As we approach the final stretch...

For years we've talked about one of the most fascinating phenomena of Arctic sea ice loss as it progresses from year to year: The moment when the ice disappears regardless of the weather. We've seen some of that, back in July 2012, when extent and area numbers kept going down steadily, even though weather conditions weren't conducive to melt and should have caused a marked slowdown on the graphs (like they did in the preceding years). Some time later I concluded that this had to do with what I now call melting momentum.

The idea is that if during May and July the ice gets hit by lot of sunshine, lots of melt ponds form and they soak up a lot of heat. This doesn't immediately cause the ice to melt, but starts to make itself felt towards the later stage of the melting season. The built-up momentum keeps the melting going. For this reason I've been focusing on things like compactness, that tells us something about the amount of water within the ice pack, whether open water or melt ponds (satellite sensors can't see the difference).

But does this still hold true? Will there be a moment when the ice disappears regardless of melting momentum as well? We saw in 2016 that there wasn't much melting momentum built up during May and June, but despite the cloudiness, a record warm winter and persistent high air temperatures were enough to make the ice vulnerable enough for a big cyclone and massive dipole to have 2016 come in second on almost all graphs (see this overview). Yes, melting momentum is important, but there are other things at play as well, of course.

NSIDC Compactness (are divided by extent) as of July 29th

Now, this year has been extremely interesting so far. For the third time in a row, the winter preceding the melting season was relatively mild, the maximum was low (second lowest on record), and the trend line on many a graph went low from the get-go. Not as low as 2016, but low. However, as in previous years, clouds moved in and things slowed down to a crawl, taking the trend line as high up as 11th lowest on record on the JAXA sea ice extent chart last week. But then this happened in the last couple of days, suggesting spectacular melting on the Pacific side of the Arctic:

So, what's going on?

The first thing you need to do if you want to say something about Arctic sea ice is compare, compare, compare. Even if we're in uncharted waters here, and the New Normal (2005-2011) quickly evolved into the New Abnormal (2012-now), we still need to compare current data as much as we can to that of the past, especially if that data was gathered in a consistent way from year to year.

Before looking in more detail at the weather conditions during this melting season so far, I'd like to dig a bit deeper into the melting momentum aspect of things. Above, the NSIDC compactness graph shows how the 2018 trend line swooped down at the end of June, but picked up again after that, where years like 2012 and 2016 continued to go down. And before it went lowest on record, during those crucial months of May and June, it was actually relatively high. This would suggest very little melting momentum was built up during that time.

But NSIDC compactness is far from perfect as a melt pond measurement, because it is also heavily influenced by open water between floes, patches of open water inside the ice pack. Luckily, David Schröder, a scientist from the University of Reading, developed a model together with his colleagues that calculates melt pond fraction. And even more luckily, Dr Schröder is so kind as to share some of his model output with me. Here's how May and June look according to the model:

Edit 31-7: The legend is correct for May, but should be ten times higher for June, running from -20 to +20.

Caveat: This is a model result, and so the distribution of melt ponds doesn't necessarily reflect reality.

2012 really stands out and 2018 doesn't come close to it, but it does seem that during June more extensive melt ponding took place than in 2016 and 2017. It moved Dr Schröder to lower the CPOM June prediction for the SIPN Sea Ice Outlook by half a million km2 compared to the June prediction. The submitted prediction also had this: 

While melt pond fraction has been generally low in May, the area covered by ponds in June is above the mean (the last 10 years) in the Central Arctic and below in the Siberian part.

So, let's have a closer look at weather conditions during May, June and July. For that, I have downloaded weather maps, both for sea level pressure (SLP) and surface air temperature (SAT), from the Daily Mean Composites website, based on NCEP/NCAR Reanalysis data and other datasets. Mind you, the two maps on the right run to July 25th, not 31st:

First, SLP. It's always good to remember that green, yellow and red represent high pressure, while blue and purple represent low pressure. Basically, this tells us something about how cloudy the Arctic has been, because in high pressure areas clouds are less prevalent than in low pressure areas (or cyclones). As we can see, high pressure dominated over the Central Arctic during the second half of May, but after that clouds took over, for a large part explaining the slowdown in sea ice extent/area decrease.

Of course, temperatures play a role as well. In July they were anomalously high over the Siberian side of the Arctic, but on this front things slowed down too during July. It's interesting to note that CPOM's melt pond fraction model had below average melt ponding in the Siberian part in June (which is why I underlined it in the quote), whereas skies were reasonably open during the first half of June, and anomalously warm throughout the whole month. I'll get back to why this is important at the end of the blog post.

But first we compare, compare, compare some more. It's interesting to see how things have evolved, pressure- and temperature-wise, this melting season so far, but without context it doesn't mean much. And so I've downloaded the SLP and SAT maps for every year since 2007, or actually, re-downloaded them because the format had changed, cropped them, and combined them into the largest images I've probably posted on the blog since starting it (click for full size, SLP left, SAT right):

Don't try to take it all in. Just compare the different periods of this year with those of previous years. For instance, on the SLP side (left), 2018 can't keep up with 2010 and 2012, the big volume droppers. And as for SAT (right), 2011 and 2012 stand out much more. It's all about those periods of intense sunshine and cyclones. The longer they last, the more impact they make. They usually don't last longer than a week, although there are exceptions, like 2007. This year did see two feisty cyclones which stirred things up a bit, and then quickly fizzled out. But no real sunny spells to speak of.

There's one more measure we can look at and that's sea surface temperature. For that I use DMI's SST anomaly maps:

And here too it seems 2018 can't keep up with the Joneses. But is that it? You can never know with the Arctic. As we saw in that animation of UB SIC maps huge swaths of lower concentration suddenly appeared. Now, this ice won't melt out in the coming week, but it's a sign that the ice in these regions may not be making it to September. In fact, we might be seeing some of the flash melting and detachment that made 2012 legendary.

As to how low 2018 may still go: It has caught up quite a bit in the past few days on the JAXA sea ice extent chart, almost as the past two years now:

This probably has to do with the warm winter, the intense cyclones, and if the CPOM melt pond fraction model has it right, more melting momentum than in 2016 and 2017 as well. All of this is based on the tools I have used for years, tools that allow one to compare to previous periods. Over on the ASIF, commenters use many more sources, like different thickness models, but it's difficult to ascertain their reliability, and they often don't allow one-on-one comparisons with the past.

Also, we can't know what may happen in the last 5-8 weeks that this melting season has left, but we can look at what the next week has in store for the coming week, according to the ECMWF weather forecast model (source: Tropical Tidbits, click for a larger version):

Now, that's some serious high pressure, and though not positioned over the Beaufort Sea, it is covering a large part of the Central Arctic and western Siberia. This is going to diverge the ice pack some more, bombard it with solar radiation, and push the vulnerable parts towards the Pacific.

And this brings us to why what's happening on the Siberian side of the Arctic is so important. According to PIOMAS the ice is thicker there than it was compared to the 2011-2017 period (red means thicker, blue means thinner):

To be more precise, here's Wipneus' comparison with 2012 and 2017, the lowest volume years as of June 30th:

If those red zones near the Siberian coast have disappeared on the next PIOMAS update, this 'average' melting season without any real ice melting weather to speak of so far and no Fram Strait export whatsoever, could still end up among the lowest on record. Does that mean we may be witnessing another moment where ice extent keeps going down steadily, regardless of the weather, regardless of melting momentum even? Or will August weather present a couple of surprises that cause this year to crash down to a top 3 position, making it even more difficult to make heads or tails of dynamics?

Lots of questions, and I guess we're about to find out. Either way, this has been the longest prelude to a PIOMAS update I've ever written.

Comments

[NeilT]

Thanks Neven that is a huge amount of work to try and analyse an arctic which refuses to conform to any norm.

I'm guessing that the Arctic is transitioning state so fast now, as it moves to a new blue ocean state, that prior methods of analysis are going to give different end results.

Even if this year doesn't make top 3, pretty soon almost any melt season, conducive or not, is going to produce top 3 results.

Essentially the poles of the world have a cooling budge and a warming world has almost spent the Arctic one.

[Alan Clark]

The plural of phenomenon is phenomena. Apart from that, a great article.

[Neven]

Thanks, Alan, fixed now. There's also a mistake in the melt pond fraction image: The legend is correct for May, but should be ten times higher for June, running from -20 to +20.

[Neven]

In fact, we might be seeing some of the flash melting and detachment that made 2012 legendary.

Already happening.

[John Christensen]

I will need more time to digest, but what a great overview - amazing Neven!

And yes; it will be very interest with the high approaching. I suspect (Always with the optimistic inclination) that since the high will be closer to the Pole than was the case with the flash-melt of 2007 in Beaufort, then the radiation influx will be notably less significant, but let's see - it's a cliffhanger!

[wayne]

It is not too hard to understand Neven,

You don't need "melting weather" when sea ice is thinner, the melt is already in the system. Has occurred by less accretion from one dark season to the next. Particularly since September 2012, a new kind of sea ice exists, the summer response in 2013 was key. Thinner sea ice gives more favorable conditions for Cyclones to exist in the midst of peak melting solar forcing conditions. It does not imply a sea ice recovery, it means greater albedo by clouds during summer, nevertheless sea ice gradually thinned year by year. To 2018 point when the first year ice is so thin that it melts under lesser favorable weather. When thinner sea ice does vanish and is not compacted it leaves the older sea ice pans loose, available to circulation to crush them and turn them to mush, this mush takes very little energy to melt. What we also see is new compaction zones, a reflection of this new circulation system, again, the compaction zones are not proof of sea ice recovery, but rather a pile up of thinner sea ice which gives the impression of recovery. Therefore they linger longer during the melting season, but eventually crumble rapidly under direct assault from more open sea water. 2018 has had great compaction towards the CAA this spring, its apparent sea ice appears strong to this date, but will fail as soon as it decompresses. The melt pond maps do not reflect the circulation process, as your essay shows, they imply greater solar forcing areas. Again I am not at all surprised by what has happened to date because I predicted this in April, therefore a certain degree of understanding was achieved. Looking for alternate explanations may be helpful, but I doubt that you will find them satisfactory.

[John Christensen]

Hi wayne,

You stated:

"nevertheless sea ice gradually thinned year by year."

Has Arctic sea ice thinned year by year since 2012?

Do you have data on this?

[wayne]

Yes John

Look at Today's JAXA map, no favorable weather conditions and look at all this melting. Your data is PIOMAS I presume? Look at map presented above, nearly all the greater sea ice thickness zones are open water.

Furthermore, your logic trying to protect an experimental model fails in an elementary way, near all time low extent maximas , automatically imply lesser volume. You see volume has 3 components, no surface sea ice means no volume at all. And I know your data is irreproachably correct, yet fails explaining anything of current value.

[Neven]

The data that Wayne has, is the work he has done on Extremely High Horizon Refraction (which is the name of his blog). The problem for me, as someone who tries to make things more understandable for a general lay audience, is that it's difficult to explain, partially because I'm somewhat limited scientifically.

Wayne, what do you still expect for this melting season? Do you see it getting close to 2012, come September?

[wayne]

Hi Neven,

EH2r.com seems complex, because it has relativistic implications,
in many interesting ways, especially with regards to speed of light issues.

But in essence the greatest work is a lonely one, I measure the sun disk sizes in order to determine the temperature of the entire atmosphere, which is much much more precise than any satellite platform to date. Spring 2018 had a very cold Upper Atmosphere for the CAA, in fact the coldest since 2002, the platforms missed a good chunk of that. This was an obvious La-Nina cooling response as always measured post major El-Ninos, however the post 2016 La-Nina was never as strong as post 1998, so the cooling was limited to a narrower upper air region of the Arctic, causing a massive CAA vortex, sadly forcing North Atlantic air over the near entire Arctic Ocean all dark season. Therefore less sea ice accretion.

AS far as minima 2018 is concerned, I am bound by my prediction in April stating slightly more extent than 2012, but that prediction seems on shaky ground now. It is definitely 2012'ish, say the 1.5 million km2 East Siberian ice peninsula goes, it will be less than 4 million for sure, the rest depends on decompression of the main pack against North America and especially if the Pacific super warm sst's will spread out towards the Pole.

[John Christensen]

wayne,

So you must then fundamentally be in disagreement with the PIOMAS model, since PIOMAS shows 2011, 2012 and 2017 having the thinnest sea ice and 2018 being thicker:

http://psc.apl.uw.edu/wordpress/wp-content/uploads/schweiger/ice_volume/Bpiomas_plot_daily_heff.2sst.png

Is that correct?

[Neven]

John, that's for the end of June. Maybe wait with that question until PIOMAS data is updated next week. Although Wipneus did do a mid-month update and said: "The volume and volume-anomaly graphs show the melt in July is higher than most until now".

A lot of the zones that are red on the end of June map that Wayne referred to, will probably be gone on the next update. Depending on what happens with the other zones, 2018 could close the gap significantly.

[wayne]

Yes Mr John

Isn't PIOMAS experimental? There is a fundamental difference between my work and PIOMAS, I acquire rock solid data , not subject to errors, PIOMAS is a presentation from data sets, in other words , I interpret from one set of rock solid data, PIOMAS is an interpretation from a mish mash of data bases, which probably needs more work....

[John Christensen]

Neven,

Yes, and given the changes in area and extent I agree that July should show increased melting rates. DMI somehow has not captured that, but could be a flaw in the DMI volume assessment.

And wayne; that works for me - you are right of course that the models all are experimental.

My ongoing assumption for any of these metrics, including e.g. the DMI 80N 'average' temp, is that they do not necessarily reflect the reality accurately, but at least they can be used to track relative changes.

That is - of course - unless the changing Arctic environment cause these models to become increasingly inaccurate with a bias towards everincreasing overstatement of Arctic sea ice.

[wayne]

Right John

DMI 80 is not necessarily wrong but there should be a pure sea ice surface temperature reading DMI 83.5. There is still languishing much colder air in one sector of the Pole.

[Hans Gunnstaddar]

R. Scribbler has a new YouTube video linked below, attributing the recent Arctic melt to above normal temps along Western North America, including East Alaska, Beaufort coast.

https://www.youtube.com/watch?v=o8k9K-jhIT4

[wayne]

Hi Hans

Alaska has been mainly warm all past winter, but the flow of air hitting the sea ice area in question came largely from NE Russia for weeks. It is rather monstrous sst and continuous smashing of sea ice which made great contributions. The latest above normal air temperatures cited might have helped the melting further. But the process took a long time to fruition.

[Hans Gunnstaddar]

Yes, Wayne, I'm sure other factors as you describe helped contribute to that melt event.

One thing worth noting from the video is at 1:24, freeze that frame and you can see the anomalously warm waters of the Pacific feeding right into the middle of the above normal heat band on the Western coast of North America. The ocean feeds the hot weather, ending up contributing to melt in the Arctic.

[wayne]

Well Hans

Scribbler explanation is too simple, the 7 day NOAA charts have warmer lands but average sea surface air temps, and as an example amongst many, does not explain what is going on near Wrangle Island right now. We remember you bringing up +30 C weather on Russian side a while back which did not do much to the sea ice near East Siberian Sea.

In addition, there is an amazing ongoing 100 K melt in one day near Wrangle, I am sure, the whole Arctic lost a lot of sea ice on July 31, some of the broken ones scattered, may be bringing down the total result. But I would say 150 to 200 K km2 lost on the last day of July...

http://eh2r.blogspot.com/2018/07/monster-hot-sst-from-pacific-finishing.html

Neven ,

Wish to express my thanks in advance to JAXA , arigatou Japan and Vielen Dank to good people in Bremen. I am sure they will do a major press release in next few days.

[Hans Gunnstaddar]

Well, Wayne,

We'll have to agree to disagree. Warm air pouring in over the Arctic as Scribbler describes would have a contributing effect on sea ice melt. How could it not? Also, warm air over the ESAS might have had a contributing effect but not as noticeable as what happened recently in the Beaufort.

[wayne]

Hans

Actually I agree with you,

but the recent contribution of warmer air came late, is in a small sector, might have added some melting but did not do the major damage as we see unfolding right now. My point over ESAS is similar, it did contribute to over all melting but did not have an apparent immediate visual impact.

[Hans Gunnstaddar]

Ok, accepted, and now we will see what August has in store, as today is Aug. 1st. I have no prediction this year, just fascinated by the unfolding situation.

[EthonRaptor]

Has anyone else notices the polyannas (or worse open water) between Greenland and Ellesmere Island and between Greenland and the central Arctic Ice pack.

Any ideas about why these are developing or if they are real

https://seaice.uni-bremen.de/data/amsr2/today/Arctic_AMSR2_nic.png

[John Bilsky]

I'm generally just a lurker but will comment anyway since the cause is relatively obvious to me. The ice is in bad shape to start with and the recent shift to a clockwise circulation is spreading already broken ice apart. Here's a fun site to watch. Be sure to have plenty of popcorn to watch this slow motion train wreck. :-)

https://earth.nullschool.net/#current/wind/surface/level/overlay=temp/orthographic=200.94,92.16,805/loc=96.248,89.958

I think that Wayne is really on to something with his analysis. Time will tell huh?

[John Bilsky]

Here's another good one:

https://worldview.earthdata.nasa.gov

Be sure to click on the upper right corner icon for the Earth and pick the Arctic from the drop down selection.

JB is going back to lurking (with popcorn)

[Susan Anderson]

Thanks for a great article and fascinating discussion/argument, that's very informative and useful. Thanks also for the links. Neven, your overview is solid gold.

One thing to add. Right now, there's a tropical storm in the Pacific (just formed) that is heading straight north past Japan. Oops - just double-checked and it looks like more trouble for Tokyo, but still the energy is pushing north, and there's a lot of it.

[zakelwe]

Wayne said July 31st

"Look at Today's JAXA map, no favorable weather conditions and look at all this melting."

You do realise melt occurs from the bottom as well as the top Wayne?

I find people always tend to favour their pet area when it comes to melting over others, ie melt ponds etc etc.

This can lead to bias and incorrect predictions due to each melting factor not being correctly weighted ( a tricky enough problem as it is).

[zakelwe]

John Christiensen wrote on 31st July

"DMI somehow has not captured that, but could be a flaw in the DMI volume assessment."

Their volume map currently lags reality compared to satellite photos and therefore I assume their volume graph does too. As long as it finally catches up of course not an issue. Will be something to watch closely at the minimum time period.

Even if it overestimates as long as it overestimates same each year not too much of an issue.

As a note Bremen ice extent map closer to reality.

[John Christensen]

Hi zakelwe,

Yes, especially in the Golf of Bothia and Foxe Basin in southern CAA the sea ice thickness seems overestimated on DMI charts.

[Susan Anderson]

Thanks sakelwe, good point. Further to my own comment, I see it's not as pushy as I thought it might be, but here's a way to track that push:
https://earth.nullschool.net/#current/wind/surface/level/overlay=temp/orthographic=176.95,58.52,442/loc=-152.155,3.612

[wayne]

Hi Zakelwe

Yes sst's are extremely important, but you have to keep in mind that the winds either bring the warmed seas North, or the colder seas South, as well as brings the sea ice towards extermination, or saves it from pushing it away from warmer waters. Ice pack mobility becomes ever so important in this geographic equation in the not so distant past, pack ice was dense and stubbornly moved one way after weeks of wind pummelling in one direction, now the pack is so fluid it changes direction in a matter of hours. On top of that you got sst and atmospheric anthropogenic warming making sea ice extremely vulnerable, ie not necessarily subject to massive warming during solar peak radiation, but rather subject to any weather whims, making it appear either vulnerable or very robust, thin sea ice can give this illusion.

[wayne]

I usually don't do revisions of my projections, because it is a bad habit, is better studying why a forecast was wrong than to change it. But it looks very much like my April projection was wrong, 2018 is poised to surpass 2012, despite the great cyclone event this date 6 years ago:

http://eh2r.blogspot.com/2018/08/6-years-later-what-about-80-degrees.html

What matters is strictly in the Central Arctic Basin, so we study there....

[NeilT]

Don't be too quick to write it off Wayne. The NSIDC chartic is now showing a step away from 2012 at the point that 2012 was very rapidly dropping.

The thin, weak, ice in the Beaufort and the ESS may have vanished quickly, along with the peripheral ice which should never have still been there.

But we don't have a GAC and things are trending slightly back to the norm.

Time will tell as to just how much the thin ice and the mobility of the ice will push 2018 over the edge.

More interesting watching to go.

The ice is so thin and weak that surprises are bound to happen. It is, now, virtually impossible to say X happened then and Y will result today.

Every model is now a moving target and even direct comparison won't yield the same results because we simply don't have enough quality in the data that we use to compare.

One thing is for sure. 20 years after the ice has gone for the first time, there will be a perfect scientific article which tells us why it happened....

[wayne]

Heads up to top regulars!

I wonder where Jim has been???

There is a totally unfamiliar very wide sea ice formation North of Greenland and Ellesmere. I actually worked in that area in 1984, been looking at remote sensing pictures since 1985. Never seen anything like this.

http://eh2r.blogspot.com/2018/08/totally-unfamiliar-wide-region-of-sea.html

[NeilT]

Nares running backward, CAA melting inside but not outside very much, Beaufort ice near the coast intact but the CAB on the Beaufort side taking a beating. Still having ice in the Hudson and the Baffin ice is slowly melting away.

It is extremely odd. However, unless something changes radically, it's still unlikely to make a new record. Although the Atlantic side is now beginning to show a rather feathered shape as though it is taking a significant beating without giving up too much extent.

[wayne]

"unless something changes radically, "

How bout entire pack veering clockwise?

http://eh2r.blogspot.com/2018/08/laptev-sea-surface-temperatures-bites.html

Hi NeilT

The change is already in place, the shift in entire pack rotation did cause a scattering spread which simulates a slowdown in melting. If you look at the JAXA loop, scattering occurs in tandem with incredible melting. The entire pack shoreline is shrinking or where the flow is towards open water, remaining steady, implying melting as soon as ice is hitting wider water...

I count September or October Maxima less than 4 million for sure as days go by will see how much less than 4 will be.

[AnotherJourneybyTrain]

I've always wanted to ask, if I may:

Regarding Andrew Sleters 50 day forecast, what does the light blue line called "Anom. Forecast" mean in a functional way?

[Zakelwe]

My 4.1 extent forecast for 2018 was wrong by a long way out. Has it has been for some years and correct on others.

Best never to guess the yearly result and instead watch the trend.

Onwards to 2019

:)

[John Christensen]

Hi wayne,

I wanted to follow up on our exchange from end of July above, where you stated:

"your logic trying to protect an experimental model fails in an elementary way, near all time low extent maximas, automatically imply lesser volume. You see volume has 3 components, no surface sea ice means no volume at all. And I know your data is irreproachably correct, yet fails explaining anything of current value."

So I have used area, extent and volume models combined with Arctic weather oversight in general to make my prediction, while you have acquired:

"rock solid data, not subject to errors"

This data apparently caused you to revise your prediction on Aug. 5th for Sept average minimum SIE from slightly above 2012 to surpass 2012 levels and bring new record low SIE.

How is your data holding up currently?