« NSIDC Arctic sea ice news August 2011 | Main | Retrospective 3 »


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

Ian Allen

If people expect negative feedback to rescue the summer ice, maybe they should ponder that strength varies with -4th power of thickness for a floating slab. We seem to have about a quarter of the ice of 1979, if this occupies half the area then we have already have half thickness thus 16th strength of before. swells will grow as fetches grow in this round ocean, and eventually start prevent the thin skin reforming in the centre with its deep water.


Hi Neven,

(Sorry to keep so quiet. I've been offline)

In your last paragraph, that's September 2010 that is below the exponential downward trend. September 2011 is a projection and is thus bang on the line. I think.


You're right, Idunno. It's a recovery! :-P

Just kidding, I was looking too hastily to get the post out. Fixed now. Thanks.

Chris Biscan

If the euro is right we will see a record low Volume maybe under 4,000km3 and for sure an extent under 4,000,000km2.

We would have record September exent drop around 1,000,000km2.

This is not the time for a Severe compaction event and 14-16C 850 warm anomalies wide spread.


Idunno is correct: September 2011 will be below the trend if anymore melting will occur in the September month.

The new thing with this graph is of course the specified confidence which should be taken in to consideration when talking about projected ice free conditions.

Here is on for the August monthly average:


Indeed, Chris. I'm seeing it too. If what ECMWF projects at the moment comes about, we are in for a prolonged melting season, with some heavy melting to boot. I will have to revise my expectations. Exciting!

But let's discuss this further in the latest SIE update comment section. :-)

Wipneus, thanks again for your graphs!


31 Aug 2010 value was 4.646
31 Aug 2011 value is 4.275
So loss in the last year is only .371

At that rate it would take 9 years to get down to 1000 km^3.

Using one year that could be unusual like above is a bad way to extrapolate. 2007 was unusual and there was a 2 year period recovering so excluding those years

2010 decline 0.371
2005-6 min-min decline 0.166
2004-5 min-min decline 0.722
2003-4 min-min decline 0.359
2002-3 min-min decline 0.552
average decline 0.434
This gives a 7.5 years to reduce to 1000 km^3

A potential problem with this extrapolation is that is ignores possibility of a year like 2007 which might be a 1 in 20 year event so the chances of another such year occuring within 7.5 years is not very small.

Volume per piomas may not be accurate so +/- 2 years on these extrapolations is quite possible.

Another extrapolation is that in 2010 we lost .218 from 31 Aug to min (.192 in 2007) so 2011 minimum is likely to be around 4.06.

Lord Soth

I was talking yesterday to one of the support staff for the Louis which is in a joint survey with the Healy.

They occasionally hit a thick patch, but most of the ice is light in thickness. They are encountering thicker ice above 88 North now.

They like to run at 4 knots, as this is the optimal speed for surveying, so don't take the speed of the ship as an indication of ice thickness.

They are not going to the pole, as their is no interesting bathymetry there, however they may cross over to Lomonosov Ridge (but not at the pole) to do some extra work, as things are progressing so well, due to light ice thickness conditions.

This is in line with PIOMAS which gives an average ice thickness of 0.91 meters, with the thicker areas being cancelled by the slush puppy areas.

The Navy Arc Ice thickness chart more and more appears to be a work of fiction.

L. Hamilton

Uni Bremen SIE today stands less than 60k above the 2007 minimum.

Many keen eyes on this forum, it should be interesting what they see in the ice/water/air over the next few days.

Seke Rob

Horrible, and yet another record about to be broken, weeks earlier [than 2007].

Reading about Igor Semiletov, who per a post here seemingly rushed off from Vladivostok to see for the methane plumes, he wrote in March 2011:
“It is High Time to Warn People”: Igor Semiletov and the Methane Time-Bomb (Feedback, Part 2)
PETM 2.0 prologue?

Seke Rob

Sorry, that's about Semiletov's work. There's no fridge door we can close on that [clathrate] gun.

Kevin McKinney

Going back a link from Seke Rob's comment:


Trouble--but then, we already knew we've got that.

Kevin McKinney

Neven, I have to say I like L. Hamilton's graphs where you've moved them to--not only do they merit it, it gives us something to look at while CT loads!

Also--and speaking of CT--I note that SIA this morning was at 3.002 million km2. Looks to me that sub-3 is pretty much a given this year.

Ned Ward

It's absolutely worth keeping an eye on methane emissions from permafrost and the Arctic continental shelf.

But so far, at least, there's no indication that global methane levels are taking off (see, e.g., here).

There certainly appears to have been a slight increase in the atmospheric concentration, relative to a few years ago when it seemed to have stabilized. But so far there's nothing like the huge increases we saw in the 1970s and 1980s.

L. Hamilton

"Looks to me that sub-3 is pretty much a given this year."

CT has sort of hit that mark several times, falling minutely below 3.0 on 8/24-25 and again on 9/1. Although the 9/4 level is just 82k above 2007 minimum, that level hasn't changed much in the past two weeks.

Chris Biscan

jaxa had a near 12,000km2 upwards revision this morning.


North East Greenland:

We got a niece piece of ice leaving her dock at Joekelbugt she is aprox. 11000km2, more than 15 times bigger than Singapore, I am wondering how long she will survive?


Regards Espen

Bob Wallace

no one here is expecting the trend line to continue falling like that

I'm really interested in knowing why one would not expect the current observed trend to continue.

In my simple-minded way I see lower beginning season volumes paired with higher air temperatures continuing to drive the line downward.

I look at Ian's undulating ice comment (top of this discussion) and I see more and more major high altitude fires pouring more soot into the melt system.

I see the feedback of lessened albedo and earlier, more extensive exposed water to soak up heat for late season melting add their influence.

I see a heavy boot coming down on the remaining ice, especially as I see curves calculated in previous years being recalculated downward.

Then I see almost hopeful comments of possible negative feedbacks which could stop this plummet and I wonder what they might be. And I wonder why they have not yet shown their faces.

Can anyone point out a reason to hope for something than a total summer melt-out in the next few years? Describe a physical process which will protect the last bastions of ice?

Seke Rob

Re: L. Hamilton | September 07, 2011 at 16:05

And again below 3.0

4-Sep-2011 247 2011,6768 -1,7537520 3,0018246 4,7555766
5-Sep-2011 248 2011,6794 -1,7693100 2,9764333 4,7457433

(Read commas as decimal point).

L. Hamilton

Yes, I just updated the CT bar graph (though it looks just the same).

Honestly, I can draw other stuff besides bar graphs, but I like how they work for this 1-day minimum theme.

As a novelty I sketched yet another graph depicting the global anomaly:


Can anyone point out a reason to hope for something than a total summer melt-out in the next few years? Describe a physical process which will protect the last bastions of ice?

Bob, maybe I'm describing this all wrong, but there is something about large expanses of thin ice radiating a lot of heat out into the atmosphere. This cools the water off so much that it makes the ice thicker, or something like that. Of course, this might also screw up NH weather patterns, but hey, the sea ice is saved (for another decade).

Or clouds, maybe clouds, lots of clouds, due to the evaporation.

Or lake snow effect. Due to all that moisture in the atmosphere huge amounts of snow get dumped on the NH land masses, staying on the ground longer during Spring --> Albedo.

Sorry, when it comes to these things, I'm not much better than the average WUWTian. But at least I'm willing to learn!

Patrice Monroe Pustavrh

By being no better than average WUWT, I can also think of changes in ocean currents due to huge fresh water coming from Greenland. But, it is mere speculation.

However, according to Tietsche et. all and Kay et. all, ice decline is somehow "overtaking" "expected" loss. According to Maslowski, it is not.

Bob Wallace

Thanks, Neven, but let me play the graph's advocate. A version of the Devil's.

I don't quite understand the thin ice cooling water more concept, but for our purposes let's suppose I do. ;o)

Now, since we're approaching a 2014 September melt-out (I'm assuming a re-configuring of the black curve) shouldn't we see some of these negative feedbacks peeking in over the horizon? Wouldn't one expect deceleration rather than acceleration? The curves starting to be re-figured upwards rather than downward.

We've already stuck a lot more moisture into the atmosphere. Just look at our record floods and massive winter snows. That moisture should be increasing clouds and snowfalls in the Arctic already, should it not? (Perhaps it is there and slowing down what would otherwise be an even sooner melt.)

Now, I'm not predicting a total melt-out in a couple more years, but I'm looking at the curves and seeing Wile Coyote well over the cliff and plunging toward the desert floor. I can't imagine what might keep him from making a coyote-shaped hole in the sand.


Bob, what else can I say than repeat my motto: Nothing in the Arctic is a dead certainty. :-)

Maybe the oceans all of a sudden stop transporting so much heat towards the Arctic? Some flip in the PDO, AMO, etc? I've looked into that a bit and didn't find it convincing. But who knows?

I honestly do not know, but I wouldn't be surprised if something pops up and makes that last 3000 km3 of ice more difficult to disintegrate.

Bob Wallace

I haven't spent time on WUWT but my impression is that many people there grasp for every straw and inflate that straw into a bale in attempt to prove "the greenies" wrong.

Suggesting reasons why the ice might not melt does not put one in their camp. Let's toss all the ideas on the table and let the data show us the winners.

It's the scientist way. (For Riders in the Sky fans.)

Peter Ellis

Neven: Unless my eyes deceive me, then according to the penultimate graph, the August figure is indeed below the exponential trend, so you can reinstate your original closing sentence and point at that instead :-)


Espen I may be wrong but I think that it is the broken ice between that large mass and the mainland that is moving I think that large "island" is still landfast.


Ah well, just on it or just below it...

Since Wipneus made both graphs, maybe he can tell us which description fits better.


Hi there Neven, Bob, et al,

May I make an observation about your comments about possible impediments to the flush out of Arctic Sea Ice once it gets below a certain threshold?

My (albeit very limited) understanding is that there is a negative feedback mechanism linked to the increased oceanic heat loss out to the atmosphere (and hence eventually out to space) once the insulating layer of ice has gone bye-byes. (Think of igloos.)

Obviously this would be working in direct competition with the Ice-Albedo Effect (due to both greater ocean surface exposure AND surface melt ponds) and with other +ve feedback mechanisms such as fracturing events leading to increased surface area in contact with relatively warm waters.

It never fails to amaze me how much one hears nonsense in the deniosphere (and I'm sure Chris Biscan knows exactly where I'm talking about) about "positive feedbacks can't exist, otherwise we'd have a runaway greenhouse effect" (Even though there is no such thing as GHE, and even if there is - then it's very small, and even if it isn't - it's just natural cycles.)

Apart from the normal chaotic behaviour, the system will be driven by a myriad of feedbacks (in both directions). I'm sure that time will reveal certain feedback processes that are currently not even recognised, or whose importance is currently grossly underestimated.

After all, that's why modelled projections of Arctic Sea Ice behaviour from a few years back were so far out - the melt dynamics were simply not well-enough understood.


I am not sure that I understand what is asked exactly. I hope the data below is what you want.

Graph: https://sites.google.com/site/arctischepinguin/home/piomas/piomas-trnd6.png

The yearly minimum values, with a trend based on years 1979-2010 data points. For 2011, the minimum so far is plotted as well.
Next month, when the melting season should be over, the graph will be finalized with the 2011 figure.

Red dot, piomas latest value (2011/9/1): 4275 km3

A trend based on yearly minima 1979-2010, calculated for that day: 4156 km3

graph: https://sites.google.com/site/arctischepinguin/home/piomas/piomas-trnd8.png

The August monthly mean values, with a trend based on years 1979-2011 data points

piomas: 5050 km3
trend : 5439 km3

Bob Wallace

Thanks for the input Bill.

I'm not sure I understand this correctly...

...a negative feedback mechanism linked to the increased oceanic heat loss out to the atmosphere (and hence eventually out to space) once the insulating layer of ice has gone bye-byes. (Think of igloos.)

I can see that once the Sun has quit warming the exposed water then the water's heat will radiate out and refreezing could happen quicker than if there were an ice blanket.

But - if the water is exposed while the Sun is shining the water is going to be absorbing more heat.

I suppose someone can or might have calculated the difference between increased absorption and increased radiation. Intuitively I think increased heating will win over increased radiation.

Wish we had a guru to consult. Perhaps we could recruit someone from inside the science to be our Ask Dr. Science. A graduate student might be a good resource.

(I'm not sure igloos are a good example. I suspect they work by preventing air exchanges.)

Anyway, we're collecting some ideas. Increased radiation of heat from open water, changes in circulation patterns, increased snowfall, and increased cloud cover.

We are't seeing signs of them changing the rate of melt. If they're are going to save Santa's Ice Island they better kick in soon....

Steve Bloom

Bill, one thing to look forward to (as it were) is the effect of ice loss on the cold freshwater lens that keeps saltier, warmer water from reaching the surface. Without a steady input of fresh water from melting sea ice, how long will it last? There's also the runoff from the rivers and land ice, but the distribution of those inputs will be limited by currents. The first big storm to hit an ice-free central basin may be enough to disrupt the lens, at which point we may be looking at a long ice-free autumn. I'm sure ice will ultimately form in the winter darkness, but will it be so thin as to be vulnerable to a fast melt the following summer? The models would have to shed some light on this process, but I don't recall seeing a discussion of it. Anyone?


Wipneus, thanks for the short explanation. I get it now.

I've changed the text to: "Wipneus has one last graph that shows a little bit more clearly that the current number actually is right on the projected exponential downward trend (although the minimum probably hasn't been reached yet)"

Thanks, everyone, for helping create the perfect sentence. :-p

Bob Wallace

Steve, that freshwater lens...

I'm pulled back to the Boxall, 2007 study that found much more thermal exchange occurring than was expected. They found much colder, saltier water at 200 meters below the surface.

They stated "the older sea ice is, the less salt it contains". This sounds like what they were observing is the melting and sinking of first year ice.

As we loose multi-year ice it would seem that we would loose any 'just under the ice' fresher water protective layer. First year ice isn't fresh water (accumulated snow, I would assume).

As for annual freeze and melt, unless we see an increase in mutli-year ice over years then one must assume that with increased warming any refreeze following a total meltout is going to melt most summers.

Bob Wallace

Wipneus/Neven -

Would I be far off the mark if I read the different values for "Red dot, piomas latest value (2011/9/1" vs. "August monthly mean values" as the red dot happening to show a high point day in a noisy distribution with the likelihood that that one data point will be averaged out to a lower September average?


Oops, I didn't proof read my earlier post correctly. It's annoying when your fingers don't type what your brain is thinking.

When I talked about ... a negative feedback mechanism linked to the increased oceanic heat loss out to the atmosphere (and hence eventually out to space) once the insulating layer of ice has gone bye-byes. (Think of igloos.)

I was thinking of late in the melt season when the Ice-Albedo Effect is phasing out for the year. However, as just demonstrated, meaning to type something isn't the same as actually doing it. Sorry about the confusion - welcome to my world.

The point I was trying - and failing - to make was that late melt-season radiative/convective loss to the atmosphere should act as a -ve feedback mechanism that would serve to postpone (for a limited period) the unwelcome prospect of a seasonally ice-free Arctic.

There was a study conducted earlier this year (I'll try to find the link) which concluded that a "permanent" tipping point was unlikely. If I understood it correctly (always a big "if") the scientists in question thought that the Arctic could quickly - i.e. within a few years - return to the levels of a few decades back. However, I believe this was really dependent upon atmospheric compositions reverting to pre-industrial levels. (Not something I am ever expecting to see again.)

Interestingly, on November 17th 2010, Rear Admiral David Titley, the chief Oceanographer of the US Navy saw fit to warn the House Committee on Science and Technology about the need to readdress naval readiness in the light of imminent (10 – 20 years hence) periods of effectively ice free conditions in the Arctic. It appears that the science community within the US Navy also has grave reservations about the accuracy of the claims pertaining to any supposed "recovery" of Arctic Sea Ice.

Admiral Titley’s testimony to the House was a mere 8 pages in length, and was obviously tailored to be understandable even to those Representatives on the committee that were lacking a scientific background. Members of the public so taken in by the “less than scientifically robust” musings being spread on certain blogs could do far worse than to read Admiral Titley’s paper.



Bob Wallace, thanks for putting that into words. It is what confused me (entirely my own fault, I'm bad at reading graphs).

Steve Bloom

Bill, it's my impression also that all hope of recovery depends on a lessening of the forcing. I'm doubtful about the recent one postulating that there could be substantial decadal recoveries on the way to an ice-free summer period.

Bob, AIUI as sea ice ages it becomes fresher through a process of ejecting salt, becoming fresh enough to drink after several years. But then when it actually melts (at any age) there will be fresh(er) water in its place. IOW these are different things.

I'll look up the study you mention. Of course there's cold water under the warm(er) water under the cold fresh water, but 200 meters sounds a bit deep for fast mixing.

Chris Reynolds

Bob Wallace,

At the risk of sounding like a stuck record, this is an issue I've been trying to start discussion on, I agree that we can only make headway by discussing this issue, which I see as vitally important. So let me put an opposite view. I don't say the opposite view, this is merely my own amateur musings.

Firstly; PIOMAS shows substantial losses in recent years. Whilst one might argue about the actual numbers I'm fairly confident that the overall picture painted by PIOMAS is correct. Here (figure 5) I show that from PIOMAS the greatest losses have been at the start of the melt season, not in the late melt period (when there's a lot of open water). I am at a loss as to why this is happening. However it does suggest that the sea-ice/ocean albedo effect is not responsible.

Secondly two points with regards the extrapolation by trend implied by Wypneus' graph (IIRC Wipenus has cautioned against simple extrapolation of trends). PIOMAS projections of future sea-ice do not suggest an early transition (this decade) to a seasonally sea-ice free state, see here. So if it's reasonable to accept what PIOMAS shows with regards volume trends why is it not reasonable to accept what it shows for the future? And if one accepts the extrapolation of volume's implications for summer, why not for winter - i.e. a year-round sea ice free state by the second quarter of this century. The physics models in PIOMAS that are (IMO) serving us so well tell us that other factors will come into play to reduce the rate of volume loss.

And my third and final opening point. We can be fairly certain that the volume loss is mainly happening at the start of the melt season. So what is happening to the energy that's being gained throughout the rest of the season when energy gain should be at it's maximum? - either due to high insolation, or later due to large areas of open water. In my opinion the answer is that since 2007 the 'Tietsche effect' (as outlined in that paper) of large losses of heat to the atmosphere from the ocean due to anomalously large areas of open water in the autumn is active now. I've explained in more detail here.

2007 saw a loss of 1/4 of the sea-ice area, that has been maintained, we have not seen further losses, this is despite substantial gains of energy. We know that the volume loss from PIOMAS is not directly due to energy gains in the period June to September. Something must be happening to the energy gained each year that explains the relative stability of extent/area since 2007's massive crash. In my point of view Tietsche et al tells us what is going on - losses of energy in the autumn are acting to partially decouple each year's energy gain.

Bob Wallace

That "could quickly - i.e. within a few years - return to the levels of a few decades back" paper struck me as someone needing to add another count to the publication list on their resume.

I got nothing from it other than if conditions were to reverse (with no arguments why they might) then we would see more freezing (duh!). It seemed that they plugged the most extreme values into their model and came up with a very small probability event and pumped it into a publication.

And the paper weaseled out at the end saying that any temporary decrease in freezing would likely be only temporary as the globe continued to heat.

Any of us who have looked at the graphs at all understand that there is noise in the system and there is no steady march lower each successive year.

I, for one, would not at all be surprised to see a melted out September followed by one or more ice-present Septembers. Variation exists.

A return to 1970s, 1980s ice? Count me as one who would be amazed....


Bob, he lower value for he red dot simply reflects he melting that occurred after mid August. Next month we Will know he final value, which must be he same or lower.

Bob Wallace

Thanks, Chris. I'm going to have to chew on all that for a while as my head can't seem to get out of second gear this morning.

One point, volume is dropping every year. We are starting with less ice almost every beginning of the melt. Incoming heat, we are told, is increasing. Without mitigating factors one would expect greater-than-straight line losses.

That early season rapid melt might simply be a confuser. Not knowing how ice melts, let me speculate that ice might tend to hold together until a certain amount of heat is absorbed. Then it rapidly changes state. (A tipping point, if you will.)

Something like that could explain the observed early season higher volume loss. Then after lots of the ice has achieved slush puppy/rotten ice status more orderly melting dominates.


Like you I'm a musing amateur. And in discussions I sometimes find myself taking a side, not because I hold firm beliefs, but because I want to see the issues on the table. And, gosh, someones got to open their pie hole if it's going to happen.

I'm very quick to drop a hypothesis once data shows it faulty. (But I'm sometimes slow to understand the data. ;o)

Peter Ellis

Chris R: We can be fairly certain that the volume loss is mainly happening at the start of the melt season.

Not quite. Look at the third graph down: melt follows a distorted sinusoid pattern this year and every year.

The early-season dip shows up predominantly in the anomaly graph. What this means is that volume loss is following a normal seasonal pattern, but shifted a few weeks earlier than the historical average. Thus there is a big drop in the anomaly at the new (early) start of the melt season, and the anomaly then holds more or less constant until the freeze-up.

A second-order effect is the "rebound" we've seen in the last couple of years after the initial anomaly drop each year. This seems to correspond to the time Hudson Bay ice fully melts out each year, so it may simply indicate that Hudson Bay is a large volume of ice that has now transitioned to a new seasonal norm of melting out several weeks ahead of schedule.

Patrice Monroe Pustavrh

Peter, I make some objections on Hudson Bay melting faster than usually. Actually, this year, the melt in Kara sea was the place which melted really, really fast (HB melted fast too). Otherwise, I quite agree with you.

Bob Wallace

Good point Peter. The anomaly graph simply says that the rate of melt changed somewhere for some reason.

If we hit a critical volume/starting heat for ice in lower latitudes then we would see a temporary change followed by a 'new normal'.

We'd probably need more localized melt data to downgrade the sea-ice/ocean albedo effect. The SI/OA effect could be alive and kicking, just kicking in at different times in different places.


Hi Bob,

As regards the "no tipping point" paper I mentioned, I do tend to agree with your assessment.

My purpose in mentioning it was to further reinforce the lack of understanding as to the processes actually occurring in front of our eyes. On the other hand, the results of these process are pretty self-evident - worryingly so.

Chris Reynolds


I said I was clueless about those periods of ice volume loss, well I have an idea but not a firm one.

It is interesting that these sea-ice volume losses happen around the time of maximum insolation. The recent 2 springs have been marked, but it's seen to be a pattern in previous years. Going back further in the past 1993 shows this pattern, with a spring melt as great as 2010/11, but the years around don't show it.

Gerald Roe did a paper in which he showed that the rate of change of ice sheets during the ice ages correlates well with insolation changes at 65degN. This also brings to mind Hansen's "Trace Gasses and Climate Change" in which it's stated:

"We suggest, however, that spring is the critical season for terminations, because the albedo feedback works via the large change in absorbed sunlight that begins once
the ice/snow surface becomes wet, after which the surface albedo remains low until
thick fresh snow accumulates. A spring maximum of insolation anomaly pushes the
first melt earlier in the year, without comparable shortening of autumn melt, thus
abetting ice sheet disintegration."

Now that's about ice-sheet disintegration not sea-ice, but it's at base level a similar process.

Actually when I write this suspicion down it seems rather more convincing. Maybe in the post 2007 with a younger thinner ice cover insolation driven melting is becoming a much greater driver, possibly implying the sort of rapid response Hansen is concerned about.

I don't want to be confusing by posting things that appear to argue against my position. I have a tendency to do that because reality is more important than my 'position'.

Bob Wallace

Chris, post both sides. Otherwise we become the people we don't like - those people 'over there'.

Bob Wallace

OK, is this saying that in the 'big dip' years there was a lot of ice just waiting to be melted?

Then with the big insolation onset this thin ice quickly disappears.

The emergence of big dip years came with the disappearance of multi-year ice and, possibly, thinner-than-usual first year ice.


Espen That big island may have snagged but it looks like the whole top corner is shifting.
Bob et. al. I'm thinking that once the early seaon dip happens the fresher top layer establishes which then oscillates between freezing cold and frozen soaking up heat every time it defrosts then the evaporation into the warmer air cools it til it freezes again.
As far as total ice loss goes my guess is that there has to be a 'cold dump' established, just somewhere that tropical heat is moved to. The most sheltered spot for it is to the west of greenland so whilst i anticipate an ice free arctic basin i expect prodigious snowfall between greenland and the great lakes to create the kind of frozen desert that the 'cold dump' needs to be. That means eastern greenland would have warm[ish] currents carrying much of the the excess heat to a more temperate arctic basin,and remaining ice free. The warm arctic ocean is going to evaporate rapidly into the arctic night and this may prove sufficient to accelerate the atlantic drift.
Of course all the lands not protected by northern mountains can expect much more snow, as can those northern mountains. Before this, or an alternative, settles i think erratic may not be a strong enough word to describe the weather.

Peter Ellis

Patrice: I agree that Hudson Bay didn't melt out as early as last year - and the downward spike in PIOMAS anomaly wasn't quite as pronounced as last year. That's one reason I suspect that this particular spike is governed by Hudson Bay dynamics.

Note also that the Kara Sea is about 0.8 million km^2 (going by Cryosphere Today) and loses about 0.6 million of that in the first two months of its melt. Hudson Bay is 1.2 million km^2 by the same metric, and lost nearly a million of that in the first two months of its melt. Thus, even given the less severe Hudson Bay melt this year, it still contributed almost twice as much to any potential anomaly spike.

Hudson Bay is huge, and because of its enclosed nature, melts rapidly across its entire area once breakup starts.


Chris R,

I don't see the problem with 2007 that you do. A certain amount of melt each year will be caused by random weather patterns. 2007 was a truly exceptional weather year.

When you smooth out the random weather variations on the volume/area/extent curves they all show a steady pattern of accelerating decline.

Now I'm not saying that random weather/simple trend lines necessarily explains everything, but it seems close enough to what you would expect that there doesn't seem to be a huge need to invoke some unknown third factor.

Kevin O'Neill

Chris R says:

2007 saw a loss of 1/4 of the sea-ice area, that has been maintained, we have not seen further losses, this is despite substantial gains of energy.

Chris, if one only looks at a 2-dimensional map, this is true. But if one takes into account the ice has 3-dimensions, then we've seen major losses since 2007; volume has decreased by 40%. It takes a lot of energy to melt all that ice.

Extent this year central basin bordering the Beaufort-Chukchi-East Siberian Seas was prolonged and the ice never melted out the way we expected it to. Yet those areas had large amounts of MYI. The volume losses had to be tremendous.

When I first read Tietsche et al, I thought, "Well, I guess we won't see a quick melt out after all." But another melt season has come and (almost) gone and we're still seeing those absurd trends continuing. Maslowski's 2016 +/- 3 prediction is one year closer to fruition.

After reading Rampal et al, I realized that Tietsche's paper relied entirely on the GCMs available and, as Rampal points out, they're horrible at 'doing' sea ice. We know these models have significantly missed the mark - and then we turn around and quietly accept the Tietsche results that are entirely based on these same models.

I do not doubt the negative feedback effects that Tietsche describes. These are physical processes fairly well understood. 2008 did see a 'recovery' in extent/area - but not in volume. That's not a recovery, just the appearance of one. The negative feedbacks are real - but they haven't been enough to reverse (or really even pause) the trends.

Andrew Xnn

2008 and 2009 had greater volume than 2007 (at least most of the time). So, there was something of a recovery.

It just wasn't sustainable.

Kevin O'Neill

Andrew, you're right. I meant that 2008 and 2009 lost more ice than was lost in 2007. Over the entire period since 2007 we've seen volume decrease by 40%.

If 2012 starts the melt season with more than 25kkm^3, then we may have reached an oscillating stage with the MINs and MAXs determined by overall temps. If it starts with 23kkm^3 or less - then I think Maslowski will be proven correct. Between 23 and 25 we'll have to wait and see how the 2012 melt progresses.


You stated that volume recovered in 2008 and 2009. If you calculate thickness using either extent or area to divide by (drum roll) ice thickness decreased both of those years.

Andrew Xnn

In the winter of 2007 to 2008, an exceptional (record) amount of ice formed in the Arctic. The reason is that there was an exceptional amount of open water. This is the feedback that may help stabilize the ice.


Sorry I just double checked ice thickness was very slightly higher in 2008. Since then it has gone down very fast.


The problem I have with the Tietsche et al, paper is that if you read it carefully you will see that they spend time defending that one GCM is pretty good and only slightly overestimates the ice but they didn't use that GCM they used a low resolution GCM and there is nothing in the paper saying how the GCM they used performs.

Kevin O'Neill

Andrew - despite those feedbacks volume has fallen 40% since 2007. I agree there are negative feedbacks in play - but they haven't been enough to stop the decline.

Remember, 2006 ended the melt season with nearly 9kkm^3 of ice. The 'recovery' years never got higher than 7.1 -- that's not recovery, that's treading water, albeit frozen water :)


Yet another amateur musing. The mere, obvious fact that volume is so steadfastly sticking to its steep decline indicates that the major influence is _something_ that constantly melts ice.

May I introduce ocean heat content as the leading contender in the lineup of suspects. If we acknowledge that 90% of heating has gone into the oceans, the fact that currents may vary, insolation varies, storms, winds and other weather varies from year to year - it makes no difference to the fact that beneath the ice, water is constantly, all day every day, warmer than it was 10, 20, 30 years ago.

I look at the ice maps because they're fascinating. But the one to watch is the SST anomaly map. And that's just the surface.

What records and maps do we have for water temperature beneath the ice for the last 30-60 years? Anything easily accessible and useful for graphing or mapping? I've not seen one.

Bob Wallace

Good question about water temp. One I've wondered about myself but only found anomaly maps which don't tell history.

Among driving forces, I would suggest people give serious attention to Mark Jacobson's piece on soot. I've followed Mark's work in renewable energy for some years and have found him to do insightful, high quality stuff.

Daniel Bailey

@ Ned Ward

Earlier you made mention about atmospheric methane levels:

    "But so far there's nothing like the huge increases we saw in the 1970s and 1980s."

Out of curiosity, what is your source for that statement?

Ethan O'Connor

Daniel --

Take a look at http://www.esrl.noaa.gov/gmd/webdata/iadv/ccgg/graphs/ccgg.MLO.ch4.1.none.discrete.all.png

The Mauna Loa data only goes back to 1983, but you can see CH4 levels rose from about 1.63ppm to about 1.76ppm in the decade from '83 to '93, but flatlined or even decreased slightly from '99 to '06. Since then, however, the rate of increase has returned to just below that of the 80s. To what degree this is attributable to a clathrate gun scenario is unclear to me.

Daniel Bailey

Thank you for pointing that out, Ethan.

However, Ned's linked graphic went back to 1985, but his statement covered the 70s and the 80s. It's that 15 year window of data he refers to that I'm interested in.

Timothy Chase

Daniel, I believe the following page might have what you are looking for, 1970-2005...

Methane (includes major trends in CH4 emissions from 1970-2005)
E(mission) D(atabase for) G(lobal) A(tmospheric) R(esearch)

Ethan O'Connor

Timothy, I believe that the page you linked only contains information about emissions, and not concentration/mole fraction. I have found reference to Kitt Peak spectrographic estimation of column methane starting in 1977, but I don't know of any continuous time series measurements of ground-level methane from that time period. There were individual measurement campaigns carried out, however, but I believe most of the data for those is to be found in published papers and not in most of the online carbon cycle data repositories.

Law Dome ice core and "CH4 in archive tanks from Cape Grim Station, Tasmania" data are however in strong agreement and highly supportive of the assumption that the 70s had a high growth rate:

CH4 in archive tanks from Cape Grim Station, Tasmania, deseasonalised:
(year A.D.) Ratio (ppb)
1978.3 1468.3
1979.1 1490.9
1980.7 1507.7
1984.4 1572.9

Law Dome:

1965 1261.3
1967 1285.2
1967 1278.4
1967 1288.5
1970 1351.7
1971 1357.2
1972 1380.3
1972 1379.9
1973 1385.8
1974 1421.5
1975 1420.4
1977 1458.1
1980 1479.7
1980 1479.3


Seke Rob

AGGI (Atmospheric Greenhouse Gas Index) will update soon I think, to add the 2010 data.

(right click on graph and select view image to see in full)

Since 1979, there's 1.0 Watts/M^2 added to the system, which already exceeds the mean variation of the sun which swings in the [smoothed] range of 1.2-1.4 Watts/M^2 between minimum and maximum i.e 0.6-0.7 Watts up or down. Agree with previous posters that OHC is the one to watch, each year, even in the Arctic, a little more staying behind, and a little bit more coming in via the ocean currents. Gulf of Mexico was record matching hot this year. How long for the stream to reach Barents and up... 3-4 months? It goes at 4 miles per hours average, distance to travel?

Alan Clark

An observation about the last of the graphs at the top (minimum ice volume): in recent years the points are mainly within 1 Standard Deviation of the trend line, and never much more. However, in the earlier years there are many points which are well over 1 SD from the trend line.
This means that the variation is getting smaller in absolute terms but not so much in percentage terms (unsurprisingly). So if the trend continues then zero will almost certainly be reached in 2014 or 2015, and the range of possibilities is much less than the graph suggests.

Daniel Bailey

Thank you, Timothy, Ethan and Seke Rob for your efforts. I remain interested, however, in the source for Ned Ward's statement about:

    "But so far there's nothing like the huge increases we saw in the 1970s and 1980s."
I resume my patient vigil for a reply...

Artful Dodger

Chris R | September 07, 2011 at 20:28:

"... the greatest losses have been at the start of the melt season... I am at a loss as to why this is happening. However it does suggest that the sea-ice/ocean albedo effect is not responsible."

Not true, Chris! First year sea ice has an albedo around 0.6 while multiyear ice is around 0.8 Thinner ice of the same age also has a lower albedo.

So the proportion of first year ice (over 67% in April 2011) significantly affects albedo, especially at lower latitudes where the Sun first heats the Arctic in Spring.

Do the math, mate. It works out very closely!


Remko Kampen

"By the time volume really gets low some negative feedback might kick in and slow things down."

Whence this speculation? I can only think of two positive feedbacks that will quicken the process to a summer zero:
- Vast albedo decline, a summer sun warming up the arctic sea pronto for up to a half year per year;
- Some more water vapour in the air over the Arctic Sea with some extra greenhouse effect.

Ned Ward

Daniel Bailey writes: I remain interested, however, in the source for Ned Ward's statement about:

"But so far there's nothing like the huge increases we saw in the 1970s and 1980s."

I resume my patient vigil for a reply...

Thanks for your patience.

Very long-term data on methane are only available from ice cores. From these it's clear that methane concentrations increased even more rapidly than CO2 in the years after 1750, more than doubling from their preindustrial value of around 750 ppb.

In the 1960s and 1970s, there were sporadic measurements of methane published in the literature. I don't think there are digital data online anywhere, but you can find them in older papers. A nice compilation is provided in Figure 4 from the following:

Rasumussen, R.A. and M.A.K. Kahlil. 1981. Atmospheric Methane (CH4): Trends and Seasonal Cycles. J. Geophys. Res., 86(C10): 9826-9832.

A scanned copy of Fig. 4 is available here. Based on these data, Rasmussen and Khalil calculated an average rate of increase around 1.7% per year over the period 1960 to 1980.

That was a pretty alarming rate of increase! But through the 1980s and into the early 1990s it gradually declined. From the early 1980s onward you can find digital data from the World Data Center for Greenhouse Gases. I randomly picked Point Barrow as an example (hey, we like the Arctic here...) and by the late 1980s the increase had dropped to around 0.6% per year.

By the mid-1990s it was ca. 0.2% per year, and in the early 2000s it was under 0.1% per year.

It's a bit worrisome that CH4 levels appear to have begun increasing again. In the past couple of years, the rate has averaged around 0.3% per year.

As for the "clathrate gun" hypothesis, I agree that anecdotal evidence of new methane plumes from the Arctic Ocean seafloor is worth paying close attention to. But in terms of the net effect on the atmosphere, the rise today is, as I said, still quite a bit smaller than the rate of increase 30 to 50 years ago.

Of course, it's entirely possible that some other negative feedback has been stripping out CH4 from the atmosphere and thus (temporarily?) mitigating the rise from a new clathrate source. In that case, perhaps the post-2006 rise is a sign that this hypothetical negative feedback is being overwhelmed, causing the net flux to increase again. This is all obviously just speculation, however.

Disclaimer: I'm not an atmospheric chemist, and I don't even play one on TV.

Daniel Bailey

@ Ned

Thanks for the information; it's in-line with what I've read elsewhere.

Evidence for the Clathrate-Gun exists elsewhere than the Arctic (the biggest evidence lies in the southern hemisphere); discussion here:

Disclaimer: I co-wrote the article, but am not an atmospheric chemist either. But I have appeared on TV and have stayed in a Holiday Inn Express...

Ned Ward

Back when I was still involved in Skeptical Science, I thought about proposing a blog post on atmospheric methane. What they have over there is a bit scattered and some of it is outdated (e.g., this).

There is a nice discussion of the recent increase in methane concentration at Tamino's website.

[Ha! I just noticed that Daniel himself actually linked to Tamino's post, while wearing his moderator's hat in the thread at Skeptical Science.... What a tangled web the internet is.]

Ned Ward

Yikes, talk about comments crossing in a most inopportune way!

Sorry I hadn't seen your post "Waking the Kraken". I missed it when I was searching the site (... the connection between Krakens and methane is not intuitive ...)

In that case, I take back what I said about a lack of up to date information at SkS. I'll read your post there right now.

Ned Ward

OK, that's a great post about clathrates, Daniel. Very well done.


There is one rule on this blog and that is not to link to the Yooper's scary Kraken-thing. ;-) :-(

Ned Ward

So, basically, you're telling me I just broke 100% of the blog's rules in one comment?



Yes, this is what I'm telling you. You're just lucky that there are no repercussions for breaking the rules. Not for you, that is. The combination of methane and Kraken makes me go whimpering under the table for five minutes. Like any good alarmist should!

I had a comic book with a Kraken in it when I was a kid. That's why it scares me extra.

Seke Rob

"Sphere" (I'll not link to a pic of Samuel L. Jackson reading thé book, and the consequences for others)

Chris Reynolds

Hello D,

I see 2007 as more than just 'weather' it was an outcome of an intense +ve index Arctic Dipole event. I agree with the acceleration of loss, the question that interests me is what happens next, a continuing acceleration until sea-ice free in summer, or a tempering of the loss. I don't see an unknown factor, I see reason to doubt the energy gains of each year post 2007 are being carried over to subsequent years.

Kevin O'Neill,
(and those who don't like models...)

I know it takes a lot of energy, I've calculated the energy implications of the PIOMAS loss of volume on my blog (see Playing around with PIOMAS). I've been trying to calculate the energy gains implied by the maintenance of the 1/4 loss of area but can't come up with a satisfyingly robust method. Obviously that woud help compare the energy going into melt with that gained by late season open water.

Tietsche's key result does not depend upon modelling sea-ice trends. I agree that the overall trend of loss in Tietsche et al is way too long - I think next decade for the first sea-ice free minima. However the loss of heat to the atmosphere is basic physics and is already being observed e.g. Overland lower tropospheric warming - see here, that's from NCEP/NCAR reanalysis.

Consider the observation that most of the PIOMAS volume loss is actually from the early part of the melt season, not the late after all the water is exposed. So whilst a lot of energy is going into volume loss at the start of the season it isn't going into volume loss later in the season. So what is happening to that energy? If it were staying in the ice-ocean system why have there not been subsequent crashes?

Tietsche et al merely offers a theoretical framework to see the power of losing this energy. The autumn open ocean and NCEP/NCAR warming profile is evidence showing that the 'Tietsche effect' is real and is happening now.


So whilst a lot of energy is going into volume loss at the start of the season it isn't going into volume loss later in the season. So what is happening to that energy?

Maybe it has a harder time following the ice as it retreats to higher latitudes?

Chris Reynolds


Increases in ocean heat fluxes along with mixing of warm surface water from the open areas of water is a significant issue. In August NSIDC sea ice News there was a graphic of top and bottom melt, showing that for July most of the melt was surface melt, not bottom melt. However this can change. In the last two years there have been substantial losses in volume (PIOMAS) during April to June. I have been unable to find information about top and bottom melt over this period specifically.


I said that the "sea-ice/ocean albedo effect" (an 80% gain) is not responsible, the extent of open water from April to June (when the volume loss occurs) shows that statement to be correct. I am aware of the impact of the change from MY to FY ice (e.g.) and it's impact on albedo. But this does not explain the substantial early melt season PIOMAS volume losses in 2010/11.

The years prior to 2010/11 do show an early season volume loss that isn't as great as the last 2 years, although going back to the '80s and '90s and this is only an occasional feature, suggesting in the past it was due to particular weather. Whereas now it is due to some other factor. Changes in sea-ice albedo in the period around maximum insolation is as good a candidate as any.

Bob Wallace

Here's my topic for one of this winter's webinars...

"A description of how ice melts at different points in the season. With attention paid to location and thickness."

Now if we could find someone to give that one....

Aaron Lewis

I suggest that some heat is going to melt out channels inside the ice, i.e., the formation weak or "slush puppy ice" as described by Dr. Barber, and as indicated in the flash melt events of last month. At this point, I do not think much of the ice is solid. I expect that much of it is mostly granules weakly bonded together, with a significant void volume. Having significant void volume would affect the ice's mechanical strength, its thermal conductivity, and its transmittance.

In the old days when the ice was solid, the ice would raft up, and we would lose ice extent and gain ice thickness. Now, the puppy slush just "flash melts" when there is a storm.

Does anyone know how PIOMAS deals with void volume within the ice?

And, as I look at AMSU-A surface brightness, I think some heat goes to vaporize water.

Kevin O'Neill

Aaron, good points. Despite having read Rampal I completely neglected to take into account the (likely) weakened mechanical strength. We could see more compaction in the Beaufort area than I initially thought.

Artful Dodger

Chris R: Okay. Show us your math.

Yooper: Read up on JAXA's GOSAT: (Greenhouse Gases Observing Satellite)
Launched in 2009, provides regional CO2 and CH4 measurements, including the Arctic


Peter Ellis

Consider the observation that most of the PIOMAS volume loss is actually from the early part of the melt season, not the late after all the water is exposed.

This just isn't the case. The anomaly spikes in the early melt season, but that's not the same thing. Look at the actual quasi-sinusoidal melt graph: the vast majority of the drop occurs in June and July, the height of the melt season.

Looking at the graph with all the years superimposed shows this very nicely. 2010 and 2011 are both way below all the others - but note also that the difference between them (i.e. the anomaly) increases rapidly in May. This is because previous years showed little volume loss in May, while both 2010 and 2011 had significant May losses (though not nearly as much as in June or July.

As I've said before, I'm sure that the dominant factor in this is the timing of Hudson Bay melt - it is a huge area of ice at comparatively low latitudes that tends to melt out in one "spike" - perhaps because it's enclosed and once the melt is under way it can't be "topped up" by ice advected from other regions.

David Gould

Re no-one thinking that the melt will continue parabolically, I do - to a point.

The reason I think that it will continue is this: if the amount of ice melted each year increases linearly while the amount of ice the freezes each year remains constant, the plotted graph is an parabola (2nd order polynomial).

And there is a reason to think that ice melt will increase linearly: the temperature rise is basically linear.

At what point will this change? My guess is that it will change when there is basically open water (in other words, zero ice)at the end of summer. Heat will escape, meaning that the temperature rise will slow, changing the dynamic.


Alan Clark: This means that the variation is getting smaller (...) So if the trend continues then zero will almost certainly be reached in 2014 or 2015, and the range of possibilities is much less than the graph suggests.

Not quite. This range does depend on 1) the variation in piomas data as you mentioned, but also 2) on the range of curves that are consistent with the data to this day.

The second uncertainty dominates the range of possible timings for "zero ice" and is the cause of the rather wide range of the confidence range on the high side.

When the 2011 data can be included next month, this range will be smaller though.

(here totally ignoring the possibility that a future departure from the current exponential trend can not be ruled out)

Alan Clark

Wipneus: If you did the calculations with only the data from, say, 2000 I think you would find that the standard deviation is less (although the set is now rather small). The year 2007 is supposed to be an exceptional year, and yet it is only ~1.5 SD below the trend, which indicates that there is room for improvement in the way the algorithm calculates the variability. Probably not worth worrying about it too much though!

Kevin McKinney

"Re no-one thinking that the melt will continue parabolically, I do - to a point."

Me, too, though that's strictly FWIW, as I have zero quantitative analysis to justify that expectation.

But I'm thinking more and more that as thickness continues to drop--and I don't see any reason it won't--the ice will be more and more subject to the heat-transferring effects of increased mixing. The most open areas of water will be most vulnerable, I think--which intuition makes me the anti-Will Crump, I suppose.

Under this scenario, the 'point' where the dynamic shifts again is the point where ice that is sheltered in some way by land dominates the population. Deep North-facing fjords will stay iced for a long time to come, I think.

But as I say, it's all just FWIW.

Kevin McKinney

Oh, forgot to add my appreciation for Peter Ellis pointing out the difference between anomaly and actual decline as regards seasonal decline.

The CT regional graphs tend to illustrate this interesting relationship nicely; whenever a sea melts out completely 'ahead of schedule' you see a 'tail' during which the extent falls to zero as the anomaly *rises* to zero in a sort of distorted mirror image.

And yes, that's clearly visible in the Hudson Bay graph:


Kevin McKinney

Come to think of it, one of the most striking visual illustrations of how far we've fallen from the ice climatology is the fact that the proportion of time spent by the curve in any basin is so small. Or it would be, had we not come by now to take it utterly for granted.

There's no reason in principle why we couldn't see quite a lot of time 'above the line', even though there are hard upper limits to the actual extent for many of the regions. If we were in a cooling regime, you'd get earlier freezeups and later melts, with consequent positive peaks at each time of year, fading to zero anomalies as complete freezeup was reached. So in a way, the sorts of shapes we see are a sort of diagnostic, too.


Too bad you have to understand the relationship before the striking aspect becomes, well, striking.

Peter Ellis

Wipneus: Can you confirm what's going on with the "spaghetti" graph with all years superimposed? The end-of-year value for any given year should be the same as the start-of-year value for the following year, but they don't seem to be matching up quite right. Am I missing something?

Bob Wallace

I have strong "Me, too" suspicions.

I look at the 'death spiral' graph, all the way back to 1979, at the yellow April line and I see volume declining.


All along the line there's 'noise' as annual conditions melt a little more or let a little more freeze. 2007 was just a big hunk of downward noise, but by the time 2007 had rolled around volume was low enough that a "more melt" within normal variation year opened up a lot of water.

We've now cut volume down even lower. Unless we have a string of "more freeze" years I can't see how we don't have a meltout.

We've lost volume 21 out of 32 years. Stringing a bunch of "more freeze" years together is not statistically likely.

The only thing that saves the ice, IMHO, is that unknown unknown we've been trying to identify. The one which seems to not have turned up in noticeable amounts to date....

Chris Reynolds

Artful Dodger,

Chris R: Okay. Show us your math.

I've already explained that I've been unable to come up with a satisfactory method of accounting for energy gain due to exposed ocean. So how do you think I'd be able to account for changes in ice-albedo (a trickier problem as the numbers will be smaller and angle of incidence more of an issue).

Frankly the lack of numbers on the problem helps neither of our positions. The door of uncertainty swings both ways. The last person I had to point this out to was an AGW denialist.


Achtung/ Important information from Bremen Uni:


Regards Espen

L. Hamilton

Espen, we got it covered:

The comments to this entry are closed.