice in the Arctic Ocean during that time period. "

    In January the NSIDC noted that a negative Arctic oscillation, which had in the past operated to preserve multi-year ice, had operated in 2010 in a manner that increased melt of the multi-year ice. This may have prevented the pool of more than 2 year old ice from increasing in 2010. Per the NSIDC:

    "2010 started out with a highly negative phase of the Arctic Oscillation, an atmospheric pattern that in the past has favored the survival of old ice through the winter, and more ice at the end of this summer. But this tendency seems to be changing. A recent study led by Julienne Strove of NSIDC showed that while wind patterns linked with the strongly negative Arctic Oscillation winter of 2009-2010 transported much old ice into the southern Beaufort and Chukchi Seas, most of this ice later melted. It may be that with a warmer Arctic, old rules regarding links between the atmospheric pressure patterns and sea ice extent no longer hold. So far the winter of 2010-2011 has also had a strongly negative Arctic Oscillation, but it is not yet clear how this pattern will affect summer sea ice."

    I take it that you expect the ice volume anomaly to continue to show an increase in the negative anomaly in 2011. I am not certain it will play out that way . The long term trend is for less ice, but that would not prevent individual years from showing an increase in ice volume from the prior year.

    Sorry about using the incredulity argument in reference to your graph. I am certain you put a good deal of effort into the graph, but the idea of an "ice free" Arctic in December of 2023 was unexpected given that most of the Arctic basin regions reached their maximum area in December of 2010.

    Thanks again for all the information and direction you have provided.

    Sorry Will, I was a bit sweeping with that - ICESAT was intended to return thickness data, but was a troubled mission. One of the main instruments failed early in the mission, limiting the accuracy of the data and the duration of the mission. While it was active for ~6 years, for most of that time it only returned data intermittently. But I should have said, we don't have reliable long term reasonably complete satellite data.

    WRT to PIOMAS fluctuations, your data points are a bit out. The max anomaly (-11,000) was at the end of June, not in September (indeed, that was one of two questions that led me to attempt to pull data from the PIOMAS graph). And while the recovery was quick, it also makes sense when considered in a bit more detail, particularly in comparison to extent.

    The maximum anomaly, as I say, was at the end of June, when extent numbers were going through the floor. It corresponds to the maximum negative Cryosphere Today anomaly too. Bad weather in July-Sept the slowed things down - note that the anomaly recovered, even though absolute volume was still dropping. The numbers averaged for the month were:
    June: 14,700 (-10,600 against an average of 25,300)
    July: 8,800 (-10,500 against an average of 19,300)
    August: 4,900 (-9,700 against an average of 14,600)
    September: 4,100 (-9,300 against an average of 13,400)
    October: 5,800 (-8,700 against an average of 14,500)
    November: 8,900 (-8,200 against an average of 17,100)
    Figures are my interpolations from PIOMAS graphs of anomaly + average as at the mid point of the month (which is not the same as a true monthly average but was close enough for my purposes). So all we can really say is that we lost a lot of ice in June and a bit less than average in subsequent months, with a modest recovery following minimum. Which is pretty much the same as every other dataset.

    The other question that led to my PIOMAS re-graphing was that to my eyes, overlaying a linear trend on the anomaly data they are presenting is questionable. I mean, does it look linear to you. To me, it's as plain as a Bulgarian pinup that that is a quadratic curve, not a straight line. I can only presume that the 2015 pictures are premised on that straight line extrapolation, which I just think is wrong. In effect they are saying that the state of the ice in five years time is going to be better than in four out of the last five years. In 15 years, winter ice will be healthier than it is now. I wish it were so, but I'll believe it when I see it. I don't know how the "heff" forecasts were made, but I presume that Maslowski (who was still using linear trends) was extrapolating over a shorter period, thereby deriving steeper gradients on his trend lines.

    Ice Age: The key factor in your question is whether ice extent for 2011 will be greater than 2010. I don't think so. My first projection has a pretty wide margin for error with the upper bound being similar to 2010. I'd give it maybe 10% chance of being above, 90% below. Do I think volume will "return to the blue trend line"? Not a prayer. Will the anomaly decrease (get closer to the blue line)? Possibly, guesstimate 30% likely. From the data above, the anomaly could on average be lower this year while still achieving a record low volume - it depends on the timing, and I'm nobodies idea of a weather forecaster. I would predict 2011 will be lower on extent and volume, but its a bit early to give hard numbers. But therefore, I think old ice extent will be lower this year than last. Of course you are correct in referring to single year fluctuations, but the last couple of months has not augured well.

    I would touch on one of my own hobby horses - the NorthWest Passage. I've said it above, but its worth repeating. Its rarely if ever been open before, but it was this year, and it creates different conditions for advection. The conditions so far in Baffin Bay so far increase the likelihood of a relatively early opening of the NWP (from the Atlantic end at least). If that goes early, the channels above McClure Strait are also more likely to open early, greatly increasing the opportunity for old ice to be exported. Firstly because the typical conditions will push old ice into the archipelago to be melted, rather than simply push it up against the islands to raft over and thicken up. Secondly because these channels are available for export in either of the dominant AO conditions (as far as I can tell), while the Fram Strait, for example, in only available in one state. This is a qualitative change that is not reflected in past trends, because its never been available before.

    Once again, I don't regard that graph I did as authoritative in any way, and you have illuminated some key weaknesses. My intent was to alleviate my own grumpiness at what I saw as an oversimplification, and to show how much difference fitting a quadratic curve to the data made compared to a linear extrapolation.

    But my graph is just playing the numbers, not the physics. Nature will do what she does, not what I tell her to.

    :-)

    Thanks for the PIOMAS numbers. I saw a source that indicated the negative anomaly at the minimum was 9,400 km3 and I was just guessing at the other numbers from the graph.

    Is there a link that will take me to the monthly anomaly figures like the ones you have above?

    Volume, if PIOMAS is correct, has dropped faster than expected and that supports the lines you have drawn. Based on some numbers you posted on Neven's open thread it looks like the ratio of area to volume has increase to 980 to 1 at the 2010 minimum. It looks like the October and November ratio of area to volume is over 1,000 to 1. I know there is some concentration index of area to extent that either you or someone else is maintaining, but I have not seen an area to volume index.

    Do you know if anyone has put together a graph that tracks the area to volume ratio? This might be a way to show how the ice has thinned. To a certain degree, I think this supports my view that area will persist even if volume declines as area does not decrease as fast as volume.

    Based on the NSIDC description of the melting out of the over 5 year ice in 2010, I have some doubts that 2011 can produce the same year to year decline as 2010. Since the over 2 year ice was at a minimum in September 2010 and it appears there is a significant amount of 1 to 2 year ice that could thicken and become over 2 year ice by September 2011, I think that it is more likely than not that the September 2011 minimum volume will be higher than 2010 and that 2012 minimum will be lower than 2011, but this is a rather soft guess as I do not have a number or other analysis to back this up. Your suggestion that 2011 will show a decline may be correct.

    I thought there were situations in which volume and extent do not move in the same direction.

    At this point, you have exhausted my understanding. I will have to do some work before I can contribute in a meaningful fashion to our discussions.

    Will,

    I quite agree about area to volume ratios. I don't know of anyone who has done it, but I'll give it a whirl. I'll let you know if I get any useful results. I was maintaining an area to extent index, but only for the melt season, so that's in recess at the moment. Artful Dodger at Neven's place maintains an index he calls CAPIE (Cryosphere Today Area Per IJIS Extent). It's more accurate than mine, and more complete, but I suspect it doesn't compare like to like. Doing Area to Volume required Volume numbers, which I've only finished deriving recently. The PIOMAS data can be found here: http://snipt.org/wkuj

    Note that these numbers are published by PIOMAS - I have derived them from the two PIOMAS graphs, which show the running anomaly (resolvable at ~10 day intervals on my copy of the graph) and the average volume for the middle of the month. I've smashed these together, and although they produced the information I wanted, they are neither perfectly accurate, nor even exactly comparable. So you know, rough numbers, not hard data.

    The top row is the long term average for each month, and the succeeding rows are the monthly data for each year. To get back to the anomaly for a given month, just subtract the average for that month from the volume for the month you are looking at (which just reverses the process I performed to get the absolute volumes in that dataset).

    I produced my data by pixel counting on a bitmap image of the PIOMAS graph. At Neven's, a guy named wipneus did the same, but used a vector graphics editor to derive numbers for him. His method was probably a bit more accurate, but our numbers agree closely (mostly). See: http://neven1.typepad.com/blog/2010/12/open-thread-3.html

    You are correct about volume and area moving in opposite directions at times. If ice volume is increasing during the freeze, but weather causes the pack to compress (floes rafting over each other) reducing area. Alternatively, we have seen periods over the last few years where as the pack has broken up and thinned, area has increased while volume decreased. But generally, and over the long term, they track each other to a fair degree. Area does fall more slowly than volume, but as the pack thins there is less chance for that to happen. Eventually the area melts through and the area metric crashes. That is what I think we will see over the next five to ten years. Area will decline slowly, volume will decline more quickly and in perhaps five years we will suddenly lose a very large area which thickness hits zero.

    I'll post again here when I have completed the volume to area graphs.

    cheers

    Thanks for the data.

    Wow, the decline looks worse in the data than it does in the anomaly graph.

    Has anyone posted about the lack of month to month upticks when comparing to the prior year in the volume data other than late 2008 compared to 2007?

    I do not know what to make out it? But I have today studied some modis images from Ross Ice Shelf and Sea (Antarctica), there are some very big cracks in the surface the ice shelf, it looks like situation before a massive calving. Can I get some answers from one guys on this?

    FrankD:

    Thanks for the data you posted on Neven's thread 4.

    Work intrudes, so it may be awhile before I can get back to you.

    FrankD:

    Thanks for the graphs and the other information you have posted here and on Neven's thread.

    The decline in volume is rather sharp, but you have not convinced me that the trend line analysis using volume data for the Arctic as a whole is a valid exercise.

    I note in the NSIDC September age of ice chart that 50% or more of first year ice was able to survive at the September minimum in 2010. If that much thin first year ice can survive a melt season as extreme as 2010 after the severe melt in 2007 and subsequent volume declines, I see no reason to believe that none of the first year ice created after September of 2010 will survive until September 2011 or that it will all disappear as fast as the volume graph indicates. Since there is very little of the thick ice older than 2 years remaining at the 4,000 km3 level for September of 2010, I do not know what source will feed future declines in September volume predicted in the graphs. I suspect we are nearing a leveling out of the September minimum volume level. and for now I am predicting that 2011 minimum volume will be higher than 2010, but the lack of month upticks from the same month in the prior years does concern me and clearly supports your position. While it is likely that the first year ice which survived in September of 2010 was thinner than it was in prior years, I can not tell if it is thinning at a fast enough rate to support a predictions of its wholesale disappearence by 2016.

    Just because thick ice has disappeared does not mean ice extent will decline. We have had several years of volume decline since 2007 without producing a new low in extent. The volume drop from 2007 was 1,963 km3 and ice extent increased by 620,000 km2 from 2007. The PIOMAS chart you have in the link above shows September 2006 ice volume as 9,672 km3 compared to 2010 volume of 4,079. That loss in ice volume of 5,593km3 (Wow!)has resulted in a loss of extent of "only"1.0 million km2 (Per NSIDC September average of 5.9 v. 4.9 million km2 the minimum was 5.7 v. 4.6 million km2). The 2007 volume loss of 3,630 km3 from 2006 resulted in an extent decline of 1.62 million.

    I can calculate that average thickness has decreased from 1.63 meters in 2006 to .83 meters in 2010, a 50% drop, but I am reluctant to extend this trend and say that the ice thickness will reach zero in 4 more years. I do not believe first year ice that survived at the Septembet 2010 minimum was only half as thick as the first-year ice in 2006. If you can graph the trend in first year ice thickness in the Arctic Basin and show that it is declining fast enough to cause it to disappear by 2016 then you will have won me over on an "ice free" September, but for reasons indicated below I do not trust the October through December predictions made by the volume graphs.

    The trend lines you are drawing for September using data for the Arctic as a whole are reasonable if the ice in the central Arctic Basin is melting at the same rate as ice outside the Arctic Basin. Based on area and extent data, the ice in the Arctic Basin does not melt at the same rate as ice outside the basin. Part of the reason is that ice is transported into the Arctic Basin from surrounding regions. The trend line method using the data base of the Arctic as a whole would be valid if the Arctic was a static flat surface and all areas of the Arctic received exactly the same amount of heat input from the sun and ocean currents.

    I am challenging the reliability of the volume trend line analysis because the central Arctic Basin receives significantly less heat from solar input as it is above 80 degrees north, than even the surrounding regions that are in contact with the central Arctic Basin, which fall between 70 and 80 degrees north (I think this excludes the Chukchi Sea). The central Arctic Basin should also receive less heat input from ocean current sources than these same regions. Including regions in the trend line analysis which fall below the arctic circle, such as Hudson Bay, the Bering Sea, and the Sea of Okhotsk or have no significant contact with these regions, such as Baffin Bay will almost certainly produce a false trend line. Although the data is not available, it would be interesting to sea the volume trend line analysis broken into three segments. One for the central Arctic Basin, another for regions between 70 and 80 degrees north which are in contact with the central Arctic Basin (this excludes Baffin Bay), and a third for the remaining regions. The significant variations in these three trends lines would illustrate why a trend line using data for the Arctic as a whole is not valid.

    Even if you plot trend lines for the central Arctic Basin, I believe the lines for October, November and December will overstate the rate of loss as the trend lines fail to take into account the physical process of the expansion of the ice cover after the minimum. Try plotting the monthly expansion of ice volume after the September minimum on a month by month basis for October, November and December and see what happens.

    A quick look at current conditions in a few regions shows why the Arctic as a whole does not produce a valid trend line. For example, Hudson Bay in 2010 had a negative anomaly in December of 400,000 km2. (Cryosphere Today chart) Hudson Bay has rapidly been increasing in extent since this low point and currently has a negative anomaly of 150,000 km2 and is likely to reach its normal maximum extent in February.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.13....

    The Sea of Okhotsk currently has a negative anamoly in excess of .250,000 km2.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.14....

    The negative anomaly for Baffin Bay has steadily increased since the minimum, when it was zero, and it currently has a negative anomaly of 400,000 km2.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.4.html

    Since ice from these regions is not transported into the Arctic Basin regions and their points of contact with the Arctic Basin regions north of 70 degrees are nearly non-existent, including data from these regions in predicting the future of the ice in the Arctic Basin regions north of 70 degrees skews the trend lines. If there is no correlation between the anomalies reported in these three regions and the various regions in the Arctic Basin, then the data from these regions needs to be excluded. I do not believe it is cherry picking to exclude these regions from the analysis since they show no correlation and have significant physical differences.

    The December 2010 plots done by the NSIDC show that regions north of 70 degrees are not reporting any problems in getting ice extent to its maximum extent (with the exception of Baffin Bay, which I have excluded from this group).

    http://nsidc.org/data/masie/masie_plots.html

    The trend line analysis based on volume that you are citing suggests that the regions north of 70 degrees will be ice free in October, November and December within the next decade or so. Previously I challenged this result without providing any analysis of why I disagreed with it. Focusing on how the ice has melted up to the September minimum is a false predictor of an ice free October. Instead of looking at melting, the analysis needs to look at how and why the ice expands after the September minimum. Even if the September minimum goes to zero, ice extent will quickly increase after that date, just as it does now.

    The best evidence to undermine an ice free October prediction from the volume trend lines is current observations. If you plot the date the Arctic ice starts its expansion of extent you will note there is very little change in the date. In order to achieve an "ice free" state in October, you will have to move the date of the expansion of ice extent to October 31st. The mismatch in the trend for the date when the ice begins its expansion and the prediction of an ice free Arctic in October is a strong indicator that reliance on the volume trend line is not valid for months after September. While some of the regions in contact with the central Arctic Basin reached very low volume and extent levels (zero in some cases) at the minimum in September, observations still show rapid expansion of ice extent in late September following the minimum. While I do not have figures by region, 2010 extent observations for the Arctic as a whole (based on JAXA) reported an increase in extent of 750,000 km2 in September after the minimum was reached and an additional 2.5 million km2 in October. In order to have an ice free October, you would need to eliminate both the 4.6 million km2 of sea ice at the September minimum and stop the 3.25 million expansion of sea ice after the minimum. It ain't going to happen in the time frame of the volume trend line prediction for October and certainly not December as the ice extent in 2010 expanded by 2.0 million km2 in November and 1.8 million km2 in December in spite of several regions reaching their maximum possible extent in early December.

    The volume in this expansion is no doubt extremely low, as it consists of ice that is at most, three and a half months old, but low volume at the September minimum does not prevent the expansion from occurring. It only takes a volume of 1,000 km3 to produce an ice extent of 4,000,000 km2 with a depth of 1/4 of a meter (at this thickness, is it correct that only about an inch of the ice would be above sea level?). Yes the ice will be very thin, easily transported, and broken into small pieces, but it would still exist. The ease with which a layer of thin ice can be formed undermines attempts to use the volume trend line as a predictor of post September ice extent.

    It is the loss of heat input from the Sun to areas north of 70 degrees north that drives the ice expansion. Changes in sea temperature below the surface do not trigger ice expansion. Winter time air temperatures still show sufficient declines by early September (day 250) for the process of ice expansion to begin.

    http://ocean.dmi.dk/arctic/meant80n.uk.php

    The process of ice expansion occurs even though sea temperatures in the Arctic are above the freezing point. Additional future heat input from ocean currents into the Arctic will not stop the ice expansion process after the September minimum as the decline in air temperatures is more than sufficient to cool the sea surface and cause it to freeze. I can't find the citation, but I remember reading that currently a thin layer of cold water that insulates the ice from the heat of the sea below. The volume trend lines for October through December are suspect as they do not reflect the physics behind the ice freezing process and expansion. The September volume level will not go below zero.

    Thank you for charting the ice thickness factor. It showed that ice extent can be maintained even though volume declines by reducing average ice thickness. Even these trend lines overstate the timing for a complete collapse of the ice. Sorry to ask, but could you try charting ice volume expansion by month for post minimum September, October, November, and December and see what kind of trend line develops? I suspect the trend lines for the monthly expansion of extent will not match the rate of decline in volume as it does not take much volume to cause the ice extent to increase. This last factor may be the primary reason why accelerating trend lines for volume decline are not a valid predictor of future ice extent for October, November and particularly December. There will be a minimum threshold level of ice extent and thickness that will be very difficult to make disappear after the September minimum date as long as the air temperature drops below freezing in winter. If you want to show no ice growth in October, you have to predict that air temperatures will stay above freezing in October. No one has issued such a prediction.

    The anomaly reported on the Cryosphere Today charts for the regions above 70 degrees, with the exception of Baffin Bay, do not currently report any significant negative anomalies. The Barents Sea, which has an edge in direct contact with Atlantic currents, has a negative anomaly of 180,000 km2 or so, but is reporting higher extent levels than it has for the previous four years per the NSIDC graphs in the link above.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.6.html

    While I applaud the effort at graphing trend lines for volume declines, observations of actual conditions and processes leads me to the conclusion that this analysis is deeply flawed for periods after September. I still contend that using a volume decline graph to predict an "ice free" Arctic is not a valid model as the "model" does not work for months after September. By next September, we can test whether this model is accurate in its September prediction once data from Cryosat becomes available. We can then drive ourselves crazy with a new set of numbers. Until then, we can continue our discussion.

    Will,

    Firstly, just a reminder that I am not, and never have been trying to provide an accurate model of when we will reach zero ice (reread my original post from Nov 6th). Your observations about the limitations of my approach are largely valid. Indeed, early in the thread I remarked that playing a numbers game was a problem in the absence of physics, and that is all either of us are really doing. And my original point was to observe that extent-based trend analysis gives unrealistically long time frames. If I have managed to convince you of that (even if my numbers are unrealistically negative) then I have little to add - I'm neither a statistican nor a physicist, so I can't improve on the analysis I have posted here. But I have never claimed to have an accurate model - only attempted to show that considering three dimensions gives very different results that the debatable approach of considering only two dimensions.

    But as I say, I am happy to concede some of the points you make. I don't think they will preserve ice into the time periods you originally referred to (2030's and beyond), but I do accept they introduce "problems" to volume-based predictions.

    Your point about first year ice is encapsulated in this diagram I originally posted at Neven's. http://img403.imageshack.us/img403/9463/arcticicethicknessrecen.png
    While the thickness for each month is declining, off a short dataset there seems to be some resistance to dropping below ~1.1 metres. I attribute the thinning to the loss of thick multiyear ice, and the "resistance" to the fact that once it is gone, first year ice does not thin further (it either survives at ~1 m thickness, or it doesn't). SO I agree that, although individual months (say Apr-Jul) may drop quickly in the next few years, overall then thinning will slow (esp Aug-Jan). It should be remembered however, that this is average thickness (volume / area) and situations where volume and area are declining at the same rate will see thickness remain the same, without that being a good result).

    While I agree in principle with your point about detached basins, I think you overstate it. Hudson Bay and Sea of Okhotsk are connected to the Arctic via the atmosphere. If warmer, late freezing water in HB is helping to create the anomalous high south of Greenland seen in recent years, that slowed ice production across the whole Atlantic side in Oct-Dec (to mid-Jan this year), then it is intimately involved with the overall health of the ice. In the case of Baffin Bay, even now there is considerable ice being advected though Nares Strait, (blink compare ASAR images here: http://ocean.dmi.dk/arctic/kennedy.uk.php), so I would say that Baffin Bay remains an integral part of the system. This advection has no impact on extent in the central basin, but obviously does on volume. You advance a strong line of reasoning based on the supposition "if there is no correlation" - perhaps you could look an see if there is, is not or is a negative correlation. It's a very important point to rest on "if".
    Edit -> Ahh, I see from your next post that your have also picked up on advection into Baffin. This happens every year, I have no framework for comparing one year to another, other than to observe that in 2007 a bridge formed across it. This did bad things for extent numbers, as Baffin Bay was completely clear, but overall was better for the health of the pack as less ice was advected out of the central basin.

    Growth trends from September through January behave exactly as you suggest - Sep-Oct trend is flat, with each successive month trending up by a bit more. This is obvious physically from the winter "recovery" of extent, and numerically from the volume trends I posted way back (but I checked anyway). It is worth noting that on that original graph, these were also the least realistic curves. This is given away by the fact that the volume trend for these months is climbing at the start - physically wrong, and the main problem with the curve fitting I used. This was disclaimed from the start, and both here and at Nevens, I have recommended people with better software and skills than I should apply better curves ogives rather than quadratic).

    However I think it is rather heroic to assume that October growth from a September zero would be the same as October growth from a September minimum of a few million sq kilometres. This, again, goes to non-linear respnses. With a September minimum of, well pretty much any amount, the sea surface temperature at the pack margin will be close to zero and there is no impediment to ice forming. Once September falls to zero, sea surface temps can rise above zero, which creates several issues. It has to be cooled to zero before ice can form (this may only take 5% of the energy required to turn sea water into ice, but it is not nothing). More mixing can occur, vertically and horizontally - this would result in later and later ice formation with rapid recovery through the winter (which is what we are seeing now). There are probably other factors. But while the sun is above the horizon, open water is absorbing hundreds more watts per sq metre than ice would have. This ice albedo feeback must heat water further, and delay October growth, however much we might want to just assume that whatever September gives us, October will be the same as previous years. I predict October growth will remain flat until September zero, then will nose down. Sept-Nov change will decline from slight growth until October zero, and then it will nose down. And so on...

    But at this point, you seem to be gaming the extent numbers - a crumbly thin pack spread out like crushed ice might on some level invalidate volume based projections, but it hardly reflects well on the state of the ice. It's also a point I've conceded repeatedly, as I've said the curve is more probably ogival.

    And if volume gets below 1000 km^3? Can you have 4 million sq kms of 10 cm thick ice, or 1 cm thick ice? As mentioned here and at Nevens, once you get below a certain thickness, ice becomes translucent, and the albedo feedback effect kicks in, even if there is still some nominal ice coverage. If you want to claim a "win" on the basis of 2017 bottoming out at 3 million sq kms of brash which for all the good it does the physics or the ecosystem might as well not be there, then I'll happily concede the "bet", whils also knowing that those claims that we'll still have good ice in 2030 were even wronger.

    And I still maintain the first statement I made on this thread, that extent projections do not seem to take any non-linear feedbacks into account: "I think there is a danger in the numbers game - the assumption that this year is only a bit different from the last few years. I don't think that that is necessarily the case."

    FrankD:

    Radar images for north of Greenland are at:

    http://ocean.dmi.dk/arctic/modis.uk.php

    From Nord in the northeast , Morriss Jessup due north to Lincoln in the northwest the ice appears to be in a very broken up condition, so it must be thin. I do not know how this compares to images of this area prior to 2005.

    If you view the images of the Nares Strait to Baffin Bay from Kennedy, which includes the Petermann glacier, through
    Kane and Qaanaaq you can see the lack of ice in Baffin Bay and it does not appear that any ice bridges are likely to form in the Nares Strait.

    Well, exactly,

    Again - in 2007, ice stopped advecting through Nares because of a big thick chunk blocking the entrance. Bad for extent, good for volume.

    But thinner ice lacks the structural strength to form such a block, and I think in future we will see continuous advection through the Strait. Good for extent (as Baffin stays above zero). Bad for volume, as ice is leaving the central basin.

    FrankD:

    As always thanks for responding and I appreciate your graphing skills and analysis, although some items you are using to support your position, such as increased heat from Atlantic waters, while correct in a general sense, do not reflect the complexity of actual conditions in the Arctic which prevent this heat input from coming into contact with sea ice.

    You have helped refine my focus to the volume of first year ice in the central Arctic Basin rather than just looking at extent, but there is no data base to chart the trend in ice thickness of first year ice for this region, so I will stick with extent as a second best solution until such data becomes available.

    You keep saying the ice is thinner and that is easy to see and undeniable, but which ice is thinner and how much thinner is it? Show me data for the trend in first year ice in the Arctic Basin! A claim that we will see future declines in central Arctic Basin ice extent that outstrip past declines is not supportable. The sharp 2007 decline was 500,000 km2 when it dipped to 2.1 million km2. You need five consecutive years of decline in extent in the central Arctic basin of 500,000 km2 to create an ice free Arctic. Only 2007 reported a decline of this magnitude and in the three years since, the ice extent has expanded in spite of all the factors you have mentioned. How do you explain this and how does it figure into your prediction of an "ice free" Arctic?

    I am unconvinced, that drawing trend lines based on Arctic wide data for either volume or extent produces a reliable model or basis for predicting when the Arctic will be ice free and you have not provided a supportable position that refutes this. The post September volume trend lines are not supportable and you appear to agree.

    Your graphs are very illuminating, but I will still challenge the data base that is being used to make the graphs. I like your ice thickness chart, but would still like to see how it works for just first year ice in the 4.25 million km2 of the central Arctic Basin. It is this ice that will determine when the Arctic is "ice free", not trends for multi year ice or ice in other regions of the Arctic. Show me the decline in thickness of first year ice in the Arctic Basin ice and you will have convinced me that the central Arctic Basin will be ice free on the schedule suggested by Maslowski.

    Until then, I will stick with the estimates of an ice free Arctic made by the scientists at PIOMAS, who are producing the volume data you are using to support your position. What is it about the prediction being made by this group of experts that makes you believe their prediction is wrong and your position or Dr. Maslowski's of an ice free Arctic by 2016 is better? Certainly they are aware of the ocean warming and albedo loss factors you keep citing to support your view. If you can find an article discussing their view of Dr. Maslowski's prediction, I would be most interested in their opinion.

    I think you have a basic mis-understanding of how the additional heat input from Atlantic waters is being distributed in the Arctic and this may distort your view of when the Arctic will be ice free.

    More heat is building up in the Arctic Basin waters, but the water volume and depth of the central Arctic sea leaves a lot of places for that heat to be stored. The water temperature in the central Arctic Basin is not the same for all depths. The heat from Atlantic waters that you are relying on is 100 to 200 or more meters below the surface and does not come into contact with surface ice! This undermines a major support for your position. Surface waters are several degrees cooler than this water. To support your position, you need to show that surface waters have warmed dramatically in the central Arctic basin to the point that ice creation can not occur. There is no data set that supports this view, just as there is no data set supporting that air temperatures for the region above 80 degrees north have reached a point that does not support ice creation after the September minimum. Changes in the surface air temperatures and sea surface temperatures for the Arctic basin do not appear to be providing sufficient heat input to the Arctic to create the ice free conditions you are describing.

    http://upload.wikimedia.org/wikipedia/commons/9/90/Arctic_sea_temperatur...

    The quote below is from a more complete description of the Arctic sea at:

    http://www.waterencyclopedia.com/Oc-Po/Oceans-Polar.html

    "Arctic Water Masses.
    There are three distinct marine water masses located within the Arctic Ocean: the Arctic Surface Water (0–200 meters); the Atlantic Water (200–900 meters or 650–2,950 feet); and the Arctic Deep Water (900 meters–seafloor). The Arctic Surface Water is divided into three layers: the surface, subsurface, and lower surface layers. Each of these water layers has distinct salinity and temperature characteristics.

    The Atlantic Water (AW) is located below the Arctic Surface Water (ASW) and above the Arctic Deep Water (ADW). The average temperature (3°C [37.4°F]) of the AW is warmer than both that of the ASW (−1.9°C to −1.0°C [28.6°F to 30.2°F]) and the ADW (−0.8°C to 2.0°C [30.6°F to 35.6°F]). The AW has a higher salinity range (34.8–35.1) than that of the ASW (28.0–34.0). The ADW, with a salinity range of 34.9 to 34.99, represents approximately 60 percent of the Arctic Ocean total water volume and is comprised of the Norwegian Sea, Greenland Sea, Eurasian basin, and the Canadian basin deep waters"

    I also do not think the albedo argument you are making works as well for the region above 80 degrees north as it does for regions below 80 degrees north, particularly for the period after August. For the period before August, ice loss is not as great for the area north of 80 degrees, so albedo loss is not having as great an impact as it does for other regions.

    http://creativeenergyalternatives.com/solar/how_does_a_polar_sunchart_va...

    How much solar heat input is there to latitudes above 80 degrees north after August?

    I have no doubt that loss of albedo is impacting areas below 70 degrees latitude, but these areas do not matter as they are already at a zero volume and extent level at the September minimum.

    Observations indicate that first year ice at the minimum is not showing the degree of decline necessary to produce an ice free Arctic. I know you prefer volume, but there is no volume data for first year ice. The decline since 2000 at the minimum appears to have occurred only with respect to multi-year ice. This explains why volume declines have occurred at a higher rate than extent declines. The additional heat input you are relying on for your analysis is not showing up in its impact on the formation of first year sea ice in the regions above 70 degrees north other than the Barents Sea, which is in direct contact with Atlantic waters. If your presentation is correct, we should see a significantly later freeze up in the Arctic Basin regions.

    Observations for the central Arctic basin do not support your theory of addtional heat from sea currents and albedo causing the demise of the sea ice. The 2010 anomaly at the minimum was .75 million km2, which left 2.5 million km2. That rate of decline for this region does not support your position.

    http://www.arcus.org/search/seaiceoutlook/2010/pre-release

    What is wrong with using the graph in figure F at the above link to predict future ice decline?

    What age and thickness of ice is going to generate the kind of future volume declines you are predicting? How does this compare to the age and thickness of ice that has contributed to the 2010 decline to 4079 km3? If you need different types of ice to generate the future decline than the types which produced the observed decline, then the prediction is suspect.

    Please note that the expansion of sea ice in the central Arctic basin begins just before the minimum is reached. The Arctic Basin increased to 3.0 million km2 by October 1st. It had a zero anomaly by Novemeber 1st when it reached an extent of almost 4 million km2 and reached its maximum extent by December 1st when it hit 4.25 million km2.

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

    My focus has been on refining the data base for drawing the trend line in a fashion that will produce a more accurate prediction of future ice conditions and I have since added looking at multiple data sets for confirmation of any such trend line. I am using Maslowski's focus on the central Arctic Basin, and I have included the entire time period of satellite data. The region I am focussed on represents more than 80% of the ice at the minimum, You have not provided any support showing a correlation between the Arctic wide numbers you are using and the central Arctic Basin. Without a correlation, the Arctic wide numbers are interesting, but prove nothing.

    As for advection through the Nares Strait, while it is visually dramatic, the actual amount of the increase in ice advection in recent years is not that significant. I thought the increase was less than 100,000 km2 and the maximum advection is limited by the narrowness of the strait. The real action on ice advection is through the Fram Strait. How much larger is ice transport through the Fram Strait than the Nares Strait, 10 times or more? Like Baffin Bay, ice volume and extent in the Greenland Sea is heavily influenced by transport from the Arctic, thus this region is a poor indicator of melting conditions. I would not expect to see significant future declines in ice extent in the Greenland Sea below current levels (which are not that significant at the minimum) due to this transport even though it is below 70 degrees north.

    As for lack of ice in Hudons Bay and Baffin Bay increasing air temperatures in the Arctic, I need you to show that prevailing winds go from these areas into the central Arctic. Based on ice flow south through the Nares Strait and Baffin Bay, the winds appear to go away from the Arctic. I thought the major inflow of air into the Arctic came from Siberia and Asia and not Canada. That is why the ice was pushed away from Asia in 2007 and why the thickest ice piles up in the Canadian Archipelago and Greenland.

    I have seen the graphs for temperature increase for 60 degrees and above latitude, but these include too great a latitude range to be useful in predicting the future of the Arctic basin. Please show me surface temperature increase data for the region above 80 degrees north that suggests it will be ice free by 2016.

    In the Antarctic winds move in a circle pattern and appear to keep global warming in the atmosphere at bay, although the extent numbers there have recently shown declines. The Arctic, due to land masses has a more complex wind pattern, but I don't remember seeing that winds go north from Hudson Bay and Baffin Bay to the Arctic on any consistent basis.

    While the support you offer for your position sounds plausible, on further investigation, these factors do not operate in a fashion that will cause an ice free Arctic in the time period Dr. Maslowski has predicted. I will continue to follow the model results from the PIOMAS scientists of when the Arctic will become "ice free" rather than rely on trend lne analysis based on volume or extent for the Arctic as a whole.

    Yes the volume decline in the PIOMAS model is dramatic, but I think your focus on this modeled data set to the exclusion of other observational data sets is providing you with a misleading impression of when the Arctic will be ice free. I would find your position more supportable if the result you are predicting was supported by more than this one data base. Yes the Arctic ice appears thin in all the satellite and radar pictures, including the ones in the link I provided above, but the ice is still there. It will not be gone in 2011, 2012, or 2016.

    FrankD:

    You may be able to put together information about the survival of first year ice using the charts below, but based on some initial numbers I am not sure there is a significant decline trend in first year ice. The numbers appear to bounce around more than the ice extent figures for all of the ice.

    The September minimum of 4.6 million km2 was about 50% first year ice per the figure below:

    http://nsidc.org/images/arcticseaicenews/20101004_Figure6.jpg

    Looking for trends in first year ice over the last 5 minimums is confusing. As a result of the record setting minimum in 2007, the minimum in 2008 consisted of 60% first year ice so at 4.52 million km2 for all ice, first year ice was 2.7 million km2 at the 2008 minimum. If 2008 is compared to 2010 the result is that first year ice decreased by 400,000 km2 and older ice at the minimum has increased by 500,000 km2. This result appears to be driven be the extreme conditions of 2007 which created a greater area for the expansion of first year ice into 2008 rather than a trend of decline in first year ice as the analysis from other years below shows.

    The 2006 minimum was about 35% first year ice, so at 5.7 million km2 for all ice, first year ice was about 2 million km2. Based on this, first year ice has increased by .3 million km2 since 2006 and older ice has decreased by 1.4 million km2. While this makes sense, it does not show that first year ice is disappearing at a rate that would produce an ice free Arctic.

    First year ice at the September 2007 minimum of 4.2 million km2 was only 30% of total ice, thus it was only 1.3 million km2; and older ice declined to 2.9 million km2. Thus, the 2007 dip resulted in an almost equal loss of first year ice and older ice. However, since 2007 first year ice has come back and increased by 1.0 million km2 at the minimum and older ice has continued to decline as it decreased by .6 million km2.

    I thought a trend might become more apparent if I looked at the data for 2000. Ice extent at the minimum appears to have been about 6.5 million km2. Only 30% of this was first year ice, so about 2.0 million km2 of first year ice existed at 2000., In looking at 2002, the percentage of first year ice was 40% and total ice extent was about 6.1 million so first year ice was 2.4 million km2. Since 2010 was above 2000 and about equal to 2002, there does not appear to be a trend showing a decline in first year ice in this decade. All of the decline appears to be reflected in the extent of older ice. There appears to be a great deal of variability in the mix of the ice, but there is no clear trend that first year ice at the minimum will cease to exist by 2016.

    I think this shows that while the ice is thinning, it is still capable of generating significant extent numbers at the minimum even if the minimum consists of an increasing percentage of first year ice.

    While it is intuitive that first year ice should be declining as volume declines and the "protection" afforded to first year ice from multi-year ice decreases, observations show that first year ice can maintain itself even as older ice declines. This is just another factor which makes the ice free by 2016 prediction suspect based on volume trend lines.

    Please let me know if you come up with a different analysis when looking at the persistence of first year ice.

    Will,

    I have no particular insights into your first year ice analysis, beyond what I have said in my long screed above (ie that once September does reach zero is will start forming later, from a lower baseline. I suspect it will grow faster, but I don't know whether that will offset the first two factors.

    It is not at all intuitive (to me) that first year ice should decline in lockstep with overall volume. I would argue previously that this has been disproportionately taken up with melting thick multi year ice. Of course thin ice is also melted, but replaced by older ice "spreading". There is now so little of that left, that volume loss will not, in a few years, come from that source, unless there is some dramatic turnaround in conditions that allows the old core to regenerate.

    So, we have one of two possibilities (and I think there are only two).
    1. Volume loss stops. I guess thats possible, but I would ask: where does the excess heat go?
    2. Extent loss kicks in with a vengeance. With no more thick ice to melt or spread, if the heat isn't dissipated by some other mechanism (radiation into space?) then it must carve into extent. Any other option would seem to violate thermodynamic laws.

    Once again, the key factor that you are not considering in your trends is simply that we are getting close to entering non-linear land. Charting extent trends does not suggest that, but volume trends do. If some feedback does not kick in to stop volume loss - what? - then extent must start crashing in a few more years - departing from any trend you care to project.

    I agree (in anticipation) that some of those feedbacks act to preserve extent (and volume). For example, the rate of first year ice formation, which IS dominated by two dimensions (surface area dropping below ~-2), is an aspect that does so. But there are a number of factors I have observed that would more likely accelerate ice loss:
    * thinner pack lacking structural strength is advected more easily.
    * thinner ice is translucent, lowering albedo and increasing warming for a given amount of sunshine.
    * thinner ice is more easily broken up by wind and wave, increasing surface area: volume ratio, and accelerating melt.
    * more open water -> more cloud (this is complex, but I suspect on balance the increased heat trapped outweighs the decreased insolation).
    * final demise of multiyear ice meaning "spreading" will no longer "hide the decline" but lead directly to extent loss that has hitherto been masked.
    * earlier melt, later freeze - by 2.5 days per year on average, and by much more at the margins.
    * more rapid melt and freeze - based on IJIS extent, time less than 13 M sq km, is increasing by less than 1 day per year, but time below 6 M sq kms is increasing by nearly 4 days per year. More open water for more days = more heat retained
    * weather feedbacks seen over the last few years that pump more heat (in Dec-Jan) into the Arctic

    These are mostly effects that kick in on when certain conditions are met, and introduce a non-linearity. Your analysis of extent trends has from the start assumed that tomorrow will be like yesterday. At some point in the near future, that will stop being the case. Prepare to be surprised.

    cheers

    FrankD:

    I do not believe we need to worry about violating thermodynamic laws as long as the heat from Atlantic waters remains more than 100 to 200 meters below the ice surface. That means it has zero impact on surface ice.

    Show me an increasing air temperature trend for the region above 80 degrees north sufficient to cause all the ice to melt in this region and I will agree with your prediction. Where is the loss of albedo impact on these temperatures? Ain't no sunshine there in the winter time.

    I am considering that we have entered a non-linear trend period. It is just that my non-linear trend is a slower sloping curve rather than your accelerating curve rate. Real world experince with systems approaching a zero level and the rule of diminishing returns suggests we will see a slowing, not an accelerating trend line.

    Thin is not zero.

    The 2.5 day a year number sounds high over the last 30 years this would have increased the melt season by 75 days, please provide a cite, I do not believe it. I thought the refreeze date had only shifted about 10 days over 30 years. That means the Arctic will not be ice free at the end of October in this century. When are you predicting an ice free Arctic at the end of October?

    http://earthobservatory.nasa.gov/Features/ArcticIce/arctic_ice3.php

    Why don't we look at the number of days ice extent is below 4.0 million km2 (how about central Artic days below 2 million km2) as a predictor of an ice free Arctic rather than 6.0 million km2? That number is ZERO!

    If you are predicting ice free conditions, why use 6.0 million km2 as an indicator, isn't that a bit high?. That is more than 3 Alaska's. When was the last time the minimum was above that level? You can do better than this. Why not look at the number of days volume is below 7,000 km3. That will at least create some data over the last 4 years, but 7000 km3 is a long way from zero. As for the maximum, the central arctic hit its maximum by December 1st, and all other regions above 70 degrees north have hit their maximum except the Barents Sea. What do I care if Hudson Bay, Baffin Bay, the Sea of Okhotsk, the Greenland Sea or the Bering Sea fail to reach their maximum? How much do they contribute to the minimum? Almost zero.

    Frank, forget about the regions below 70 degrees north and the regions outside the basin as they show no correlation with the amount of ice at the minimum, why include them in future ice predictions if they show no correlation to current ice conditions at the minimum?

    The Arctic basin constitues more than 80 percent of the ice extent at the minimum. The Arctic will not be ice free until this region of 4.25 million km2 (2.5 million km2 at the 2010 minimum) is empty of ice. Provide data for this region alone that shows how it will go to zero in the time frame you are describing if you want to convince me. I do not care that the volume analysis shows that the rest of the Arctic goes to zero at the minimum in the future, it is nearly there now.

    Frank D:

    Is Maslowski's ice free prediction a volume level below 4,000 km3 or 400 km3?

    His extent lines were October November figures not September, No way we are going to be ice free in October by 2019.

    Another source from January 27, 2010 on the days of melt indicates the figure you supplied of 2.5 days per year is not correct::

    "New NASA-led research shows that the melt season for Arctic sea ice has lengthened by an average of 20 days over the span of 28 years, or 6.4 days per decade."

    http://www.nasa.gov/topics/earth/features/longer-melt-season.html

    Check out this image:

    http://www.nasa.gov/images/content/419940main_arctic_melt_trends_lg.jpg

    Based on this image above, the more central areas of the Arctic do not show that great a change in melt season. This supports my view that the central Arctic basin is not showing as fast a rate of melt as the Arctic as a whole. (While the image suffers from the donut hole, it is clear that melt days closest to the hole show the least change.)

    Where did you get the 2.5 days per year figure?

    Climate progress had a post in September of 2010 that shows the increase in area at the maximum made up of first year ice. The additional heat inputs you cite are devastating multi-year ice, but first year ice has expanded its extent. This is not multi-year ice that is being spread, it is new growth ice.

    http://climateprogress.org/wp-content/uploads/2010/09/wintericeage1.gif

    The article has the following graph showing that the volume of first year ice is increasing, but I have not found anything tracking trends in ice thickness for first year ice at the September minimum.

    http://climateprogress.org/wp-content/uploads/2009/07/Arctic-Ice-Volume-...

    The full article is at:

    http://climateprogress.org/2010/09/14/exclusive-scientists-track-sharp-d...

    The more data sources and research papers I read the more support I find for treating the central Arctic Basin as a distinct region that must be looked at seperately from the rest of the Arctic What sources can you cite that show that Arctic wide data has a correlation to data for the central Arctic Basin?

    The graphs you are drawing show the dramatic decline in thick multi-year ice, but do not capture the increasing volume of first year ice. This is a serious flaw if you are using a trend line to predict ice free conditions. The trend lines you are drawing show that the Arctic is becoming free of multi-year ice, but they do not prove that the Arctic is going to become ice free in the time scale shown in the graph. To do that, you need to chart the volume trend for first year ice and show that it will disappear too.

    FrankD:

    I have not ben able to post this to Neven's Thread 5. Perhaps you would be so kind as to post it, including any rebuttal points you would like to make.

    The volume trend line graphs previously provided in Neven's Open Thread may only be showing when the Arctic will be "ice free" of the oldest and thickest multi-year ice, they do not necessarily show when the Arctic will be "ice free" of first year and second year ice.

    In order for the trend line analysis to be valid for predicting an "ice free" Arctic, we need to show that the rate of volume decline of first and second year ice is the same as the rate of volume decline for older multi-year ice.

    Information posted in a September 14, 2010 post at Climate Progress suggests the rates of decline are not similar for these ice classes.

    The extent of first year ice has expanded in the Arctic since 2000 and based on ICESAT measurements the volume of this ice has increased. Please see the information at Climate Progress at the link below:

    http://climateprogress.org/2010/09/14/exclusive-scientists-track-sharp-d...

    Accepting the PIOMAS volume figure of 4,079 km3 for September of 2010, you have to ask how is ice extent being maintained in the face of such steep volume declines. The answer is that first and second year ice has become a larger percentage of the September minimum as illustrated by the chart in the link below posted by the NSIDC showing that such ice is 85% of ice extent.

    http://nsidc.org/images/arcticseaicenews/20101004_Figure6.jpg

    Even a volume level of 2,000 km3 can maintain an ice extent of 2,000,000 km2 of 1 meter thick ice. It may be "rotten" be it will still be there. Obviously, the thinner the ice, the less volume is needed to produce extent. Certainly at a zero volume point, ice extent will collapse to zero as well, but when will we get to zero volume for first and second year ice?

    Heat input from Arctic waters is not likely to cause the first year ice to disappear as these waters reside 100 to 200 meters or more below the surface, although heat from Atlantic waters does appear to be nibbling at the edges of the Barents Sea. Since ice from the Barents Sea does not drift into the Arctic basin, loss of ice from this region may not have a significant impact on the September minimum for the Arctic basin.

    http://www.waterencyclopedia.com/Oc-Po/Oceans-Polar.html

    The increase in temperature of sea surface water does not appear sufficient to slow the winter growth of ice extent in the Arctic Basin regions. The current anomaly is produced by regions outside the Arctic basin and regions below 70 degrees latitude and the Barents Sea, which does not contribute in a significant fashion to the September minimum.

    The increased heat from loss of albedo in not showing up in temperature data for areas north of 80 degrees latitude in a manner sufficient to cause the dramatic change in ice extent predicted by the volume trend lines although there is a steep rise in the average temperature for 60 to 90 degrees that may be affected by loss of albedo.

    http://ocean.dmi.dk/arctic/meant80n.uk.php

    While we could disagree over the accuracy of the DMI temperature data 80 degrees north, the support that the heat input has not been sufficient to clear the area above 80 degrees north is in extent observations for this region for September. (Volume data is not available for the region, so extent data is the best I can do). The slower rate of loss of ice extent in the central Arctic Basin region can be seen in the links below which chart the historic ice area average for September - see Figure F:

    http://www.arcus.org/search/seaiceoutlook/2010/pre-release

    or in current year area anomaly analysis for this region by Cryosphere Today:

    http://arctic.atmos.uiuc.edu/cryosphere/IMAGES/recent365.anom.region.1.html

    Neither chart t reflects a rate of decline sufficient to create an "ice free" Arctic in the time frame predicted by the volume trend lines.

    Clearly the Arctic is rapidly becoming "ice free" of older ice, such as ice that is at least five years old. The NSIDC noted this when it reported:

    "Older ice tends to be thicker than younger, one- or two-year-old ice. Last winter, the wind patterns associated with the negative phase of the Arctic Oscillation transported a great deal of multiyear ice from the coast of the Canadian Arctic into the Beaufort and Chukchi seas. Scientists speculated that much of this ice, some five years or older, would survive the summer melt period. Instead, it mostly melted away. At the end of the summer 2010, under 15% of the ice remaining the Arctic was more than two years old, compared to 50 to 60% during the 1980s. There is virtually none of the oldest (at least five years old) ice remaining in the Arctic (less than 60,000 square kilometers [23,000 square miles] compared to 2 million square kilometers [722,000 square miles] during the 1980s)."

    The loss of the older thicker ice has skewed the usefulness of the volume decline trend lines as a predictor of future ice conditions. The central Arctic Basin made up more than 80% of the remaining ice at the September minimum. It would take an annual rate of collapse of 500,000 km2 a year to clear the 2.5 million km2 of ice that existed at the September minimum in this region by 2015 to clear it out. Only 2007 with its wind conditions produced a single year September drop this large when the central Arctic basin was reduced fom 2.6 million km2 from 2006 to 2.1 million km2 in 2007. In the three years since 2007, this region has stayed above 2.4 million km2 at the September minimum in spite of the volume decline figures for the Arctic as a whole published by PIOMAS.

    The accelerating trend lines in the volume analysis are not likely to occur because there is an insufficient amount of older thicker ice left to be lost to generate this rate of decline It is more likely that the volume trend line will flatten out and decline at a rate just slightly faster than the rate of thinning for the remaining first and second year ice which was 85% of the 2010 minimum. I have not been able to find a data source showing the rate of volume decline for first and second year ice so I can not provide a specific rate prediction other than to say it will be slower than the trend lines for the Arctic as a whole indicated in the volume graphs.

    While the Arctic is headed for an ice free state at the minimum, the volume trend line analysis, while interesting, does not appear to be reliable indicator of when this will occur.

    FrankD:

    The following from 2003 provides an explanation of sea ice freezing. Some of the info about multi-year v. first-year ice may be dated, but the description of the ice expansion process is still valid.

    How Does Arctic Sea Ice Form and Decay?

    Peter Wadhams
    Professor of Ocean Physics, Scott Polar Research Institute, University of Cambridge, UK
    (from 1 Jan 2003: Dept. of Applied Mathematics and Theoretical Physics)

    "Cooling the water down
    Consider a fresh water body being cooled from above, for instance a lake at the end of summer experiencing subzero air temperatures. As the water cools the density increases so the surface water sinks, to be replaced by warmer water from below, which is in its turn cooled. This creates a pattern of convection through which the whole water body gradually cools. When the temperature reaches 4°C, the lake reaches its maximum density. Further cooling results in the colder water becoming less dense and staying at the surface. This thin cold layer can then be rapidly cooled down to the freezing point, and ice can form on the surface even though the temperature of the underlying water may still be close to 4°C. Thus a lake can experience ice formation while considerable heat still remains in the deeper parts.

    This does not apply to sea water. The addition of salt to the water lowers the temperature of maximum density, and once the salinity exceeds 24.7 parts per thousand (most Arctic surface water is 30-35), the temperature of maximum density disappears. Cooling of the ocean surface by a cold atmosphere will therefore always make the surface water more dense and will continue to cause convection right down to the freezing point - which itself is depressed by the addition of salt to about -1.8°C for typical sea water.

    It may seem, then, that the whole water column in an ocean has to be cooled to the freezing point before freezing can begin at the surface, but in fact the Arctic Ocean is composed of layers of water with different properties, and at the base of the surface layer there is a big jump in density (known as a pycnocline), so convection only involves the surface layer down to that level (about 100-150 metres). Even so, it takes some time to cool a heated summer water mass down to the freezing point, and so new sea ice forms on a sea surface later in the autumn than does lake ice in similar climatic conditions."

    Full article is at:

    http://www.arctic.noaa.gov/essay_wadhams.html

    Patrick:

    Where are you?

    Hope you are well.

    Oh for heavens sake, Will. I'm not a climatologist (perhaps you are?) and this isn't Nature. I don't "need" to substantiate anything. If you read my posts with interest, fantastic. If you disagree with them, fine. But your latest batch of posts have perhaps a dozen statements of what I do or don't need to establish or correct - things which I have already conceded, disclaimed in advance or commented on, or which we both know lack the data to confirm either way. And, honestly I didn't come here for an argument. Really, I was just sticking around to see if anyone would put a hand up to *fit better curves* to the data (another point you belabour), and since no-one has, I can't add to anything that I have said here. I maintain that the curves I have fitted to the PIOMAS data are a better reflection of the past trend than the simple linear trend on their graph. But I have never maintained that they produce accurate forecasts for the future (quite the opposite!).

    I do hope what I have said here gives cause for people to reconsider the timeframes we are talking about, because frankly I consider projections using area trends to be hopelessly facile. You have made some good observations about some area data that would be better but which doesn't actually exist, and I agree with a lot of what you have said about 1st year ice. But here's the thing, it doesn't matter who is right and who isn't, or if neither are, or both (unless you have money on it, which I don't). Nature will do what she does, regardless of what two amateurs think she *should*. And, with a certain grimness, I'm content to continue watching her do it, regardless of whether it takes 4 or 40 years.

    And truly, I too hope Patrick is well, and look forward to discussing the whole situation further when he breaks out a new reel of thread. This sort of back and forth could be had on a hundred different blogs, but only Patrick ladles out the yummy servings of historical insight which brought me here in the first place. So stay well, PL.

    kthx, bye (for now)

    FrankD:

    Sorry, I have been battling a miserable cold and it made my response overly testy.

    Please forgive any excess.

    We will just have to wait and see how the next decade unfolds with respect to Arctic ice as you and I have exhausted almost all available data in our discussions.

    Thanks for your patience.

    logicman
    Thanks to Will and Frank - and all others - for keeping my blog alive with your insightful debate and contributions.

    Thanks also for your kind thoughts.

    I have been a little unwell - as noted in my most recent blog: I'm Back - With an Apology.  I hope to start a new discussion thread soon with new arguments in support of my thesis that the Arctic ocean will soon have effectively ice-free summers.

    In passing:
    In my previous Arctic article I commented on Ekman transport and suggested that ocean mixing would be a significant factor in future ice extent.  Check out -
    Rainville et al. Observations of internal wave generation in the seasonally ice-free Arctic. Geophysical Research Letters, 2009; 36 (23): L23604 DOI: 10.1029/2009GL041291

    Hat tip to
    Science Daily.
    Patrick, good to see you back. For a while there I thought maybe we lost you.

    I have noticed that the Polar Science Center http://psc.apl.washington.edu/home.php hasn't published ice volume data since December 31, 2010. Do you know what is happening with them?

    Also, has any of the Cryosat II data been analyzed and made available?

    Relieved,

    Vaughn