As mentioned yesterday, the ice extent recently experienced losses at daily rate in the upper 10% of those observed in the combined JAXA/GSFC record. The event was sufficiently extreme that I waited until CT reported areas before posting the current extent, the current best forecast based on my weighted model and announcing the which bettors appear to have taken the lead. Based on my model forecast Steve T and Chris Brown are tied! I don’t know which got his bet in first, but numerically they are neck and neck:
Meanwhile, my current best estimate for the minimum is 4.55 million square kilometers, with a ±2σ range from 4.12 to 4.99 square kilometers:
Reading discussions at Neven’s blog I was motivated to try to include the most recent ice volume levels reported by PIOMASS in the fit. However, so far none of the single or multiple regressions involving either volume or (volume/ct) area have passed my criteria for including those predictions in the weighted model I use to creat forcasts. Some of the regressions would have passed my criterion for AICc having AICc’s suggesting they are at least 1/20th likely as the most like regression of those I tested; in fact some appear quite likely. However, of those all had at least on regression coefficient that was not statistically significant so they are rejected as candidate models for my weighted model. Had I ignored the lack of statistical significance in the regression coefficients, my best estimate for the upcoming extent minimum would be 4.51 million square kilometers and my uncertainty intervals would be larger than those associated with the model used to create the current forecast. I will be looking at regressions with volume a bit more because it seems to me that volume or thickness may somehow matter, but it may be that something like “volume^ 3/2” would result in better predictions. (Note: volume ^3/2 has units area; that’s why I’m going to look at it. )
Next week we’ll see if a new leader has pushed SteveT or ChrisBrown out of the lead. It’s hard to predict who is going to win the quatloos. This is quite a horse-race.
Here is a link to the
July 6August 1 post that lists the entries. (Lucia closed out the pool with 4.569.) Chris Brown’s bet of 4.55 was quite a bit earlier than SteveT’s wager on the same. With 5 and 4 quatloos, respectively, on the table, there won’t be many leftovers for us also-rans… assuming the horses keep this order ’till the first nose touches the tape.To mix metaphors.
CT area had a big drop in the most recent update. With over two weeks to go to the minimum of the average, Arctic ice area at 3.0651 Mm² is already the third lowest in the 32 year record. The average loss to minimum gives a projected minimum of 2.73 Mm², far below the 2007 minimum of 2.919 Mm². Unless the area takes a fairly big jump upward, 2011 will be below 2007 for the same day tomorrow.
DeWitt–
There’s a good chance for a CT area minimum. That’s for sure. I haven’t been applying my forecasting algorithm to that– but it does look like just average melting would do it.
Lucia,
From your latest numbers what are the chances this years minimum extent will surpass 2007?
HR– The chance the upcoming minimum for the 7-day average will be lower than 2007 is 11%. The probabilities are in the lower right hand corners of the figures.