Arctic Ice June 2010
    By Patrick Lockerby | June 2nd 2010 05:21 PM | 13 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 June 2010

    This article describes the current state of the Arctic sea ice and makes some predictions for the month ahead.  The predictions are based on my visual analysis of hundreds of satellite images. 

    I begin with some background and explanatory materials and then show how satellite images can be analyzed to show the mechanisms of  ice consolidation and fracture.  In the concluding part I show some images of the Arctic ice at the time of writing and point out what I think will happen to that ice during this month.

    It is one thing to feed a stream of data into a computer and output a map or graph, but quite another to make use of the finest instrument known to science: the Mark 1 eyeball.  To be entirely fair to the many scientists engaged in Arctic studies, they do not have the budgets to pay for the man-hours needed to analyze hundreds of images daily.  I, on the other hand, being unable to do much physically, delight in studying those images for hours at a time.

    This article follows on from Arctic Ice May 2010 and  Arctic Ice May 2010 - Update.

    In the first of those articles I wrote:
    I have no doubt that by the end of this month, May 2010, there will be much less sea ice than there was in May 2007.

    I have no doubt either that the anti-science propagandists will continue to insist either that the ice is recovering or that Arctic melt is perfectly normal.

    I claim that I was right on both counts.  Visible spectrum satellite images show a substantial reduction in both the quantity and quality of sea ice.  I shall shortly explain what I mean by those two terms.  As to people denying Arctic melt, or denying that it is unusual, there are actually few global warming deniers wishing to stick their necks out.  I have already written about a pair of propagandists in Desperately Denying Arctic Warming.  Meanwhile, others are trying desperately to pick holes in graphs.  Lord Monkton makes very artistic use of graphs.   But why look at graphs when you can look at the actual ice?  That said, I shall start with some graphs.

    Here is the graph from April 04 2010 that had some bloggers and media talking of Arctic sea ice recovery, but which I discounted as a minor blip:

    This next graph shows ice extent as of June 01 2010 with my annotations in red.  My projection for June ice loss is based on the current state of the entire Arctic ice and on past regional loss rates.  I consider my projection to be, if anything, over-optimistic.

    The above graphs showing ice extent were sourced from the National Snow and Ice Data Center

    The next graph shows ice volume anomalies as for June 02 2010.  I have reduced the image size, but it is otherwise unaltered.  The image is sourced from the Polar Science Center

    Whatever the reliable source of data, the year-on year trend is comparable: Arctic sea ice is shrinking.  Short-term trends of a few days or weeks in graphs can be discounted.  Trends in actual ice loss in Arctic summer can, however, be treated as regional weather trends within the larger Arctic region.  The reason should be clear: ice which has melted in summer will almost certainly not re-freeze until the approach of winter.

    The Polar Science Center explains the purpose in plotting sea ice volume as follows:
    The purpose of this page is to visualize recent variations of total Arctic Sea Ice Volume in the context of longer term variability.  Arctic Sea Ice Volume is an important indicator of climate change because it accounts for variations in sea ice thickness as well as sea ice extent.  Total Arctic sea ice volume cannot currently be observed continuously.  Observations from satellites, Navy submarines, moorings, and field measurements are limited in space or time.  The assimilation of observations into numerical models, currently provides one way of estimating sea ice volume changes on a continuing basis.

    Sea ice quantity and quality

    For purposes of showing past trends it hardly matters what is being measured, as long as the method remains consistent year - on - year.  But ordinary quantitative measures of ice do not readily lend themselves to predictive modelling.  Sea ice is not a simple and uniform material.  It is more like a matrix of aggregate and cement.  Like many materials, when ice fractures, small particles prevent perfect closure of the fracture.  In winter, every fracture in the ice, large or small, is bridged, if at all, with fresh ice.  Fresh sea ice is salty ice.  Salty ice is weaker ice.  Normal build-up of sea ice rejects salt into the ocean below the ice.  If the mass of old and new sea ice is not consolidated by the actions normal in the Arctic in winter then the fractures will remain weak. 

    Quantification by area, extent or volume does not model the weak points in the ice - the points where the ice will melt, fracture, spread out and thence either melt slowly as bits in situ or be dispersed into warmer waters where melt is more rapid.

    The quality of sea ice may be considered as a set of engineering materials properties.  Ice is weak in shear, that is why thin ice is easily punctured by a human foot.  Ice is weak in tension.  This shows up when a shore-fast ice sheet is subjected to a strong offshore wind or current.  The ice fractures at the shore, producing a shore lead.

    When large ice floes move around in water which is freezing over, the ice between the floes tends to become a mix of bits of ice all frozen together.  A mixture of thicker and thinner floes frozen into a sheet is called an ice mélange, adapting a term from geology.  Typically, the largest floes in a mélange show a top surface pattern of fissures matching the ice tongue from which they calved.

    Ice melange, from Jacobshavn glacier tongue, to right.

    A variety of ways in which sea ice can fracture and re-freeze is shown in the next images.  If fractures are relatively uncompressed they tend to show as shades of grey in true color images.  Where sheets of ice are compressed together compression ridges form.  These are thicker than the sheets, more compact and more free of salt.  The ridges show as white lines against bluish ice.  Note that the color patterns of the ice mirror the albedo patterns.  It is often possible to predict in the short term where the ice will fracture next.

    Ice in Kane basin, Nares Strait, August 2007.
    A - blue color of water shows through thin ice, white lines are thick, compressed ice.
    B - thicker ice separated by recently frozen thinner ice.
    C - compacted and re-frozen fragments at least 1 year old.
    D - Humboldt glacier tongue calving point.

    E - Humboldt glacier, ice fairly immobile, apparently grounded.
    F - margin of coastal melt zone with moulins.


    Detail near location of image above, August 2007.
    A - A
      shows a line of grey cloud.
    B - blue color of water shows through thin ice, white lines are thick, compressed ice.
         Some blue areas also have melt pools above the ice.
    C - ice type B blends into type D.

    D - thicker ice separated by recently frozen thinner ice.
    E - ice caps and ice sheet.

    Images sourced from an 'image of the day' for August 30 2007:

    Discrimination problems

    There are two problems with using computer analysis and modelling of sea ice.  The first is a problem of measurement, the second a problem of identification.

    Ice thickness measurements take a sea level datum and measure the height of the ice above datum.  The problem with this method is that the ice is 9/10ths submerged.  Any error margin in measuring the ice height above datum may be taken to be 9 times that error below water.  Data publishers need to make clear whether their error bars apply to the directly measured height or the computed thickness.  If the error margin is not carefully specified then it would not be unreasonable for anyone to assume that it is the above-datum error only - and allow for a larger-than-published error.

    The problem of identification is a problem of determining exactly what a satellite is recording.  As can be seen in the two images above, there can be subtle changes across ice types which might lead to errors in determining sea ice extent.  There is also the problem of coastal contamination - a situation where land ice is seen as sea ice and vice versa.  Allowances are made for such problems in data analysis and modelling, but direct observation - preferably at ground level - gives far greater accuracy.

    Recent Arctic sea ice images

    The most recent ice extent map from NSIDC is shown below, followed by detail images from MODIS satellites which highlight some of the regions in the map.

    Sea ice extent June 02 2010
    Image source:

    May 31 2010 Bering Strait area.
    The ice between the red lines is now open water, but is not readily visible in more recent images due to cloud cover.  The two areas indicated by the green lines are areas of broken and mobile ice.  The Beaufort Sea, linked in my final image below, is off-image at bottom left.


    May 31 2010 - Greenland - Svalbard - Franz Josef - Novaya Zemlya
    The green line shows the approximate limit of continuous ice.  Ice south of the line is highly fragmented and mobile.


    June 03 2010 - Franz Josef Land and Novaya Zemlya
    There is little ice in the Barents Sea area shown here.  The ice at the top of the image is highly fragmented and mobile.  In the Kara Sea between Novaya Zemlya and the mainland the ice is highly fragmented but trapped, with only a small strait as outlet.


    June 03 2010 - Greenland Sea Scoresbysund / Ittoqqortoormiit.
    Apart from the small area indicated by green lines, there are few large floes left along the coast.  Dense ice-smoke and cloud indicates that some ice is still melting in the area, but it is much less than the 25% extent indicated by the NSIDC map.


    June 03 2010 - Baffin Bay.
    The Jacobshavn glacier is arrowed.  Under the heavy cloud is open water.  The main body of pack ice can be seen clearly towards the top of the image.  To the south of Davis Strait - off image - the water is very much open and ice free.


    June 03 2010 - Nares Strait.

    The Nares Strait remains open.  Ice is channeled from the Lincoln Sea, right, to Baffin Bay, left, melting along the way in the warm currents of the Nares Strait

    June 03 2010 - From Lincoln Sea to Fram Strait.
    The entire Lincoln Sea area remains highly fragmented and mobile, but this fact is obscured by hazy cloud.  The highly fragmented area at right continues to lose fragments into the Fram Strait, bottom right.  The two areas of fragmented ice are connected by a strip, approximately 10 miles wide of highly fragmented coastal ice.  That strip is mainly pressed inshore, but leads develop frequently and the ice appears to be drifting mainly towards Fram Strait.


    May 31 2010 - North West Passages.
    The ice between Parry Channel and McClure Strait - between the melt fronts indicated by the green lines - has persisted for some time.  It is likely, in my opinion, that opposing tides or currents have caused a consolidation here as an ice bridge.  Until that ice is gone, the movement of ice fragments in the channels will be very restricted.  I expect more melting in this area over the next month, but expect the ice bridge to remain for about 2 to 4 weeks more.

    May 31 2010 - Beaufort Sea.
    From the McClure Strait we proceed into the Beaufort Sea.  The lobe of open water at left is the Amundsen Gulf.  It leads into the two still frozen routes around Victoria Island.  Cloud at top right continues to obscure the ice.  Much of the ice under that cloud is highly fragmented.  It connects to the Bering Strait.


    In estimating what will happen in the Arctic short-term, I suggest we should first try to determine the mobility of ice in a given area.  If ice is highly mobile and likely to soon enter warmer waters then discussions of its thickness or age are somewhat moot.

    In order to determine if ice is likely to become mobile any time soon,  we should try to estimate its past and current degree of fragmentation and re-consolidation.

    Credit: all of the Modis images used here are courtesy MODIS Rapid Response System -

    I hope to post an update to this article in about 2 weeks, or sooner if anything special happens.

    Related/recommended reading:


    Very interesting read, Patrick! Thanks so much. I have started watching those MODIS images too recently, so the extra explanation is very helpful. The other years had quite a bit of heavy melting the coming few days (2007, 2008 and 2009 all had 1 or 2 days of >100K melt, what I call 'century breaks' ;-) ), so let's see if 2010 can maintain its lead.

    ps I spotted two typos: 'snall particles', 'a siruation', and I'm still not sure about sheer vs shear.

    Neven: thanks for the feedback.

    ps I spotted two typos: 'snall particles', 'a siruation', and I'm still not sure about sheer vs shear.

    Sheer - often refers to vertical, as in a sheer drop. Nautical: to veer or swerve.
    Shear - to separate by parallel and opposing forces.  As in force of foot on ice against force of buoyancy surrounding the foot.

    I'll also add these links as related/recommended reading:

    edit: typos fixed, many thanks.  Two were 'adjacency errors' where the wrong keyboard key gets pressed.

    The spelling of 'sheer' where 'shear' was intended was a plain and simple, good old-fashioned brain malfunction.
    I suggest blink-testing tiles of the MODIS true-colour mosaic over several consecutive days, using separate browser tabs. I was just doing this with the northern-Greenland 1km tile, for instance,
    Loading several of those tiles from consecutive days (...2010151.terra, ...2010152.terra, etc) into adjacent browser tabs and using Ctrl-Tab to cycle through them, it is very easy to see ice motion and to gauge its approximate speed. I was surprised to discover that the ice north of Greenland is moving *West*, towards the Beaufort Sea, at approximately 20 km per day, not east to the Fram Strait.
    One can also use this technique to easily gauge the rate of retreat of ice fronts, for instance in the McClure Strait.
    Of all the ice this year, I am most concerned with the apparent condition of the ice north of Greenland and Ellesmere Island. This is the last refuge of the really thick multi-year ice, and it's all on the move. There are large white areas - scores to hundreds of km - with darker areas between, but there is enough mist ("ice steam" ?) to obscure whether the darker areas are actually open leads. I don't have the expertise to tell. But it definitely looks less solid than in previous years (substitute e.g. 2009 for 2010 in the same URLs to see for yourself).

    Nick: thanks for the feedback - much appreciated.

    In response to your points:

    I suggest blink-testing tiles of the MODIS true-colour mosaic over several consecutive days ...
    I go for several months, but the animated files tend to get huge.  :-)
    See my recent article: Interpreting Arctic Satellite Images And Data #2 - Animations

    Try blink-testing same-day aqua vs terra Arctic mosaics to help pick out clouds.
    Blink 7, 14, 21, 28 days apart images to see areas which have greatest ice-cover changes.

    I was surprised to discover that the ice north of Greenland is moving *West*, towards the Beaufort Sea, at approximately 20 km per day, not east to the Fram Strait.

    I guess you were witnessing the main pack.  My reference in the article was to a fairly narrow band close inshore.  From a point midway between the Nares Strait and the very tip of Greenland, the prevailing ice motion splits east and west.  Close inshore where fragments are small and leads are more open, the ice tends to flow from the edge of the Lincoln Sea towards Fram Strait.   The arrows in this image demonstrate the most common main pack flows:

    Of all the ice this year, I am most concerned with the apparent condition of the ice north of Greenland and Ellesmere Island. This is the last refuge of the really thick multi-year ice, and it's all on the move.
    I agree entirely.  I have been studying that area intensely to try to discover any trend.  There appears to be a trend to a continuous band of fragmented coastal ice.   That is not good.  The region is the only part of the pack that has remained shorebound in recent years.  A free-floating expanse of relatively young sea ice is going to be very vulnerable.

    This image from MODIS/Terra 2010/123 05/03/10 17:50 UTC shows earlier fragmentation.

    Those images are from Arctic Ice May 2010.

    A tip:  if you see a series of curved cracks, a line pointing through them away from the convexity will show the drift direction.  It's the opposite sense to a bow and arrow.  You may wish to check this for yourself.

    If you have any more comments, queries or suggestions, please feel free to fire away.
    Help to keep me honest!  :-)
    The main pack, yes. But it looks to me that it is moving west, even NE of Cape Morris Jesup.

    But it looks to me that it is moving west, even NE of Cape Morris Jesup.

    You may well be right, Nick.

    This melt season is totally abnormal.  At this time of year the ice should be fully shore-bound somewhere between the Nares Polynya and the Fram Strait.  The most common drift pattern is towards Cape Morris Jesup where the streams divide.

    What I am seeing right now is solid ice around Cape Morris Jesup which looks like either an ice shelf or embayed thick ice.  Outside of that, and all along the coast the ice is highly mobile.  It may be that, lacking a shore-bound anchor, the point where the ice movement divides is shifting around.

    I guess I'll have to make myself a hi-res animation of ice motion in that region and see what gives.

    The 250m resolution image for today is less obscured by cloud, so you can see how the Lincoln Sea and Fram Strait are linked by a lead filled with mobile ice.

    The NSIDC Arctic Sea Ice News&Analysis 2007 archive is well worth a visit for comparison with current ice conditions.  Watch out for figure 3 - it's a huge 139 Mb.
    If I can remember, the Mcclure Strait-Parry Channel does not normally fracture until mid August. The fact that it has been reduced to the status of an ice bridge, that will be gone in two to four weeks is amazing.

    I am one that also checks the MODIS images every few days, and the condition of the ice is very bad. The fact that consolidated ice is few and far between, and so early in the year, will practically guarantee a new minimun.

    I can still see the WUWT headlines in 10 years however "Summer ice minimun holds steady at 0 sq km for five years in a row, an indication that recovery is on its way" :)

    If I can remember, the Mcclure Strait-Parry Channel does not normally fracture until mid August.
    If you go back in history, most of that route was impenetrable.

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

    NSIDC 2007

    The main, deep channel of the Northwest Passage (Lancaster Sound to M'Clure Strait) has been open, or nearly ice-free, for about five weeks (since August 11, approximately).

    In 2007 most of the ice loss in all areas of loss was post June.  At the start of winter, much of the east coast of Greenland was still fringed with pack ice.  At the time of writing, June 04 2010, there is no great amount of pack ice south of a line drawn through the narrowest part of Fram Strait.

    I can still see the WUWT headlines in 10 years however "Summer ice minimum holds steady at 0 sq km for five years in a row, an indication that recovery is on its way" :)
    Viscount Melville Sound can remain filled with pack ice for weeks or months while other channels in the passage open up. Just look back over the last few weeks: the ice boundaries haven't shifted for almost a month (although the loose ice outside the pack has dispersed and melted). The state of the passage is very weather-dependent. Watch the skies!

    Nick: thanks for the info.

    If you look at satellite images of the area in previous years you can see that even when the channel was open there was a lot of - presumably thick - shorefast ice remaining for winter ice to build on.

    This year has seen thick cloud cover there most days, and a lot of what looks like fog banks.  As far as I recall, the area has only been cloud free on occasional days.

    For anyone not familiar with the names, Viscount Melville Sound is the location of what I refer to as an 'ice bridge' in my article.  There are only three routes for ice, tides and currents into and out of Melville Sound: McClure Strait, McClintock Strait and Prince of Wales Strait.  Prince of Wales Strait has been showing signs of ice thinning, so may open soon.  Once open, it can feed warmer surface water into Melville Sound and increase the ice loss rate there.  My projection of 2 to 4 weeks in the article allows for the influence of Prince of Wales Strait.

    Today's image shows much cloud cover.  The arc at the ice edge in McClure Strait has not changed its location for some weeks, indicating an ice bridge which is probably not being subjected to warm water melting.  The other ice edge is creeping extremely slowly along the channel.

    This April blip interest me; as I've already noticed a seasonality to the anomaly graphs.
    The most negative values are tending to occur in September (after sunset),
    while the ice then tends to "recover" to near normal values by the winter maximum.
    At least this seem to have been the pattern over the last 3 years.

    Could this "recovery" or blip instead be a reflection of a generally more mobile and
    expanded pack of ice as opposed to historical norms of density?

    Could this "recovery" or blip instead be a reflection of a generally more mobile and
    expanded pack of ice as opposed to historical norms of density?

    Andrew: I agree in general.  An ice pack which has been highly mobile and spread out in summer leaves much interspersed open water. During the first part of winter darkness thick water gives off some of its heat - to a limited depth - to the atmosphere.  The atmosphere, due to the greenhouse effect, slows the rate of heat loss to space.

    It is not widely known that the formation of new ice also gives off heat, albeit a relatively small amount.  Taken all together, these heat inputs to the atmosphere delay the start of the build up of ice.  In effect, true Arctic winter is getting shorter.

    I have started to talk of quality and mobility because I feel that the terms 'extent', 'area', 'volume', 'thickness' and 'density' are not properly understood by the lay public.  During the past three decades we have seen a dramatic loss of ice quality and a rise in mobility which do not show up in graphs.  The term 'rotten ice' is apt, but still doesn't allow for mobility.

    I hope to write an article soon on the Arctic ice formation mechanisms, focused on how quality and mobility affect the progress of the thawing and freezing seasons.

    Thanks for the comment.  I always appreciate intelligent feedback.
    Dear Patrick,
    I'd just like to say thank you. I look at the Bremen graphs regularly and you have provided a very interesting detail well beyond those.
    Best Regards