The 2014 Melting Season

[viddaloo]

Again, take the numbers with a grain of salt.

[iceman]

... The next two central-Arctic cyclones will cause some Ekman pumping, ...

....
The sense I have of it is, that 2014 has been astonishingly "static".  There has been very little movement of ice across the CAB.  Without that movement, we have neither (1) export of MYI to warmer water or (2) disruption of thermo and haloclines which in previous years provided heat to assist melting.
....

                   nor (3) much in the way of heat advection from lower latitudes.

... Atlantic and Pacific currents are still steadily pumping heat into the basin from lower latitudes. 

I would have expected the Arctic Ocean to be fairly efficient at cooling off the Atlantic and Pacific currents.  ....

(understanding that your comment pertains more to the freeze season)
There's been a considerable pool of warmth in and around the Bering Sea for most of the summer, but atmospheric circulation patterns have not pulled much of it into the Arctic as either sensible or latent heat.

[jdallen]

... The next two central-Arctic cyclones will cause some Ekman pumping, ...

Someone (I forget who) in another discussion with me noted correctly that the heat to melt *all* of the ice is already present in the Arctic Ocean.  It is only by way of lack of convection that it doesn't happen.  The upshot *I* would say is, that means the heat is still there.  Atlantic and Pacific currents are still steadily pumping heat into the basin from lower latitudes.  If it isn't melting ice, and it isn't re-radiating out of the top of the atmosphere, it means it is stacking up at depth.  I expect we will meet that heat somewhere in the near future.

I would have expected the Arctic Ocean to be fairly efficient at cooling off the Atlantic and Pacific currents.  Partly this is shown by the greatly increased winter surface temperatures, particularly in October -- the increased heat that is arriving in the current is being moved from the water to the air.  As a thin layer of ice forms, the water just beneath that ice concentrates salinity and sinks, creating a convection current.  The ice at the edge of the pack where the incoming currents are warmest, has cracks which prevent that ice from being a good insulator.  The incoming Atlantic current should be rapidly cooled as it enters the Arctic in winter.  Or, at least, that's what I've been thinking...

It would be, if the heat transferred efficiently.  Transfer from the ocean surface to the top of the atmosphere is much more efficient than transfer from depth, in part because of convection.  Ice and haloclines put a lid on that.  I expect most of the heat lost matches up pretty closely with heat gained during the melt season.  Once you have a skin of ice, and more critically, snow, the transfer rate will drop considerably.  Open water in the arctic *will* produce far more precipitation.

[Lord M Vader]

Viddaloo: if you want to look at the ECMWF forecast a.k.a the "EURO" you find it here: http://old.ecmwf.int/products/forecasts/d/charts/medium/deterministic

[viddaloo]

Viddaloo: if you want to look at the ECMWF forecast a.k.a the "EURO" you find it here: http://old.ecmwf.int/products/forecasts/d/charts/medium/deterministic

Thx, Vader! Bookmarked. On that same note, does anyone know where the Uni–Bremen went after Sep 5th? I see no update there since.

[viddaloo]

Closing in on 2007.

[viddaloo]

Uni–Bremen 7–day animation Aug31–Sep06.

[Rubikscube]

Delta maps for 2007, 2012 and 2013 are not changing abruptly from week to week as melting is nearing an end, but one thing worth to notice is how large the red areas on the Pacific side has grown in the 2013 comparison. In addition, export through Farm still appears to be lagging behind. The last attachment is a comparison with 2009, a year which ended up close to 2014 in both extent and volume.

[jonthed]

Vidaloo, the rate of 2014's melt in the last few days contrasts sharply with the other years which all slowed in a similar fashion on these dates. A later and stronger end to the melt season seems likely no?

[viddaloo]

Absolutely, jonthed. September shows a general upward trend, despite the decline of July & August. September's melted 226641 km² in 9 days, compared to 2009's 151624 km². Remains to see exactly how big a volume it's been though.

[viddaloo]

This animation of the first half of the week from Uni–Bremen certainly seems to imply that there is quite an extensive amount of melting and thinning going on in the CAB. My guess is it will be lower than the 2007 volume of 6458 km³ before the end of the melt.

Plus, I wouldn't be surprised to see 2014 break the old September melt record from 2008:

September 2008 melted 751 km³ from Aug31 to minimum on Sep19. September 2014 needs to melt the 2008 record amount of 751 km³ plus another 11 km³ in order to get from 7220 km³ (Aug31) to 6458 km³, which was the 2007 minimum (also on Sep19).

[Frivolousz21]

Refreeze showing up on Jaxa in the CAB but AMSR2 Bremen as well as SSMIS Bremen show nothing of the sort yet.

This caused Jaxa to go up 9k+ off setting losses elsewhere.  If this is real and not a false positive Jaxa may have hit the min but there will be solid compaction over the ESS for a little while.

If this vanishes tomorrow then Jaxa will probably drop another -100K or more.

[gciriani]

I've asked this question in a different thread but probably I should have asked here:
Does anybody know the meaning of the different  shades of white-blue in the IJIS map at U.of A. Fairbanks? I thought any color meant SIC > 15%, but I'm not so sure now. Perhaps the fainter color is SIC < 15%?

[iceman]

I don't have a good feel for the magnitude of Ekman pumping, but looks like it's going to contribute to melting in and near the Barents over the next few days.  Anyone know how long the effects of halocline disturbance persist after the low pressure system passes over?

[gciriani]

... If this vanishes tomorrow then Jaxa will probably drop another -100K or more.

Do you mean a drop of -100K in a day, or in total to the minimum extent for the year?

[Frivolousz21]

It's a bit obscured by the clouds.  You can see a couple areas where the ice is consolidated but for the most part the ESS has almost entirely melted out.

Should get a better visible view later on this evening.

[Frivolousz21]

Here is a much better image of the ESS region.

It's hard to see between the clouds in some spots.  But it is still very easy to make out.  Towards the top center right there is that little patch of consolidated ice that is easy to see on Bremen.

All around that area the ice is mostly wisps, swirly mush floes.



Jaxa uses the top image. You can see the blueish green which is frozen ice and snow.



You can see channel 89ghz which is what bremen uses is prone to cloud interference not temperature.

[Shared Humanity]

With the assault on the ESS, could we see a  late minimum for SIE?

[NeilT]

The NW passage is making a late run for the deep water channel too.  Quite interesting.

[Rick Aster]

The “snowflake” of sea ice in the western Beaufort Sea is 1/100 of what it was two weeks ago and looks like it could disappear from the Bremen map tomorrow.

A thousand kilometers to the west in the East Siberian Sea, the block of high-concentration ice continues to hold together. It has shrunk by maybe 20 percent while holding roughly the same shape and position. This contrasts with the ice around it which halfway melted in two weeks.

[viddaloo]

From my calculation algorithm, the 2014 ice volume seems pretty stable in-between 2009 and 2007.

[viddaloo]

Now closer to 2007 than 2009.

[viddaloo]

72% of this Winter's ice volume is now melted.

[jonthed]

Vidaloo, how does that compare with other years? Don't suppose you have a graph for that

[viddaloo]

Vidaloo, how does that compare with other years? Don't suppose you have a graph for that

You're funny!

[NeilT]

I posted that in tabular form back here, from 2000 onwards.  I also included the Max, Min and Vol lost, not just the % of volume lost.

Note that as the %loss increases, the max volume is falling.  So the actual total volume of ice lost is far less than the % curve shows.

Another interesting thing is that although the %loss for 2008/9 is lower than the %loss for 2007, the actual volume of ice lost is greater.  Due to the larger start ice volume.  So this picture is actually misrepresenting the situation somewhat.  As we know the 1979 volume was much higher than the 2000's volume, significantly less % of ice volume loss could mean exactly the same actual volume lost than in the 2000's.  If you follow me.

As importantly, the % ice loss in 2012 was 3.17% more than 2010.  Yet the actual volume of ice lost overall, was almost exactly the same....

Statistics are great, but percentages can hide very different interpretations.

The chart in volume loss.  Physical, not %, is:



Which is a totally different representation and a very different trend line.  Adding 2014 into this will give a clearer picture.

What this shows to me is that the PIOMAS volume overall loss depends heavily on how much volume the model calculates has been re-created in the winter.

Which is why I am challenging the 2012-2013 winter re-growth specifically.

[gciriani]

...
What this shows to me is that the PIOMAS volume overall loss depends heavily on how much volume the model calculates has been re-created in the winter.
...

Is it possible to follow a similar reasoning for the ice extent, and come up with more meaningful numbers?

[jdallen]

Vidaloo, how does that compare with other years? Don't suppose you have a graph for that

You're funny!

Thanks viddaloo. So. It looks like we are near parity with 2007, 2008 & 2009.  2010-2013 spike up.  2007 itself was a moderately sharp uptic than set the tone for the following trend.  Stair step melting?

[jdallen]

...
What this shows to me is that the PIOMAS volume overall loss depends heavily on how much volume the model calculates has been re-created in the winter.
...

Is it possible to follow a similar reasoning for the ice extent, and come up with more meaningful numbers?

Extent is more volatile.  You can lose an awful lot of thin ice without significantly changing volume loss.  I suspect the correlation would be quite a bit looser.

[viddaloo]

Which is why I am challenging the 2012-2013 winter re-growth specifically.

Neil, what does it mean that you challenge the 2012–13 refreeze? Do you assume that less ice was built during Autumn and Winter than what PIOMAS states? I think others have explained well how a record area of open water also implies a record refreeze, sort of a negative feedback.

[jdallen]

Which is why I am challenging the 2012-2013 winter re-growth specifically.

...I think others have explained well how a record are of open water also implies a record refreeze, sort of a negative feedback.

Yes and sort of.  A record refreeze isn't necessarily a negative feedback.  More like a fire given more fuel to burn.  That there is more open water doesn't encourage more freezing - the contrary in fact - there simply is more surface to freeze.

[viddaloo]

Another interesting thing is that although the %loss for 2008/9 is lower than the %loss for 2007, the actual volume of ice lost is greater.  Due to the larger start ice volume.  So this picture is actually misrepresenting the situation somewhat.  As we know the 1979 volume was much higher than the 2000's volume, significantly less % of ice volume loss could mean exactly the same actual volume lost than in the 2000's.  If you follow me.

I follow you, absolutely. But when I wrote this code about a month ago, I was primarily eager to see what year all ice would be gone, and then the bar or target could easily be set equally for all years to 100%. The goal is not to melt 20.000 km³ if there's only 19 available etc, so to have a meaningful target for all years, 100% meltdown was coded as 100% of the Winter Maximum.



The same graph also shows clearly that more and more ice is put into play each year; from about half the ice when I was a kid to probably about 85% as a mean for the 2010s when the decade is over. This of course increases the possibilities of a total meltdown immensely.

jdallen: Yup, I don't believe we'll see sub 70 percentages again any time soon. That hasn't happened since before the iPhone Gap 

[Bob Wallace]

I posted that in tabular form back here, from 2000 onwards.  I also included the Max, Min and Vol lost, not just the % of volume lost.

Note that as the %loss increases, the max volume is falling.  So the actual total volume of ice lost is far less than the % curve shows.

Another interesting thing is that although the %loss for 2008/9 is lower than the %loss for 2007, the actual volume of ice lost is greater.  Due to the larger start ice volume.  So this picture is actually misrepresenting the situation somewhat.  As we know the 1979 volume was much higher than the 2000's volume, significantly less % of ice volume loss could mean exactly the same actual volume lost than in the 2000's.  If you follow me.

As importantly, the % ice loss in 2012 was 3.17% more than 2010.  Yet the actual volume of ice lost overall, was almost exactly the same....

Statistics are great, but percentages can hide very different interpretations.

The chart in volume loss.  Physical, not %, is:



Which is a totally different representation and a very different trend line.  Adding 2014 into this will give a clearer picture.

What this shows to me is that the PIOMAS volume overall loss depends heavily on how much volume the model calculates has been re-created in the winter.

Which is why I am challenging the 2012-2013 winter re-growth specifically.

Seems that what you've got here is half the story.  Plot the annual starting volume on the same graph and we can watch them converge.

What viddaloo is doing is plotting a combination of the two, decreasing annual start volumes and increasing annual melt volumes.   Doing a single line that combines both numbers hides the details but still shows how we're moving toward zero.

[viddaloo]

Tuesday was only 36 km³ behind 2007.

[Neven]

This right here should end the melting season, right?

[Neven]

Well, looking at Wipneus' thread it has already ended.

[Jim Hunt]

According to the NSIDC's latest Arctic Sea Ice News:

"The end of this year’s Arctic sea ice melt season is imminent and the minimum extent will be slightly lower than last year’s, making it the sixth lowest extent in the satellite record.

Arctic sea ice extent for September 15 was 5.07 million square kilometers (1.96 million square miles). This is only 30,000 square kilometers (11,600 square miles) below the same date last year, yet sea ice extent remains low compared to the long-term 1981 to 2010 average. As is typical for this time of year, weather conditions near the ice edge heavily influence the timing of the minimum, which has occurred as late as September 23. We are now a day past the 1981 to 2010 average minimum date of September 15"

[JackTaylor]

Well, looking at Wipneus' thread it has already ended.

Neven,

Regardless of the exact date/time end of the melting season,
congratulations and Thank-You for making the blog & forum
available again this year.

I truly enjoyed the coverage provided by everyone 

R,
JT

[viddaloo]

Well, looking at Wipneus' thread it has already ended.

Neven,

Regardless of the exact date/time end of the melting season,
congratulations and Thank-You for making the blog & forum
available again this year.

I truly enjoyed the coverage provided by everyone 

R,
JT

Ditto. It was both fun and educational

PS: I still think we can beat 2007, though! As Cartman says: "Gentlemen, we can win this battle." (Playing the South, reenacting the Civil War.)

[NeilT]

Seems that what you've got here is half the story.  Plot the annual starting volume on the same graph and we can watch them converge.

What viddaloo is doing is plotting a combination of the two, decreasing annual start volumes and increasing annual melt volumes.   Doing a single line that combines both numbers hides the details but still shows how we're moving toward zero.

Yep understood.  I was talking on a slightly different line.  I'll redo my chart to show what I mean.

[BornFromTheVoid]

The melt rate is nearing the 2σ level, with the loss over the last 7 days beaten only by 1991.

[viddaloo]

Lower than 2007 yesterday, but also lower than 2007's minimum.

[viddaloo]

How did 2014 end up in a tête-à-tête with then record low 2007? Here are raw volumes for the 365 days ending in September 17th for 2007 and 14.

Hopefully, the graphs can tell us something about the changing dynamics of sea ice melt. Note for instance the steeper refreeze in November–December (caused by lower previous autumn minimum), the lower Spring maximum starting point and the more relaxed July–August melt. These 3 things are trends, meaning they will likely continue and be more expressed in the future. Another such trend is the continued melt well into September, that will also likely increase in coming years.

A final note is the fact that 2014's been lower in volume than 2007 for an entire year, except the short period of 58 days from July 21st to September 16th, caused by the decline in July–August melt, one of the strongest trends in the PIOMAS data set period.

[Rubikscube]

I didn't post any delta maps this week, instead, since we have reached the minimum (I presume), I made this animation of all minimums since 2002 where reds indicate decline from previous year while bules indicates increase (the exact minimum dates are from CT, which is not entirely correct since this is Uni Bremen, but the difference should be miniscule). It becomes a little bit hard to determine the exact size and shape of the pack on each individual year, but the purpose is more to give an impression of the size of year to year fluctuations.

Click to animate

[iceman]

... the purpose is more to give an impression of the size of year to year fluctuations.

Quite dramatically so.  From eyeballing the graphs I can't tell whether internannual variability of the shape or distribution is on the increase.  It does, though, appear to be increasing in a broadly circumscribed central area of the ice pack.  Could be yet another indication of how susceptible the ice is becoming to atmospheric circulation patterns.

[Rubikscube]

Similar maps to these may already exist, though, since I haven't found them yet I made a couple myself. The first one shows the average concentration of all minimums between 2002 and 2013 (2002-2011 is AMSR-E, while 2012 and 2013 are AMRS-2), while the second one is a comparison with the 2014 minimum (as in previous post all minimum dates are delivered from CT) using the traditional colors and layout.

As one can see the Laptev region is the one that stand out the most, both in terms of how large the area of anomaly is and how much it departures from the 2002-2013 average. Furthermore, and perhaps quite unsurprisingly, the Barents region is showing a large positive anomaly, and so is CAA as well. The absent of export through Farm is also noticeable, while the Pacific side in general seems rather mixed and perhaps slightly above normal.

[viddaloo]

Great maps, Rubikscube!

September 20th 2014 was the first post–2010 day to have a 365 day mean volume bigger than that of the same date in 2010. The yearly mean — average volume for the year than ends on the date in question — volume for Sep 20th 2010 was 14874 km³, and for Sep 20th 2014 it was 14877 km³. This means that from yesterday and probably until the end of the year, we are 'recovered' back to before 2010 territory, mean ice volume–wise.



(Further info for this graph: At the end of the year — Dec 31st — it shows the average volume for 2014, or for any other year. Further to the left in the graph, it shows the year or 365 day average for one year cycle until that date. This removes seasonal fluctuations and focuses on whether the year in question is packing ice, or melting ice, compared to the last year. As a bonus, you can also see where we are, on average, during a year, currently at about 15 000 km³ sea ice. When the 2002 descent that ended in the 2007 all–time low minimum began, the average volume was 20 207 km³.)

[Magma.]

Very nice maps indeed, Rubikscube. I especially like the first with the preceding 12-year average at minimum. It is both informative and intuitively clear.

What software did you use to create them?

[viddaloo]

Yearly average sea ice volume is back to pre-2010 levels as I've laid out above and in this thread, but at least now we have a fairly accurate way of predicting when and if new all-time low Autumn minima will occur.

As already mentioned, every descent of the average for the last 2 decades or more has ended in one or more new all-time lows. Further, the descents often last many years (2002-2008) and span from one decade to the next (2009-2011), so my bet is we're looking at at least 1000 days of average volume descent when the graph (2014 or 2015) finally turns to go downward again.

Now, the new Autumn minimum record may not come during the first year of the next descent, but during the descent we can expect one, two or even three new Autumn minimum records, following each other, and chances are the latest of them will be an ice-free minimum of zero km³.

[Rubikscube]

Thank you all for your appreciation. I'm unfortunately no programming guy that can take the raw data and run it through some fancy algorithm, so I just screenshot the Uni-bremen false color maps and do the rest in photoshop 7.0. The weekly maps don't take long when you got the adjustment layers ready, but the average map costed me quite a few hours, which is as shame since I would love to make that particular one more often than once a year.

Like that you make separate threads for some of your neat graphs Vindaloo, as long as it don't get too many to keep up with.