The undeath spiral

By Steve Goddard

Over the last three years, Arctic Ice has gained significantly in thickness. The graph above was generated by image processing and analysis of PIPS maps, and shows the thickness histogram for June 1 of each year since 2007.

The blue line represents 2008, and the most abundant ice that year was less than 1.5 metres thick. That thin ice was famously described by NSIDC as “rotten ice.” In 2009 (red) the most common ice had increased to more than 2.0 metres, and by 2010 (orange) the most common ice had increased to in excess of 2.75 metres thick.

We have seen a steady year over year thickening of the ice since the 2007 melt season. Thinner ice is more likely to melt during the summer, so the prognosis for a big melt looks much less likely than either of the previous two summers. More than 70% of the ice this year is thicker than 2.25 metres thick. By contrast, more than half of the ice was thinner than 2.0 metres in 2008.

So why did 2008 start out with so little thick ice? Because during the summer of 2007 much of the ice melted or was compressed by the wind. During the winter of 2007-2008, much of the remaining thick ice blew out into the North Atlantic and melted. So by the time that summer 2008 arrived, there was very little ice left besides rotten, thin ice. Which led to Mark Serreze’ famous “ice free North Pole bet.

Can we find another year with similar ice distribution as 2010? I can see Russian ice in my Windows. Note in the graph below that 2010 is very similar to 2006.

2006 on the left. 2010 on the right.

2006 had the highest minimum (and smallest maximum) in the DMI record. Like 2010, the ice was compressed and thick in 2006.

Conclusion : Should we expect a nice recovery this summer due to the thicker ice? You bet ya. Even if all the ice less than 2.5 metres thick melted this summer, we would still see a record high minimum in the DMI charts.

Mark Serreze has a different take for 2010:

“Could we break another record this year? I think it’s quite possible,” said Mark Serreze of the National Snow and Ice Data Center in Boulder, Colo.

Bookmark this post for reference in September.


The report of my death was an exaggeration

– Mark Twain


Addendum By Steve Goddard 6/3/10:

Anyone betting on the minimum extent needs to recognize that summer weather can dramatically effect the behaviour of the ice. The fact that the ice is thicker now is no guarantee that it won’t shrink substantially if the summer turns out to be very warm, windy or sunny. Joe Bastardi believes that it will be a warm summer in the Arctic. I’m not a weather forecaster and won’t make any weather predictions.


newest oldest most voted
Notify of

What size perturbation in summer albedo causes runaway glaciation? Maybe we should be up there spraying carbon black!!


Bookmarked. I’m still with you, Steve, and have been since late March.


It looks like a new ice age is fast approaching. We should have a geoengineering brainstorming session on how to mitigate the approaching catastrophe.
I guess aerial burst 100-Meg hydrogen bombs over the Arctic? Cheap and efficient, and probably the only way to do it. Anyone who has seen dirty snow knows that the particles melt their way through the ice, and the top refreezes. Not to mention a fresh snow layer that covers it up. So carbon soot would be too short lived.

had a quick look at the Serreze post. They don’t like you very much do they? One of the comments accuses you of only useing data sets that show the least warming/mostice extent. As distinct, presumably from the reverse which seems to be what they do!! Ah well time will tell>


Did anyone run these numbers past the Catlin Arctic Survey Team? They’ll be shocked, I’m sure.
(Sorry. Couldn’t resist.)
I’m all for an ice free Arctic, but from the looks of things above we won’t be getting one any time soon. Thanks, Steve.


I think we can rule out an Arctic ice cover tipping point.

Xi Chin

Is the area under the curves (the volume of ice) a conserved quantity? It looks a bit like that.

Layne Blanchard

I’d like to see continued recovery, and your points are well taken. It would be nice to shut down some of the clamor. But I’m not hanging my hat on any outcome. I don’t think it (extent) means anything beyond some combination of regional parameters.


Great pioneering work steve. You gonna integrate the thickness curves to get volume?


This article is an independent data point reaching the same conclusion as my own expectation that this will be a 2006-ish year.
Lets see how it goes.


I agree with Buffon–I see nothing to rejoice about runaway freezing. Sure, some of that ice will melt again this summer but come next winter, if this trend continues, lower temperatures will make life more difficult as the cold spreads south (and no, I don’t believe the “But this month is the HOTTEST on record” meme, seeing how they’re throwing away all the thermometers that give the real picture until they’re left with just one, which just happens to be in somebody’s shirt pocket).


Steve Goddard writes,
“Conclusion : Should we expect a nice recovery this summer due to the thicker ice? You bet ya. Even if all the ice less than 2.5 metres thick melted this summer, we would still see a record high minimum in the DMI charts.”
I’m impressed by how unequivocally and often WUWT has committed to the proposition that Arctic sea ice is recovering, in disagreement with most Arctic researchers. As you say, these will be pages to bookmark.


i’m a fan of the site guys. keep it up.
i do have a slight issue – and excuse me for being a layman – but how can climate progress and WUWT be looking at ‘real’ data and come with equal and opposite views for the current state of ice in the arctic, let alone future predictions ??

Excellent analysis. Simple yet effective comparisons.
However, if we’re back at a 2006 area distribution, I guess that allows doomsayers to say we could still go back to 2007 levels next year.

John Blake

Static analyses deriving from seasonal snapshots have little historical interest or validity. Arctic winds and currents form a complex dynamic system subject to many variables such as seafloor depth, coastline configurations, global weather patterns up to several years’ duration. Projecting climate/weather components requires depicting cyclically varying oscillations super-posed on long-term trend-neutral baselines. Even then, necessarily chaotic fluctuations in amplitude and frequency, regardless of linear slope, will tend to render annual extrapolations a casino game. “Climate science” PhDs purporting to resolve Nature’s ineluctable uncertainties might better study random-walks in equally arcane financial sectors.

R. Gates

Another intersting and well done analysis, based on that wonderfully erroneous PIPS 2.0 data. GIGO…and so I must once again, very humbly disagree with your conclusions, despite how convincing they appear, the fact that they are based on garbage data is cause for concern. In response, I would ask readers to go through the following very excellent presentation, given just this past March of 2010, at the State of the Arctic Meeting held in Miami Fl.:
Now, I have fairly cautioned you that the PIPS 2.0 data is erroneous at best, and that much better models exist, such as PIPS 3.0 and PIOMAS, both based on CICE modeling.
But you have made your case in an excellent manner, and you’ve stated quite plainly what you expect (a summer low that is higher than 2008 or 2009) and why you expect it. I take the opposite view, and base my case on what I consider to be more accurate models than PIPS 2.0, meaning that I think that PIOMAS model, and the summary given in the link I’ve given above are reasons why 2010’s summer minimum will be less than 2008 or 2009. One of the things that is clearly stated in the report I’ve given in the link above to is the ocean heat content. There is a lot of warmer water coming in at depths below the surface into the Arctic from both the Pacific and the Atlantic. (i.e. you can’t just rely on SST’s!!!) The report states:
“Oceanic heat has contributed critical preconditioning to sea ice melt in the western Arctic since the mid-1990s”
So we await September. Everyone should bookmark this discussion, as suggested…and as always, if your analysis proves the more thorough and your data more valid than…mea maxima culpa!

Mike Lough

The ice “Melt” each summer isn’t caused by the ice melting do to local temperatures. It is caused by wind and currents pushing the ice out of the artic. With thicker ice and everything else equal there will be a much higher minimum in the artic ice area this year. But everything else is never equal.

It’s the back-to-life spiral.

Robin Kool

Makes perfect sense.
Isn’t it true that a prediction of summer ice extend is incomplete without a prediction of what the wind is going to do.
What happened in 2007 might happen this year too.
Can arctic wind patterns be predicted for this summer?

Mike McMillan

What image processing program are you using, Steve?


Plenty of ice without a drop of whiskey!

Mac the Knife

WOW! The arctic ice thickness trend is turning up like a hockey stick chart! Could this inconvenient truth have real consequences also? Apparently, Al Gore and wife have reached (ahem..) a Tipper point in their 40 year relationship, with many icy years predicted to follow! I’ll refrain from references to gorebal cooling….
A bit more seriously, a big THANK YOU to Anthony Watts, contributors like Steve Goddard, and the long suffering Moderators, for providing this venue for publishing, perusing, and pontificating on our planets many climatic influences! The open forum and open access to data and analyses presented are delightfully refreshing! You are my daily “must read” and my first reference to others seeking current topical knowledge.


WUWT is really hanging it’s hat on the Arctic ice recovery. Will be interesting in August.


Anthony et al – OT. Maybe something for a new thread.
I am very concerned about this year’s snow melt as a result of the record snow pack. And now, two “Pineapple Express” systems are hitting this week. This could be a real disaster.
REPLY: Yeah, could be. Me thinks right now there may not be as much precip as models predict. – A


Looking at the data presented by Mark Serreze, it seems hard to predict with confidence ANY major deviation (+/-) from a notional average of Arctic ice extent. With these types of websites in general, the usual side issues always come up in reader comments, but I am struck also by a lack of attention to the ambiguity of the data trend on that site, which seems so obvious.


Ha ha ha!
Well…if somebody told me ten years ago that I today would dayly be monitoring Arctic sea ice statistisk with nail biting thrill and intrest. I would just laugh!!
But here I am and thanks to you Steve (or because of) im been adicted!!
I even got intrest in painting and locations of weatherstation in alaska. Well I can tell you Ive never expected to become a “Nerd” within this field. But “nerdiness” is intresting and benificiary in many ways. Thanks guys for letting us laymen participate.


Those poor Polar Bars (:). Because of rapid ‘climate change’ they will be TOOOO far from the water and will not be able to feed, breed or take happy snaps on icebergs.
They will be separated from lunch like AL & Tipper. 🙂


Well if the pattern of the recent glacial/inter glacial periods holds to form, we are in prime time for the beginning of the next ice age.
That would be hilariously funny at the expense of what is left of environmental credibility . . an ice age starting after 20 years of relentless fear mongering and hysteria about CO2 causing global warming.


You may recall I suggested you contact the researchers and ask them for help. Well, it looks like another blogger stoll my idea:

Mike McMillan
I was founder/lead developer of a VC backed video/effects software company west of London a few years ago, and we developed some cool software for doing image processing, effects, animation etc. It is not publicly available yet.

Come September, someone or other is going to have some serious egg on their face.
It is fun watching Romm and fellow travelers rant in the meantime. I’m still waiting for the ice free Arctic and pole we were promised in 2008.

UK John

I suggest we all return in September, in the meantime we should get on with our lives.

Here are the integrated volumes, in units pixel-metres. I haven’t worked out a conversion factor to metres^3 yet, though it should be pretty straightforward. Pixels seem to be about 1.5 km^2 in my maps.

2006    40,164
2007    32,818
2008   24,253
2009   26,505
2010   36,746

Sorry, meant to say that pixels are about 700 km^2


PIOMAS data is bogus, man, bogus:
1) They report the difference from “normal”, i.e. the average that they arbitrarily set in the coldest cycle of the 20th century, then set fixed SD bars. This is not kosher. The average they subtract from has no error bars!. Besides, using a differential rather than total ice mass, is a short way of inflating a result (by deflating an error bar). The smaller they can make the error bar, the greater any outlier point can wander from the SD interval, and look like an “event”.
2) Even if you subtracted from an average, including the SD uncertainty of the average (then it is a mean), how do you do it? They don’t do the former, they just represent “normal” with an average number for ea time point, not a mean or median. Example: subtract 10 ±5 from 20±10. Is it 10±5? [No, it is 15±10 – the SD just doubled!].
3) Also, if these were true values, with uncertainty at each point represented by an error bar, would the error bars be equal throughout the trend line as shown in PIOMAS? It appears that PIOMAS used a fixed error, a statistical blunder when trying to deduce a probable trend.
They are not stupid. They have the big $$$ to pay for the best actuaries and statisticians to present things in the most dramatic way, so as to hook sycophants like certain true believers on this board!


No fair using histograms and doing statistical analysis… you are only supposed to look at the extent like NSIDC does and make WAGs about what will happen this summer…


“WUWT is really hanging it’s hat on the Arctic ice recovery. Will be interesting in August.”
Jakers, look at it another way:
2008 minimum ice increased from 2007. 2009 increased from 2008. What evidence do you have that the trend of the past two years is going to reverse? All data indicate that the inventory of old ice is increasing. Old ice is harder to melt than new ice (note, I am not talking about thickness, I am talking about age).
If you have something that suggests that the trend of the past two seasons is set to reverse this year, please share it.

Murray Carpenter

Whats the difference between these two images?
Not much, except the 20 years between them. and I think more ice in 2010 than 1990.

sunderland steve
Serreze has little patience for the “breathtakingly ignorant,” such as myself.

Perhaps the researchers should contact WUWT for help? They would have done better in 2008 and 2009.

Ian W

Mike says:
June 2, 2010 at 12:52 pm
“You may recall I suggested you contact the researchers and ask them for help. Well, it looks like another blogger stoll my idea:

Its interesting isn’t it that your reference has such a short timespan? So it gleefully shows a trend line from the 1978 just after the decade when everyone was concerned about a new ice age and then runs it to current time when everyone is talking global warming.
Get a sine wave with a wavelength of say 30 years and then you can cherry pick dates with a 30 year simplistic straight trend line to prove whatever you want. Look at the same sine wave for say 200 years and you may consider that there is no trend but some cyclical changes.
If you are talking CLIMATE what length of time is appropriate?

Tom P

Whatever numbers your are deriving, they are not thickness. The average values from my analysis of the same data come out at around 2 m, which indicate massive melting if they were true June values. And as I have mentioned before, these derivations give a negative correlation to the PIP 2 team’s published ice volumes.
Anyway, you have produced a derivable number, let’s call it a Goddard (I don’t think Gd is taken) , and you are now predicting a recovery in the minimum ice extent as the ice values are rising from 2.0 Gd in 2008 to 2.2 Gd in 2010.
However, the June 2007 value is even higher that that for 2010, 2.3 Gd compared to 2.2 Gd according to my image analysis. So your prediction is based on a number that is inversely correlated with published ice thicknesses, and with high June values associated with low minimum ice extents. I think you’re quite brave to insist on a recovery this year on the basis of your analysis.
Whatever happened to that software company of yours?


bubbagyro says:
June 2, 2010 at 1:17 pm
Example: subtract 10 ±5 from 20±10. Is it 10±5? [No, it is 15±10 – the SD just doubled!].

No way! It is 10±11

George Kominiak

Hey Guys!
Take a look at this one:
It claims the Greenland ice is melting to such an extent that land levels are rising!


PIOMAS is a worthy endeavor, but much too immature to be laying any significant credence on.
If you look at this: It doesn’t pass the smell test. 2009 lower than 2007? I don’t think so.
But then I start looking closer and a few things occur to me.
1). “November ice volume anomaly”. *November*? Are you kidding me? Look at November here: You can throw a hat over the whole lot of them. November and December are just as uninteresting as May and June. If you want to maximize your ability to say something that has meaning, look at March or September (and much preferably September).
Yes, I know –“volume is not extent”. That doesn’t change the fact that when looking for signal you want to reduce the noise as much as possible, and there is a whole lot of volume (and extent) in May, June, November, and December that is just “noise” so far as trying to discern meaningful long-term trends. If that wasn’t true, those extent graphs for those months would not be so tightly grouped. The day will hopefully come in another 20 years or so when we’ll have enough historical data to make something meaningful out of those months, but in the meantime when searching for a needle in a haystack, the wise sleuth first reduces the size of the haystack as much as possible.
2). Look at the error for the last ICESat calibration, for 2007. Pretty significant. Extrapolate that to error bars for PIOMAS results for 2007, 2008, and 2009 on that graph and all three become meaningless relative to each other. You could, within the error bars, come up with pretty much any relative relationship you wanted between the three, including having 2007 the highest.

Tom P
I get an average thickness in 2010 of 2.4 metres and 2.1 for 2007.
I’m still developing the software, without external funding.


Tom P says:
June 2, 2010 at 1:52 pm
It just does not matter what facts you throw at Steve–he is clearly in DENIAL about Arctic ice extent and thickness.

richcar 1225

If the NAO goes negative the ice volume will build fast. This a warmists worst nightmare.


crosspatch says:
June 2, 2010 at 1:19 pm
If you have something that suggests that the trend of the past two seasons is set to reverse this year, please share it.
Joe Bastardi – who once was quite popular on here…