What caused the September 2012 record low?

[Steven]

This is a recurring topic in other threads.  It may be useful to have a dedicated thread for it.

I think multiple factors played a role:


1) Sea ice volume in Autumn 2011 was, and still is, the lowest on record according to CryoSat.  That may have preconditioned 2012.

2) Buoys in spring 2012 (most of them were in the Beaufort) showed only a thin snow layer on the ice, allowing melt pond formation to start early.

3) There was early widespread surface melt in June 2012, which continued strongly throughout July.  I think that was the main reason for the record low extent in September 2012.

4) Atmospheric conditions in summer 2012 were hostile to the ice, with a dipole with high pressure on the American side of the Arctic Ocean (Beaufort high and Greenland ridge) and low pressure on the Siberian side.

5) The GAC in August 2012 clearly played a role, but I think its importance is often overstated on this forum.  The main area hit by the GAC (in the Central Arctic between Laptev and Beaufort) was already far advanced in the process of melting out and separating from the rest of the ice pack before the GAC hit that area between 6 and 9 August 2012.  A study showed that the GAC contributed a loss of only 0.15 million km2 to the September sea ice extent.  So there would have been a record low extent in September 2012, even without the GAC.  That seems right to me.

6) Also, there were records all over the Arctic in 2012.  Greenland land ice loss in summer 2012 was also the greatest on record, a record that still stands.  And the Northwest passage was already as good as open at the end of July 2012, also a record I think.  None of that was related to the GAC.

[nadir]

The persistence of the Greenland high, contributing to the worst possible atmospheric circulation setting for the ice, might be another factor. Hence the record melting over Greenland ice sheet as well.

Tons of discussions about the Greenland High that year, including how it contributed to drift Sandy toward the US coast as late as September.

[Paul]

A very warm Beaufort sea could also be a contributing factor , it may of lead to thinning of the ice over the Chukchi and the lows during July I recalled were around the 980/990MB range so fairly deep which would of contributed towards the dispersion of the ice pack.

I do agree with the thoughts a record probably would of happened without the GAC but what happened after that during August weather wise was nothing out of the ordinary yet extent still dropped sharply, so did the GAC caused warm water to rise to the surface then? Something I'm not an expert at in answering.

September also saw a compaction event on the Atlantic side with ice reaching near the 85 degree north mark. Only positive with 2012, because the ice did melt out in the Chukchi quite late, there was chilly SSTS  so refeeeze was quite rapid unlike what we saw in 2020. Unsurprisingly refereze in the Beaufort was quite slow though.

[Richard Rathbone]

A whole lot of things, I'd particularly pick out 3 and 4 on Steven's list, all lined up together.

I look at it as the weather rolling 5 dice and scoring something like 6,6,3,5,6. At the time I was thinking the dice had become a lot more weighted than they've actually turned out to be and it was a harbinger, but I now think it was a freak in how far below trend it managed to get.

That sort of score is likely to take a long time to beat when you only get to throw the dice once per year.

[gerontocrat]

I thought you might want to see a closer look at what happened before, during and after the GAC of 2012, (August 5th to the 9th?)

So I attach graphs of NSIDC sea ice extent, area and concentration, and PIOMAS volume.
Note the graphs only include the 7 central seas of the High Arctic.

The area, extent, and volume graphs show that before the cyclone sea ice losses were high and a record low minimum was certainly on the cards,

At the time of the cyclone the downward kink in the graphs also seem to show a definite acceleration in sea ice losses.

After the cyclone sea ice losses continued at a high rate to a later than usual date of minimum.

The concentration graph shows a very low concentration before the cyclone, what looks like a sudden drop during the cyclone, and then a rapid increase to more normal levels after the cyclone.

Make of it what you will - I for one have no answers as to the relative importance of the drivers of this extraordinary event, unique so far in the 43 years of the satellite record from 1979 to 2021. That is why in my view it should be regarded as an outlier.

I will say that 2012 was likely to see a record low with or without the cyclone.

[Jim Hunt]

1) Sea ice volume in Autumn 2011 was, and still is, the lowest on record according to CryoSat.  That may have preconditioned 2012.

I only have the combined CS2/SMOS numbers to hand, but they give me cause to quibble.

2020/21 volume was below 2011/12 for (almost) the entire period, and was considerably lower by mid April.

What's more volume on April 15th 2012 was almost identical to April 15th 2013.

Predictive power for the subsequent minimum extent seems negligible?

[Bardian]

Volcanic ash cloud shuts down parts of Greenland airspace

https://www.smh.com.au/environment/volcanic-ash-cloud-shuts-down-parts-of-greenland-airspace-20110523-1f0rd.html

reduced albedo, Svalbard was also affected. Ash traveled probably also to high arctic

[Steven]

1) Sea ice volume in Autumn 2011 was, and still is, the lowest on record according to CryoSat.  That may have preconditioned 2012.

I only have the combined CS2/SMOS numbers to hand, but they give me cause to quibble.

2020/21 volume was below 2011/12 for (almost) the entire period, and was considerably lower by mid April.

You're right, I had been looking at the CPOM sea ice thickness graph (which has 2011 as lowest in autumn) rather than the volume graph (which has 2011 as second lowest, behind 2020).  Either way, the data suggest 2011 was a strong melt season in terms of volume loss, which may have played some role for 2012.

By the way, I find the autumn data more relevant than the spring data, as the volume gains in winter depend also on the thickness of the insulating snow layer on the ice: a year with less snow would tend to have more volume gains in winter but would also have an earlier onset of melt ponding and albedo drops in the next melt season, which may cancel out the extra winter gains.

[be cause]

Hi gero , your Fram export also merits adding here .. it was significant and near continuous in 2012 ..

[Paul]

1) Sea ice volume in Autumn 2011 was, and still is, the lowest on record according to CryoSat.  That may have preconditioned 2012.

I only have the combined CS2/SMOS numbers to hand, but they give me cause to quibble.

2020/21 volume was below 2011/12 for (almost) the entire period, and was considerably lower by mid April.

What's more volume on April 15th 2012 was almost identical to April 15th 2013.

Predictive power for the subsequent minimum extent seems negligible?

Too counteract that, is there extra volume because there was more multi year ice and/or thicker ice towards the CAA. From what I remember of the 2011/2012 winter season was the complete struggle of the Barants and in particular the Kara seas to freeze over hence the early melt out in these regions. I think another surprise was how quick the Beaufort retreated away as seemingly this area had the colder conditions to counteract the poor ice forming weather on the Atlantic side.

[binntho]

Volcanic ash cloud shuts down parts of Greenland airspace

https://www.smh.com.au/environment/volcanic-ash-cloud-shuts-down-parts-of-greenland-airspace-20110523-1f0rd.html

reduced albedo, Svalbard was also affected. Ash traveled probably also to high arctic

Very interesting point. If volcanic ash fell on the ice during freezing season, the following melt season could see vastly increased albedo.

The Icelandic glaciers have experiended the same effect - volcanic ash that falls during winter is covered in snow, but once the annual snowmelt reaches the ash layer, total melt increases drastically. Eventually the ash is washed downhill and the effect is negated.

On sea ice, with minimal snow cover, the ash would remain on top or very close, and would remain on top more or less unchanged until the ice melted out completely. The added albedo could be a significant factor in adding to total melt, particularly if insolation was strong.

NB a uniform thin ash layer does not necessarily show up clearly in satellite imagery, and volcanic ash is not always black - it can be almost white depending on the mafic constitution of the magma.

Bardian: The link does not work, can you fix it?

[Jim Hunt]

Either way, the data suggest 2011 was a strong melt season in terms of volume loss, which may have played some role for 2012.

Quite so. 2007 melted a lot of the "oldest, thickest" ice, and 2011 pretty much finished the job. This graph of sea ice age is from March 2022:

[gerontocrat]

Hi gero , your Fram export also merits adding here .. it was significant and near continuous in 2012 ..

I have produced a couple of graphs and a table from Steven's data to supplement his graphs on the PIOMAS thread.

The first graph shows the daily Fram export for the 1980, 90's, 2000's and 2010's, 2012 and 2022.
And it is a mess. So I did it again using the 7 day trailing average - which looks a bit more useful.
I also attach a table of yearly totals.

From a quick look - in 2012 Fram export was high from mid-February to mid-May. In december it was low.
The total Fram export for the year was somewhat higher than for the 2010's average, close to the 2000's average, and lower than to 1990's and 1980's average.

ps: The trend shown in the FRAM export 10 year average totals is of a decline in Fram export. The 2010's average is 26% below that of the 1980's. Does this reflect overall thinning of the ice sheet and especially the loss of thick MYI?

[Jim Hunt]

Too counteract that, is there extra volume because there was more multi year ice and/or thicker ice towards the CAA.

Here's the AWI CS2/SMOS thickness map and NSIDC sea ice age map for the middle of April 2012:

[be cause]

sorry gero , I had a pic in my mind of your name next the graph .. but yours was the next post .. so thanks to Steven /Wipneus for this annual Fram export graph showing 2012 was an impressive year : https://i.imgur.com/KReS6ql.png
 
 

[nadir]

Given that the average thickness is reduced by a 30% or so since the 80’s, what Gerontocrat numbers say is that the average area exported has remained constant, or even slightly increased, consistent with a more mobile ice pack.

[Stephan]

Here is a comparison of Arctic sea ice volume and the relative proportion of the Fram Export

decade      avg. ASI volume kkm³    Fram export kkm³ (percentage)
1980s                 23.248                              2.141   (9.2%)
1990s                 21.659                              2.038   (9.4%)
2000s                 17.999                              1.903 (10.6%)
2010s                 14.056                              1.590 (11.3%)

The relative proportion increases with time.
Please note that the average volume contains all Arctic ice (from Hudson to Kara to Chukchi to Okhotsk), many areas do not matter at all when it comes to Fram Export.
A better analysis would only take CAB, ESS, and Laptev ice volume into account. I do not have the regional volume data.

[Steven]

Here's the AWI CS2/SMOS thickness map and NSIDC sea ice age map for the middle of April 2012:

It's interesting how that CryoSat April map shows a band of relatively thin ice between about 75°N and 80°N all the way from Beaufort to Laptev.  That is exactly the main area that melted out in the first half of August 2012.  The thicker ice further South in ESS and Chukchi around Wrangel Island thereby got separated from the rest of the ice pack, and eventually melted out some days later.

FWIW, below is an animation of SMOS images from mid July to mid August 2012.  It shows how that region gradually weakened and melted out.  (As always, interpretion of SMOS images in summer is tricky, SMOS in winter shows thickness of thin ice but in summer the measurements get heavily affected by surface wetness and open water between floes etc, but I think it is still useful for showing where the vulnerable ice is that is on the brink of melting out.)

This is the kind of thing that makes me believe that by early August 2012 (and maybe even earlier), a record low September extent was already guaranteed, even without the GAC.  There was a lot of low-hanging fruit (thin ice) when the GAC arrived on 6 August 2012, and I think most of it would have melted out even without the GAC.

[Bardian]

binntho: the link works fine for me, maybe there is a geoblocking?

REYKJAVIK: A portion of the airspace over Greenland has been closed because of ash from an erupting Icelandic volcano, Danish air traffic control officials said yesterday.

A Naviair spokeswoman, Helle Kogsbell, said that authorities had partly closed Greenland airspace late on Sunday ''due to the spread of ash from the Icelandic volcano to the east coast of Greenland''.

She said the airspace would remain shut until at least 10pm Sydney time yesterday.

''The space, up to 6000 metres of altitude … remains closed to air traffic,'' Ms Kogsbell said.

Naviair is responsible for the airspace over Greenland only up to 6000 metres. Canada and Iceland have control over airspace above that.

Denmark was the first country besides Iceland to close any airspace after the Grimsvotn volcano began erupting late on Saturday, but European aviation authorities have been closely monitoring the giant column of ash amid fears it could, like last year's eruption, wreak havoc on air travel across the continent.

European safety experts have warned that the ash could reach Scotland by today before sweeping across Britain to hit France and Spain two days later.

Meanwhile Norwegian airport operator Avinor said it planned to halt flights between mainland Norway and the country's Arctic archipelago of Svalbard.

''The cloud of ash will reach the waters separating continental Norway and Svalbard,'' an Avinor spokesman, Ove Narvesen, said.

''There will therefore not be any flights to or from Svalbard today,'' he said, adding there was also ''a certain risk that western Norway will be affected, but it is too early to say''.

The junior transport secretary of France, Thierry Mariani, warned yesterday that flights would be cancelled if the ash cloud blew over Europe, sending airline shares tumbling. ''One thing that is certain … is that if Europe is affected then flights will be cancelled,'' Mr Mariani said on Europe 1 radio, adding that it was too early to say for certain if it would.

''If the ash isn't noxious then the planes will fly. If the ash is noxious or presents a risk, then the planes won't fly,'' he said.

In April last year Iceland's Eyjafjallajokull volcano erupted, spewing a massive cloud of ash that caused the planet's biggest airspace shutdown since World War II, with more than 100,000 flights cancelled and 8 million passengers stranded.

This video is from 2010 showing the ashcloud of Ey...kull eruption

[El Cid]

There's always talk on the melting forum about how a june HP preconditioning is necessary for significant melt later on. I compared 2007,12,16 and 20 to climatology averages for June. Other than 2007 it was not very convincing (as you can see 2007 was really outstanding with constant high pressure anomalies, 2012 had HP over Greenland but not much of the Arctic, while 2016 and 20 had LOW pressure anomalies!):

[Jim Hunt]

Via Friv on the melting season thread:

2012 ended May and started June with a huge full dipole anomaly. 

Almost the entire Canadian Basin was sunny in early June for days.

The Beaufort was the warmest on record ever. By a ton. 

There is was 20C ssts at the mouth of the Mackenzie river and beaufort.

Two separate warm low pressure systems set up to bring unbelievably hot weather into the Canadian Basin

The lead up to the cyclone saw unprecedented melt taking place in the Western Canadian Basin.

That's just a small sample of the different ways 2012 was set up to be a huge year. 

It would have broken the record low without the cyclone for sure...

[uniquorn]

<>
2) Buoys in spring 2012 (most of them were in the Beaufort) showed only a thin snow layer on the ice, allowing melt pond formation to start early.
<>

Would like to follow up in that if you can provide a link

Meanwhile:

Huge waves measured for first time in Arctic Ocean
Hannah Hickey  UW News   July 29, 2014
https://www.washington.edu/news/2014/07/29/huge-waves-measured-for-first-time-in-arctic-ocean/

As the climate warms and sea ice retreats, the North is changing. An ice-covered expanse now has a season of increasingly open water that is predicted to extend across the whole Arctic Ocean before the middle of this century. Storms thus have the potential to create Arctic swell – huge waves that could add a new and unpredictable element to the region.

A University of Washington researcher made the first study of waves in the middle of the Arctic Ocean, and detected house-sized waves during a September 2012 storm. The results were recently published in Geophysical Research Letters.
“As the Arctic is melting, it’s a pretty simple prediction that the additional open water should make waves,” said lead author Jim Thomson, an oceanographer with the UW Applied Physics

The open access paper has more detail
https://agupubs.onlinelibrary.wiley.com/doi/pdf/10.1002/2014GL059983

extract

There were two notable wave events in 2012: a storm early August and another in mid-September. The first storm occurred before the in situ observations began (and when satellite observations suggest that ice cover was still significant at the mooring location); hence, only model results are shown for that storm.Although the first storm was significant in forcing ice retreat [Parkinson and Comiso, 2013], the waves associated with that storm were not modelled to be as large as the September storm. This is likely because the fetch available for wave growth was much smaller in August (when the ice edge was at approximately 72◦N)than in September (when the ice edge was at approximately 80◦N)

[Steven]

<>
2) Buoys in spring 2012 (most of them were in the Beaufort) showed only a thin snow layer on the ice, allowing melt pond formation to start early.
<>

Would like to follow up in that if you can provide a link

It seems the links to the old buoy website are broken.  But found an archive here:

http://imb-crrel-dartmouth.org/archived-data/

I just browsed through the data.  It looks like in general, the buoys on the Pacific site of the Arctic Ocean had snow depths somewhat lower than normal, except buoy 2011K.  Whereas the buoys on the Atlantic side had somewhat higher snow depths than normal, except buoy 2012C.

Anyway it seems the 2012 snow depth in spring was not as unusual as I had remembered.  So maybe this did not play much of a role after all.

Also found this study, which uses a snow model and compares it to the Operation IceBridge snow data.  This too suggests that there was lower snow depth than usual on the Pacific side and higher than usual on Atlantic side.

[uniquorn]

Thanks, may do some buoy analysis though it will be on the back burner.

Found this;
Autonomous observations of solar energy partitioning in first-year sea ice in the Arctic Basin
March 2014 Journal of Geophysical Research: Oceans 119:2066–2080
Caixin Wang1 , Mats A. Granskog1 , Sebastian Gerland1 , Stephen R. Hudson1 , Donald K. Perovich 2 ,Marcel Nicolaus 3 , Tor Ivan Karlsen 1 , Kristen Fossan1 , and Marius Bratrein 1
DOI:10.1002/2013JC009459

Abstract
A Spectral Radiation Buoy (SRB) was developed to autonomously measure the spectral incident, reflected, and transmitted spectral solar radiation (350 to 800 nm) above and below sea ice. The SRB was deployed on drifting first-year sea ice near the North Pole in mid-April 2012, together with velocity and ice mass balance buoys. The buoys drifted southward and reached Fram Strait after approximately seven months, covering a complete melt season.
 At the SRB site, snow melt started on 10 June, and had completely disappeared by 14 July. Surface albedo was above 0.85 until snow melt onset and decreased rapidly with the progression of snow melt. Albedo was lowest on 14 July, when the observed surface was likely a mixture of bare ice and melt pond(s). The transmitted irradiance measured under the ice was largest in July, with a monthly average of 20 W m-2, compared to less than 0.3 W m-2 pre-melt. Under-ice irradiance peaked on 19–20 July, with a daily average around 35 W m-2.
 From mid-April to mid-September, the solar energy transmitted through the ice into the ocean contributed about two thirds of the energy required for the observed bottom melt (0.49 m). The energy absorbed by the ice after snow melt was enough to melt an additional 0.1 m of ice. Solar energy incident on open water and melt ponds provided significant additional heating, indicating solar heating could explain all of the observed bottom melt in this region in summer 2012.

my emphasis

[kassy]

Climate change: One year of warm currents fundamentally changed Spitsbergen’s sea ice coverage

...

What happened in the winter of 2006 then, which led to the dramatic change?

In January 2006, the atmospheric low pressures were lined up from the North Atlantic to the Fram Strait between Svalbard and Greenland, heading directly towards the Arctic. The normal route for these weather systems is east of Svalbard ending up in the Barents Sea, but the low pressures this winter were west of Svalbard, bringing warm air all the way from Spain to Svalbard.

Southerly wind west of Svalbard gave rise to three dramatic impacts on the sea ice; 1) warm air and ten January days with plus degrees, 2) a push of the existing sea ice around Svalbard northward into the Arctic, and 3) these southerly winds changed the current systems west of Svalbard. The result was a flooding of warm Atlantic water on the continental shelf along the west coast of Svalbard.

In February, the wind turned northerly again, and the air temperature dropped to -20°C. Despite the biting cold, the sea ice in Isfjorden, Kongsfjorden and the fjords along the west coast of Svalbard, dissolved and disappeared due to inflow of warmer water from the shelf outside. The fjords were kept ice free the whole winter and due to the large heat capacity of the ocean water, this excess heat was also able to heat up the atmosphere.

...

https://sciencenorway.no/arctic-ocean-climate-change-opinion/climate-change-one-year-of-warm-currents-fundamentally-changed-spitsbergens-sea-ice-coverage/2040786

[Steven]

Moving this comment here from the latest JAXA poll thread.

     It has been 12 years since the 2012 debacle and I am still mystified.  I know about the GAC, and I should re-read the post mortem analyses, but what the heck happened in 2012?  It still is a shocker.  Loose recollection tells me that even up to mid-July, the brewing disaster was not apparent in the usual melt progression charts.

 But by September, except for an honorable mention in 2020, no other year has come within striking distance of 2012.  An amazing outlier event.  Even so, I think its stand-alone status won't last much longer.

There is a thread about this here, discussion about 2012 may fit better there:
https://forum.arctic-sea-ice.net/index.php/topic,3796.0.html

TLDR: the August GAC played a role but I think the main culprit was the strong preconditioning in the months before.  NSIDC area for June 2012 is still the lowest on record for June, and July 2012 is second lowest for July, just a hair above July 2020, so there was a lot of surface melt going on in those two months.

[Bardian]

There was a huge sea ice export event through the Fram Strait in July 2011, boat traffic was affected for many days. Maybe i`m wrong about Fram, people claim that ice came from east coast of Svalbard, around the archipelago but from high arctic anyway

https://www.youtube.com/watch?v=HZKTJNeYyC8&ab_channel=YorickBroekema

https://www.youtube.com/watch?v=sY9PhHpC3sw&ab_channel=DanCelius