Betting

Ice Record Getting Crushed.

28 Comments

Things are looking bad for those of us who posted “moderate” bets. My best estimate range for the September NH ice extent based on the most recent JAXA is 3.28 million square KM. The range is 3.02 to 3.55 million square KM. (Note: the computed uncertainty intervals would be larger if I used the more complicated method I used last year because that accounted for the possibility this statistical model could be wrong and other statistical models might be better. Also, if I recall correctly, fitting based on current extent was not the best method earlier in the year– but for obvious reasons did become the best method as we approached the actual minimum.)

Still, for how, here is the graph:

As you can see, even the upper value for expected range is well below the previous record. (Also, if some recall, Rob Dekker has a real money bet with William Connolley. I suspect Dekker will think he’s sitting pretty at the close of 2012!)

As for me and my quatloos: I’ve resigned myself to the idea that I’ve already lost. This afternoon, I’m going to have to gin up the histogram so we can see who remains “in the running”.

28 thoughts on “Ice Record Getting Crushed.”

  1. Amidst the ice betting (i haven’t participated) and record warming metrics this summer I thought I’d try something different (perhaps this has been done) – a daily record of earth heat content. I was just interested in the relative magnitudes of the different seasonal phenomena that we can actually measure to some accuracy. Heh, maybe I should get a blog and update this regularly.

    Starting with EXTREMELY crude calculations for TLT and Arctic sea ice – here is my first take. Conclusions first then method. It shows more of a steady heat gain than you get from just looking at TLT, proving global warming hasn’t stopped ;). It looks possible that not likely that 2012 will be the “warmest year ever” beating out the 2010 El Nio year.

    CRUDE method:
    *Used Aqua history for TLT, nothing more
    *Energy content calculated by mass of atmosphere 5.15E18 kg, heat capacity 1 kJ/kg
    *Used PIOMAS sea ice volume (daily) for same period, heat of melting 334 kJ/kg.

    huh.

  2. “It looks possible that not likely that 2012 will be the “warmest year ever” beating out the 2010 El Nio year.”
    Should “that” be “but”?

  4. Lucia,
    For what it’s worth, I still think that you are in with a chance of being close.
    The latest figure of 4.001 million square km, at August 27th, which will probably be revised upwards, is in line with the mean 2008-11 mean daily change *1.5, which produces a September mean of 3.59 million square km.

  6. Still no confirmed Northwest Passage though. I doubt long term records are being crushed. What year do the records (using the current methodology) that you are using go back?

    BTW, the average(sum/number of bets) for your late bets was 3.8544 based on the post a few days ago. Let’s see how crowd sourcing does.

  7. I dunna Ray about Lucia’s (and my) chances.

    The latest North Pole data actually shows a mild warming to +0.4°C (well about the –2°C needed for sea water to reeze), continued overcast conditions, and now the fresh water melt ponds on the surface of the ice look to be expanding.

    The situation looks grim for the luke-melters.

  8. I bet early and let it ride. It will take a pretty big volcano so save those quatloos now! Maybe an asteroid strike…

  10. Carrick,
    You are probably correct, my estimate is based entirely on statistics and not reality!
    It all depends on whether or not the temperatures produces a decline rate faster than the 2008-11 mean +50%!
    Also, if I understand it, Lucia’s graph assumes a linear relationship, which may not be the case.

  11. Maybe this is a good time to buy Arctic ocean waterfront property for future resort hotel.

  12. Also, if I understand it, Lucia’s graph assumes a linear relationship, which may not be the case.

    Yes. It assumes a linear relationship between Jaxa extent today and September Average NSDIC extent. This may not be the case.

    There are all sorts of things that happen when there is less ice. More room to move around, more broken up. Generally thinner all around etc.

    I also don’t have metrics like ice volume, area etc. which I had last year. Oddly, my method last year used this in a way that resulted in wider error bars– not narrower ones. So, it was the opposite of “over-fitting”. We’re close enough to the minimum (I hope!) before I had a chance to get to plotting all these things that it’s not worth firing everything up.

    I dowloaded the data but withouth checking for duplicate to prune out over-ridden bets and have a graph that shows my bet, the mean bet and the ±1SD . (I did dump the negative bet. )

  13. It’s amazing, when I look at the graph on the screen, I would swear that the central line isn’t pointing to the red circle, but when I checked it with a ruler, I confirmed it was. Probably some sort of optical illusion.
    Personally I don’t think the relationship is linear, especially if you exclude the largest (presumably oldest) values. There would only have to be a slight deviation from the linear to make the central prediction a lot higher.
    But I am probably “teaching my grandmother to suck eggs”.

  14. Ray–
    There is no particular reason for the relation to be linear. I didn’t try others. But the record I used doesn’t have a huge number of points. It’s a method. It’s a projection. It could be totally wrong.

    Last year, I had more metrics, so the ones with the best AIC got picked out This year… too lazy!

  15. Aslooo…. This is a big extrapolation. The current Jaxa extent is a lot lower than the previous extent at this time of year. This is easier to see on this type of graph rather than the graphs showing the time 365 day time series where your eye can be tricked into thinking it’s not that bad because the current value isn’t that far below the minimum ever. The ice level is low. As I said: I didn’t prune duplicates so some of the older “high” bets are in and I need to take them out. But currently, the mean bet is above the best estimate.

    Tomorrow I’ll also show the bets we placed back in June. We can see how wrong most of us were….

  17. Lucia
    For comparison of record extents of sea ice:

    . . .our records would correspond in the model to an Arctic Ocean sea-ice cover in summer at 8 ky B.P. that was less than half of the record low 2007 level. The general buildup of sea ice from ~6 ky B.P. agrees with the LOVECLIM model, showing that summer sea-ice cover, which reached its Holocene maximum during the LIA, attained its present (~2000) extent at ~4 ky B.P. (fig. S3)

    Science 5 August 2011: Vol. 333 no. 6043 pp. 747-750 DOI: 10.1126/science.1202760
    Cited by The Hockey Schtick

  19. David L. Hagen (Comment #102240)
    August 27th, 2012 at 3:17 pm
    ————————————–
    The sea ice summer minimum graph from LOVECLIM as shown on the Hockey Schtick (parameter set E4) shows in increase from 2.5 to about 7E+12 m^2 over a period of 7000 years. The current drop from ca. 7500 down to less than 4000 has occurred in ca. 30 years.
    To me, that time frame is staggering and points to a strong forcing.

  22. Lucia,
    No criticism intended, I was just speculating.
    My own “gut” feeling is that the forecasts will start to increase as we get nearer to the minimum and rates of decline continue to fall.
    The figure for the 27th has been revised upwards and shows the smallest decline since August 2nd.

  23. diogenes (Comment #102247)
    August 27th, 2012 at 7:50 pm
    but does it matter on a planetary basis, Owen?
    —————————————————-
    At the minimum we will see an increasing positive feedback in the Arctic Ocean arena due to the loss of albedo in the summer months of ~24 hour sun, causing continued arctic amplification. The additional absorption of energy in the arctic will contribute (in a small way) to the overall energy imbalance for the planet, but every little bit adds to the integrated heat content building up in the world’s oceans.
    What the additional heat in the arctic will do to the oceanic conveyor currents, I do not know for sure, but I would think there is a good chance they could be disrupted. Also the rate of melting of the Greenland ice pack could well be accelerated by a continually warming arctic (as it currently is being affected).
    furthermore, more oil will be suddenly available in the arctic which produces (in a way) a positive feedback as there will be more oil to burn and more CO2 to be emitted.

  24. Owen:

    At the minimum we will see an increasing positive feedback in the Arctic Ocean arena due to the loss of albedo in the summer months of ~24 hour sun, causing continued arctic amplification

    The problem with that theory is that most amplification occurs in the winter months when the ice sheet is fully covered over. There may be mechanisms by which ice loss affects polar amplification, but it’s more complicated that this (or this isn’t the explanation, put it that way).

    What the additional heat in the arctic will do to the oceanic conveyor currents, I do not know for sure, but I would think there is a good chance they could be disrupted

    There is very little evidence or even theories that aren’t just wildly speculative that suggest that ocean circulation will be “disrupted” by Arctic Ice loss. Modified in a way that can be measured.. probably. Disrupted… little to no chance.

    Also the rate of melting of the Greenland ice pack could well be accelerated by a continually warming arctic

    Again see comment that most of the amplification occurs in winter.

    That actually speaks to increased wintertime precipitation in Greenland, which of course means more snow accumulation, not less.

    The main argument in your favor you left out, which is the ice piles up on the north-wiest side of Greenland’s coast. Expose the coast to above freezing temperatures of open water, and you’ll get more melting (at low elevations).

    Of course this is what is expected by the modelers… a loss of ice at low elevations, and an increase at high elevations.

    When you have balancing forcings as with Greeland, you have a rather sensitive system. Small perturbations could swing it one way or the other in terms of net ice loss or gain. (Black carbon comes to mind as an example here.)

  25. Carrick,

    Actually, the greatest amplification on your graph looks to be for sept-nov, where heat accumulation in the summer-warmed arctic ocean is released back to the atmosphere as both sensible and latent heat. The warm jun-aug time period (see http://ocean.dmi.dk/arctic/meant80n.uk.php ) is always right around 273 K (melting point of ice) and moderated at its warm maximum by the ice (even though heat is flowing into the arctic ocean itself). Do I have that right?

    –Owen

  27. Owen thanks for the link.

    They explain the mechanism for fall/winter amplification at least, which is increased heat loss due to less ice being present in the fall and winter. It’s not an albedo effect, except in so far as it accelerates ice loss.

    With less sea ice, the refrigerator door is left open: more dark open water is exposed, which readily absorbs the Sun’s energy in summer, heating the ocean and leading to even more melt. With less sea ice there is also less insulation, so that heat from the ocean escapes to warm the atmosphere in the autumn and winter

    In summer, maybe the extra heat energy being absorbed is going into melting the ice, instead of warming the air? (Seems logical.)

Comments are closed.