« PIOMAS May 2017 | Main | Melting momentum: May 2017 »


Feed You can follow this conversation by subscribing to the comment feed for this post.


For those who like to share what they think this melting season's outcome will be, two June polls have been put up on the Arctic Sea Ice Forum (there will be new polls at the start of July and August):

JAXA 2017 Arctic SIE September daily minimum


NSIDC 2017 Arctic SIE September average

Happy voting!


Good summary of a critical metric at a pivotal time.
One thing that catches my eye is the range of trajectories in Jim Pettit's graph. If a neutral expectation for volume minimum is in the middle of the pack, a regression assumption (i.e., moderation of the current extremely low anomaly) would land somewhere above the middle: safely higher than the record low. But it doesn't feel like that assumption is valid this time. Low average thickness, thicker ice heading for the exits, and the seeming likelihood of melt ponding over first-year ice that extends well into the pack... not good at all.


In short Neven that sounds much like 2007 with less ice volume at the outset. Set up for extensive melt and extensive export.

As 2007 was the critical pivot year in massive volume loss that is not good news.


Iceman, a more realistic neutral assumption would be the 2007-2016 average which puts 2017 well below 2012. If Pettits graph only showed the past 10 years only one trajectory would put 2017 above 2012. Reversion to the mean on the current trajectory only means return to the slope which would be a negative anomaly of 6 which would still put the volume near record lows. The thickness graph suggest less than a quarter of the remaining ice extent is thick enough to be a contender for surviving the summer.

Rob Dekker

Thanks Neven.
Indeed, if we follow PIOMAS numbers, the outlook for September does not look good at all ; even for an average melting season we end up below 2012.
The only hope is that PIOMAS is wrong. There is some hope there, since Cryosat does not confirm PIOMAS numbers this year :

So who is right ? PIOMAS or Cryosat ?
I have to admit that I personally feel that, considering the very warm winter we just went through, that I think PIOMAS is more accurate than Cryosat, and FYI (the ice that will likely melt out over the summer) this year is thin and fragile.

If that is the case, we must see some evidence of that in the "extent" record as we move into June and July, in the form of a 'cliff' as the thin FYI melts out faster than other years.


Good reasoning to bias the projection toward more recent years. Archimid's graph of volume gain during the freezing season (post #1726 in ASIF PIOMAS thread) shows an increasing trend with a possible step-change around 2008. Conversely that implies even larger peak-to-trough losses since then. Also iirc, Chris Reynolds argued for a regime change in ice dynamics following the extreme 2010 volume loss.
Eyeballing Wipneus' PIOMAS chart, there seems to be a tendency for the largest (negative) volume anomalies to revert more. But it's not at all consistent, and doesn't necessarily happen before the melt-season minimum. No sign of reversion at all so far this year.


@Rob the last comparisons that were shown at the Forum do not yield big differences between Cryosat and Piomas


except for the concentration and location of that thick-ice blob, and the thickness in Greenland sea (if that is what makes the departure greatest, then the difference is irrelevant)

Rob Dekker

Thanks Navegante,
Poster "Michael" is doing amazing work on the ASIF regarding the differences and similarities between PIOMAS and Cryosat.
The latest difference plot
indeed suggest that the differences are minor, although the blue (more ice volume in Cryosat) still appears to dominate the field.


This is the latest graph from the PIOMAS team comparing their April results with those from Cryosat-2:


Since January, the difference has narrowed. However, it still exists.According to PIOMAS, the volume in April 2017 is the lowest of the time series. On the contrary, according to Cryosat-2, the volume is the third lowest, very close to April 2012, 2013 and 2016.


Rob, looking at Chartic, Extent is now dropping fairly fast, just dipped below 2006 and sitting almost right on top of 2010 with only 2010 and 2016 below it.

However 2016 went sideways from today onwards, for a week or so and with the periphery collapsing rapidly June is more likely to be an express drop in extent than any kind of pause.

What interests me most is the Bremen concentration maps which show _signigicant_ melt ponding right on top of the pole spiralling out to the Atlantic open water incursion.

Whilst there may be some open water there (but I believe more likely to be melt ponding), the whole concentration picture since May has been weak, broken, ice, all over the CAB with incursions of melt ponding in many different areas.

One to watch.

John Bilsky

FWIW, every time I check out the NASA Worldview and cloud cover has moved, I see new ice conditions that are unprecedented. Leads are everywhere and it appears that the entire Arctic is cracked to bits much more so than I've ever seen in the past. While I do understand that volume & extent are important, I can't help but believe that the physical integrity of the ice is going to play a huge role in this year's melting outcome. *sigh*

Rob Dekker

Diablo, thanks !
The interesting thing about that April comparison between PIOMAS and Cryosat is that Cryosat seems to project even less ice volume than PIOMAS, while in the Jan projection it was the other way around....

So what to believe ? :
When I look at Nico's FDD measures :
that suggests that the 2016/2017 winter accumulated some 4200 FDDs and prior years hovered around 5000.
That is a fraction of 4200/5000=0.84.
Ice thickness is approximately the SQRT(FDD) which means 2017 ice thickness is SQRT(0.84)=0.91 of prior years. Which suggests 10% decline in FYI thickness.

Now when I look at that PIOMAS/Cryosat April image you linked to, PIOMAS April 2017 has 20k km^3, and prior years are about 22k km^3. Which is indeed a 10% decline.

Cryosat April (and also Jan) seem to be less consistent, and in fact 'all-over-the-place'.

This makes be have more confidence in the PIOMAS numbers than in the Cryosat numbers.

Rob Dekker

Yes. If we are to believe PIOMAS (and there is no reason not to), then we can expect (even with an average melting season) about a 10% steeper decline of 'extent' over June/July than in prior years.
And that should be amplified by Arctic amplification.
Let's see what will happen, but the low volume numbers suggest its going to be a hell of a ride...


Rob, the one thing to add to the above is that all melting seasons are not equal.

So if we were to have 10% less volume growth but a 2013 melting season, then it would not matter very much.

However if we have 10% less volume growth and a 2007/12 melting season, then it will matter a LOT.

Add Arctic Amplification on top and... Not a pretty picture.


Posting for the sake of old friendship…

I haven’t done any work for months. In a way, I said all I had to say for a couple of years. I’m still watching though…

Thanks Neven, for keeping us all informed.

In March it looked like NASA had clipped my favourite means (the MODIS tiles) to study ice quality. But they came back. With some annoying re-aligning for me to get the new orientation in my CAD-format. I don’t feel like I’m going to do this as often as I used to.

As for what I saw, enhanced in CAD and comparable to June ’16 and ’12, I support what John Bilsky wrote above.
The whole Pacific side is in process of fragmentation, wind and current have easy grip. The ESS and Laptev look extremely vulnerable.

There still is a 1,8 Mkm2 mesh-grid pack. Until ’16 that was a remarkably stable area, in August (!). And it used to remain stuck against the fast ice lining the CAA and Greenland. As it was in June ’16. But it is detached now, leaving a mean 200 km wide stretch of mobile, broken ice with lots of open water.

Two parameters seem at least to be favouring the pack compared to the situation on 6th June ’16. One is the lack of melt ponding. The second is the better condition on the Atlantic rim.

Whether that will be enough to prevent a dive well under the ’16 minimum remains to be seen.

Rob Dekker

While average volume melt suggests that a record low is in the works this melting season, land snow cover is on the other extreme.

May land snow cover is out, and it is a whopper :

May left the highest snow cover since 1996, an astonishing 4 million km^2 more than last year.
That has got to leave a mark on Arctic amplification, and my regression formula (based on snow cover and ice extent/area)
suggests that September SIE average will be some 5.4 M km^2.
Note that May is really a bit too early for this method, since the SD is not much better than for a linear trend.

But since volume is pointing radically one way, and snow cover is pointing radically the other way, this melting season is going to put some theories to the test...;o)

Jim Hunt

Werther - A recent image of the "stretch of mobile, broken ice with lots of open water" north west of Greenland:


For comparison purposes here is the same area on the same date in 2012:

Bill Fothergill

@ Jim H
Thanks for that side-by-side comparison Jim - I'm sure to sleep better tonight! ;-)

@ Rob "... September SIE average will be some 5.4 M km^2 ..."

Wow. That is bigger than any of the NSIDC September averages since 2006.

We both know that SIE projections prior to June/July tend to lack skill, but that suggests that your algorithm perhaps places undue significance on snow cover - at least when that variable is appreciably at odds with the SIE conditions at the time of year.

I'll watch with interest how your SIPN values evolve over the coming months.

@ everyone. I know this is OT, but the BBC Horizon programme will be doing a special this evening concerning the enforced relocation of the BAS Halley station on the Brunt Ice Shelf.

The title will be...

Horizon: Antarctica - Ice Station Rescue

The programme should be available on line shortly after its transmission this evening.


Thanks Werther, I thought the Arctic Mosaic was gone for good.

I notice the Beaufort is one complete mess of floes with open water around them. There appears to be no contiguous ice at all.

At the current rate extent should drop into the lowest levels around the beginning of July. If it goes faster it will hit the lowest on record within a fortnight..

Still a lot of melt ponding in the centre of the CAB.

Interesting watch anyway.

Hans Gunnstaddar

"Still a lot of melt ponding in the centre of the CAB."

Hi NeilT, can you assist with the best graph/chart to assess melt ponding? Increased MP would seem likely since snow melts easier than ice, so cannot concur with Dekker's 5.4 km2. I'm at 2.5 km2 so one of us is way off or there is a happy medium it will end up.

Great comparison pics from JH. The ice this year already in a much weaker state than 2012?! Or, certainly at least in the compared region. Maybe there are other areas that can be compared?

Each melt season is different but due to record low volume and thickness, slushing around like it is in many areas, if there are widespread melt ponds in the CAB, that combination would seem to point towards great potential for a new record low. In any case, by July 1 the overall picture should be much clearer.


Sorry Hans, I only do this with Mk1 eyeball from the Area/concentration maps. One of the nice things about the area/concentration maps is that you can usually tell when it's melt ponding by the rapidity with which these maps (Bremen/NSIDC), show open water; which is suddenly ice again. As we know it is very unlikely for large patches of the pack to flip between open water and solid ice (at least at this time of year), it is a good indicator of melt ponds which either drain or freeze up again.

I'm sure there is a much better way on the forum, but I've never really been so keen to get into hard numbers when watching this.

I find that the deeper you get into hard numbers the less you notice the big picture changing around you. Tends to cause a few surprises.


Neil, hi,

The only melt ponding I see on MODIS is now in the Southern CAA and around Pevek/Ostrov Wrangelya.
There's still a way to go for this melting season to kick off. Even though 'winter power' was awful once more. The Arctic has a lot of tricks to pull...


I've just seen on the forum melting thread Neven using a rule of thumb of 3 days persistence for melt ponding on Bremen concentration overheads.

I'm sure I didn't do that so I'll keep that in mind.

Dean B

Given the large abnormality of snow coverage over the hymalayas this year, how much of the abnormal snow coverage for the northern hemisphere is due to that and how much to the arctic. Clearly there is way more snow coverage than last year but then last year was exceptionally early snow melt.

Susan Anderson


"I find that the deeper you get into hard numbers the less you notice the big picture changing around you. Tends to cause a few surprises."

Another useful rule of thumb.

fwiw, I have looked for the last few weeks at the overall surface temps of the entire Arctic ocean area, and they are very close to 0. I don't know if that's normal for this time of year, but it seems a bit high for the beginning of the melt season.

Hans Gunnstaddar

"I've just seen on the forum melting thread Neven using a rule of thumb of 3 days persistence for melt ponding on Bremen concentration overheads."

NeilT, will look to see if I can see those. Would be nice if there was a specialized imaging system that pinpointed melt ponds with tiny neon pink or green dots, then kept a running tally on number of them and average size, then as years of data compile, compare to previous years in bar charts. But I suppose everybody wants more detailed info. of one kind or another.



Somewhere on the forum is a discussion about melting. If you take a glass of water filled with ice cubes and heat the water, the water temp will rise to 0C then remain there until the ice has melted. Only once the ice has melted will the water temp start to rise again.

As I understand it, the same is true of the surface temps where ice is melting. It raises to 0C and then sits there until the ice is gone.

So it's not unusual to have 0C surface temps, it just means that if there is heat input, it is melting somewhere.

I've noticed that every time I go to the Barrow Webcam it is sitting below 32F. But the water on the ground is liquid and the coastal ice is melting with fresh water melt ponds which are not freezing. The energy is getting to the surface and the ice is melting.

Susan Anderson

Yes, I realized during the unfreezing (or freezing) process the temp would need to stay near the freezing point. As soon as you get near land the temps are mostly lower (exception, Alaska). Thanks.

Susan Anderson

"Seawater freezes at about 28.4* degrees Fahrenheit"

* -2C

Rob Dekker

DeanB, regarding where the May land snow anomaly resided :
The excess snow in May seems to mainly form a ring around the Arctic, with most excess on the Siberian side.
Explanation : It was just a rather cool May in the Arctic.

Bill said

but that suggests that your algorithm perhaps places undue significance on snow cover

Perhaps. Or maybe not. Check out the pan-Arctic situation here on MODIS :
and notice the pristine white ice and snow covered land all along the Siberian coastline.
Compare that to the situation last year :
where a snow-free Siberia and blue ice suggest that melting was in full swing all along Siberia.

It is still early to make predictions. But if that land snow anomaly holds through June, then melting onset in the Arctic will be seriously delayed, which makes 5.4 left over in September not an unlikely scenario.

Let us see what June brings, but I for one am hopeful that this land snow anomaly will compensate for the record low ice volume numbers.

Rob Dekker

To add to that last sentence : If the snow cover excess and the volume deficit even out, we may end up in what Hans called the "happy medium".

Jim Hunt

Hans - I don't think this is a very exact science, but see for example this graph from Wipneus:


Based on JAXA's "melt ice concentration" data:


Hans Gunnstaddar

It's an interesting possibility, Rob, the way you lay it out, especially since 2016 was the height of an El Nino. We shall see.

Thanks for the links, Jim. I guess the idea was more in line with something like the Bremen false color with tiny neon dots representing melt ponds, but I suppose those can be quite small in comparison to that big view, so likely just a fleeting thought. Meanwhile plenty of graphs and data to pour over.

Hans Gunnstaddar


That article is from Weather Underground, Bob Henson, 6/9/2017 (today!) Opening photo work the link click alone.

'Arctic Sea Ice Primed for Phenomenal Melt Season'

"Jim Pettit (who goes by the handle Neapolitan at WU) shared similar concerns in an email on Wednesday: “With the current record-low volume coupled with the not-record-low extent, we can assume that the bulk of the ice is thinner than usual. Satellite images seem to bear this out; it appears that much of the ice has already cracked up, and that which is still together is nevertheless frangible and fractured, and not likely to last until mid-September. The next 100 days or so are going to be very interesting indeed.”

Hans Gunnstaddar

Correction: ...worth the link click...

Hans Gunnstaddar


May, the annual high water mark for CO2 monthly average came in at 409.65 a whisker shy of 410.

Hans Gunnstaddar

Rob Dekker, in that article posted above from wunderground is one prediction that would be in the 'Happy Medium' range of 4 km2 between my 2.5 & your 5.4

"For 2017, the just-completed CESM-MPAS outlook is calling for an ice-depleting pattern dominated by surface ridging in June and July and more frequent surface cyclones in August. The outlook calls for a minimum extent of around 4 million sq. km. which would be very low but still above the 2012 record of 3.62 million sq. km."

Jim Hunt

Rob - It may not have seemed terribly relevant before this year, but have you considered adding "remaining volume" or some such to your regression formula?

Rob Dekker

Jim, yes, I did consider using PIOMAS to find out FYI thickness, and use that as an additional variable that affects the melting season.
However, I did not have time yet to set up the data stream for that and validate that variable (check how strongly it really shows up in the September extent data).

Rob Dekker

Another problem with incorporating 'remaining volume' or 'FYI thickness' in my linear regression method against 'extent' is that there may be non-linear effects : if ice melts, it adds to 'extent' as long as it has ANY thickness. After that it disappears quite suddenly.

Jim Hunt

Thanks to the sterling work of Wipneus here’s a regional breakdown of PIOMAS Arctic Sea Ice volume for the month of May:


Click through for a bigger version.

Rob Dekker

A bit dated, but since PIOMAS numbers are now released mid-month as well (thank you Dr. Zhang!) we had a PIOMAS update June 15.

Wipneus reports on the PIOMAS forum thread :

the volume gap with 2012 has been deminished from about 1. 48 down do 0.38 [1000 km3]

Which is quite surprising, but explains a lot :

It appears that 2017 has lost much of its lead in volume loss due to moderate melt in the Arctic over the first weeks of June.

I can't suppress the notion that this is caused by the lack of melting momentum that Neven already talked about, and I believe that this is at least partially caused by the land snow cover anomaly, which keep things cool in the Arctic.

We are not out of the woods yet though. The latest PIOMAS numbers are still record low, and although a significant land snow cover remains, it is still too early to exclude an extreme outcome in September.

If this (persistent) land snow cover is an indication, 'extent' should really start to level off over the coming weeks, away from its current position among the top three.

We will see...


The mid-month data looks like there's more ice volume now than in 2012 in Barents, Kara and Laptev, while there's less in the Central Arctic.

That could still mean less ice in the end, as it would mean more energy would go into melting the ice and less into warming water far away from the ice.

We'll have to wait and see...

Jim Hunt

The mid-month regional volume breakdown:

There's less volume than in 2012 across the Pacific side of the Arctic in the Chukchi and East Siberian Seas. On the Atlantic side there's also more volume in the Greenland Sea and Baffin Bay.

Amongst other things we'll have to wait and see when (or even if?) all that extra ice on the Atlantic periphery melts away to nothing.

Verify your Comment

Previewing your Comment

This is only a preview. Your comment has not yet been posted.

Your comment could not be posted. Error type:
Your comment has been posted. Post another comment

The letters and numbers you entered did not match the image. Please try again.

As a final step before posting your comment, enter the letters and numbers you see in the image below. This prevents automated programs from posting comments.

Having trouble reading this image? View an alternate.


Post a comment

Your Information

(Name is required. Email address will not be displayed with the comment.)