PIOMAS Arctic sea ice minimum daily volume in 2014: August poll

[slow wing]

This poll will run for 10 days. You can change your vote.

PIOMAS volume minima for previous years:
Year --- Min (km3)
1979 16900
1980 16100
1981 12600
1982 13400
1983 15100
1984 14500
1985 14500
1986 15900
1987 15200
1988 14900
1989 14600
1990 13700
1991 13500
1992 14900
1993 12200
1994 13600
1995 11200
1996 13700
1997 13200
1998 11500
1999 10900
2000 11000
2001 12200
2002 10800
2003 10200
2004  9900
2005  9200
2006  9000
2007  6500
2008  7100
2009  6800
2010  4600
2011  4300
2012  3700
2013  5400



Daily volume plots by year...

(Plot and link courtesy of Wipneus.)

See this thread and post for the thickness graphs of July 2013 and 2014 and the difference (also supplied by Wipneus).

This thread has the equivalent poll for last month: July 2014.

[OSweetMrMath]

The daily PIOMAS data for July is available, but not the monthly average. Because I'm impatient, I'm using Wipneus' computation of the monthly average, which is 11900 cubic km. My SARIMA time series model for PIOMAS volume had predicted 11000 cubic km for the July average, so this is a large miss, well outside the 95% (nominal) confidence interval.

This large error produces a large change in all future predicted values, so my predicted value for the September monthly average is now 6900 cubic km, up from 5400 cubic km, with 95% CI of 5900-7800. Note that the current predicted value is outside the 95% CI for all previous predictions this year, starting with the April data.

As I've stated previously, the daily minimum is typically around 100 cubic km below the September monthly average, so my vote is for the 6750-7000 bin.

As an aside, my current prediction for next year's April monthly average is now 23600, which is still below the maximum for any year prior to 2011.

[ChrisReynolds]

2001 to 2013. Take losses from 31 July to minimum for each year, apply to 31 July 2014 value. Exclude 2002, 2004, 2008, as unlikely large losses for this year's weather and ice conditions. Average is 6.7312, which fits with similarity with 2010 (who's losses from 31 July to min project 6.752). Std Dev of projections is 0.20, so calcluated range 6.5 to 6.9k km^3.

Equally likely to be in 6750 and 7000 km3, and 6500 and 6750 km3, bands. Voted for the lower of the two but may range remains  6.5 to 6.9k km^3

[Neven]

Thanks a lot for posting this new poll, slow wing. I hope a lot of people vote. I will vote later today.

[slow wing]

You're welcome, Neven. 

[Steve Bloom]

Vote early, vote often.

[ibelox]

I expect a tie with 2007. Volume loss will catch up a bit first half of August, but as big part of the thick ice seems well protected I do not see how it should go much below 2007.

[Rubikscube]

If the anomaly drops back to where it was on July 1th PIOMAS will almost catch up with 2010 again by September, just saying. Though, I have my doubts that will go through, even with the kind of amazing weather we are seeing right now. I put my vote in the 5,5-5,75 bin, but I really have no idea what will happen during the next couple of weeks because this looks to me like a rather unique situation which makes it difficult to compare with other years.

[wili]

Eyeballing the current uni-bremen concentration map, about 2 million square k is now pure slush ready to vanish with a few days sunshine, which we are apparently about to have. That will leave little outside the little ~1 million circle above 85 degrees, except for the parts toward Greenland and CAA, and a tongue of thicker ice pointing toward Alaska, which itself will probably melt once everything around it is gone, and under the onslaught of the continuous insolation predicted for that area. Hence my lowball estimate. (Hey, somebody's got to cover the outlier case!)

[Neven]

Last month I voted 5000-5250 km3. I'm going up one bin. I would've gone up 2, maybe even 3 bins, if it weren't for that map wili just posted (hadn't done my rounds of ASIG yet).

[realitybytes]

It's being called 'poof' this year, but two years ago it was called 'flash melting'. Will it or won't it?

I'm not even up to amateur rank, and I know conditions are different, but it looks to me like the ESS is lacking but 2 weeks and a GAC away from


Those two weeks mean a lot though...

"Weather," as they say.

[Neven]

To be fair, and speaking as someone who is and was involved with making these scientific definitions better known  , 'flash melting' is when a cyclone goes through an area and lots of the ice disappears in a 'flash', because it melts, or gets washed over by water, or the sensor is confused because of all the clouds etc, and then some of the ice flashes back.

'Poof' is when an ice floe (or floes) are experiencing weather that's conducive to melting. Extent- or area-wise the floes don't show much change for quite a while, but then 'poof', disappear completely in just a few days.

[slow wing]

Voting just about to close... 

[slow wing]

Forty-five people voted and the median bin chosen for this month's poll, taken in July 2014, was "between 5750 and 6000 km^3".


For the second month in a row, the median bin jumped up by three bins and 750 km^3 from the median bin for the poll of the month before.


Last month's update surprised everyone with the comparatively low drop in volume from the month before. This month's update will hopefully be out within the next couple of days. We should be fairly close to the volume minimum and it will be very interesting to see where we are!

[viddaloo]

slow wing, as July set a new record for All–Time Low July volume melt, I expect August 2014 to do the same. I expect August to end up way below the 2000 km³ line for melted ice. That means only one poll participant will say yes, I told you the minimum would be «more than 7000».

[viddaloo]

I expect August to end up way below the 2000 km³ line for melted ice. That means only one poll participant will say yes, I told you the minimum would be «more than 7000».

Clearly, I was wrong, and there was a much bigger melt in August than I anticipated. Yet, with a total melt of 2355 km³, 2014 had the second lowest August melt ever, after 2013's 2221 all–time low.

[OSweetMrMath]

As I said earlier in this thread, my time series prediction for July had a large miss, with the actual value coming in 1000 cubic km above my predicted value. But this month, my corrected August prediction was dead on. The difference between my prediction and the measured value of 8150 amounts to a rounding error.

As a result, my prediction for September is essentially unchanged. The current prediction is 7000 cubic km, with a 95% confidence interval of 6500-7500 cubic km. The difference from last month's prediction of 6900 is due to rounding. I don't make an exact prediction for the daily minimum, but the best category in the poll for the daily minimum is still 6750-7000 cubic km.

[Richard Rathbone]

Actual was 7.22

[ChrisReynolds]

OK.

Regressing loss from 31 Aug to min against volume at 31 Aug, R2 is 0.06, little relationship. i.e. how low volume is has negligible impact on the loss from 31 Aug to min.

So what do the wider stats look like? Average loss is 0.267k km^3, stdev is 0.124k km^3. Interestingly 2008 stands out, a -6 sigma event! Losses seem to be a bit lower after 1997, the average for 1997 to 2013 is 0.305, stdev of 0.147.

Using that latter period and +/-1SD I get; 7.06 to 6.77k km^3 for the daily minimum.

But the nearer you walk to the bucket the less skill there is in dropping the ball into it. 

[iceman]

I'll hazard a guess at a relatively late volume low of 18th Sept, owing to two successive cyclones (13th Sept near Svalbard and 17th north of Greenland) that will hit some of the thicker ice. 
    That would shorten the odds of a volume loss for the month of September, an atypical occurrence but I think it will happen this year.

[viddaloo]

piomas14prog: -144 (-846) — 6374 km³

If my calculation algorithm is fairly correct, we've just set a new all-time high for September melt: 846 km³, beating the 2008 record and continuing and strengthening the trend of gradually bigger September melts.

[viddaloo]

Looks like 6810 km³ was the yearly minimum volume for 2014.

[Neven]

Looks like 6810 km³ was the yearly minimum volume for 2014.

Last month I voted 5000-5250 km3. I'm going up one bin. I would've gone up 2, maybe even 3 bins, if it weren't for that map wili just posted (hadn't done my rounds of ASIG yet).

Wow, I should've gone up 6 bins!

[OSweetMrMath]

The monthly average for September was 6973 cubic km. This is essentially indistinguishable from my prediction based on the July data.

The bad news for my predictions is that my prediction based on the April data was 4700, with 95% CI of 3100-6300. This was a clear miss. Things look better going month by month. The April prediction was fairly accurate for May and June, but badly missed July. The July prediction was extremely good for August and September. I'm looking forward to seeing how long the prediction will continue to be good.

Speaking of, now that we've reached the minimum, it's a good time to predict the maximum. My prediction for the monthly average in April is 23700 cubic km, with 95% CI of 21800-25500 cubic km. The current prediction for the next September monthly average is 5700, but I'm in no way convinced of that prediction's reliability.