Poll: PIOMAS Arctic sea ice minimum daily volume in 2013

[slow wing]

This poll will run for one week. You can change your vote.

PIOMAS volume minima for previous years:
Year --- Min (km3)
1979 --- 16,900
1980 --- 16,100
1981 --- 12,600
1982 --- 14,000
1983 --- 15,100
1984 --- 14,500
1985 --- 14,500
1986 --- 15,900
1987 --- 15,200
1988 --- 14,900
1989 --- 14,600
1990 --- 13,700
1991 --- 13,500
1992 --- 14,900
1993 --- 12,200
1994 --- 13,600
1995 --- 11,200
1996 --- 13,700
1997 --- 13,200
1998 --- 11,500
1999 --- 10,900
2000 --- 11,000
2001 --- 12,200
2002 --- 10,800
2003 --- 10,200
2004 --- 9900
2005 --- 9200
2006 --- 9000
2007 --- 6500
2008 --- 7100
2009 --- 6900
2010 --- 4400
2011 --- 4000
2012 --- 3300
(Values from 2003 copied from Volume Predictions thread by Dave C; private communication from Wipneus for values prior to 2003.)

PIOMAS Volume at 31 July 2013: 7104 km3
(Reference: Latest PIOMAS Update thread by Wipneus)

Daily volume plots by year...

(Plot and link courtesy of Wipneus.)

[Glenn Tamblyn]

Basically I voted on the assumption that HYCOM thickness isn't totally crap.

[ChrisReynolds]

About 4.4k km^3, which is basically 2010's value, because 2013 anomalies have been close to 2010 since about the solstice, 21 June.

[slow wing]

Steven, thanks for your comments on the binning. It turned out though that the votes are spread over quite a few bins, so I'm now comfortable with the granularity chosen. I always prefer to err on the side of a wide enough range that few, if any, voters are relegated to the underflow or overflow bins.

[Steven]

Slow wing, thanks for the interesting poll.  I removed my comment on the binning (the suggestion to use step size 150 km3 instead of 250 km3) just before you posted your last post, since there were already many votes being made in the meantime, so it seemed too late to change the bins.  BTW, it appears that the vote results are only visible once you're logged in and made a vote.

My vote is 4250 - 4500 km3.

[icebgone]

I wanted to believe that we could best 2010 but it just doesn't look that way right now.  Should be close though.  I expected a range from 4425-4575.  the middle is 4500.  I took the slightly more conservative bin.  I might change my mind if a new ice mixer shows up in the next 4-5 days. 

[Alistair]

I have gone for 4000-4250.

ChrisR is normally on the money but my eyes are telling me that while the Extent and Area are higher than last year, there is not much multi-year ice left and an awful lot of that extent is very thin and mushy.  Something similar to 2011?

[epiphyte]

My guess of 3.25-3.5 (and I'm happy to admit that a guess is exactly what it is) is playing a hunch that :

• a) The models aren't used to dealing with large expanses of relatively uniform FYI,
• b) when uniform ice gets thin enough to start breaking up over a large area, the dynamics of the interaction between ice, sea-water, fresh water, wind, waves, currents, etc are likely less well understood  than in an environment where the ice/water boundary is relatively stable, as has been the case until recently, and
• c) It seems to me that if there are mechanisms in play that are not accounted for in the models, they are more likely to accelerate the melt than retard it.

Whilst recognizing that the third point above is purely speculative, one observation I'd make in support of it is that during the time I've been lurking I've seen several references to "rotten" ice, which doesn't seem to be rigorously defined. I have, however,  been poring over wordview long enough that  think I recognize it when I see it, and there seems to be an awful lot of it around right now.

One question I have in connection with this is,  is it reasonable to assume that fractured, rotten, uniform sheet ice will still show up as 100% extent, >80% area, and (again, at a guess) 0.5-0.7M^^3/M^^2 volume, even at the point where it is so fragmented and low in the water that seawater is starting to wash over it and its 'bottom' melt rate suddenly doubles?

Again, I'm the first to admit I'm so unqualified that I'm not even 100% certain I can spell meteorology. (I did set up a supercomputer at ECMWF about 20 years ago, though - so maybe I'll get a pass for that )

[Neven]

Thanks for setting this up, slow wing. 

[LarsBoelen]

I voted for 4-4.25, but we may get into a discussion on what is actually counted as "Arctic Sea Ice". Looking at recent NASA EOSDIS Worldview images (peeping through the clouds) it looks like a bowl of ice cubes and does that count as "The Arctic Ice Sheet"?

Question on that : is it possible that there are so many clouds this year because there is so much open water between the cubes. It almost looks like the relativly warm water evaporates and condenses into clouds above the colder ice....

[Richard Rathbone]

August melt volume is pretty consistent from year to year. It'll end up close to 4.4. It might just sneak into a bin either side of 4.25-4.5 but there's no case for any other bin. I'd put it at about 70% chance on 4.25-4.5, 5% on under and 25% on over.

I'm more interested in how high the next maximum will be, but its a couple of months too early to poll that. The minimum's pretty much set already. The last 3 years were all pretty similar. Will 2014 see a higher maximum as a result of this year's low melt?

[Shared Humanity]

I voted for 4-4.25, but we may get into a discussion on what is actually counted as "Arctic Sea Ice". Looking at recent NASA EOSDIS Worldview images (peeping through the clouds) it looks like a bowl of ice cubes and does that count as "The Arctic Ice Sheet"?

Question on that : is it possible that there are so many clouds this year because there is so much open water between the cubes. It almost looks like the relativly warm water evaporates and condenses into clouds above the colder ice....

In the winter time, the west coast of Lake Michigan is often blanketed with clouds that snow over the lake even as Chicago is basking in the winter sun. I'd be very surprised if the open water on the arctic does not have the same effect.

[slow wing]

The voting has closed. Thanks to those who voted and/or commented.

I picked the bin for between 4000 and 4250 km3.

As a central value and sigma, I would pick 4200 +/- 400 km3. (I.e. about a 2/3 chance of lying within those limits.)
So that is a drop of 2900 +/- 400 km3 from the known PIOMAS Volume at 31 July 2013 of 7104 km3.

 I based that on historical guidance mainly from 2010-12 as the Arctic sea ice volume was much higher in all years prior to that. The corresponding drops for those years have been eye-balled from Wipneus' graph as 2600, 2400 and 2700 km3 for 2010, 2011 and 2012 respectively. I have guessed that the loss will be a bit higher this year given the poorer state of the ice on the eastern side of the Arctic Basin. On the other hand, the Arctic has been cooler this month than it was last year. So the chosen uncertainty also brackets the loss amounts for 2012 and 2010.

[ChrisReynolds]

Not long now until August is out...

Just about ten days...

       Ooh! Ooh! I'm so excited!     

[Neven]

Indeed, Chris. It feels like wanting to know the end result of some important football match. I'm really curious to see whether 2013 has crept any closer to 2012, 2011 and 2010 (probably not, especially not 2012).

[ChrisReynolds]

Neven,

CT Area and NSIDC Extent for 2013 are tracking around 2010. PIOMAS Volume for 2013 end July was near 2010, but for volume 2010 at end July was 2.9k km^3 below 2009. So I think volume will be close to 2010, with no prospect of being near 2009 or 2012.

Over winter we'll see this increase in volume and the greater area/extent of ice lead to an increase in thicker ice, but it won't see an increase in thicker ice of the sort that typified the years 2007 to 2009 (and earlier) because one winter of ridging isn't enough to build ice many metres thick. Thermodynamic growth limit is around 2m thick over a freeze season. The increased area/extent doesn't mean this ice will thicken to much more because the surviving ice won't thicken by 2m on top of the remnant thickness in September - the 2m limit is set by thicker ice insulating better as it thickens. This autumn will see much less volume growth than 2012, probably more like the volume growth of 2010's autumn.

The following figure shows growth rate as a function of thickness for the central Arctic from Thorndike 1975 (figure 5)



This behaviour also goes some way to explaining why the Beaufort fracturing of February turned out to have no detectable impact by the summer (although weather had a role there the fractured region showed no sign of weakness).

[Oyvind Johnsen]

As a layperson with no understanding of ice dynamics - other than some bits and pieces I pick up from following the blog here - all I can do is "curve-spotting".
From this it seems quite obvious that volume minimum this year will be about the level of 2010, the curve is tracking 2010 quite close.
What I also can see is that volume maxima 2011-2013 are about the same. The differences in the three melt seasons seem to occur in May. So from this you might expect volume maximum in 2014 to be roughly the same as 2011-13, and that the weather in May 2014 will decide the minimum next year, somewhere in the range of 2010/13-2012.
These reflections on the last years seem to indicate a stability in volume, established since 2010. Looking at the whole set of curves since 1979, this of course seems unlikely.
But this is just curve-watching.  The physical questions seem to be:
What can cause an end to the volume stability of 2010-13?
What caused the great volume loss in the spring of 2010?   

[ChrisReynolds]

Oyvinnd,

I've blogged on the 2010 volume loss event.

Why the 2010 volume loss was important.
http://dosbat.blogspot.co.uk/2012/09/why-2010-piomas-volume-loss-was.html

What caused the 2010 volume loss, part 2, which links back to part 1.
http://dosbat.blogspot.co.uk/2013/03/what-caused-volume-loss-in-2010-part-2.html

Schweiger finds it is a volume loss event outside of the uncertainty range of PIOMAS, I argue that it is as important as 2007, and that it was a real event, not just something that happened in a model.

[ChrisReynolds]

Continued from the NSIDC thread.

Helorime,

When you start to dig into the basic research into sea ice physics you find that a lot of it was done back in the 70s and 80s. A major financial driver for research into sea ice at that time and earlier was the tactics of the cold war, the quickest strike against the USSR would have been from submarines under the sea ice where they could not be observed and could hit the Soviet Union within an hour. Critical to that strategy was an understanding of sea ice. But it's given a legacy of basic ice physics and data such as the Data Release Area upwards looking sonar transects - before you can surface to fire you need to know how and where to find the pockets of thin ice. But there has been ice research before then, particularly by the Soviet Academy of Sciences, there's a wealth of papers derived from their many decades of sea ice camps - most of it in Russian unfortunately.

One problem sea ice modellers face is computer power, PIOMAS runs a 43200 grid point grid (120 lat by 360 lon). The ice is two layer, the atmospheric driver supplied from daily NCEP/NCAR data, and the ocean model is 25 layers deep. The original programme was designed before 2001 (described in a 2001 paper) but the programme is designed for parallel computing. I suspect that if they had the funding they could probably do more to take advantage of subsequent advances in computing, they may already have done so to a degree. But more money always helps.

[werther]

On the Pole CAD count in the “ice-free North Pole”-thread…

The digitized 50 km radius quadrant at the Pole yesterday:



What does this mean for volume, at least, in the tiny 1954 km2 quadrant?

I suppose the larger floes, >4km2, are remnants of the original “mesh-pack” MYI rhomboids. These must still reach app. 2m thickness.
Anything smaller in size originates from the broad lead fill rubble and has been much more prone to melting. Like in my N.Greenland CAD-count June ’11, this can’t be averaged higher than 60cm.
The fade-zone area uncertainty contains small rubble, nilas etc. Its thickness is almost irrelevant.
The larger floes can be guessed at 320 km2 and 2 m, so 0,64 km3. The rubble  at 660 km2 and 0,6 m, adding 0,4 km3. The quadrant then contains app. 1 km3 ice.

Can this be extrapolated over all low concentration/structureless extent? It is not justified, but an interesting game. If, say, 2000 km2 holds 1 km3, the whole ‘splinter-zone’ shouldn’t be far off that average.

This is a digitized version of UB 26 August:



As shown, the low concentration zone including what UB presents as a “Siberian side-high concentration pack” holds app. 2,2 Mkm2. The game: 1100 km3 ? (must be more, my gut feeling is its about 2000 km3)

What about the high concentration side? 2,9Mkm2 x 1,2 m average? 3400 km3 ?
The low estimate could be just over 4000 km3. I think it is reasonable to count in the weather this summer for preserving some 700 km3 more than last year. The high estimate could be over 5000 km3.

What will PIOMAS present? What’s the quality?

[Richard Rathbone]

Thats about another 0.5 I didn't see staying solid in August.

Its going to take a big September storm to turn the near surface heat into last minute melt rather than delayed refreeze or this is another poll on the forum I am way too low on.

[Steven]

I was also too low on this poll.  4500 km^3 seems unachievable now.

We lost about 2000 km^3 of ice volume in August 2013.  This is less than in any other August month in the PIOMAS record.  The volume loss in August in recent years is about

2000  (2013)
2200  (2011)
2400  (2010)
2500  (2012, 2007 and 2006)
2600  (2009)
2700  (2005)
3400  (2008)

[ChrisReynolds]

Steven,

Thanks for reminding me (by email).

Day 12 to Day 212 PIOMAS data has been changed earlier in the year. It is important for people updating spreadsheets with only recent data to update their spreadsheets with the entire PIOMAS series.

[jdallen]

I was also too low on this poll.  4500 km^3 seems unachievable now.

We lost about 2000 km^3 of ice volume in August 2013.  This is less than in any other August month in the PIOMAS record.  The volume loss in August in recent years is about

2000  (2013)
2200  (2011)
2400  (2010)
2500  (2012, 2007 and 2006)
2600  (2009)
2700  (2005)
3400  (2008)

Less thick ice to start means less to loose . its also a basic characteristic of systems; the closer we get to an absolute limit (zero), the frequency of relative large percentage variations will increase, while their absolute amplitude will tend to decrease.  No surprise here, especially considering the good weather for preserving ice.

[ChrisReynolds]

Now the really interesting question is - to what degree will the volume increase of this summer be passed through to a volume increase (from 2010 to 2012 - all similar) for the start of volume loss season in April 2014?

[Steven]

Less thick ice to start means less to loose . its also a basic characteristic of systems; the closer we get to an absolute limit (zero), the frequency of relative large percentage variations will increase, while their absolute amplitude will tend to decrease.  No surprise here, especially considering the good weather for preserving ice.

Well, the PIOMAS volume anomalies in the post 2010 period drop very quickly in May/June, and they rise in July/August.  So there is a slowdown of the volume loss in July and August compared to earlier years.  That's partly because the sea ice area at the beginning of July or August is much smaller in recent years than it used to be.  Therefore there's a smaller area available where ice volume melt can take place.

2013 was atypical since the sea ice area at the beginning of August was significantly higher than in the 2010-2012 period.  And there was a record amount of FYI.  So there were reasons to think that the ice volume in August 2013 could keep up with 2010.  But apparently the weather has led to an early slowdown of the area/volume melt in large parts of the Arctic.  The storm from late July caused a spectacular slowdown, the storm from the second week of August did more damage to the ice, but with the persistent cold temperatures in most of August the melt never really resumed.

[Pmt111500]

Inquiry: does PIOMAS do the volume analysis also on Antarctic Ice?

[Wipneus]

Inquiry: does PIOMAS do the volume analysis also on Antarctic Ice?

No, but there is a similar (and older) calculation from the same people called GIOMAS (G for global) that includes the Antarctic.
When I contacted Dr. Zhang earlier this year about the calculation  he wrote back that he was in the  process of re-evaluating the data. I expect the new data to be published soon.
 

[Pmt111500]

Inquiry: does PIOMAS do the volume analysis also on Antarctic Ice?

No, but there is a similar (and older) calculation from the same people called GIOMAS (G for global) that includes the Antarctic.
When I contacted Dr. Zhang earlier this year about the calculation  he wrote back that he was in the  process of re-evaluating the data. I expect the new data to be published soon.

Thanks for the info. Anyway isn't Antarctican Sea Ice is almost all FYI so some sort of near-linear extrapolation of area numbers might be applicable?

[Jim Hunt]

The Polar Science Center reports that "September Update Delayed:  Required data not available due to US Government shutdown.". The daily numbers still only get as far as day 243.

[ChrisReynolds]

Jim,

It's because NCEP/NCAR reanalysis isn't working.

[Jim Hunt]

NCEP/NCAR reanalysis isn't working.

The Tea Party have a lot to answer for!

[slow wing]

Wipneus, on another thread, provides the minimum as 4942 km³:

Updated, monthly graphs are in the top post.

The minimum 2013 value was 4.942 [1000 km³] at September 7.

Yearly minimum fit to exponential trend:



Yearly minimum fit to linear trend:



Yearly minimum fit to Gomperts trend: