First Ice free day poll

[crandles]

Since 2013 and 2014 minimums were so high, I thought it might be an idea to have another poll to see how much later if at all people think ice free conditions will occur.

Other long term outlooks are welcome. Feel free to describe what you think will happen. If you wish that can include downloading the attachment and drawing what you think is more plausible.

To explain the attached graph:

Blue data is April Average PIOMAS volume
Yellow data is September average PIOMAS volume

Red and green are 4 parameter gompertz fits.

The straight blue line is intended to show acceleration as MYI over 2m thick collapsed due to less area, less age and thinner MYI mainly as the MYI stopped making it around the Beaufort Gyre. Once there was little MYI over 2m thick this accelerated phase dies out so that the thinning was only at a rate consistent with just FYI thinning.

The purple line descends more quickly than the blue line again aiming to be consistent with past. If there is less ice at the start of the season then albedo will be lower allowing more open water to form and more energy to be absorbed so that more ice melts. I have the difference between the two straight lines increasing from 17.8 K Km^3 in 2014 to 18.8 K Km^3 when we reach ice free.

This purple line first goes below 1000 km^3 in 2028.

Also note that the difference between 2012 and 2014 minimum volumes is nearly 3200 Km^3. Therefore while the purple line only gets down to 1000 Km^3 by 2028, it could occur any time after the purple line gets below 2600 km^3 which could be as early as 2022. I am therefore going to average 2022 and 2028 to arrive at 2025.

[viddaloo]

Since 2013 and 2014 minimums were so high, I thought it might be an idea to have another poll to see how much later if at all people think ice free conditions will occur.

Crandles, if you seriously believe that some people on this forum are thinking the ice–free day will never occur, you should make a vote option for the out-of-closet global warming denialists, IMO.

[crandles]

Crandles, if you seriously believe that some people on this forum are thinking the ice–free day will never occur, you should make a vote option for the out-of-closet global warming denialists, IMO.

I think a small fraction would suggest later than 2200 if I gave that as an option. I think I am limited to 9 options wrong more options possible. So even though I don't think many will really believe it perhaps I should give an extra option for after 2050 which I think former modeller William Connolley believes.

Is this better?

[viddaloo]

You already have/had a post–2050 option, so I was thinking of the «if at all people» that you mention in your next to top post.

[Neven]

I always thought this decade was possible, though not very likely. After 2013 and 2014 I don't think it's going to happen. Mainstream opinion is, I believe, somewhere after 2030. This means that as an alarmist I need a more aggressive prediction, and there's just one decade left: the 2020s.

But then I feel I have to err on the side of least melting, and so I go for somewhere during the latter part of the decade, but not the final part, of course (not alarmist enough): 2025-2027.

[crandles]

You already have/had a post–2050 option, so I was thinking of the «if at all people» that you mention in your next to top post.

Originally the latest option was '2035 or later' I think but I added more.

> «if at all people»
When I wrote that I was thinking 'if at all later than what was thought by people here a couple of years ago. I didn't want to imply that people must change their opinion to a later date if they didn't think it appropriate.

[viddaloo]

Edit: Yup, I see that now, Crandles, sorry for misreading you!

I'm waiting a bit to see if this purple line swings downwards. Everything depends on that, IMO.



PS:

• Can you change your vote later?
• Are votes invisible until the March 20 deadline?

[crandles]

PS:

• Can you change your vote later?
• Are votes invisible until the March 20 deadline?

Anyone should be able to see the total votes for each category but only after they have voted. Can you not see them after voting?

I would like to make it so people could delete their vote and vote again. I cannot see how to do that even though I have been able to do so on other polls on this forum.

I thought I would give a long deadline before votes close - hoped this would encourage discussion.

No-one gets to see who voted what unless they post their vote (optionally with reasoning) - at least as far as I know.

[Neven]

I would like to make it so people could delete their vote and vote again. I cannot see how to do that even though I have been able to do so on other polls on this forum.

Crandles, I've edited the poll and ticked the box that enables people to re-vote.

[ChrisReynolds]

Crandles,

Clarification on two points.

Totally ice free? As in no sea ice at all in the Arctic Ocean?

And is this the first year after which 60% of following years have likewise no ice?

Once I'm a bit more clear on those points I can vote, but to be frank I'd not really thought of such a scenario. I think below 1M km^2 regularly will be in the late 2020s to early 2030s. Totally ice free? I may have to vote for the 2040s or even 2050s.

Even with the experiment where PIOMAS is artificially thinned by 1m in early summer there's a significant residual of sea ice at the end of summer. Totally ice free implies very thin ice off the CAA, and that implies substantially warmer winters with thin ice rafted and mechanically deformed off the CAA into fairly thin ice. Such a situation is many decades away.

EDIT - re-reading I think I'm being overly cautious in asking for clarification, so I've voted for 2041 to 2050, but my PDF extends after 2050.

[wili]

CR (as usual) brings up good points. "1M km^2" used to be the definition around here for 'essentially ice free,' and I got the impression that when most people casually talked about 'ice free' in discussions, they were generally talking about this kind of 'virtually' ice free conditions.

It is much more difficult to decide when there will be absolutely no ice in the Arctic at all.

If nothing else, ice will continue to calve into the Arctic and thereabouts from GIS and from CAA for a very long time. Such calving and run off will also both freshen the surface waters, and it could also possibly lead to regional cooling around mid-century. (I think Hansen, among others, have made these predictions.)

So I'd be much more comfortable predicting a 'virtually' ice free date (probably by the end of this decade, but not likely later by the end of next decade) than saying anything about completely ice-free day.

[viddaloo]

I tend to agree with that. Otherwise we'd have to exempt sea ice that stems from calving, which seems troublesome. (And if we don't exempt calving, then ironically the 'if at all' option becomes relevant again!)

[viddaloo]

According to this collapse (hi, Bob!) graph based on yearly minimum volume percentage of winter maximum, I'd have to say earlier than 2023, because of the documented inherent conservatism of such curve–fitted trendlines. My options are then '2019 to 2021' or earlier. And as the last year has been a rather extreme outlier on this graph, a more balanced trendline next year or after the 2016 minimum will probably end at a year before 2022. To compensate for the inherent conservatism of this 2nd order polynomial estimate, I then have to go down a couple years from the anticipated more balanced trendline crash of 2020/21, so my guess will probably be '2017 or 2018'.

[ChrisReynolds]

I think you can exempt calving pretty easily. During summer MODIS (or similar in the future) could easily be used for that purpose. Failing that, with a seriously thinned ice pack, large bergs would stand out on an EM flight over the CAA Arctic coast (I think).

Calving isn't the issue.

For as long as there is ice over winter there will be compaction against the CAA. The PIOMAS thinning experiment isn't realistic in that respect as the ice was thinned in early summer. However thinner ice is more easily ridged and transported by the wind, so the compaction against the CAA will produce ice that is thicker than the majority of the pack, whatever the thickness of the bulk of the pack. I am not convinced it will be easy to remove that ice totally. I suspect that not only will substantial winter warming be needed to get it thin enough to melt late season (as the rest of the pack has to melt out to allow the ice edge to get near it), but substantial ocean warming will also be needed.

I can post graphs if needed but the greatest source of heating is in the autumn, and that is heat venting from open water warmed during the summer. So using trends of annual average temperature to get a handle on future prospects is flawed because a large part of the annual warming is due to a heat loss process. Summer warming is pegged low because of all the melting ice, spring warming would be highly relevant, winter warming is to a lesser degree than autumn contaminated by the heat loss signature. It's hard to get a handle on how winter thermodynamic thickening will be affected by future warming. Yes it will go down, but how fast.

To examine the difficulties involved I want to use the relative robustness of the PIOMAS rate of change plot I've done, robust that is in the face of the changes seen since 1978. What it shows (peaking of volume loss rate around the solstice) is the dominance of the sun in the seasonal cycle.



Taking the rate of loss from the transition from gain to loss of volume in April I can accumulate the tend day increments of volume. Using the post 2007 period the cumulative sum of volume loss is 18.1k km^3. That's the amount of ice melted by available solar radiation. Therefore to get zero ice in theory you'd need a winter peak volume of about 18k km^3. That volume corresponds to a thickness of about 1.9m average for the whole pack (assuming an area of around 9.65M km^2), i.e. say 1.5m around the peripheral seas and maybe 2 to 3m thick off the CAA.

That doesn't sound a lot to lose. But the PIOMAS thinning experiment tells us that all is not well with this reasoning, because in PIOMAS there is still a substantial residual of ice even when the pack has thinned by 1m early in the season (from current levels that's a pack thickness of a bit over 1m thick average for the pack away from the CAA).

What I presume is going on is that the extra energy that would have gone into melting ice then goes into warming the ocean, but much of this is then vented to the atmosphere in autumn. During the summer there is no strong mechanism to mix this warmed water effectively with the remnant ice pack off the CAA, storms mix laterally but they also mix vertically in the water column, reducing the temperature of the warm surface water and thus reducing the effect on the ice (heat flux into the ice that melts ice depending on temperature difference between ice and water).

The whole matter is so complex I think only physical modelling can get a handle on it. And we've all seen by now the results of the study that used PIOMAS to project into the future.



Even I don't think winter ice will stay so level for so long (2026 to 2050), but add in the effect of ridging and ice piling up against the CAA and I can't see how we'll have a totally ice free Arctic Ocean much before the middle of the Century.

[Neven]

I'm quite sure crandles is referring to virtually ice-free, or ice-free for all practical purposes (better definition).

[jdallen]

Great analysis, Chris.

Viddaloo, your graph I think shows nicely how much more important short term weather will be in what metrics we see.  Less volume will translate into far more volatility.  The plots of the last five years illustrate that pretty well.

[crandles]

Perhaps we should have two polls, one for volume excluding any calved glaciers or ice shelves being frequently less than a thousand km^3 and one for frequently less than 100 km^3?

Suppose
20x1 < limit
20x2 > limit
20x3 > limit
20x4 < limit
20x5 > limit
20x6 < limit
20x7 > limit
20x8 < limit
20x9 < limit
20y1 > limit
20y2 onwards < limit

I don't think 20x1 should be considered the first year as it is followed by 2 years higher than the limit. So considering x1-x3 results in 33.3% which is more than 10% below 60%.

20x4 could be considered to be the first year. Although considering 2 years (x4 & x5) or 4 years (x4-x7) results in 50% that isn't more than 10% below 60% and considering other numbers of years results in over 60%.

I would like to avoid that level of complexity needing to be built into opening post. Therefore is it simpler and clear enough to say something like

When is the first year that the 5 year double smoothed centred trendline for September average volume goes below the limit?

[viddaloo]

I would like to avoid that level of complexity needing to be built into opening post. Therefore is it simpler and clear enough to say something like

When is the first year that the 5 year double smoothed centred trendline for September average volume goes below the limit?

Yeah, for that's super–simple! Not.

Why can't we just have a poll for the first virtually ice–free summer?

[Gray-Wolf]

I think we will see plenty of this type of debate from 'the other place' once sea ice drops below 1 million sq km. (LOL)

As it is I'm with the camp that sees the first 'ice free' year as a result of the weather conditions over that melt season ( and , Yes, this depends upon the ice conditions over the years prior) but there after is much harder to predict ice conditions due to the impacts of a basin of purely FY ice?

We are told that the perfect melt storm synoptics come around  on a ten to twenty year cycle with the previous two having a ten year period. Should we expect this behaviour to continue and would another '07' prove enough of a forcing to drop ice below the 'practically ice free' threshold ? I'm leaning toward a 'Yes' in answer to this which is why I plumped for the 22' to 24' option?

I figure that the next 'perfect melt storm will either do the job or leave the basin in such order that an 'average' year ( akin to 2012) would be enough to have area go below the million mark.

Flies in the ointment?

Well weather is changing. I believe that the shortening of the spacing between 'perfect melt storm' years ( to the ten years) was a result of the changes running through the 80's and 90's but we are beyond those forcings now and under a different regime. Who's to say that Mother N. does not have another 'stalling mechanism' as She struggles to resist the GHG forcings and keep climate stable? Could this 'relax' from the 2012 low be the start of such an episode with 'export' , or lack of it, being her key weapon of choice?

On the other hand could the 'hiatus years' have been Her best shot at keeping global climate stable and we are about to see Her switch up to the next 'stable climate setting ( with the Nino helping flip the -ve naturals back to positive and have us enter a period of rapid warming into the 20's?) This might impact the heat available both in the Arctic, and available for transport into the Arctic, bringing the ice free basin to us before the start of the next decade but also impacting winter formation?

As it is I have no doubt that we are headed for an ice free basin and, along it would seem with Neven, thought that before this decade was out we'd have such. The current rebound also has me jittery though as I cannot recall seeing synoptics like we have over the past two years ( esp. the low export component?) and so wonder whether we are seeing something more than 'weather' at play over the basin post 2012?

As ever time alone will tell all!

[werther]

The events during summer ’07, ’10 and ’12 really took me by surprise. All parameters showed values way under most or all trend modelling available then. The last two lower-melt summers have not influenced me in my opinion that a sudden crash under unfortunate circumstances is possible any year now. I never had the idea that crash would be complete; the <1 Mkm2 is what I have in mind for ’10-’19 if that ‘black swan’ situation arrives.

Nevertheless, the bifurcation-theory I learned from the blog during ’10-’11 still sounds convincing to me. The first <1 Mkm2 ’black swan’-event could trigger/initiate this bifurcation. In that case, I’d expect complete melt-out during summer some years later, and a year round icefree Arctic Ocean as a result within ten years from that stage.

I know it is hazardous to make any predictions on ‘black swan’ events. As I think the energy-imbalance in the system is now almost at the threshold (oceanic heat content), I’d choose 2017 as a possibly eerie year. That could make FI 2023 the ‘Crandles’ year. As a matter of speakin’…

PS without considering a methane-burp and the large statistical possibility that the smooth modelled decline will after all be right (which is bad enough).

[crandles]

Simplified the question

Should I reset votes?

Votes don't seem deletable again - I still don't see how I do that or affected it by changing the poll question. Neven could you add this again please?

Why can't we just have a poll for the first virtually ice–free summer?

With a lot of noise about the trend the first year could be well before the trend. Some people might adjust for this and others not. Therefore I want it to be clear that you shouldn't adjust for that.

[viddaloo]

When will be the first year that trend goes below 1000 km^3?

Much better question, Crandles! Thank you 

I'll still vote for the first occurrence of virtual, though, mainly because of what Werther refers to as 'bifurcation'. I'm thrilled that he talks about all–year virtually ice–free within ten years of that first ice–free day, as that is exactly what I've found inherent in the observational data (as little as six years, in fact).

When centering the annual average volume on the data point, with half a year before the date and half after for the annual average for that date, I get this frightening collapse graph (which nevertheless is still probably too conservative, as explained earlier):

[Memes]

What is the meaning of the poll question?  Is it about a trend line or the measured ice?  if one means the trend line there are two answers (1) "Today" - The trend line points to a time of less than 1000KM^3.  (2) "2028" - Where the trend line points.

[werther]

For the 'new' Forum friends, it might be convenient to repeat some info on this 'bifurcation'. It was discussed often in the early start of Neven's blog. I credit good old 'Lodger' for having referred to this often.

Eisenmann & Wettlaufer, 2009
Eisenman, I. & Wettlaufer, J.S., 2009. Nonlinear threshold behavior during the loss of Arctic sea ice. Proceedings of the National Academy of Sciences of the United States of America, 106(1), p.28-32.
In light of the rapid recent retreat of Arctic sea ice, a number of studies have discussed the possibility of a critical threshold (or “tipping point”) beyond which the ice–albedo feedback causes the ice cover to melt away in an irreversible process. The focus has typically been centered on the annual minimum (September) ice cover, which is often seen as particularly susceptible to destabilization by the ice–albedo feedback. Here, we examine the central physical processes associated with the transition from ice-covered to ice-free Arctic Ocean conditions. We show that although the ice–albedo feedback promotes the existence of multiple ice-cover states, the stabilizing thermodynamic effects of sea ice mitigate this when the Arctic Ocean is ice covered during a sufficiently large fraction of the year. These results suggest that critical threshold behavior is unlikely during the approach from current perennial sea-ice conditions to seasonally ice-free conditions. In a further warmed climate, however, we find that a critical threshold associated with the sudden loss of the remaining wintertime-only sea ice cover may be likely.

[werther]

When will be the first year that trend goes below 1000 km^3?

Well, even put like that, it doesn't change the line of arguing in my first post on the topic. I guess given my description that 5-year-smoothed trend would get below 1000 km3 about the year I suggest the first completely ice free Arctic Ocean.
Not initially through a 'black swan'-event in '17, of course.
After '23, the trend would be hovering around the 1000 km3 benchmark for a span of time (as it would even suggest some ice to linger on for 5 years after the bifurcated perennial-ice-free state around '33).
I'm speculating. Of course. Maybe I overestimate the 'rubber-band' effect of already baked in consequences. Maybe a dent will be put in the relentless spill of GHG's through unilateral negociations  (USA-China) and/or those in Lima/Paris.
It's Equilibrium Climate Sensitivity and Ocean Heat Content that concern me most.

[ChrisReynolds]

New question? Why? The bands are too tight for me to answer the question with a vote. I could do a monte carlo and produce a PDF, which would probably give a decade. But the trend issue means there's more maths involved than I can be bothered with for a poll so I'm bowing out. My previous vote can be scrapped.

Sorry Crandles.

[crandles]

We are told that the perfect melt storm synoptics come around  on a ten to twenty year cycle with the previous two having a ten year period. Should we expect this behaviour to continue and would another '07' prove enough of a forcing to drop ice below the 'practically ice free' threshold ? I'm leaning toward a 'Yes' in answer to this which is why I plumped for the 22' to 24' option?

I figure that the next 'perfect melt storm will either do the job or leave the basin in such order that an 'average' year ( akin to 2012) would be enough to have area go below the million mark.

Yes stepping down each time there is a near 'perfect melt storm synoptics' does seem a plausible interpretation of the data that gets mentioned not infrequently.

That makes trend analysis difficult so my 2nd attempt at the poll question may well be inappropriate for this people in this class of thinking.

[crandles]

Nevertheless, the bifurcation-theory I learned from the blog during ’10-’11 still sounds convincing to me. The first <1 Mkm2 ’black swan’-event could trigger/initiate this bifurcation. In that case, I’d expect complete melt-out during summer some years later, and a year round icefree Arctic Ocean as a result within ten years from that stage.

Mathematical catastrophe certainly makes trend expectations difficult to give. So my 2nd attempt at a poll question is increasingly looking wrong.

fig 1 of
http://authors.library.caltech.edu/29296/1/Eisenman2012p17064J_Geophys_Res-Atmos.pdf

suggest we only get bifurcation once we get to seasonally ice free but I suppose that could extend to less than 1m km^2 minimum area.

I can image 3 stable states:1) ice ball earth where albedo keeps things cold
2) partial like current climate
3) perennially ice free. If there is a huge amount of heat in the upper 50m of Arctic ocean then over winter a sea fog would likely develop which would slow down the radiative loss of heat to an amount that could be recouped during summer with an open ocean.

I can image an abrupt change from 3 to 2: stormy weather disrupting the sea fog so that much more heat is lost that is recouped during summer.

However my gut reaction is that going the other way from 2 to 3 requires so much heating of the top 50m of ocean that it cannot be an abrupt process.

I guess the above thoughts is just too simplified if Eisenman 2012 linked above has referenced GCM models that do all types of bifurcation shown in fig 1.

Certainly looking a long way ahead is difficult. Perhaps the trends have some validity up to the point of reaching seasonally ice free.

[Nightvid Cole]

I think around 2020. The extrapolation suggests a bit later, but of course you're going to have variation between high and low years from weather. Thus the first year with virtually no volume I expect around 2020. I also expect the last "not virtually ice free" summer to be around 2030, for a similar reason. Roughly half the years in the 2020s should be "virtually ice free".

[jdallen]

I think Arctic Ocean heat content needs to rise considerably before we see a melt down to that level.  I say 2029 at the earliest.

[viddaloo]

When will be the first day that this ice volume poll goes below 1 phrasing change * day^-1?

[ChrisReynolds]

OK, I've had the time to look at this.

When will be the first year that ice volume per PIOMAS goes below 1000 km^3?

Attached is a plot of September vs April volume (PIOMAS) for the Arctic Ocean in km^3. I don't like extrapolating mathematical functions but I'll do so here, there is no gaurantee that fit will hold. Using that relationship my estimate for April volume implied by a September volume of 1000km^3 is 16.7k km^3.

I've previously calculated the annual average thinning of thermodynamic growth as about 1.5cm/yr between 1986 to 1995 and 2009 to 2013.
http://dosbat.blogspot.co.uk/2014/04/what-is-future-of-arctic-sea-ice-part-2.html
See text around large blue scatter plot with red lines.

Such a decline amounts to 0.145k km^3 per year.

For April 2015 I estimate an April volume of around 20.7k km^3, based on differences between sept/oct/nov for 2014 and the April 2014 maximum volume.

Applying the 0.145k km^3 annual volume loss due to thermodynamic thinning to 20.7k km^3, while assuming that winter growth means the end of the massive volume loss from 1995 to 2012 (which is due to loss of multi year ice)...

I get 28 years to 1000k km^3, that's about 2043. Frankly I don't believe that.

I expect the future warming to cause an increased rate of thermodynamic thinning from the 1986 to 1995 and 2009 to 2013 period that I calculated the annual volume loss for. If I double the thinning rate, which seems not unreasonable, then I get 14 years to a summer of 1000k km^3, that's 2029. I'm going to round that up to the next bucket which means I vote 2029 to 2033, but I think about 3 years either side of that would be reasonabe wiggle room which means my preferred range would be 2026 to 2036.

[ChrisReynolds]

Interesting double peak in the distribution.

[ghoti]

It looks to me as if there are two domains - Sept volume less than 8000km3 and Sept volume over 8000 km3. The April volume for years with Sept volume appears independent of the Sept volume. For years with Sept volume greater than 8000km3 there does appear to be a linear relationship.

Do these "domains" represent pre and post 2007?

[ChrisReynolds]

Yes, excel is a bit crap. I can't get it to label years. Arguably post 2007 looks to me like a levelling, but the data is too short to be strident about that, so I favour extending the linear trend from earlier data.

Of course I am biassed because choosing the linear trend gives a similar result to that I've previously put forward for 'virtually ice free'.

[pikaia]

It looks to me as if there are two domains - Sept volume less than 8000km3 and Sept volume over 8000 km3

That could be just an illusion. If you plot a series of random points you will get some grouped together while elsewhere there are gaps. The points will NOT be evenly spaced.

[viddaloo]

It looks to me as if there are two domains - Sept volume less than 8000km3 and Sept volume over 8000 km3. The April volume for years with Sept volume appears independent of the Sept volume. For years with Sept volume greater than 8000km3 there does appear to be a linear relationship.

Do these "domains" represent pre and post 2007?

Yes, none of the pre–2007 years went below 8000 in September (and all the post–2007 did).

I find what you say interesting and to the point: After the 2007 tipping–point just about anything can happen during the summer, and it's less linked to the April or maximum volume. As low as we are now — and both 2013 and 2014 are way lower than 2007 in terms of annual average volume — another 2007 or 2012 can happen any time. And when it happens we'll probably go below 1000 and 'virtual ice free', and then the countdown starts for perennial open seas.

[Pmt111500]

Assuming smokes of the Iceland eruption landing this winter to the ice present creating a double side mirror under the snow falling this winter, and guessing this weak El Nino will be followed by a larger one 2015, I'd say 2017 is still on the table, but since my defeatist text was censored in SkS, I decided to move the exp curve of mine pointing at 2021-2024 ;-).

Of course the Icelandic eruption may blow stuff up high too at some stage, currently though the fissure is going through a see-saw in activity. Maybe it'll blow up in 2021, so I'd have to change the vote to 2025 ;-)

The good thing about this is that we'll get a whole lot more clouds here in the north, so summer heat might not be rising to inhospitable levels of 100s F for a bit. No stats from meh for now, I should first update everything since spring, and that may be too much for my comp.

[James Lovejoy]

I went for the 25-28 bucket without much confidence.  Random noise requires a much wider band.

If you just kept the poll open for re-voting until March 2020, I'd have a much better chance of a good guess.

[DavidR]

I've gone for 2019-2021. The relatively cool past two years mean that  a record before 2017 is improbable as it seems to require a run of three hot years to create a new record. 
Temperatures in the Arctic are rising by about 0.6 degrees/decade. If this continues the 5 year average in 2020 will be higher than the hottest single year yet recorded, in 2010 . 
These temperatures will produce significantly  more melt than we are currently experiencing every year and less winter volume. 
A run of hot years, in the Arctic,  as occurred in 2010-2012 will, I  expect, take the remaining ice below the level in the question before 2022.
There have been 3 occasions in the past thirty years where the anomaly in the Arctic has rebounded upwards in one year by  more than 1 degree, but no  cases where it has rebounded downwards by the same amount.  Each of these three 1 degree rises has led to  a record minimum extent within 3 years.

[crandles]

 
These temperatures will produce significantly more melt than we are currently experiencing every year and less winter volume. 

Are you sure it isn't mainly less winter volume causing higher winter temperatures? Perhaps also extra heat built up in melt season gets vented to atmosphere causing higher temperatures? It isn't unreasonable to believe higher temperatures cause less winter volume even if there is also cause and effect the other way so I don't have much objection to that part.


Regarding higher temperatures causing more melt, are you aware temperatures in summer seem lower not higher:
eg hottest year 2010:

[jdallen]

 
These temperatures will produce significantly more melt than we are currently experiencing every year and less winter volume. 

Are you sure it isn't mainly less winter volume causing higher winter temperatures? Perhaps also extra heat built up in melt season gets vented to atmosphere causing higher temperatures? It isn't unreasonable to believe higher temperatures cause less winter volume even if there is also cause and effect the other way so I don't have much objection to that part.


Regarding higher temperatures causing more melt, are you aware temperatures in summer seem lower not higher:
eg hottest year 2010:

Theory regarding those slightly cooler summer temperatures... Are they driven by heat taken up by phase change? Increased over water implies more evaporation, and would pull temperatures back to zero.
Where might we find dew point or absolute humidity info?

[crandles]

Theory regarding those slightly cooler summer temperatures... Are they driven by heat taken up by phase change? Increased over water implies more evaporation, and would pull temperatures back to zero.
Where might we find dew point or absolute humidity info?

What has previously been suggested:

Phase change with ice or melt water pond (in contact with ice) is likely to keep temps near 0C.
Phase change with ocean water (with ice nearby) is likely to keep temperature close to somewhere between 0 an -1.5C depending on salinity of ocean water. More open water means more held close to lower temperature of freezing ocean water.

Seems like a sensible explanation to me.

It is the heat flux into the ice that matters not the temperature. Whether the lower temperature means anything in terms of heat flux into the ice is less clear to me. Perhaps a little less heat flux into the ice? However even if so, I suggest this is completely insignificant compared to effect of albedo causing more energy to be captured rather than reflected.

[jai mitchell]

Cran,

This is consistent with increased cloud cover dude to intrusion of midlatitude moisture into the arctic.  This is increasing as the arctic is warming faster than the midlatitudes and increased water vapor in mid pushes into the arctic.

From 4 years of observations from Barrow, Alaska, it is shown that the cloud radiative impact on the surface is a net warming effect between October and May and a net cooling in summer. During episodes of high surface haze aerosol concentrations and cloudy skies, both the net warming and net cooling are amplified, ranging from +12.2 Wm-2 in February to -11.8 Wm-2 in August. In liquid clouds, approximately 50%-70% of this change is caused by changes in cloud particle effective radius, with the remainder being caused by unknown atmospheric feedbacks that increase cloud water path. While yearly-averaged, the warming and cooling effects nearly cancel, the timing of the forcing may be a relevant control of the amplitude and timing of sea-ice melt.

http://onlinelibrary.wiley.com/doi/10.1002/2014GL062015/abstract

[crandles]

Cran,

This is consistent with increased cloud cover dude to intrusion of midlatitude moisture into the arctic.  This is increasing as the arctic is warming faster than the midlatitudes and increased water vapor in mid pushes into the arctic.

From 4 years of observations from Barrow, Alaska, it is shown that the cloud radiative impact on the surface is a net warming effect between October and May and a net cooling in summer. During episodes of high surface haze aerosol concentrations and cloudy skies, both the net warming and net cooling are amplified, ranging from +12.2 Wm-2 in February to -11.8 Wm-2 in August. In liquid clouds, approximately 50%-70% of this change is caused by changes in cloud particle effective radius, with the remainder being caused by unknown atmospheric feedbacks that increase cloud water path. While yearly-averaged, the warming and cooling effects nearly cancel, the timing of the forcing may be a relevant control of the amplitude and timing of sea-ice melt.

http://onlinelibrary.wiley.com/doi/10.1002/2014GL062015/abstract

Yes I would expect more water vapour in warmer temperatures. Regarding cloud effects, the following seems to find different observations:

Kay and colleagues have also analyzed satellite observations of Arctic clouds during this same 15-year period. Kay’s research shows summer cloud amounts and vertical structure are not being affected by summer sea ice loss. While surprising, the observations show that the bright sea ice surface is not automatically replaced by bright clouds. Indeed, sea ice loss, not clouds, explain the increases in absorbed solar radiation measured by CERES.

http://www.nasa.gov/press/goddard/2014/december/nasa-satellites-measure-increase-of-sun-s-energy-absorbed-in-the-arctic/#.VKbTkSusXtx

[jai mitchell]

The water vapor is moving into the region from lower latitudes this is why the temperatures go up (+10 watts/meter^2 in winter!)

Your link has nothing to do with this fact.  Your link is regarding the difference of reflectivity in areas where the ice has disappeared, and consequent to that disappearance.

The parts I am talking about are happening at the very beginning of the melt season, before significant ice loss occurs.

[crandles]

Jai I was quite clear that I accepted that there would be increased water vapour.

Without seeing the paper in your link, I am not sure whether it is limited to an area around Alaska or limited to episodes of high haze or both or might more generally apply across Arctic.

The part of the link I was highlighting was the bit saying

summer cloud amounts and vertical structure are not being affected by summer sea ice loss.

This isn't to do which differences in reflectivity. It also appear to apply to Arctic rather than a region near Alaska.

It potentially seems quite surprising - more open water might be expected to provide more humidity.

You seem to be saying there is more water vapour moving in from lower latitudes. Presumably this air cools (except in peak summer time when insolation increases towards the poles). So this should presumably result in increased cloud cover at other times of the year?

Thank you for explaining you are talking about "the very beginning of the melt season".

Now, I can understand that in winter there would be increased temperatures, more water vapour and cloud cover and therefore radiative warming in winter; perhaps up to  +10 or +12.2 Wm-2. However, if the effect (of haze???) is positive in winter and negative in summer, what does this tell us about the situation at "the very beginning of the melt season"?

I just don't see how your link is supporting what you are saying but maybe I need to see the whole paper rather than just the abstract?

The water vapor is moving into the region from lower latitudes this is why the temperatures go up (+10 watts/meter^2 in winter!)

Temperatures wouldn't go up if this was happening previously, so are you saying that previously jet stream was more zonal? Or previously it was cold enough for the moisture to precipitate out whereas this isn't happening now? Is there a problem arguing this if there is no change in summer cloud amount?

You could be completely correct. However, I would like to see some supporting evidence or arguments. A possible alternative or additional explanation might be:

Thinner ice and more open water as the ice melts faster reduces albedo so more heat is captured in melt season rather than being reflected. This extra built up heat is vented to the atmosphere and then to space during the Autumn and Winter as there is less ice acting as insulation reducing the heat flux from the ocean.

Why is your explanation to be preferred to the above or why isn't both a better explanation for increased winter temperatures than either?

[jai mitchell]

your misinterpretation is that the clouds are being formed in winter due to warmer temperatures, in fact it is quite the opposite.

your link talked about measuring reflectivity over regions with no ice to see if clouds made up the difference.  It is a totally different animal than what we are talking about.

You could be completely correct. However, I would like to see some supporting evidence or arguments. A possible alternative or additional explanation might be:

well, the paper I shared had direct measurements. . .

the breakdown of the polar cell began around 2012, based on observations.  It could have been a bit earlier but this really synchs with dr. Francis' work on the irregular jet stream.

[crandles]

well, the paper I shared had direct measurements. . .

the breakdown of the polar cell began around 2012, based on observations.  It could have been a bit earlier but this really synchs with dr. Francis' work on the irregular jet stream.

Can you share a link to the full paper? (Otherwise it just looks like it is only about haze)

[jai mitchell]

Cran,

The paper is about Aerosols and cloud interactions.

During episodes of high surface haze aerosol concentrations and cloudy skies, both the net warming and net cooling are amplified