Sunday, July 05, 2009

Arctic sea Ice pool: 4 weeks to go

With 4 weeks left to guess, the graph for the 2009 Arctic sea ice minimum gaussian guessing game is looking like this:


Contestants have solid, colored curves. The thick black/grey curves are the Ensemble 1 and 2 outputs from the Wegener Institute’s June 2009 Sea Ice outlook. The collective contestant’s pdf is the dotted light grey line (click to embiggen). It has grown from a bimodal distribution around the 2007 and 2008 minima, to a trimodal distribution with a third peak around 4000. Nick Barnes (4700 ± 200, pink) still needs to take another guess, if he so desires.

The current value (as of Friday) is 9,500 thousand km2.
at

9 comments:

  1. Divalent05 July, 2009 10:42

    5325 75

    ReplyDelete
  2. Nick Barnes06 July, 2009 07:27

    Does 4700 +- 100 get me back in the game? Can't be bothered to do the sums. If not, I'll go for 4700 +- 75, or 50.

    ReplyDelete
  3. C W Magee06 July, 2009 11:14

    Of those options, your best bet is +/- 50, so I'll put that in unless you object.

    ReplyDelete
  4. Anonymous07 July, 2009 05:27

    Can you put me down for 4000 +/- 100

    Cheers, Alastair.

    ReplyDelete
  5. Robert Grumbine09 July, 2009 21:36

    Chuck: Thanks for the invite (over at my blog). But I'm much more confident about monthly averages than the minimum (which, I'll caution, sometimes happens in August, rather than September). One good storm system can shove the ice pack around pretty well, and destroy a lot of thin ice (mixing up the warmer water from below).

    For monthly average, I'll go with the 4920 +- 500 I submitted to ARCUS.

    I'll also put a non-gaussian step function of 'strictly less than 4920' (thousand km^2) for the minimum day(s) from NSIDC. Say uniform between the 2007 minimum and the 4920.

    ReplyDelete
  6. Brian Dodge11 July, 2009 08:01

    4325 +-25

    College dropout - art/ceramics
    Rockhound
    described as a "triviologist" by a geologist friend

    ReplyDelete
  7. C W Magee11 July, 2009 23:48

    Just a reminder that all guesses must be Gaussian, just to keep the playing field level. Or bell-curved. Or whatever. And I think that step function is mathematically eliminated anyway.

    ReplyDelete
  8. Unknown06 July, 2010 04:15

    I'll do 4250 +- 110.

    ReplyDelete
  9. Unknown06 July, 2010 04:17

    I will guess 4250 +-110. The 4250 is a result of ruler on graph analysis of the deviation from average, the deviation is slightly larger than the standard 100.

    ReplyDelete