« 2012 minimum global sea ice area | Main | A first clue »


Feed You can follow this conversation by subscribing to the comment feed for this post.


so difficult when things are covered in cloud. I cannot see anything either way in that day 43 shot that makes me want to comment further.


Welcome back, Espen!

Michael Fliss

In this MODIS tera 250m Bands 3-6-7 image from day 72(vector option coast only) Arctic_ r02c05 Subset, one can see tendrils of ice reaching across open sea towards a partially iced Kara Strait.


If you are not sure of of the meaning of the bands (I'm sure many are), the colors of the image are explained in the link below.



Hi Espen,

Re North of Svalbard


Not really like this year, but it is 1985!

When I looked at end of January it was common for there to be open water north of Svalbard; near one year in three. However, it seems to be much rarer for early/mid March. Really just 1985 and 2006 that are anything like 2012.

Kevin McKinney

"This maximum is a weather event on the end of a low ice winter. The graphs tell you that at a glance. Smooth curve is climate sudden blip is weather. How hard is that to grasp?"

Ah, but wishful thinking is powerful. We saw the same thing last year, and no doubt we'll see it again. We know that extent will crash just as quickly as it peaked, but we can't very well save the short-sighted from their disappointment.


Comparing Uni Bremen SIC March 12th to 11th I see quite a bit of retreat, especially in Kara and Baffin. I wonder what effect that will have on CT SIA.


Neven wrote:

especially in Kara ...

As the sun climbs up to 15 degrees in that zone now we can assume ice there is definitely on his way back, which would be spectacularly early in the year.

Neven wrote too :

... and Baffin.

Really? To me it looks very much the same.

Moreover, at the Eastern part of Greenland - to the south of Iceland - in a couple of days coastal ice has nearly doubled in surface. Which is a quite normal phenomena in this time of the year anyway. Remember, ice there "normally" vanishes away in the first half of Juin.


I had another look (via the new Uni Bremen archive and I definitely see a retreat in Baffin, but not very marked, I admit. That could be the Sun's influence, as the ice edge is positioned at 60N. It can't be the winds, because they come from the North.

There is ice growth in the Greenland Sea, and a bit in Bering. I don't if that can fully compensate the retreat I see in Kara, Baffin and Okhotsk.

I'm not seeing anything conclusive in the ECMWF forecasts, so I guess we'll just keep an eye on CT SIA. The maximum could go higher today.


Regarding Baffin, is it highly broken up and being moved south by winds more than normal this year? If so, does this have an effect of the ice further north being thinner than usual? Or has it just been cold enough that the replacement ice quickly thickens up to normal thickness?


Crandles wrote:

If so, does this have an effect of the ice further north being thinner than usual?

Let's have a look at todays conditions:

At Baffin Island, the whole coast from the NW Passage down to Hudson street temperatures are around -35 °C. At Alert -38 °C. Greenland's Baffin Bay coast temperatures are from -38 °C at Thule, -16 °C at llulissat (Disko Bay) to -8 °C at Nuuk.

IMHO in these conditions the ice field only can grow thicker and expand itself. As it does since the second week of February.

Remember, remember, the melt season in that region only begins in June, seldom in the last week of May.

Bottom line, we better keep our heads cool. :-)


Drop of 61k to 98k below this years max so far 6 days ago.

3 of last 10 years have rises of more than 98k from this .1918 date.


...though it should be noted that he day before yyyy.1891 is meant to be the peak date for Arctic sea ice, as per the long-term average.

OTOH, recent years have shown higher ice concentration after the peak than before it.


@Crandles: Regarding Baffin, is it highly broken up and being moved south by winds more than normal this year?

Difficult question to answer fully, but for simplicity:
Comparing MODIS pix for Day 72/2011 to Day 72/2012, I would say Baffin Bay was more broken up last year than this. In particular last year had greater intrusions either side of Devon Island and reaching up towards Nares Strait (Caveat - there was an ice arch last year which exaggerated the appearence of ice clearing below it, while actually blocking the passage of ice down Nares).

Comparing the change between Day 71 and Day 72 for the two years, it appears that where ice is clearly moving south, it was moving faster last year than this. (Caveat - this year appears to be slower, but moving more consistently. Last year saw something of a gyre in Baffin, with ice moving rapidly down the coast of Baffin Island, but moving laterally or even northwards on the Greenland side).

So, briefly, definitely probably not, but maybe a little sort of... :^P

Honestly, I think Baffin will take longer to clear than last year. This years early action will be be on the Russian Front, obviously.


Looking at Modis it seems like a large part of the Baffin ice if fragmented and being pushed out by strong winds if the clouds are an indication?



Drop of 61k to 98k below this years max so far 6 days ago.

Call me crazy, but one more drop tomorrow and I'm calling the CT SIA maximum.

I'm seeing things on ECMWF in 4 days from now that will definitely reduce ice in Barentsz/Kara and Bering.


Kane Basin (Nares) looks a lot thicker and more compact this year, compared to last year.

And here's a radar image for 2011 May 23rd (the ice bridges collapsed a couple of weeks later).

I remember that a late cold spell in the region had solid ice forming in front of the plug. It took a while, longer than previous years, to fall apart. I'm not seeing any solid ice so far this year, but there's still 2 months+ to go. And again, things look more compact and thus thicker now than last year.


Given the previous posts on continental warming and albedo, I thought a revisit to the current CONUS "snow drought" might add to the discussion. This is as of today.

The problems are at least two fold. First the open areas are helping hold the heat, second, with the early snow melt, later this summer US farmers and the midwest are going to be running dry on snowmelt for irrigation.


Chris Biscan


Definitely another sizable drop coming.

I am calling the Max after tomorrow if this translate.

Jaxa's SSMI image also shows another large drop.

Mike Constable

I see the Barrow Mass Balance site has not updated since 5th March - possibly the polar bears have upset their equipment again?
A pity we won't have these clues as to the state of the ice this year unless it is repaired again.

L. Hamilton

No sizable drops, but DMI extent on 3/13 is about 120k below the high point reached on 3/5.


As at Svalbard the third hotspell has come to it's end we can have an oversight now.

Be aware the records on the 4th and 4th of March were more than 4 °C above the previous records.

As a reminder first the oversights of the hotspells in early January and and January- February



Mike wrote:

Barrow Mass Balance site has not updated since 5th March

As they say in red, they don't succeed in calibrating the instruments.

Frankly, 1,40 m thick an ice sheet early March looks very suspicious to me, as hitherto that thickness only has been reached in the last week of May. IMHO even the extended cold spell in Alaska can't explain for that figure.


Since 24 hours the Southern part of the Hudson Bay has been invaded by warm air currents, temperatures are still positive there at present.

The warm air mass is stretching to Labrador, New Found Land, Hudson Strait and even the Northern half of Baffin Island.

We might expect some substantial changes there anytime soon.

Mike Constable

Barrow Mass Balance Site:-
They were getting readings after repairs to the site, but I got the impression that snow depth figures were unreliable. Now no data is being added so I think the polar bear has been back.
The ice thickness is assessed from readings from instruments frozen in the ice so should be out of reach of even bears!?


Mike has stated:

The ice thickness is assessed from readings from instruments frozen in the ice so should be out of reach of even bears!?

Never underestimate polar bears' muscle power. :-)
Remember, they are able to break a whole into a 50 cm thick ice field.

Anyway, the site is announcing for over 5 days "Current ice growth rate: 1.2 cm". Nevertheless, the ice thickness is reported all the time at the same level: 1,40 m.

So, one thing is absolutely sure, that is not right.

Incidentally, immediately after the "repair" the ice thickness remained for 8 days announced at a level 1,37 m whith an all the time a 1,2 cm daily growth.

Bottom line, to bad, but clearly the data coming from Barrow aren't affidable at all.


Stupid me wrote:

break a whole

That should have been/must be "... break a hole" of course.


Speaking of breaking a hole, have we notice how much ice is on the move from Chukchi to Bering. The last few days MODIS pix show some fairly dramatic changes: apidfire.sci.gsfc.nasa.gov/imagery/subsets/?subset=Arctic_r06c03.2012072.terra.1km

Cycle through from Day 72 to the present.


Ooops, busted link. Try http://rapidfire.sci.gsfc.nasa.gov/imagery/subsets/?subset=Arctic_r06c03.2012072.terra.1km

Kevin McKinney

Those with an interest in early climate-related science may be interested to check out my debut post on Skeptical Science. It sets forth the life, work and times of the remarkable William Charles Wells (1757-1817.)

Among other things, he observed backradiation from clouds beginning in 1811, as part of a research program which correctly attributed the formation of dew to radiational cooling:




Wonderful article about an unsung (or at least I'd never heard the melody) pioneer in the field.
You wield a sharp pen!!



BBC's Richard Black reports more on that meeting of scientists and MP's last week:


An eminent UK engineer is suggesting building cloud-whitening towers in the Faroe Islands as a "technical fix" for warming across the Arctic.

Scientists told UK MPs this week that the possibility of a major methane release triggered by melting Arctic ice constitutes a "planetary emergency".

The Arctic could be sea-ice free each September within a few years.

The article contains a graph familiar to the readers of this blog.


Hmm, where did they get that graph from? I'm seeing 'piomas project' in the bottom right corner.

Kevin McKinney

Terry, thanks! Much appreciated, and glad you enjoyed it. I think Wells is comparatively little-known, though he was clearly an extraordinary thinker.


I went delving to see where Richard Black got the mystery graph. One thing not completely clear from how Black has worded his article is that this site has now entered the official records of the British Parliament (which puts us one up on a certain British Lord, I think!).

The reference to UK MP's brings us to the Environmental Audit committee, a House of Commons Select Committee, which has been taking submissions on HC 1739 - "Protecting the Arctic". The written submissions have been colated into a rather large (164 page) pdf, in which we see Professor Peter Wadhams (University of Cambridge) source several remarks to this blog and to arctischepinguin (Wipneus's own site, yes?) and include a Larry Hamilton special. Footnote 3a in the written submissions is the link to the version of Wipneus'graph hosted here (without the "PIOMAS project" footer). I assume that Black grabbed a copy and tweaked the caption a bit, which is a bit naughty if that's what happened.

There are also a couple of days of verbal submissions. Whole thing is here: http://www.parliament.uk/business/committees/committees-a-z/commons-select/environmental-audit-committee/publications/
(in amongst other material).

Now, I have always been a bit guarded about the usefulness of that quadratic extrapolation into the future (while remaining totally convinced that it is the best description of what has happened in the past 30 years). But it seems that some bona-fide scientists are recognising that linear fits don't actually fit so well.

L. Hamilton

FrankD, thanks for finding that, I never would have known. Reminds me of why I put the data: source and graph: me at lower left in each image, thinking they might travel who knows where in the world.

Mike Constable

On February 14, 2012 I wrote "I am wondering - if the N. Atlantic water is not being cooled enough to produce ice and the extra saline deep water - is it just flowing under the ice and resulting in a surface flow out of the arctic between Siberia and Alaska?
That might also help explain the extra ice in that area, when other parts of the N. Pacific still have less ice than usual."

Now the arctic.io maps are showing that huge ice outflow into the Pacific - something must be driving it (I would not know if it was wind) but if it is a current (is that being measured by anyone?) it is making room for a lot of warm Atlantic surface water in the Arctic Ocean!

If the water flowing out is also relatively fresh that would solve the "problem" of whether a build-up of fresh water in the Arctic would block the formation of the cold deep saline flows at the end of the Gulf Stream. Northern Europe would then keep its mild climate (unless open seas in winter create winds from Siberia?!).

Many thoughts, questions!?

Mike Constable

But perhaps the ice is not continuing to flow into the Pacific very fast?
I did try to look up the Wales web-cam and radar but both are out of action (for a long time?) - when they might have been useful?


Mike wrote:

the Wales webcam

That cam chas changed ownership, it's now here:


Anyway, hitherto we only could see there some movement from the second week of May on.
Still a mass of time to go.


FrankD: Thank you very much with finding that information.

About the exponential extrapolation, my biggest doubt is that I know of no clear physical explanation of the exponential curve. The albedo effect seems to be too weak. Therefore I find it encouraging that a professor in ocean physics has this to say:

If we think in volume terms instead of area terms, the downward trend is more than linear, in fact it is exponential, and if extrapolated it gives us an ice-free summer Arctic as early as 2015. Others have talked of later dates, like 2030-2040, but I do not see how the trend of summer ice volume can possible permit this. Those who agree include W Maslowsky, a leading ice modeller (Naval Postgraduate School, Monterey), and the PIOMAS project at University of Washington which generated the data shown below.

iv) (below). Minimum volume of Arctic sea ice in midsummer, based on areas observed from satellites and thickness trends inferred from submarine observations. Extrapolation leads to a zero volume in 2015.

This comes close to my feelings. Taken in context with other clues, it is a genuine piece of evidence.

John Christensen

I know that CT is the reference for Arctic SIA, but has anyone noticed the discrepancy compared to Roos, where they are headed towards a new max potentially on 3/17, seemingly 400-500 km2 higher than CT on 3/16?


Why does the professor speak of midsummer (instead of September)?


John, AFAIK Arctic ROOS deviates from the other extent/area graphs most of the time. I don't know what the reason is (resolution, algorithms). What I don't get, is that they have never corrected the 2008 trend line. But that doesn't mean I don't take it seriously. And CT seems to lag the other graphs. I'm expecting some upticks today and tomorrow.


We do have already the "uptick":


Only a fistfull km² above the previous max, but still above.
So it looks like we will reach the max. in a few days.

IMHO the "culprit" still is Baffin Bay where the ice field continues to grow dispete the rising temperatures since yesterday.



I've been watching that Bering flow toward the Pacific using the


site. With the IR animation, the ice arch seem both to be widening and progressing towards the Barrow area.
No way to measure flow rate that I'm aware of, but it does seem substantial and I for one don't recall anything similar in the past.
Is this pumping enough ice to the Pacific/Bering Sea ice edge to make a measurable difference?
Is this acting as a safety valve, allowing the escape of ice that would otherwise end piled high in the Canadian Archipelago?

Arctic Ice melt has to be the greatest show on earth.

michael sweet

Cryosphere Today posts a five day average for area. Arctic ROOS uses a shorter averaging. I think it is one or two days. That makes Arctic ROOS climb faster and decline faster with changes. I expect CT to start showing gains.

I looked back a few years on ARTICIO and there are other times where a lot of ice flows through the Bearing strait. The best comparison would be to how much went through 20 years ago, but ARTICIO does not go back that far. Hopefully NSIDC will discuss this in their next update. It is possible to estimate flow by following individual ice floes.



totally different ice last year but you can still see the movement. having said that, what is happening there is trivia compared with the Atlantic side.
My gut says we see open water at or close to the north pole this September directly navigable from the direction of the East Siberian sea.


Axel Schweiger

has a post at


relevant to PIOMAS exponential extrapolation and incorrect attribution to University of Washington.

Reading the parlimentary evidence I was surprised at the lack of caveats pointing out models expect a declining rate of volume loss as we approach ice free summer and this could be correct even if not evident in the data yet. Schweiger comment seems even less impressed.

Andrew Xnn

One problem with extrapolating the PIOMAS data is that there is a possibility that the data set contains a bias that is growing over time.

The system works by measuring the time difference for microwaves bouncing off the ice and the water. However, there are also melt ponds on the ice which could be throwing off these measurements. More melt ponds could mean a greater bias.

Bob Wallace

Wouldn't there be a correction factor for increased amount of melt pond?

I'm betting that the PIOMAS data must be cross-checked by on site observations.

Every year there are teams on the ice taking "drill and tape measure" data. If PIOMAS were off significantly I would think other measurement groups would be making noise.


>"The system works by measuring the time difference for microwaves" ...
"However, there are also melt ponds on the ice which could be throwing off these measurements. More melt ponds could mean a greater bias."

Are you sure?

AIUI PIOMAS is a model with assimulation of real measurements. The correction to reality occurs only when there is less than 15%? ice cover, when, if the model has more thickness, it is reset to 0. There may also be correction where model has zero but there is over 15% ice cover.

They obviously have worked on getting the model to avoid being biased when compared to submarine, buoy and drill measurements. Links to verification work have been posted here.


Bremen is up again.

And as could have been expected, we can see that quite a lot of heath polynia have been created in the Hudson Bay due to the suddenly rise of temperature there. And temperature still is just above or just under 0 °C in James Bay, Hudson Bay's Southern "boop".

Moreover, it looks like a real heath wave has started in central America (Southern part of mid Canada, Northern part of mid USA - from about 200 km East of Michigan Lake on.

But our American friends will sure be able to elaborate on that phenomenum.


The exponential curve falls mathematically out of the physics of a simplified ice melt model.

If temperature in the Arctic is increasing linearly then it can be assumed that the amount of ice lost each melt season will also increase linearly, as that is how the physics of melting ice works.

If we assume that there is no change to the freeze - in other words, the same amount of ice freezes each year - then we end up with an exponential decline in end of summer ice volume.

Try it in excel or in R. There is no feedback necessary for this to occur. All it requires is a linear increase in temperature to cause a linear increase in ice melt.

Note that this is a very simplified model. But the fact that the simple model automatically pops out an exponential result is pretty interesting when that is what we are observing ...


The equation (not exponential - polynomial) that falls out is:

-At^2 + At + C

where A is half the annual increase in melt and C is the minimum ice value at t = 0, with t the number of years since t = 0.

As an example, starting with an annual maximum of 35,000 cubic kilometres and an annual minimum of 25,000 cubic kilometres and increasing melt by 10 cubic kilometres per year sees zero summer sea ice in 71 years.

Increasing the melt by 30 cubic kilometres per year instead sees zero summer sea ice in 41 years.


To add some more rough maths to the picture, if we assume:

1.) an ongoing energy imbalance of 0.5 watts per square metre;

2.) roughly 10^7 seconds in the melt season; and

3.) roughly 10^12 square metres in the Arctic;

then we end up with around 5 * 10^19 additional joules being pumped into the Arctic every year.

It takes around 3 * 10^17 joules to melt one cubic kilometre of ice.

This indicates that there is more than enough extra energy per year to melt around 25 additional cubic kilometres, plus some for ocean heating plus some for atmospheric heating.

And if we plug in proper figures (around 17,500 for the volume at the end of the melt season of year 0) to the equation, 25 is close to the number that matches what we have seen in the PIOMAS volume curve.


Correction to the equation: C is the *maximum* ice value at t = 0.


D'oh: right the first time on the equation ...


Now only 51k above max so far and 4 of last 10 (or 17 of 33) increases from YYYY.2028 to subsequent max were more than 51k.

A new max looks quite likely from the figures only. OTOH Okhotsk looks like it is going to get some strong warm winds from south that will melt and compact ice.


Reductionist, is that too simple? Shouldn't we account for
1) Albedo effect
2) increased rate of freeze in winter (once it does start to freeze) the thinner the ice gets.
3) Effects of different latitudes of length of time with different sun angles.

It appears to me that so far 1) has been more important than 2) and the curve is accelerating downwards. There doesn't seem much time left for that to change. Nevertheless, is it possible that 2 becomes more important than 1 as the ice gets thinner so that the curve starts to twist upwards instead?


Hudson Bay SIA is taking a bit of a nosedive.


crandles - yes, Dr Schweiger was a bit dark, and understandably. I've had a chance to read the submission in a bit more detail, and I think it comes down a little heavily on one angle (which I happen to mostly agree with) for doubtful reasons.

Just wrt to the BBC article: I emailed the BBC to ask them to correct the false impression that PSC had generated the graph (others may have done so too). They have reworded a couple of para's and cleaned off the "PIOMAS Project" label. It is still a little hazy (intentionally?) but at least its no longer actually wrong about the graph.

Peter Ellis

Reductionist: Your argument, as you noticed, favours a quadratic decline, not exponential. The problem is that you can make equally handwavey (yet convincing) arguments in favour of other functions.

For example, consider the summer ice cover: there are many factors that influence how it will melt. Its location (near or far from the ice edge, near or far from the great refrigerator of the Greenland ice cap). Ocean currents. Weather. Clouds. Ice thickness. Ice age. Salinity.

Abstracting this for a moment into a single measure of "meltability", you could use the Central Limit Theorem to argue that with so many variables contributing to the picture, then "meltability" should converge on a normal distribution. Some of the ice will be easy to melt, some very hard to melt, and the majority of it somewhere in between.

Now, once again impose a linearly-increasing forcing that promotes melt, and that normal curve of "meltability" will translate to a sigmoid drop in volume - not linear or quadratic.

In that case one would argue that a Gompertz fit (asymmetric sigmoid) or cumulative distribution function (symmetric) would be the most appropriate curve to fit.


the Hudson bay ice loss appears to be transport taking place in Hudson strait. CT blocks them together does it not?


Frank, you emailed BBC. Which is more important BBC article or parlimentary evidence. Is Dr Schweiger likely to be working on some clarification of parlimentary evidence? If so, best to leave it to him. If not, why not seek to clarify parlimentary evidence as well as correct BBC?

John Christensen

@Philiponfire; yes, Hudson Bay and Strait are combined in the CT SIA, splitting where the Davis Strait begins (Baffin Bay on CT).


R03c04 MODIS day 79
I’ve been wondering about the sharp ice boundary north of Svalbard. Since it became visible ten days ago, it remains very defined, wind direction doesn’t seem to matter much. Usually, the fringe between ice and sea is dispersed, mobile. It can be shattering in floes or dissolving in milky slush. But this looks like a massive , rafted dyke, stretching over more than 500 km. BTW, it did shift south 40 km last days. The dyke is app. 4 km wide. To the north, a broad zone of broken up floes with wide leads. App. 150 km further, the known 2010-2011 structure of the central pack still exists (irregular floes between broad, persisting leads filled with rubble).
Continuing my 14 feb argument about influx of Atlantic water, I guess water from the West Spitsbergen current forms this now 90 K wide semi-polinya. Another branch, the North Novaya Zemlya current, creates the (almost)ice free tongue between Franz Josef Land and Ostrov Vize. That’s where Atlantic water enters the deep Eurasian Basin of the Arctic Ocean through the St. Anna-through.
This is one of the best images of Barentsz Sea currents I could find.

David Gould

I do not think that it is handwaving at all.

A quadratic decline in ice volume is what we should expect to see if temperature in the Arctic is increasing linearly.

Temperature in the Arctic *is* increasing linearly.

Certainly this does not take into account other factors. But those other factors would serve to alter the basic underlying quadratic function. So the quadratic should be our starting point.

All I am saying is that there is no need to try to think of feedback effects to justify a quadratic decline in ice volume. The quadratic decline falls out of the basic physics.

More realistic and more complex physics might change that. But given the data, it doesn't look as though it is straying too far from a quadratic, so ...

Andrew Xnn

The volume of ice melt is not related by any simple mathematical expression to atmospheric temperatures. This is because, the Arctic is so large and there are many other factors as well.

For example, ocean currents and temperature play a role as does the mechanical properties of ice as it breaks up. The intensity and direction of the wind can either enhance or diminish the export of ice from the Arctic to southern (and warmer) waters.

Finally, ice itself tends to keep the ocean heat contained within the ocean. So, in some respects, melting of ice results in greater export of heat from the Arctic and cooler temperatures.

David Gould

I agree that it is very unlikely for the ice melt to be related by any simple mathematical expression to atmospheric temperatures.


1.) the observed decline appears to be a quadratic; and
2.) a direct linear relationship between temperature and the increase in ice melt results in a quadratic function.

These two facts taken together would suggest to me that all the complexities are not having a major effect to divert the ice melt from that simple mathematical relationship with temperature.

Indeed, what it seems to suggest is that all of those complexities merely have an effect at the edges, which would seem reasonable if we assume that there are a spread of effects that operate in both directions.

The reason that I started investigating this was that I did not understand how a linear increase in temperature could possibly result in a quadratic decline in ice volume. But a linear increase in temperature, in a world with no other physics, *must* result in a quadratic decline in ice volume.

I am not arguing that the system is not more complicated than that. What I am arguing is that there is no need to invoke those complications in order to explain what we are observing.


Werther wrote:

This is one of the best images of Barentsz Sea currents I could find.

Thank you.

Some remarks:

- You easely can see the Golf Stream doesn't reach the zone "behind" Nova Zembla. Something nor Barents nor Heemskerk were aware of, an ignorance which was the cause of the catastrophic results of their journeys.

- Yoy easely can deduct the Golf Stream effectively is blocked by the the Ice field's borders.
Thus we can assume if the borders would vanish this year the situation would change dramaticaaly.


Peter Ellis had some trouble with commenting, so he asked me through mail if I could post this:

Reductionist: the quadratic argument only holds if the "extra" ice lost each summer does not grow back in winter - i.e. the amount of new ice formed each year stays constant, while the amount lost each melt season increases each year.

That is not the case, at least for extent and area - the trend in the winter maximum is much shallower than the trend in the summer minimum. That is, the area/extent loss during the melt season is increasing year on year - but so is the area/extent gain during the freeze-up

As yet, the same cannot be said for volume: the trends for summer and winter are almost equal. This is likely because the "extra" ice formed each year is thin first-year ice, which figures much more strongly in area/extent figures than in volume figures.

However, as the Arctic nears a fully seasonal state, then the effect of increased ice formation will take over. While the volume melted each year will continue increasing, the volume re-frozen each year will also start to increase and partially compensate for it (because the large area of first-year ice will be a greater proportion of the total volume).

This effect is one factor contributing to the shallow "tail" of a sigmoid Gompertz-like curve. It's also why the winter extrapolations are even less valid than the summer extrapolations. We may get ice-free Septembers by 2020: we most certainly won't have ice-free Decembers by 2030.

Put another way, the total volume lost each year from peak to trough is around 18 million km^3. First year ice is around 1.5-2m thick on average. So, even if all the ice was first-year ice, you can only melt around 12 million km^2 of ice area. That means that (using your assumptions, which ultimately come down to the total thermal budget), we won't get a fully ice-free summer until the winter maximum drops below 12 million km^2.

We can get very close to it (with only a couple of million km^2 left in September, say), but we won't get all the way there yet.

Instead, what I see happening is a collapse in the next few years (I'll agree with Maslowski's 2016 +/- 3 years) to a state with ~1 million km^2 summer minimum, leaving a mostly-seasonal pack. However, in the winters we'll still be up around 12-13 million km^2. That's governed by the tilt of the Earth's axis. With no Sun, water will freeze, there's no two ways about it.

At that point, the rate of further ice loss will be governed by the re-freeze rather than by the melt season, and with such large amounts of first-year ice re-forming each season, the rate will drop right down.

Remember, first-year ice grows rapidly to its normal 1.5-2m thickness. We see that from Hudson Bay, which gets just as thick as central Arctic first-year ice despite a much shorter winter. For most of the Arctic winter, the first-year ice in the central basin is just sitting there, not getting much thicker. So there is a very large thermal "buffer" to get through before we start to lose the winter ice cover.

So, once we get to the "almost seasonal" state, with a remnant buffer of ~1 million km^2 left in September, I expect the rate of loss to slow right down - or rather, to be balanced by a much greater re-freeze.

Note: All the current climate models show a sigmoid pattern to the loss of Arctic sea ice - be that winter or summer figures. Trust them. They may be wrong on the time-scale owing to various feedbacks, but they won't be far wrong on the general nature of how the loss occurs.

Peter Ellis

Thanks Neven!

Rereading this, it strikes me that I'm actually almost re-making William Crump's argument, but with many more words.

There's a crucial difference though in that he argues that this effect will sustain the summer minimum well above zero and there won't be any dramatic collapse. I say rather that the effect only kicks in after a collapse. It's the "uncapping" of the open ocean that ramps up the heat loss during the re-freeze and provides the negative feedback governing the "tail" of the Gompertz/sigmoid curve. My envelope calculations on total volume lost each year suggest this will be enough to sustain area at ~1-2 million km^2.

Figure 1 from the Tietsche et al paper (looking specifically at the process of ice loss and feedbacks affecting regrowth) showed exactly this:


As ice is lost, there appear to be three "stable" regions where the curve levels out despite ever-increasing temperatures. These are at ~7 million, ~4.5 million and ~2 million square kilometers (at summer minimum), with the middle one of these being substantially less stable - i.e. only a temporary blip in the descent.

2007 marker the final transition from the upper "ledge" to the middle "ledge", and we've been in a holding pattern since then. We are currently sitting on that middle ledge, about to transition rapidly (over the course of a few years I suspect) to the lowest ledge at ~2 million. Then we'll sit there for perhaps a decade or so, before finally losing the last of the summer ice cover.

David Gould

That sounds reasonable. One of the things that I did when modelling was not allow the summer volume to drop below zero. This led to a linear trend in the maximum volume, which seems more realistic.

I would point out that the main thing that I am arguing is that for what we currently observe there is no need to speculate that feedbacks are somehow accelerating ice melt in some fashion.

I would suggest that at this point there is not much evidence that feedbacks, positive or negative, are occurring (or if they are all they are doing is making the linear temperature increase a steeper one and thus accelerating slightly the quadratic decline.)

Feedbacks, both positive and negative, will occur in the future. But there is not much to suggest that they are playing a major role in ice dynamics at present.


The question of which is more important is vexed. While a parliamentary committee is a more august body, my cynicism says that this hearing unlikely to achieve anything much - in any case it is premised on a strategy of despair. The BBC is important in a different way - in its reach and its likelihood of being recycled in the blogosphere.

I emailed the Beeb because Richard Black's reference to the PSC & PIOMAS contained two demonstrable errors of attribution. The original article contained the clear implication that this was PSC analysis and I felt it was important to correct that (which they have done).

I did not email the British Parliament for a number of reasons. I don't think their committee system invites general public input (expecially from non-UK residents). But critically, while Dr Schweiger might not have been thrilled with the spin put on PIOMAS data, Prof Wadhams has not erred as far as I can see. Like the Beeb, he may have been a little vague, but he did not imply that the graph of Larry's that he included, or the conclusions drawn from Wipneus's work were the out of the PSC. He clearly linked to this blog and Wipneus' website. Wadhams agrees with an exponential extrapolation but does not say that PSC produced that (as the BBC article heavily implied). Apart from some minor quibbling, I'm not sure what clarification would be of relevence...


This post edited to take out several direct links:

Hi Peter,

There's a lot in what you say that I completely agree with. Having said that, some of it might be quite sensible ;)

I would like to quibble only about the bit about refreezing: it seems to me that, as illustrated in these two area graphs...

[CT annual area graph -Laptev]

[CT annual area graph - East Siberian]

... the greatest annual area anomaly is now in the Autumn, caused by a delay in the freezing.

It seems to me that large enough open water areas in Arctic Autumn/Winter are capable of generating their own warm(er) atmosphere, due to increased heat flux from water to atmosphere than ice to atmosphere. This in turn can sustain the September period of open water well into October - giving us currently one month less of ice formation and thickening in the annual cycle.

If that "one month less" becomes "two months less" or "three months less" - well, you then have your ice-free December; and much less thick ice to melt through the following year. As perhaphs in...

[CT Annual area graph - Kara]

[CT Annual area graph - Barentsz]

Peter Ellis

idunno: Oh, for sure. The point is that let's say it takes three months for ice to thicken from the initial skim up to a layer ~1.5 m thick, and thereafter it more or less stops growing.

That means you make the same amount of ice in three months as you do in six months. Ergo, you get a plateau in the rate of decline until the length of the freezing season drops below three months. During that plateau, what's actually happening is that the freeze-up is occurring later and later, but still just about in time for the first-year ice to reach its destined thickness.

The final drop-off only happens once the freezing season has got so short that the first-year ice doesn't have time to fully thicken.


Hi Peter,

That's a good point, which I was not considering when I posted above.

It does look as if this year the ice in the Kara, Barentsz seas and part of the adjoining Arctic Basin has had less than 3 months of proper thickening.



Melt and freeze seem to have kicked upwards after 2007 but it is still far too early to tell which direction melt and freeze are now going.

[Melt = maximum daily - minimum daily PIOMAS volume in same year
Freeze = maximum daily - minimum daily PIOMAS volume in previous year]

data at


>"but he did not imply that the graph of Larry's that he included, or the conclusions drawn from Wipneus's work were the out of the PSC."...
"Apart from some minor quibbling, I'm not sure what clarification would be of relevence..."

The extrapolations are conditional "if one extrapolates...using exponetional" with the implication that submitters think those are appropriate extrapolations. So that is OK.

However they then conclude:
"Thus one has to conclude that, on current best evidence, there is a distinct possibility of a collapse in extent leaving relatively little ice this summer, and a collapse is likely by 2015."

If they had said "we conclude" that would be fine. However, as it is, the implication seems to me to be that any reasionable person would find it difficult to conclude otherwise. That is hardly the case if the head of PIOMAS disagrees.

Maybe that is over interpreting what they intended or is just a minor quibble. You could well be right.

I think misleading parliment should and is regarded as serious. Parliment should have access to good expert evidence in order to set government policy. I expect the people receiving the evidence to be pretty savvy and know this evidence is coming from a fringe group of experts. So hopefully they will be alert to the bias that is likely to be included and you are probably right that clarification of this isn't all that important.

What graph should be used as a reference for the 3B sentence certainly seems unimportant even if it is fairly obvious that it should be Wipneus's trnd1 not trnd2 graph. (Unless there is good reason to believe motive was deception rather than a slip.)

Kevin McKinney

crandles, note that they said, "a distinct possibility of a collapse."

That's where the caveat really resides.


"and a collapse is ***likely*** by 2015" or is that a distinct possibility of a collapse being likely? Hmmm., is that a contradiction of qualitative probabilities?


revised graph:

Chris Biscan

Big drop today on Bremen in the SOO.

SLP will crush/melt ice the next 24 hours before it heads to the Bering and actually rails it with a south wind for 3-5 days.

I would expect a 200-400K drop in the next week or so.


This is drifting, iceberg-like, further off topic, so I'll make this my last but wrt to crandles: "I think misleading parliment should and is regarded as serious".

Well, certainly it is for members of that parliament and their agents (eg government employees tendering evidence to enquiries). For free agents, I suspect not so much. You may recall the implications redolent in a certain famous opening: "I bring fraternal greetings from the mother of Parliaments..."


Chris Biscan


This heat wave over the CONUS/Southern Canada is by far the strongest on record for many places here, day to day, for the month.

Records are being completely shattered in ways we have never seen.

A cut off low is expected to bring us back to 15F+ anomalies from the widespread 25-40F+ anomalies.

Then we go back to 20-35F+ anomalies with 90s breaking out in the plains and 100s in Texas.

Insane heat.

Kevin McKinney

"Insane heat."

Yes. In the East it has been pretty dry, too, though with occasional dramatic outbreaks of strong thunderstorms, which have killed dozens and set some sort of record for March tornadoes, if I'm not mistaken.

A not-so-welcome corollary of this weather in Atlanta yesterday was a new all-time record for pollen count: over 9300! (500 is considered "extremely high.")

Clearly the flora are responding to the warmth... as is the human population, large numbers of whom are rubbing eyes, reaching for antihistamines, and staying indoors--though it's beautiful out, for the non-reactive among us.

Kevin McKinney

Further North, the outlook for insect pests is pondered:


Come to think of it, we had some annoying flies coming in from the back yard last night ourselves... probably not so typical of mid-March.

Timothy Hanes

Kevin, have you been getting this from people lately?
"Don't worry about why we're having all this great weather, just enjoy it!"
Sounds like this from a death row inmate, to me- " Sure, things could be better but their looking up! They tell me I get to eat whatever I want for dinner tonight, and the warden's gonna serve me himself!"

Bob Wallace

OK, I'm going to post another naive (possibly stupid) question....

It looks to me as if the 2011-12 "cold center" moved toward the Pacific and away from the Atlantic this year. Lots more ice than usual on the Pacific and lots less on the Atlantic side.

We got about the same extent/area as before (aside from the recent freezing blip) but the location changed.

Is it possible that the temperature differential between the La Nina-cooled Pacific and the every warming Atlantic drove this event?

Steve Bloom

Excellent question, Bob. One might also wonder about the downstream effect on the jet relative to the eastern NA blocking event. I'm afraid the answer is going to be that such a direct connection is over-simplifying something that's driven by multiple factors, but maybe I'll be surprised. If there's no one reading qualified to answer it, this question sounds like it would be up Trenberth's alley. Maybe just email him?

Re that blocking event, while its occurrence in early spring is not without other problems, its the prospect of a repeat in mid-summer (a la the western Russian region in 2010) that ought to be most riveting. I imagine that Chicago would see temps approaching or maybe even exceeding 110F under those circumstances. Being an optimist by nature, I can almost imagine that having to sit out such a heat wave for a week or two might focus a little attention on climate, but probably not. :(

Seriously on that last point, I think we're going to need repeated summer events in locations where it really matters.


The continuing Barentsz story...
Normally, heat would have been released in the mid-Barentsz Sea. This season, the upper layer seemed to cool in the far northern Kara Sea and NW of Franz Josef Land, in the Arctic Basin. The role of the deep Eurasian Abyss was mentioned by, I think, Kris. Deep water is formed through cooling of the upper layer. It descends and contributes to an anticlockwise gyre through the Basin. My guess remains that, while cooling wasn’t very efficient this season, a lot of heat content is driven right under the ice pack into the Atlantic Water Layer, normally between 200 and 600 m deep. It isn’t going that deep, this time. I suggested the effect of eddies, mixing up these waters to right under the ice. It sure looks like that, NE of Franz Josef Land.
MODIS shows a complicated, twisted structure of leads where both currents intertwine, where their influence should, normally, have been safely submerged.
The Kara Sea extends far north and east from the rather caged southern part. In comparison to Hudson Bay, the caged part is much more open from the north. You can see how a branch of the North N Zemlya current slips in. For the most part, currents in this cage shape up like a gyre, usually of cold nature.
Not last fourteen months. Through the Atlantic Multidecadal Oscillation, four years of exceptional sea ice loss and enduring insolation in 2011, SST’s were 3.3 dC higher than normal for the Kara Sea proper in september 2011. All that heat content has been replenished week after week last winter, until the upper layer had finally lost enough heat to freeze late February. A lot of heat was released to the troposphere (the ‘Kara Bulge’), but there was more than enough time and volume flux to force a considerable amount into the sub-ice Atlantic Layer (the Arctic Basin).

Bob Wallace

"Results show that a 1% decrease in sea ice concentration leads to a 0.36–0.47% increase in cloud cover, suggesting that a further decline in sea ice cover will result in an even cloudier Arctic."


Is this a mechanism which will slow melt over the next few years? Or trap heat at the end of the season? Or a bit of both/none?


Bob asked:

Is this a mechanism which will slow melt over the next few years? Or trap heat at the end of the season? Or a bit of both/none?

If it wasn't for the Gulf Stream we could assume melting could be slowed down.

But it is absolutely impossible to turn the blind eye onto the Gulfstream, isn't it? Denialist do, but that is their problem.

The more the ice is melting, the less the Gulfstream will be blocked, and the deeper the Gulfstream will be able to penetrate into the Artic.

IMHO the study has it right. But we shouldn't forget about another consequence: precipitation, precipitation in the form of rainfall. Basically, more clouds do trigger more precipiation. Elementary.

And, we better don't forget, rain consisted as the major factor in the 2007 record minimum.


As rain also was the cause of the absurd situation in the Barents- and Kara Seas early this year.

Moreover, this change is alreay happening and not restricted to the Arctic.

Due to the shifting of climate zones, since 2002 we get hotter and warmer Springs, Summers and Autumns around the Greenich lattitude. The Mediterranian band (Italy, Spain, South of France, New York, San Francisco ecc.... is not only becoming hotter but is getting more devastating storms too. And
that is not the future, it's already happening.


I better had published the direct link, well here it is:



And to quote the core of the testimony:

Towards the end of last week we reached our northernmost position at 88° 40N. Of course, we had expected that even here the ice would be as eroded and loose as in all other regions that we visited during the past weeks which has allowed us to maintain a speed of up to 6 kn. But a whole day of rain within 150 km of the North Pole came somewhat as a surprise! For the past few weeks, one low-pressure system after another has continuously carried warm air from northern Siberia (15°C at the Lena estuary!) towards the central Arctic Ocean. In this way the sea ice disintegrates more and more right before our eyes.

Steve Bloom

To what extent do the models take this melt mechanism into account?

Bob Wallace

Thanks. I'm not suspecting that increased cloud cover could stop the melt.

Just looking to see if there is anything emerging which might turn the trend from an "exponential crash" to a "Gompertz glide to the bottom".

Subjectively it feels like the accelerators are increasing in strength and the possible white knights are weak.


New March high temperature record for Norway 20,2 C previous record 19,8 C, we are moving up along the coast, towards you know?


Day 82
Today mean temps over 80 dN dipped under the 1958-2002 climate mean. For the first time since november. A prolonged period of more than 120 days has ended.
Tropospheric cooling has rounded up over the Arctic. It is a phase in continuing see-saw of wild swings over the northern hemisphere. Extreme jetstream configurations and blockades display confusing warming in one region, anomalous cold in another. They are features within the trend, which is always warming.
The 2010-like late freeze up on the Arctic fringes is as surprising as early heat over eastern US and Canada. Taking in more information, it becomes clear that on hemisphere scale, there is little surprise. Spring fits in ‘just fine’ on the trend. The ‘cold account’, through extra FYI, is possibly (think volume; can't be much)a bit higher than we were used to. OTOH, heat content in the deeper ocean layers is probably going to compensate enough in the coming months.


Espen - did you appear in a English documentary on the Arctic recently, by any chance? I saw an advertisement for a program the other day with a fellow that looks an awful lot like your avatar (assuming that's actually you).

BTW - does anyone know what became of the Artful Dodger? He hasn't posted for more than a month, which is an awful long time for the lodger to be quiet... :-/

Daniel Bailey

Good question; I sent him an email. His contributions are always enlightening (& fun).


The winds have turned again a bit, and it looks like Novaya Zemlya can be circumnavigated. :-)

I've also updated the Novaya Zemlya animation. The ice between Svalbard and FJL has grown, and south of FJL as well, maybe also influenced by the revived Beaufort Gyre.

The comments to this entry are closed.