Seven Months Ago

Anthony Watts is taking a poll of predictions for this summer’s minimum in Arctic sea ice. Some of you might want to join in the fun.

There’s an interesting comment from someone using the monicker “Blade”

Blade says:
May 20, 2011 at 12:56 am

Almost exactly six months ago there was an earlier thread here about this very subject …

Joe Bastardi went early in for 5.5 Mkm^2. For what it matters, I will go with him.

It’s funny to see some of the climate trolls finally mustering up enough courage to take a stab at it now, six months later. In that thread we find R. Gates bloviating profusely and criticizing JB but failing to make a single prediction. Did I miss it?

And I see in a comment above that someone says that Tamino Foster Grant finally climbed out on a limb with a prediction now, six months later. Way to go G-Man.

Perhaps some of their groupies might also report on any daring predictions from Hansen, Mann, Serreze, Walt Meier, and Julienne Strove. I’m sure it would qualify as a taxpayer funded activity, so what’s stopping them? I mean, spend the money while you still can.

P.S. did I mention yet that Joe Bastardi made a prediction six months ago?

I made my prediction seven months ago. It’s here (near the end of the post).

I get the impression that Mr. Blade doesn’t think very highly of me. Or of James Hansen, Mike Mann, Mark Serreze, Walt Meiter, or Julienne Strove.

It’s flattering to be in such good company.

P.S. Did I mention yet that I made my prediction seven months ago?


43 responses to “Seven Months Ago

  1. How many months ago did you make yours? How many months earlier than Joe Bastardi was that?

    Also while reading that post I went back to see exactly what his prediction was and he said

    My forecast for next year is for sea ice to melt only to levels we saw back in 2005, or 06. If I had to put a number on it, I think it would be around 5.5 at its lowest.

    The ice is coming back, will do so in forward and back steps, with forward defeating the back steps. I am on record as saying we will be back to 1977 levels by 2030.

    • Wow, Joe Bastardi is so hardcore. I don’t think anyone has the balls to make predictions like him. He is on the record!

      • Well, let me go on the record to say he’s an idiot. We’ll be lucky to have significant summer ice at all by 2030.

      • TrueSceptic


        Make that 2020!

      • I know what you mean, TS, but I don’t think I get to change the terms he chose in which to couch his idiocy!

        But FWIW, I would be not at all shocked (though I would/will be disheartened) to see Dr. Maslowski’s aggressive projection realized–the last year for that would be 2019. As I wrote elsewhere, we’re already 7/8s of the way there.

      • TrueSceptic

        Agreed, on both counts.

  2. The WUWT sea ice forecasting poll allows readers only a fixed set of possible answers (from “Greater than 6” to “Less than 4”) that centers on 5, or just about last year’s value (4.9). I guess there’s a strategy in that, since predicting a “nice recovery” worked so poorly last year.

  3. I’d like to know what Anthony Watts’ forecast is, but he’s probably afraid to speak out after last year’s fiasco with Goddard.

  4. For what its worth, based on the monthly sea ice extent data at NSIDC ( and the daily sea ice area anomaly data at Cryosphere today (, my estimate for the September sea ice extent is:

    Sept Extent = 1.5274*(May_Area_anomaly)+6.63 (R^2 = 0.62)

    As of 19 May 2011, the May Area Anomaly at Cryosphere Today is -0.9825, so my prediction is: 5.1 +/- 1.1 million km^2.

  5. It’s also worth pointing out that the correlation coefficient (R) of May and September sea ice area is about 0.770 and for consecutive Septembers (e.g. between September 2010 and September 2011) is 0.769. In other words, the extra information you get during the winter and spring doesn’t really tell you very much about what will happen in the summer.

    So even if Blade’s comments were true, they would still be pretty much irrelevant.

  6. I’ll stay at exactly 4’000’000 km^2 (this time with no error bars) like two last years, it’s bound to happen one of these days (to talk about faith :-)).

  7. TrueSceptic

    My guess is that this year’s minimum will be very close to that of 2007, possibly a new record, so I’m talking about 3.8-4.8 M km^2. This guess would’ve been the same had I made it 6, 7, or even 8 months ago. It’s where the trend leads us.

    Bastardi is guessing that that this year will be similar to 2005 but I think he’s wrong. He seems to believe that 2007 will prove to be the long term minimum and that we will return to 1977 levels by 2030. I can see no reason for such a belief.

    • Global Cooling! It has to be true or otherwise April, instead of being a chilly 4th hottest April in the record, would instead have been like 3rd or higher.

      We’ve bifurcated off a shifty moment in chaos. I’ve read all about it.

  8. Sharper00’s point is a good one. “5.5 at it’s lowest.” I take this to mean that this is the lower bound of the prediction. Tamino’s 4.63 plus or minus 0.9 million prediction is essentially non-overlapping. I think it’ll be 5 million or more owing to the gambler’s fallacy that one of these seasons the weather conditions will be anomalously good for retaining ice. But if that happens, and the winter starts off with greater than the recent amount of ice at the end of the melt season, watch out for next year. While the deniers will be claiming recovery, the extra heat insulated and stored under the ice will be likely to come through next summer.
    I wonder how much data one would need to find support for such a model (in which the residual from the last year was negatively correlated with the coming year’s anomaly)?

  9. I put in my wild guess at 4.6-4.8 million km squared (the range is a typo – the rest of Watts’ table is in increments of 0.1). A fluke, but I matched your prediction, Tamino. It’s one of the least popular choices (2.95% / 15 votes). Most popular choice over there is ‘greater than 5.5 mil sq km’ (16.7% / 85 votes).

    Saved the WUWT post for a September reminder.

  10. Tamino, I had forgotten that you had made a prediction last year for this one. So I redid your calculations using NSIDC September extent. That is, minus the error bars as at the moment I am not sure how I would do that for a quadratic trendline fit at the moment, but perhaps I should treat that as a research assignment. And yes, I got 4.63, or 4.6265 million sq. km.

    I noticed, however, that this gives us 176,000 sq. km for September 2029 and has us “going negative” for the first time in 2030. And yes, I know quite well that extrapolating that far is quite unwarranted.

    However, FrankD over at Neven’s calculated quadratic trendline fits for monthly PIOMAS Arctic sea ice volume. As I understand it, PIOMAS estimates are likely the best synthesis of available data that we currently have, that is, until CryoSat data is made available. And FrankD’s quadratic trendline fit for September volume has us going zero around 2016.

    Given this, I was wondering whether you or anyone else has any thoughts regarding what we should expect for September sea ice extent over the next few years.

    [Response: As sea ice declines, I expect we’ll get more into the “nonlinear” range of physical processes. Therefore I don’t expect purely statistical methods to be able to anticipate when the final collapse will occur. That said, I expect we’ll need good physical models to make predictions farther into the future. But, I don’t think physical models are yet sufficiently advanced to do so.

    Bottom line: too early to tell.]

    • Timothy (likes zebras)

      For one thing I would have thought that the last million or so square km of sea ice would be a bit more stubborn, as that’s thicker ice that’s been pushed up against the northern edge of Greenland.

      So I think that one would be better off trying to predict the first sub-1 million square km year, rather than trying to predict a zero sea-ice year.

      • Didactylos

        There are signs that the ice may not behave in the way expected from ice models. The thick old ice has been characterised as “rotten”.

        However, I expect that the call of “ice free” will be made long before there is actually zero ice in the entire Arctic basin. The first call will be when the North Pole is significantly ice free, the final call will be when the Arctic is fully navigable.

      • It’s worth remembering the terms of Dr. Maslowski’s projection:

        “By “ice-free,” Maslowski tells me he means more than an 80% drop from the 1979-2000 summer volume baseline of ~200,00 km^3.”

        –Joe Romm

        From the PIOMAS site:

        “Monthly average Arctic Ice Volume for Sept 2010 was 4,000 km3, the lowest over the 1979-2010 period, 78% below the 1979 maximum and 9,400 km3 or 70% below its mean for the 1979-2009 period. ”

        So we’re something like 7/8ths of the way there, and up to eight years to get the rest of the way before we’re outside the error bars (the time frame was 2016 +/- 3 years.) Dr. Maslowski was ridiculed by the usual suspects, but his projection today doesn’t look unreasonable at all, IMO.


  11. So Guys and Galls, what have PIOMAS and other Ice volume reconstructions got to indicate, not to speak of the multitude of other imagery available where 2011 is going. Watch the North Pole webcam… near realtime photographic feed… will they be standing in meltpools, and all that snow gone missing early surrounding continents ready to warm up… my bet is lower than 2007, and I bet very very rarely.

  12. Regarding Bastardi’s prediction for 2011 sea ice extent, Steve L. wrote:

    Sharper00’s point is a good one. “5.5 at it’s lowest.” I take this to mean that this is the lower bound of the prediction. Tamino’s 4.63 plus or minus 0.9 million prediction is essentially non-overlapping.

    I read Bastardi’s “lowest” as meaning simply the sea ice extent minimum for 2011. He isn’t qualifying it as the daily or monthly average and he isn’t giving you the data source. (Tamino is predicting the daily JAXA minimum.) Bastardi doesn’t qualify, and therefore I don’t think he even considered putting error bounds on his estimate. He is flying on gut instinct, the same gut instinct that has him stating:

    The ice is coming back, will do so in forward and back steps, with forward defeating the back steps. I am on record as saying we will be back to 1977 levels by 2030.

    Staggering around in “forward and back steps” seems to be about scientific as Bastardi gets.

    • To be fair, Bastardi was the only pseudo-skeptic who didn’t predict another recovery last year. He made a lovely series of goof-ups in his weekly videos, implying among others that NSIDC was committing fraud whereas he messed up by mixing the colours of the trend lines (that was pretty amazing), but he never predicted a recovery like Watts & Goddard did, until Watts threw Goddard under the bus just as their prediction was increasingly and visibly wrong.

      Bastardi is the only pseudo-skeptic who has a iota of credibility left when it comes to Arctic sea ice. But just that, a iota.

      BTW, has anyone heard what Watts’prediction for this year is? Still nothing? God, he’s so courageous.

  13. Michael Hauber

    My bet is that the Watts posting will trigger a flood of submissions to the sea ice outlook that are as ridiculously on the high side as that guy who predicted 1m last year was on the low side. Then in October Joe Romm gets to say that the Arctic Death Spiral is even worse than the ‘scientists’ predicted.

    Perhaps I’m being overly cynical today..

  14. Michael, you’re being overly cynical today. Have you read the sea ice outlook reports, to see what they actually are?

  15. Michael Hauber

    I read all the submissions from last year. In particular I refer to the June public submission from Wilson, which predicted 99% of people in the Northern Hemisphere would die after winds caused by a nearly ice free Arctic.

  16. Off-topic but there’s no open thread on the front page … this thread by Eli Rabett:

    would indicate to me that you might want to take a look at the Fall et al statistical work (that’s the result of Watts surface stations photo hobby project).

    Eli and John N-G are engaged in a lively discussion there which might benefit from some statistical rigor …

  17. I’ll use the average of all semi-expert’s predictions as my prediction.

  18. There is a reasonable chance that this year the NW Passage may be open in time for the Heartland conference. Already ice is thinning round the NW edge of Greenland and along the north Alaska coast. Only the Canadian archipelago seems to stand in the way.

    Give ‘em something to chat about if it is open.

  19. [Tamino: As sea ice declines, I expect we’ll get more into the “nonlinear” range of physical processes. Therefore I don’t expect purely statistical methods to be able to anticipate when the final collapse will occur…

    Bottom line: too early to tell.]
    Quite so – and this is why I have tried to emphasise that a quadratic fit is not predictive, only that it fits the existing data better than linear projections that I have seen around.

    To my way of thinking, the processes are partly dependent on volume (eg: how much heat is consumed in melting ice) and partly dependent on extent (eg: insulating effects of ice cover, albedo effects etc). Projections using volume reach zero far sooner than projections using extent, and I suspect reality will give a result somewhere in between.

    I think it is impossible at this stage to determine the relative weight of volume effects compared to area effects, at least for a mug punter like me.

    It is too early to tell when the final collapse will happen, but I think that by the time we can tell, it will have already happened…

  20. too easy, I offer another prediction test : at which time Watts and consort will mysteriously forget to follow Arctic Ice and speak instead about Antarctic ice / Climategate / the Great Satan Mann ?

    I’d venture June.

  21. The last bet on Intrade puts the chance of the lowest 2011 ice extent being lower than 2007 (i.e. the lowest on record) at 66.7%.

    What was 2007’s lowest extent?

  22. MightyDrunken

    Purely for a bit of fun I predict* 4.8 Million sq Km.

    Now Nature will most likely prove me wrong.

    * Guess with eye balling plus a bit of “optimism” preventing me from going lower.

  23. For fun, took first 5 months of daily Arctic SIA data off Cryosphere Today, covering years 1979-2011 and computed the average anomaly and this chart rolled out. First 5 months and we have the lowest coverage since beginning of the CT record:

    JAXA in ranking shows 2011 in spot 2 for June 1, just another day in weather terms, and it’s 525 km square more than the previous record low that occurred same day in 2010. Global cooling I suppose ;o)

    NB, ignore the trendline. Somehow managed to select log, when linear was meant, it though barely makes a difference.

  24. Was reading the ARCUS June Update forecast also featured at NOAA on predicted SIE low for 2011 and was interested that the poll WUWT took showed a median of 5.5 million KM square [EXTENT]. Second most optimistic, GW Ain’t True [Crowd] consensus. Since JAXA dipped below 10 million yesterday, did a little chart fetching the NSIDC/SIDADS 1979-2007 and JAXA dates 2002-2011 of the day of year and plotted it out… Seems to come earlier and earlier. And the funny thing, of course, there was a vehement hand waver when I compared the monthly data of the same 2 sources and SIDADS was dismissed as showing higher acreage than JAXA. The day of year for the overlap suggests to me that SIDADS is lower than JAXA up to 2007 for Extent averages. Yes, JAXA is favored at where I do not wish to add clicks for Watts credit.

    Don’t know, but to think that WUWT was included to make a point, later this year, unsaid of course ;P

  25. I’ve just written a blog post on the results so far of the poll held at the Arctic Sea Ice blog. Towards the end I have a bit about what could be the ever consistent Anthony Watts, though it probably isn’t him.

  26. Watching the Arctic Webcam, looks like the snow is starting to resemble a soggy state [June 19]. Also notably, the cracks and piling visible in both NOAA1+2. These developed weeks ago [Sign of thinly state?]. Last year part of one of the stations threatened to drift off “On Camera”, but then was pushed back. This NOAA page has an interesting summary of meltponds since 2002:

    Of lesser interest maybe, but the 365 day Global rolling average has just dipped to where no other year but 2007 descended. Area has a negative anomaly of > 1M KM Square. Source data:

    All this under perpetuating sluggish sun.

    PS on snow, the Rutgers daily shows extensive shortfall, such as the full stretch of Norway. They had serious flooding [YouTube]. Nova Zembla/Barentsz Sea region is very low on Sea Ice too. Is this the near record hot Mexican Golf Stream water arriving?

  27. First meltpond observed by Polar Webcam on June 20, 2010.

    A recent history suggests no earlier observation on record:

  28. PS: Could be Anthropogenic too as it centers where the windmeter was set up… flattened snow. :D