Blowin’ in the Wind

Anthony Watts posted about how Steve Goddard’s prediction for the summer minimum in sea ice extent was pretty good.

Watts says:

This graph from the SEARCH report (in entirety below) sums it up pretty well:

and shows this graph:

Sure enough, it clearly shows Goddard’s forecast as 5.1 million km^2. But I have to wonder, when exactly did Goddard make that forecast?

Back in June, Steve Goddard was convinced that sea ice has thickened lately, and that we were bound for a “nice recovery” this summer. He even predicted on June 2nd:

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.

He doesn’t quote a number, but he does claim a record high minimum in the DMI charts. A day later he added this:

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.

On June 6th Goddard got a little more specific:

Conclusion : Based on current ice thickness, we should expect September extent/area to come in near the top of the JAXA rankings (near 2003 and 2006.) However, unusual weather conditions like those from the summer of 2007 could dramatically change this. There is no guarantee, because weather is very variable.

Incidentally, the September minimum in JAXA data for 2003 was 6.03 million km^2, for 2006 it was 5.78 million km^2. The average of those two figures is 5.9 million km^2.

At least by June 23rd he had finally gotten specific:

I’m forecasting a summer minimum of 5.5 million km², based on JAXA. i.e. higher than 2009, lower than 2006.

On July 20th he reinforced his claim:

I expect to see that JAXA will move closer to my 5.5 million km² forecast for the summer minimum.

He confirmed it again on Aug. 1st:

Our PIPS based forecast of 5.5 million km² continues to be right on track.

By Aug. 8th he suggested that his 5.5 million km^2 prediction might even be too conservative(!) and that this summer’s minimum might be the highest since 2006, maybe even exceed 2005:

Conclusion : There will probably be minimal ice loss during August. The minimum is likely to be the highest since 2006, and possibly higher than 2005. So far, my forecast of 5.5 million km² is looking very conservative. Ice extent is higher than I predicted for early August.

As late as Aug. 15th he was still suggesting that 5.5 million might be too conservative.

My forecast (dashed line below) minimum of 5.5 million (JAXA) continues to look conservative. It all comes down to what the winds do over the next few weeks. If the winds keep compressing the ice, the minimum may go a little below 5.5. If the winds quiesce, the minimum may come in a little above 5.5 – which is looking like a pretty good number right now. Some people at NSIDC started out with a 5.5 forecast this year, but seem to have backed away from it since.

The JAXA extent crossed below the 5.5 million km^2 mark on Aug. 26th, with 5 days remaining before September even started. With only 3 days left in the month, the JAXA extent as of August 28th was only 5.34 million km^2 — already below his predicted 5.5 and it wasn’t even September yet — but in spite of that Goddard on Aug. 29th didn’t back down from his forecast made in June:

It continues to look like my June forecast will be close to correct, though as we have seen – this contest is a crap shoot. It all depends on the wind.

The record is clear. After Goddard settled on a figure in mid-June, he stuck to it like glue … insisting that his June forecast would be close to correct as late as Aug. 29th. So when exactly did he make that forecast of 5.1?

75 responses to “Blowin’ in the Wind

  1. And therein lies the problem with this whole ‘debate’: some people have no compunction to blatantly lie. Chess with pigeons, all over again.

  2. Tapani Linnaluoto

    How many graphs must a man break down
    Before you call him a man?
    Yes, how many years can sea-ice exist
    Before it melts to the sea?
    Yes, how many times can a man turn his head
    Pretending he just doesn’t see?

    The answer, my friend, is…

    (With apologies to Mr. Dylan)

  3. Judging by his submission, he revised his estimate sometime after the middle of August, but prior to the end of the month, else his newly fudged value wouldn’t have appeared in the August report.

  4. The last set of SEARCH Sea Ice Outlook forecasts was due around August 15. So it looks like Goddard was writing one thing on the blog, while at the same time sending in a different number to SEARCH. What’s up with that?

  5. Goddard didn’t even realize that the SEARCH “contest” was based on NSIDC, not JAXA …

  6. So he knows how to game the system. Next year I’ll make 10 predictions in various locations, all differing by 150,000 km2. After the season I’ll point to the one that came closest and pretend I know more about forecasting sea ice than Goddard.

  7. 1) The Internet never forgets, but it is good so see someone go back and look.

    2) But, did anyone else see the Ads by Google here, for Steve Millloy’s website:
    Michael Mann…
    Defamed or defined by ‘Hide the Decline’?

  8. Something doesn’t seem quite right.

    From Goddard’s 5.1 mkm2 projection:

    Verification of previous years showed that this is a highly accurate forecasting technique, with the exception of 2007 – which was dominated by unusual winds which compacted and melted vulnerable areas of ice. Until mid-August this approach appeared to be working very accurately. Since then, strong southerly winds have developed and extent has dropped below predicted values. Thus the 8% reduction from the initial forecast.

    2010 Sea Ice Outlook, August Report
    Steve Goddard, contributor at blog

    According to this he changed his projection based on data in the latter half of August. However, the chart that he is shown as having made the 5.1 mkm2 projection is here:

    September Sea Ice Outlook: August Report

    Just below the chart it states:

    Figure 2c. Distributions of Outlook estimates for September 2010 arctic sea ice extent based on July data. Observed September minimum sea ice extent denoted by the red dashed line.

    If the caption is right then he has to be basing his estimate on July data, no later. According to his submission he is taking into account what happened in mid-August. As he states in his submission, “Until mid-August this approach appeared to be working very accurately. Since then, strong southerly winds have developed and extent has dropped below predicted values. Thus the 8% reduction from the initial forecast.” He specifically states he is using data from mid-August.

    • I am wondering whether the “southerly winds” occurred throughout the Arctic. That would be truly fascinating weather. Wouldn’t want to be around the North Pole during it though.

  9. He was closer than I was. I looked at those satellite pictures and said 4,000,000 sq km tops. As the season went on I began to think that figure too high.

    PIOMAS looks even scarier

  10. 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.

    My prediction is that it could be thicker or thinner. Give me a prize.

    • Except it wasn’t thicker, and his repeated proclamations that ice volume has been “recovering” the past few years are hooey.

  11. I made this very point to Goddard back when the minimum occurred. The sheer vitriolic animosity with which he defended and lied about his predictions told me all I needed to know about him.

  12. Why are we talking about Goddard, isn’t that the only thing he wants? The Internet may never forget, but at least give it a chance to try.

  13. There was a book about an unhappy world where predictions were “corrected” afterwards, to match observations. What was its name again?

    Like Oreskes said, how come can deniers get away with being wrong nearly all the time?

    • 1984.

      A book replete with denier memes such as double-think. The correction was done by the Ministry of Truth.

    • Alexandre,
      That one’s easy: they’re telling people what they want to hear. Politicians have been doing it for centuries.

  14. Good job, Tamino. I wrote an early time-line as well.

    I’ve discussed this with Goddard over at his site, but he will just not admit that his first predictions were for 2010 to end up in the 2006 range. There’s nothing wrong with a wrong or an adjusted prediction, the problem with Goddard is that he just won’t admit he made the predicition at all. Like Heraclitus says, I think it’s best to ignore people like Goddard. And Watts too.

    • Yes, he insists that wasn’t a prediction because he didn’t give a number (“recovering to 2006 levels” not being a statement containing a number for the minimum isn’t a prediction).

      Of course he only claimed it wasn’t a prediction *after* it became obvious that he was very, very wrong.

  15. Goddard didn’t even realize that the SEARCH “contest” was based on NSIDC, not JAXA …

    And he didn’t know the number was based on average September extent, and not on the absolute minimum. Otherwise his number would’ve probably been a bit higher still.

  16. Yeah there could be entire sites devoted to his debunking… There’s no way anyone could keep up with his ridiculous 6 posts a day pace though… One wonders how much research can be done…

  17. That person is in love with himself, i.e. an incurable narcissist. It’s written all over him and his literary output. He’s probably an underachiever who hates people with more talent that he’s got in general and people in the academia in particular. Observe his moronic past battle with vitriolic one-liners against people of standing, such as with Dr W. Meier at WUWT last summer, where Meier guest-posted, e.g. on how to accurately count pixels on some projections of world maps.

    When the Goddard-fuelled row over the definition and practical implications of the triple point of water in the Arctic broke out at WUWT soon after, I needed no more proof: this person is not a scientist by no training nohow, although he routinely leads people into thinking so. Even “Steve Goddard” is very likely a pseudonym. It’s all about obfuscation from the beginning to the end, the end that we may never be fortunate enough to see.

  18. Yeah there could be entire sites devoted to his debunking… There’s no way anyone could keep up with his ridiculous 6 posts a day pace though… One wonders how much research can be done…

    He is a waste of time.

    Cranks are indefatigable, and Goddard is a -level crank. This is what made him valuable at WUWT, where the sheer volume of BS is important.

  19. Out of curiosity has goddard ever owned up to any of his mistakes?

  20. > owned up to ….?

    Found that by accident while wondering if anyone ever figured out if the guy’s using a pseudonym or is a real person. Presumably Watts knows, or knew.

  21. Does he really put out 6 posts a day??
    And does not admit to errors? (Except for that itty bitty that Hank found)
    Sounds like an advertising campaign.
    Is it possible that steven goddard does not exist as a real person and is really a pr team for a think tank or some other organization trying to spread as much mis-information as possible?
    Just asking…

    • I remember there being some question regarding just how far back his history goes. 2008, if I remember correctly. Before that nothing. But it has been a while so I decided to recheck.

      Judging from Google, the first essay he wrote that mentions global warming was:

      Is the earth getting warmer, or cooler?
      A tale of two thermometers
      By Steven Goddard, Posted in Science, 2nd May 2008 10:02 GMT

      … and that was the first essay he wrote for the Register. But Europeans tend to speak more of “climate change” or “climate disruption” than they do of “global warming”. So I tried “climate”. Before May 2, 2008, the most recent date that came up was in The Albanian, Feb. 1978. Somehow I doubt that was him.

  22. Anne van der Bom

    AFAIK Steven Goddard is a pseudonym.

    Since this ‘Steven Goddard’ name has never been publicly linked to a real, living person, it may concern one of tousands of people bearing that name and we might as well consider it a pseudonym.

    Just as every lottery has a winner, he just suffered the (bad) luck that the pseudonym generator coughed up his real name. That clears Anthony Watts of any accusation of hypocrisy that one might be inclined to throw at him:

    REPLY: I don’t know how much knowledge he really has and how much is youthful bluster, as far as I can gather, he’s an undergraduate. Perhaps a sophomore. Besides my policy is that anyone that would guest post here has to do it under a real name. We don’t accept web phantoms for guest articles any more than a scientific journal would. – Anthony

  23. I followed the whole Arctic ice minimum episode this summer on WUWT and as far as I remember the two most high pprofile predictions were indeed Steve Goddard at 5.5 M and R Gates at 4.5 M – the latter emerging the winner – disappointingly as I was hoping that Steve Goddard would be correct. SG should bite the bullet – I dont know why he now claims 5.1 as his prediction. Still – there’s always 2011.

    Bit nippy outside just now – but that’s just weather of course…

  24. So what are we going to do with this “Goddard” person posing as a legitimate specialist of — not a thing? Grind our collective teeth and endure?

    • Jarmo, first, I don’t think too many people confuse Goddard with an actual climatologist, glaciologist, weatherman forcasting CO2 snow in Antarctica or what have you. Second, Tamino uses Goddard’s baffoonery as an opportunity to educate. We use it as a chance to point and laugh. But honestly I sometimes wonder whether we might be getting a little too focused on Goddard and WTFUWT. On occasion.

    • Just ignore him and focus on the person that will do Arctic sea ice mangling on WUWT next season. It could be this weatherman who thinks the numbers are being cooked by NSIDC. Where oh where did his thorough research go wrong?

  25. Jarmo,
    I find pointing and laughing to be the appropriate response. It works in monkey houses.

  26. If you go back to when Goddard did his Register piece in 2008, Doug Bostrom posted at RC:

    Doug also pointed out some more here this year:

    • I’ll buy that Goddard isn’t being ironic in the use of his nome de plume.

      Not so sure about the person who picked it. The bit about receiving “death threats” — given how many “skeptics” are out there scribbling away about world conspiracies involving thousands of scientists, enviro-Nazis, etc. is — in my view — a bit unbelievable. Remember: he was using this name before he became associated with WUWT and there are high profile “skeptics” who are considerably more vitriolic.

      My guess? Either a ploy for sympathy that simultaneously tries to make supporters science look like kooks (an often-used strategy of projection involving accusations of ideological motivation, financial interests and statistical slight of hand), paranoid delusions, or a convenient excuse for not divulging more information — that consequently helps cover the tracks of his employer. But ploy for sympathy seems unlikely — even given its obvious utility — otherwise it would have been more widely used.

  27. Igor Samoylenko

    As I was browsing through that thread at RC where Doug posted about Goddard (provided by J Bowers above), I came across this by Goddard himself:

    “it is scurrilous to reprint a private E-mail on a public forum without the author’s consent or knowledge.”

    That was of course more than two years before the theft of the private emails at the University of East Anglia. It must be great to have principles elastic enough to suit the ever changing circumstances…

  28. Some time ago I came to think that “Goddard” name was chosen to install some confusion with NASA’s Goddard Institute for Space Studies, GISS, whose Director is James Hansen.
    The real Goddard was a .

  29. Tamino:

    The downward sloping trend line for September average ice extent provided by NSIDC may overstate the potential for future ice loss since it treats the Arctic as a single block of ice.

    Many regions of the Arctic became essentially ice free in September of 2007. These areas will therefore not contribute to future declines in extent. Since these regions went as low as possible in 2007, they will likely cause random upticks in the September ice extent figures as the regions report some ice coverage in a particular future year. The NSIDC charts below comparing 2010 to 2007 , 2009 to 2007, and 2008 to 2007 show how some of these regions “recovered” from the 2007 low (white areas in the links below) and how these “recoveries” were partially offset by declines in the remaining ice in other areas (dark grey areas in the links below).

    The core region of the Arctic ice (the light grey area in the links above) does not appear to have significantly changed since 2007. This region generally is described as the Arctic Basin on the Cryosphere Today website (see link below). This region has not had as steep a decline from the 1979 to 2008 average extent as the NSIDC chart indicates for the Arctic as a whole.

    The Arctic Basin will be the last region to become ice free. Since this area contains almost all of the remaining Arctic sea ice at the September minimum, the future trend line should follow the trend line for this region rather than the NSIDC trend line for the Arctic as a whole. This may be part of the explanation for why there has not been a downward trend in the September ice extent numbers since the severe drop that occurred in 2007. There is no longer a sufficient amount of the easy to melt or transport ice at latitudes below the Arctic Basin remaining after 2007 to generate the same declining trend line that was observed before 2007.

    In making the sea ice extent prediction, how do you take into account the different rates of decline for the various regions? Do you separately work up estimates of ice extent for each region?

    Is there a graph that shows the September average extent trend for the Arctic Basin?

    Do you think that the declining trend line for average September ice extent produced by the NSIDC will slow or flatten out for the period 2007 and forward?

  30. Will Crump, you seem to have a major misunderstanding about Arctic ice. Talking about “blocks of ice” is a clear indication of where you went wrong. The ice, relatively speaking, is wafer thin. It is more like a skin on a cup of cooling tea, and is very mobile. So, the ice in a region one year has little to do with the ice in the region next year, particularly now that we have a negligible quantity of thicker ice.

    You spend a lot of time stressing about “declines” and “recoveries” in regional ice. However, given the variability of initial conditions, and the variability of the weather during the melt season, that’s just a pure red herring.

    We have the total data, and that should be enough. We need considerably more expert tools to do anything useful with regional data. Predictions of ice extent that have included regional ice modelling have not shown significantly more skill in predicting extent than forecasts that don’t.

    Please note that 2008, 2009 and 2010 all made the downward trend MORE severe than ever before. Your talk of a trend “since 2007” is meaningless.

    You seem blind to the third dimension. If the ice reaches zero thickness before it reaches zero extent – well, I think they taught this one in high school.

  31. Didactylos :

    You are correct in terms of the phrase “block of ice” that I used, and it would have been better if I had used the word region instead.

    I am not blind to the volume decrease and I agree that the total volume has decreased since 2007, but this is not the end of the story. Please open your eyes and note that this decline in volume is not showing up in the September minimum extent figures or the March maximum figure. The 2007 minimum is the current minimum record holder and per NSIDC was 4.13 million km2 with the average September extent 4.3 million km2. The 2010 minimum per NSIDC was 4.6 million km2 with the average September extent 4.9 million km2. Three years of significant volume decline have resulted in an expansion of 500,000 km2 of the ice extent. The 2009 September minimum per NSIDC of 5.1 million km2, was 1.0 million km2 higher than 2007. At the end of March of 2010 the ice extent figures were almost at the average for that day. If there was a strong correlation between volume and extent this would be impossible, and yet, we can see it with our eyes and the data shows it exists.

    Obviously, the ice will not reach zero thickness until it also reaches a zero extent figure. Since the volume has continued to decline and the extent has increased, the remaining ice must be thinner, but it still persists, which is my point. This suggests there is something wrong with analyzing the data on a total Artic basis.

    The flaw in extrapolating the linear decline in the volume analysis is the same as using the NSIDC extent data. Both of these charts include regions that have had significantly higher rates of decline than the Arctic Basin region (as established by the Cryosphere Today analysis).

    What I am suggesting is that the use of a trend line for the Arctic as a whole in predicting the future trend for the Arctic ice is a flawed approach since it includes distinct regions which have already or nearly reached the zero volume zero extent level in September. Rather than treat the Arctic as a single region in drawing the graphs, I am suggesting that it only be done for the Arctic Basin region (as described on the Cryosphere Today site). This will be the last region to be reduced to the zero volume zero extent level in September.

    The PIOMAS information does not provide regional detail so I will use the ice area information provided at Cryosphere Today.

    The current chart for the Arctic Basin showed a September loss of about .8 million km2 of ice area compared to the average with about 2.5 million km2 of ice remaining. Please note that for substantial periods of the year, the ice extent in the Arctic Basin stays near the average for 1979 to 2008 period and is currently slightly above that level.

    I tried to find some earlier charts and so far this is what I have.

    The 2008 minimu looked like the 2010 minimum. The 2008 mid-September sea ice anomaly was negative 0.75 million square kilometers, but there were 2.5 million square kilometers more than zero. (Note this chart used the 1979 to 2000 mean for calculating the anamoly)

    The remaining extent in 2010 and 2008 represents about 77% (2.5/(2.5+.75) of the average ice extent. This shows that this area is losing ice at a slower rate than the NSIDC trend line.

    In 2009 Arctic Basin ice extent was a higher percentage compared to the average 86% (3.0/(3.0+.5).
    (I am not sure why the total number is higher for this year, but it may be due to how the mean changes over time or just the difficulty in eyeballing the graphs.)

    I will try to get more information from Cryosphere, but the point is that the volume and extent analysis should only be done for the Arctic Basin region and not for the Arctic as a whole as this introduces distortions from regions that have already reached a zero volume zero extent level in September, which will therefore not be contributing to future September declines.

  32. Shorter Will Crump:

    It’s not melting! Lalalalala!

    So much *effort* to distract from a simple and obvious fact.

  33. Didactylos :

    Intuitively I agree with you that the dramatic decline in volume should show up in the reduction of ice extent, but this is not occurring based on observations of the central Arctic Basin. The downward trend line for the central Arctic Basin September ice area is very slight compared to the NSIDC September trend line for the whole Arctic and the PIOMAS trend line for volume.

    Based on the observations of the ice extent in the central Arctic, the decline in the Arctic wide trend line of ice extent produced by the NSIDC will slow and could perhaps flatten out somewhat.

    This is why I am suggesting that making predictions based on Arctic wide trend lines for ice extent or volume reductions will overstate the rate of future decline. Focusing solely on the central Arctic Basin region should not be controversial since the central Arctic Basin is the northernmost and coldest portion of the Arctic and will be the last region to be “ice free” in September.

    The ARCUS site has a link to a pdf at the bottom of this link:

    for a paper by Adrienne Tivy that has historical regional observations and model projections for the “Central Arctic Ocean” (see figure 8). For the period 1981 to 1989, the actual September ice area for the Central Arctic Ocean fluctuated between a low of 2.5 million km2 in 1984 and a high of 2.8 million km2 in 1987. After 1989, the region has remained between 2.7 million km2 and 2.5 million km2 for all years except for the big dip in 2007 to 2.1 million km2 and the uptic to 2.4 million km2 in 2008. The region returned to the 2.5 million km2 level in 2009 and per the chart at Cryosphere Today maintained this level in 2010.

    The observations show that the central Arctic Basin ice extent can be maintained at a 2.5 million km2 level even with declining volume. The impact of the declining volume and thinner ice appears to show up as a potential for greater variability in the minimum annual ice extent. However, the declining volume has not prevented the central Arctic Basin from returning to a September minimum ice area of 2.5 million km2, thin though it may be.

    Using the period 2005 (the year of the second lowest September minimum ice extent)through 2010, the central Arctic Basin has averaged just over 2.4 million km2 in ice extent. It is this observation which causes me to doubt the more agressive predictions of an” ice free” September Arctic by 2016 based on the extrapolation of trend lines for the Arctic as a whole.

    I do not think there will be a dramatic decline of Arctic wide ice extent in 2011 to a level below the 2007 level of 4.13 million km2 unless unusually strong winds, comparable to 2007, push the central Arctic Basin ice toward the coast of Greenland and the Canadian Archipelago and down the eastern side of Greenland (this would increase the east Greenland region ice extent as observed in the NSIDC comparison images in the December 8 5:28 PM post) . Otherwise, the September 2011 minimum should be like the 2010 minimum of 4.6 million km2 and could possibly show a temporary uptic to as high as the 5.1 million km2 level that existed in 2009.

    Of course, actual observations may disprove this supposition and the wide forecasted range of 4.1 to 5.1 million km2 could turn out to be inadequate. Given the wild fluctuations observed in 2010, we may not know until late in the melt season of 2011 if this range will hold up.

  34. Will Crump, perhaps you don’t realise why I am so dismissive of what you are saying.

    That’s because what you have to say is an odd mixture of the obvious, the wrong, and the confused. I don’t plan on wasting my time trying to extricate the fact from the fiction.

  35. Will, what if I summarize what you said as: that thinning ice in the Actic Basin has a good chance of becoming driven around so that overall extent will be lower than in the previous years.

  36. Will Crump,

    I’d say Didactylos did you an injustice with his “It’s not melting! Lalalalala!” – that doesn’t appear to me to be at all what you’re saying. I also am perfectly willing to admit that the trend of decreasing extent at the annual may well slow down in future as further decreases have to come out of the last redoubt of arctic sea ice. However, I don’t think you’ve presented any convincing evidence that this must be the case and that it couldn’t be the case that the ice, considered as a whole, will continue to disappear in line with the amount of heat energy available to melt it (which might reasonably be expected to increase over time in line with warming of the climate in general and of the arctic in particular as sea ice disappears and the albedo of the Arctic in summer decreases). I certainly think you’re making too much of the trend in the Arctic Basin. I’m sure I could find some part of the Arctic Basin in which the trend of sea ice extent is absolutely flat, i.e. the sea ice extent has completely covered that region at the annual minimum for as long as we’ve been taking observations. I don’t think that would prove that the ice within that region was never going to melt. Similarly, I don’t think there’s any reason to believe the Arctic Basin as a whole will continue to follow its current trend no matter what happens to the ice around it.

  37. Thanks Jarmo.

    The September ice extent trend line will continue to decline, but not as fast as the current NSIDC September trend line indicates and not as fast as the plot lines in the volume chart produced by Wieslaw Maslowski.

    Some regions of the Arctic have already met the “ice free” standard used by Wieslaw Maslowski of an 80% drop from the 1979 – 2000 summer volume baseline.

    Plot a trend line for the central Arctic Basin ice area based on the data and see when you get to the “ice free” standard used by Maslowski. If volume data exists for the central Arctic Basin then plot this data and what do you see?

    The central Arctic Basin will not be below 550,000 km2 in September of 2016 or 2019.

    • Will Crump,
      Keep in mind that as more of the Arctic becomes ice-free, there is a positive feedback due to decreased albedo. What is more the ratio or perimeter exposed to open (and warming) water to area increases. Thus, while I agree that 2007 was an outlier, I think we could possibly even see an accelerating ice loss above that trend.

      Also keep in mind that given the rapid changes we’ve seen in the Arctic, the past may not be a representative dataset, especially when it comes to outliers—e.g. was 2007 a once in 100 years event or a once in 10 years event. All we can really say is that the trend for all regions in the Arctic is definitely downward.

  38. Ah, complicated data being dissected by inexpert hands.


    No. I tried. But I really can’t bring myself to waste any time on this. Will, you will just have to let time be your teacher. Come back next decade and tell me if you were right.

  39. ‘Some regions of the Arctic have already met the “ice free” standard used by Wieslaw Maslowski of an 80% drop from the 1979 – 2000 summer volume baseline. ‘

    Will, I think the crucial point of variation remaining here will be timing in the yearly cycle. Here in Finland we see an ice free standard in March to June depending on the location of the melting waters. Were we to judge them by the June situation they’d all look the same.

    My instinct is that once the protective edges of the Central Basin in the more shallow waters are removed faster (in the yearly cycle), the whole bulk will begin to behave more unpredictably. Future goddards will be facing even tougher challenges!

  40. Figure 8 in the link below has the 1981 through 2009 data for September for the central Arctic.

  41. This is based upon very little, so please take it with a large grain of salt, but I expect the Arctic Basin melt to accelerate, not slow, due to positive feedbacks continuing to affect the area–and, indeed, affecting it increasingly. I’m thinking particularly of albedo, specific humidity, and what I will call ice motility.

    But I’ve got zero quantitative analysis, and have no plans to attempt any. I plan to wait and see–in between public bitching and moaning about the lack of effective GHG mitigation, of course.

  42. While the central arctic is warming by at least 0.383 degrees C per decade, the temperature for the area north of 80 degrees remains below the freezing point of sea ice for a significant portion of the year. This is why the central Arctic sea ice extent is so persistent. As stated in the link below:

    “Most of the area above 80N is (currently) still covered in permanent sea ice. In the Arctic Summer when the surface ice is melting, it is known that the air temperature close to the surface is limited by this ice melt temperature to just above zero degrees C, (Rigor 2000). This is why the Summer air temperatures have not varied much over the entire instrumental period. This maximum temperature “clipping” effect is clearly seen on all arctic data sets from Arctic buoy data to individual station data to satellite data.”

    The link includes a discussion of the high Arctic temerature and a temperature chart for 1958 through 2010 .

    Current year graph is at:

    [Response: As has been mentioned before, not only is the DMI temperature data a reanalysis product (the output of a computer model), but DMI has switched from one computer model to another more than once during the period of coverage so their data aren’t even homogeneous. This, in my opinion, makes it great for weather but far less suitable for studying climate change, than other estimates of arctic temperature.

    That aside, much of sea ice melt is due to warming from the ocean below, not the atmosphere above, so the fact that the air temperature is above freezing doesn’t prevent the ice from melting. The “air temperature still above freezing” argument is a red herring.

    It’s starting to look like you’re getting much of your information from WUWT, which is not a reliable source.]

  43. “Melt ponds may explain rapid melting of sea ice

    The current climate models do not predict an ice-free Arctic for another 50 to 100 years, but observations show that melting of the sea ice is accelerating. Why are none of the models able to describe this rapid change? Melt ponds on the sea ice in summer may be part of the explanation.”

    (One famous meltpond near the North Pole, tracked by a webcam and the then resident WUWT expert, was declared to be “frozen over” 4-5 times last August when in fact it was not. Dark objects may have surprising qualities, in the Arctic as well!)

  44. I do not view the crafty disinformation of Anthony Watts or the laughable posts by Steven Goddard as a reliable source. If the links I have provided are not reliable, please point me in the right directon.

    I am somewhat surprised that focusing on the central Arctic Basin would be controversial as it appears to be widely acknowledged that this will be the last region of the Arctic to melt. I expected the comments to dismiss my post as being old news that has been incorporated in the trend analysis in a manner that was not obvious to me.

    I referred to the DMI data to support the existence of a warming trend in the high Arctic. If there is a better data set or more accurate model of temperature trends in the high Arctic please let me know, but I suspect it too will show warming and that the current air temperatures are as low as the DMI set indicates. I am not aware of any data sets or models of SST for the permanently (at least currently) ice covered central Arctic Basin, but I agree that heat from ocean currents is generating much of the volume and extent loss witnessed in other regions. While this has no doubt reduced volume in the central Arctic, it has not significantly reduced ice extent.

    Air temperature in the central Arctic is not a red herring as the ice does not freeze without cold air. Even with the additional ocean heat, the air temperature has still kept the central Arctic above 2.4 million km2 at the September minimum for all years except 2007. Apparently, cold air still beats out warm oceans in the central Arctic.

    The 2007 dip was a wind driven event that was assisted by the weakened state of the ice due to the various heat forcings mentioned in the comments. Without such winds, the central Arctic basin has returned to its previous levels at the September minimum. Future years will show increased variability in the central Arctic basin, but the average rate of decline should still be slower than the rate observed currently for other regions of the Arctic.

    The central Arctic will go back to its maximum of 4.2 million km2 and melt back to 2.5 million km2 or so at its minimum just as the chart at Cryosphere today indicates. Lower minimum levels should be expected in future years, but the decline will not be as rapid as it has been for other regions of the Arctic.

    I appreciate the sharp analysis of the commenttors who suggested that the central Arctic Basin will show an increasing rate of decline due to the loss of surrounding ice and other factors. This suggests, and I agree, that the trend line for the future melt in the central Arctic Basin should be an increasingly downward sloping curve rather than a straight line.

    Still, I am going to stubbornly insist that this analysis be done only with reference to the central Arctic Basin and not the Arctic as a whole. The central Arctic Basin has more of its original ice extent because it has fewer days of melt, is subject to colder air temperatures, receives less sunlight (TSI) due to the oblique angle, is farther from the influx of warmer ocean currents from the North Atlantic and the Pacific, and sits atop a much deeper sea basin than the areas surrounding the central Arctic. These factors are not going to change. Any analysis that does not take into account the unique factors affecting the central Arctic Basin will be flawed and will predict more ice loss than will actually occur. The failure of such predictions will only provide fodder for the disinformers at WUWT.

    Predicting a slower decline is still a prediction of decline in Arctic sea ice.

    • Re: Will Crump

      A return to summer 2007 conditions, with its strong Arctic Dipole, will spell disaster for the Central Arctic Basin ice. With the markedly thinner ice of 2010 vs 2007, a return of that strong Dipole for the majority of summer will result in significant compaction of basin ice against the Canadian Archipelago and Greenland with much subsequent ice advection out the Fram into the North Atlantic.

      While ice would then cling to the Archipelago and Greenland for a few more years, the central basin (especially the portion over the pole) would then be ice-free in summers. Without a strong core of multi-year ice, a recovery of the Arctic ice to former conditions is gone.

      Neven’s blog and Kelly O’Day’s blog are both great resources for Arctic info.

      The Yooper

  45. Will Crump,

    Regarding your statement “The 2007 dip was a wind driven event that was assisted by the weakened state of the ice due to the various heat forcings mentioned in the comments. Without such winds, the central Arctic basin has returned to its previous levels at the September minimum.”

    You admit that much of the difference between 2007 and subsequent years was a matter of what the winds were doing, not anything fundamentally different in the causes of ice loss (volume loss as opposed to how it’s distributed in two and three dimensions). Why, this being the case, you assign so much significance to this difference, escapes me.

    Also, I see every reason to believe that the number of days of melt in the Arctic Basin will increase in future and, while the distance warm ocean currents have to come from the North Atlantic and Pacific will, I suppose, stay pretty much the same, subject to any changes in ocean circulation that may occur, there being less ice between them and the Arctic Basin to absorb that heat before it can reach the Arctic Basin strikes me as likely, i.e. I don’t think the factors you cite as “not going to change” are all as immutable as you imagine.

    I’m sure no one objects to you analyzing the decline however you like and good for you if you do so and manage to improve on the predictive skill of current models (maybe you ought to submit predictions to SEARCH next melt season) but please consider that you may be a bit too much in love with your own idea in a way that, as others have hinted at above, is somewhat reminescent of Steve Goddard.

  46. Thanks for the vote of confidence Jon:

    My estimate for 2011 is 4,1 million km2 to 5.1 million km2 for the September ice extent minimum per NSIDC with a confidence level of 75%. This is + or – 500,000 km2 from the 2010 minimum.

    If you want a higher confidence level, I will insist on expanding the range.

    What is your prediction?

    [Response: I made mine here.]

  47. Hey!
    What about this:

    The global sea ice should really be what matters shouldn’t it?
    Sure, it shows a slight down tick after 1979, but we have to ask our selves:
    How would this curve look if we had data since ~1850?
    My guess is that there would be a distinct maximum during ~1850-1930, a minimum similar to the present during ~1930-1945 and that ~1945-1979 would show up as quite uneventful.
    What is your guesses?

    And also – shouldn’t we *always* compare the same things?
    Global vs global, which means that the regional Arctic ice extent is irrelevant while talking about global atm. temperatures – which is what this whole debate really is about – right?
    And we shouldn’t compare temps with sea ice further back than 1979, since that is as far back as we have somewhat reliable global sea ice data – right?
    And if we do this, where does that leave us?

    I find it disturbing that apples and pears are constantly compared in this debate.

    [Response: Here’s the difference between the approach of a denialist (that’s you) and a real skeptic (that’s me):

    You look at a graph of global sea ice with the y-axis on a super-tiny scale and say “a slight down tick after 1979.” That’s nowhere near reality — it’s just what you want to see. Then you spew some bullshit speculation about “if we had data … I bet …”

    I didn’t just look at a graph, or take somebody else’s word for the conclusion, I studied the data myself.

    Data says: you are completely wrong.]

    • Simpler answer to Patrick:

      The global sea ice should really be what matters shouldn’t it?


      Before answering, please look at polar projections of the arctic and antarctic …

  48. Shorter Patrik:

    I don’t like what the data show, so I’m going to make up some reasons to ignore the inconvenient bits.

    But I’m going to spin this as impartiality.

  49. Daniel Bailey:

    I agree that the Arctic will not recover and I do visit the two excellent sites you mentioned.

    When do you think the central Arctic Basin will be ice free in September?

    A one or two consecutive year return of 2007 conditions will not make the central Arctic Basin ice free, but will most certainly reduce it below the 2007 level. These conditions are not maintained on a year round basis as the ice area data shows that the central Arctic Basin is cose to the average for some months.

    2007 conditions reduced the central Arctic Basin September minimum by about 500,000 km2 from 2.6 million to about 2.1 million km2 based on this graph in figure 8 of the link below.

    The minimum has since returned to a range around 2.5 million, which is slightly lower than it was before 2007. A pattern of large fluctuations in the anomaly should continue, but will not make the Arctic ice free as soon as the predictions based on extending the trend line for the Arctic as a whole would indicate.

    The Cryosphere today graph below shows the current Arctic Basin (they may not be tracking exactly the same area as the ARCUS link) as being close to the 1979 to 2008 average, and indicates that the 2010 minimum of 2.5 million km2 was about 750,000 km2 below the average.

    This is similar to the 2008 chart which shows that the central Arctic Basin returned close to the average level by December of 2007 and stayed close until May of 2008. The negative anamoly increases to .75 million km2 in September when the basin dropped to 2.5 million km2, but returned to the average by the beginning of November of 2008.

    The 2009 chart showed a smaller anamoly of .5 million km2 at the September minimum of 2.75 million km2 with some months coming close to the average ice area.

    I was not able to locate the 2007 chart.

    While the central Arctic Basin ice minimum in September is declining, it is not declining as fast as the Arctic as a whole.

    As for the 2011 prediction, I thought my range of 4.1 million km2 to 5.1 million km2 was pushing the limits of providing a range and calling it a prediction until I saw the 1.8 million km2 range of the 3.73 million km2 to 5.53 million km2 in this prediction “next year’s JAXA minimum, will be 4.63 +/- 0.9 million km2”. The confidence level for this prediction must be close to 100%.

    [Response: It’s a 95% confidence interval.]

    • I suspect Maslowski had the right of it when he said 2016 ±3 years. The biggest uncertainty I see (other than the Dipole being a weather-driven phenomenon) is the warming seen this year in 2010. Despite low levels of TSI and a prolonged La Nina, 2010 will probably be the warmest year in the instrumental record (probably the “missing heat” in the ocean coming back to haunt us). Given that about 2/3 of ice melt is from below, plus record warmth (land + ocean) this gives a scenario for more rapid volume loss than the models predict.

      My WAG (Wild-Ass-Guess) is 2013. That and $1.00 will buy you a USA Today. :)

      If the sun ramps up quickly enough and if a strong Dipole exerts itself, potentially earlier. I wouldn’t rule out 2011 completely, but likelihood has to be extremely low.

      Cryosat-2 data, when released, will open a few eyebrows.

      The Yooper

      • Daniel Bailey:

        The unreliability of applying the trend line analysis used by Dr. Maslowski based on September trends becomes apparent if you apply it to months other than September as it will generate a result that is not believable.

        The following trend line analysis based on volume was done by Neven and shows that the Arctic will be ice free in December of 2023 and will be ice free from July through December. See link below:

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

        I suspect that the rate of future declines in volume are not going to accelerate in the manner predicted by the trend lines that Neven has created and will not follow the straight line paths used by Dr. Maslowski. The Dr. is no doubt a bright guy, butI do not believe this graph provides an accurate prediction of future trends. I suspect that it is far more likely that the rate of decline in volume will slow (in a fashion similar to the concept of diminshing returns). Too many regions have reached a zero volume level at the September minimum. These regions have nothing to contribute to future volume declines. The thick multi-year ice north of Greenland and north of the Canadian Archipelago is substantially diminished and may not have as much to contribute to future volume declines as it has in the past. I do not see that his straight line trend takes these factors into consideration.

        Another reason these trend lines are not believable is that they predict that north of Greenland and north of the Canadian Archipelago will become “ice free” by 2019 and that is not believable. These two areas are likely to have thinner ice in 2019, but there is no reason to believe they will be “ice free” in 2019 given the manner in which the winds and currents drive the ice from other areas of the Arctic toward these two areas.

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

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

        While I appreciate tha Dr. Maslowski is a bright guy, I think the data set he has chosen is flawed in the sense that it can be used to make the extrapolations that he has made. I would prefer that he make such extrapolations using only the ice volume data for the 4.25 million km2 of the central Arctic Basin rather than using the data for the Arctic as a whole. This will likely generate a later “ice free” date than his original projection. I also believe that the trend line should reflect a deceleration in the volume decline rather than a steady rate of decline.

  50. Here’s my prediction. By using the simple heuristics “if it works, don’t fix it”, I maintain that the present negative cycle of Arctic Oscillation will prevail this winter as well and produce a very similar outcome to 2010, complete with a “teaser” bump in the March extent that so enticed certain quarters last season. (Why that did happen is another matter altogether — the bump that is.)

    March-April will also follow in the footsteps of 2010, but with a bit less forte so that June will see an average extent close to 2006. Contrary to 2010, the summer in the Arctic will be a much more conventional one, with the traditional Beaufort Gyre circulation present most of the time. The low extent valley by IJIS will be at 4.15 Mkm2, and the greatest rate of 30-day loss will take place between July 10 — Aug 8.

    Whoever knows best what will happen to the mass of ice pushing through the Canadian Arctic Archipelago will collect the special award for this year, and for some more years as well.

  51. More WAGerry:

    I don’t think that thinner ice will lead to a slowing of ice loss–quite the reverse, I suspect–bottom melt will be thereby enhanced, as will mechanical fragmentation.

    I don’t think that extent decline will be substantially self-limiting, either, despite some proposed negative feedbacks. I don’t see why advection of warmer water into the Arctic should slow, and the albedo feedback will, I think, prove larger than the effects of increased cloudiness, etc.

    Time will tell–but I suspect it will take less than some think.

  52. Daniel Bailey:

    A correction, the trend line analysis I referred to above below was reposted by Neven but is actually the work of FrankD.

    FrankD has been kind enough to provide some additional references that show the decline in ice thickness.

    Per FrankD “[M]aps illustrating the change in old thick ice over the last two years:
    6 Jan 2009:

    6 Jan 2011:

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

    The trend line analysis based on volume (using the PIOMAS model) makes a strong argument that the Arctic will be “ice free” at the minimum in the time frame set out by Dr. Maslowski. However, there is a disconnect when the sea ice extent/area is examined. For example, the decline in ice volume does not appear to have had a significant impact on the sea ice cover for the following regions as they have generally reached their maximum possiblie extent as of the end of December 2010 as they always do – see the MASIE Time Series Plots below :

    Beaufort Sea
    Chukchi Sea
    East Siberian Sea
    Laptev Sea
    Kara Sea
    Canadian Archipelago
    Central Arctic Basin – appears to have reached the maximum extent of 4.25 million km2 earlier than in prior years

    This leaves a couple of months for this ice to thicken. The ice does not sit and melt in place in the Arctic like it does in the Antarctic, which consists of first year ice that melts every year to approximately the same point. In the Arctic, ocean currents and winds dramatically push ice from one region to another (as evidenced by bouy drift). Some of the ice will be transported from the Laptev, Kara and East Siberian Seas into the Arctic Basin as the melt season starts. Ice will be also be exported out of the Arctic Basin through the Fram strait. The balance between the drifting ice, ice that melts and ice compaction determines the amount of ice extent/area in the central Arctic Basin.

    Approximately 80% of the sea ice extent in the Arctic at the September minimum is located in the central Arctic Basin. Observations for 2010 based on the Cryosphere Today plot for the Arctic Basin show it as having 4.0 million km2 of ice area until midway through June of 2010.

    At the minimum, the Arctic Basin had at least 2.5 million km2 of ice in 2009 and 2010 and had 2.4 million km2 in 2008. The wind driven minimum of 2007 reduced the central Arctic Basin to 2.1 million km2. (All figures are per Cryosphere Today plots.)

    Thus, the trend in ice extent for the central Arctic Basin does not appear to match the trend of decline of volume since 2007. As to why extent can be maintained in the central Arctic Basin in the face of the volume declines that have occurred since 2007, I do not have an answer.

    Because of these differences, I am not convinced that the complexities of the forces causing the Arctic ice to disappear can reliably be captured using the volume trend line analysis. If Dr. Maslowski is correct, then the future holds unprecedented and dramatic declines in ice extent/area in the next few years which will wipe out the last 2.5 million km2 of central Arctic Basin.

    Before concluding that the Arctic will be “ice free” at the minimum at some point during 2013 to 2019, I would like to see the central Arctic Basin (as determined by Cryosphere Today) report a few years below 1.5 million km2 at the minimum. Based on the existing history of ice extent observations in the central Arctic Basin I would expect that this area will continue to have ice (although I expect that the extent/area will decline)even at the September minimum through 2019 and I suspect the volume declines will slow and not match the trend lines plotted by Dr. Maslowski or FrankD.

    For now, I will continue to watch the Cryosphere Today Arctic Basin graph. If this region is to become “ice free” in a particular year, I would expect to see it fall to less than 3.0 million km2 by the end of June. If this does not occur, the central arctic Basin will not become “ice free” for a particular year.

    • Going back to an old set of posts: I think Will Crump got too little credit here (especially from Didactylos): however, I still think he is somewhat mistaken.

      Taking into account the geographic nature of the Arctic is a good idea: eg, noting that a given area can’t go below zero area is a good refinement of prediction techniques. Assuming that the trend for any subregion will be constant is not. The reason is that as neighboring regions become ice free, the reduction in more central regions will accelerate, due to both positive albedo feedbacks but more importantly the ability of ice in a remnant region to move into ice-free region down-current, and the lack of regeneration of ice in a remnant region from newly ice-free regions up-current.


  53. Thanks for looking at my post.

    I am not an expert, so I could be wrong, but what I am looking for is data to prove my observation correct or wrong.

    There is PIOMAS model output, but the volume decline may only be measuring the decline in multi-year ice.

    First year ice may be able to maintain itself at the current level for some time with a slow decline rate, as the extent figures for the last three years indicate. The current negative anomaly at Cryosphere Today appears to consist of regions outside the Arctic Basin and regions that do not transport ice into the Arctic Basin, like the Barents Sea.

    What I am looking for is data on the thickness of first year ice in the central Arctic basin. Show me declining thickness in this ice sufficient to result in an ice free Arctic and I will believe it to be possible.

    Until then, I will rely on the ice experts. While Maslowski is clearly an expert, I am not certain he ran his prediction through a peer review process. I am not finding other experts that concur with his view.

    Generally the expert papers I have come across indicate a later date for an ice free Arctic by a decade or two. I have found no info from ice experts indicating the the ice will recover to pre 2000 conditions.