Peripheral Seas of the Arctic Basin Seasonally ice free state

[ChrisReynolds]

This follows on from a throw-away justification of a statement in my most recent blog post....

How do we define seasonally ice free? The ultimate definition is no ice at all, but that is too long a prospect for me. So I turn to virtually sea ice free...

The common definition seems to be total NH sea ice area extent of less than 1M km^2. But how might we apply this to the regional seas of the Arctic?

Average September extent for the 1980s in Wipneus's data set (near as damn it NSIDC Extent), was 7.219M km^2. 15% seems to be a good demarcation for working out extent, so let's try applying that to overall Arctic extent. 7.219M X 0.15 = 1.083M km^2. That's only a bit above the 1M km^2 level below which the Arctic Ocean can be considered virtually sea ice free.

So I propose that September extent of below 15% of 1980s September extent means that sea/region is virtually sea ice free.

Looking at the regions available:

Okhostk - ice free in summer most of the record - discounted.
Bering - ice free in summer most of the record - discounted.

Beaufort 1980s average is 0.34M km^2 - OK
Chukchi 1980s average is 0.25M km^2 - OK
ESS 1980s average is 0.67M km^2 - OK
Laptev 1980s average is 0.34M km^2 - OK
Kara 1980s average is 0.30M km^2 - OK

Barents - 1980s average only 0.07M km^2 - discounted.

Greenland Sea 1980s average is 0.30M km^2 - OK
Central Arctic 1980s average is 4.40M km^2 - OK
CAA 1980s average is 0.44M km^2 - OK

Baffin - ice free in summer most of the record - discounted.
Hudson - ice free in summer most of the record - discounted.
St Lawrence - ice free in summer most of the record - discounted.

Of the regions accepted as having a reasonable amount of ice in the 1980s during September, Central, Greenland and the CAA have no years with less than 15% of the September average. That leaves us with Beaufort round to Laptev, the peripheral seas of the Arctic Basin, and the Kara Sea.

The list below shows the regions concerned that do show a (virtually) ice free state in September.
1 means ice all year (>15% of 1980s September extent). 0 means ice free (<15% of 1980s September extent)


   Beaufort   Chukchi   ESS   Laptev   Kara
1979   1   1   1   1   1
1980   1   1   1   1   1
1981   1   1   1   1   1
1982   1   1   1   1   1
1983   1   1   1   1   1
1984   1   1   1   1   1
1985   1   1   1   1   1
1986   1   1   1   1   1
1987   1   1   1   1   1
1988   1   1   1   1   1
1989   1   1   1   1   1
1990   1   1   1   1   1
1991   1   1   1   1   1
1992   1   1   1   1   1
1993   1   0   1   1   1
1994   1   1   1   1   1
1995   1   1   1   0   0
1996   1   1   1   1   1
1997   1   1   1   1   1
1998   0   1   1   1   1
1999   1   0   1   1   1
2000   1   1   1   1   1
2001   1   1   1   1   1
2002   1   0   1   1   1
2003   1   0   1   1   1
2004   1   0   1   1   1
2005   1   0   1   1   1
2006   1   1   1   1   1
2007   1   0   0   1   1
2008   0   0   0   1   1
2009   1   0   1   1   1
2010   1   0   1   1   1
2011   1   0   1   0   1
2012   0   0   0   0   1
2013   1   0   1   0   1
2014   1   0   1   0   1

[Vergent]

Chris,

Unless I am misreading your post, you are proposing that we change from an area definition of "ice free" to an extent definition, both in the vicinity of 1 Mkm^2.



Here is my problem with that; The August-September difference between extent and area is a little less than 2 Mkm^2. This difference has not diminished much as the area/extent has diminished. So what will it look like when the extent gets down to 2.0 Mkm^2? As I see It there will only be 300-500kkm^2 area left. And half of that is going to be in the GS and CAA. There would be 150-200kkm^2 area in the Basin. There would only be 2-3% of the winter max area left and under your definition this would be nowhere near ice free?

Verg

[andy_t_roo]

given that 15% is typically the threshold to determine "does this contain ice" i believe this is nice, defendable, and consistent definition;
an area is ice free when it drops below 15% concentration; it doesn't matter if it is a pixel, the entire basin, or the individual regions discussed here.

[epiphyte]

I hope this doesn't come across as unduly cynical - but IMO BP and shell and Exxon will define "Seasonally Ice Free" for us. It's the point at which they are no longer impeded from showing us the meaning of positive feedback.

Drill, baby. Drill

[ChrisReynolds]

Vergent,

The definition within the literature is typically expressed in terms of extent for the Arctic Ocean, i.e. less than 1M km^2. For example Overland & Wang "When will the Summer Arctic be Nearly Sea Ice Free?" state the 1M km^2 demarcation, although they don't specify area or extent in that sentence. But the paper goes on to discuss extent, not area.
Abstract: http://onlinelibrary.wiley.com/doi/10.1002/grl.50316/abstract

Andy,

I thought so too. I can't help but wonder if the 1M extent definition for virtually ice free derives from a similar calculation. I will be following it.

Epiphyte,

Many a true word is said in (half) jest.

[Jim Hunt]

An interesting topic Chris!

This seems like the perfect place to point out that I sank a few beers in the company of a man from the Met Office on Wednesday. I described the terms of our little wager to him, and he reckoned the odds are 75:25 in your favour. Then I asked him if he knew what a "clean-up set" is:

http://GreatWhiteCon.info/2015/03/sea-ice-and-swells-in-the-beaufort-sea-in-the-summer-of-2014/

He'd never heard of such a thing, and it seems that neither have the Hadley Centre's climate models. Here's a recent example of one:

[Richard Rathbone]

I reckon 15% is too small. You'll end up calling region(s) as containing significant ice when the whole Arctic doesn't. 20 or 25% depending on whether you prefer to risk having regions with ice when the Arctic doesn't, or the Arctic with ice when all regions are virtually ice free.

Should that original post read 1M  extent? Thats always been the benchmark and an area based one would have to be a whole lot lower to be consistent.

[Vergent]

Chris,

You said;

The common definition seems to be total NH sea ice area of less than 1M km^2. But how might we apply this to the regional seas of the Arctic?

Am I reading this wrong?

Verg

[Jim Hunt]

Am I reading this wrong?

"NH sea ice area of less than 1M km^2" is what I have been wagering on with Chris and others recently. I believe extent is more often mentioned in these circumstances however.

[ChrisReynolds]

Vergent,

As spotted by Richard, my mistake, that should read EXTENT, not area. Sorry for that. IIRC I think I can edit my posts, if so I will correct now.

Richard.

The 15% is merely intended as an attempt to scale back from a criteria for virtually ice free for the whole Arctic to allow a regional picture. I am expecting that in the years to come we'll see a state where only the Central Arctic is left as the one to fall. I don't think another threshold is in line with what the commonly accepted definition is.

Jim,

Clean set up - so you're referring to something like a 'perfect storm'? I'm betting that isn't very likely, and could be so unusual as to slip into force majeur territory especially if the following years see a return to 'normal' conditions. But in the bet I haven't specified that, it's just part of the risk I am prepared to take. 2022 is 7 summers away, my reckoning of the odds is 1000:1, much steeper than the Met Office chap.

I noticed the music to that was Infected Mushroom, I never did get that band. Eat Static were amazing though.

[ChrisReynolds]

Am I reading this wrong?

"NH sea ice area of less than 1M km^2" is what I have been wagering on with Chris and others recently. I believe extent is more often mentioned in these circumstances however.

It's corrected now.

[Jim Hunt]

Chris - I shall be an extremely unhappy bunny if the action of waves/swells on the sea ice is considered as force majeure! I've been prattling on about it for years, even if nobody has been listening.

However a conversation on Twitter with Jenny Ross et al. has raised the question of geoengineering. Personally I reckon that significant solar radiation management (SRM for short) before 2022 should count as force majeure. What about you?

[ChrisReynolds]

Chris - I shall be an extremely unhappy bunny if the action of waves/swells on the sea ice is considered as force majeure! I've been prattling on about it for years, even if nobody has been listening.

However a conversation on Twitter with Jenny Ross et al. has raised the question of geoengineering. Personally I reckon that significant solar radiation management (SRM for short) before 2022 should count as force majeure. What about you?

Jim,

I said a perfect storm, not really understanding what you were trying to say by posting that video. Wikipedia defines a perfect storm.
http://en.wikipedia.org/wiki/Perfect_storm

No I don't consider just waves to be force majeur, but just waves is not going to drop CT area below 1M km^2 before 2022. I have read the paper about the ice acting as a low frequency resonant filter and from that I didn't get the impression penetration was many hundreds of km.

Force Majeur? Here's the sort of thing: This June a parallel line of highs and lows from the Bahamas up to Nares Strait, the resultant elongated dipole jets warm tropical air into the Arctic and this extremely unusual pattern persists from June to September (absolutely staggeringly unusual!). The result, no sea ice at all by this September. The next year the sea ice bounces back up to the 2007 to 2014 average for September and remains in that range until 2022.

In those circumstances I would wait to see what you said, but I would argue Force Majeur.

Yes geoengineering would be force majeur. But I seriously doubt that anyone is going to do it, for starters; no insurance underwriter on the planet would be stupid enough to underwrite the possible risks from that.

[Vergent]

Chris,

Unless I am misreading your post, you are proposing that we change from an area definition of "ice free" to an extent definition, both in the vicinity of 1 Mkm^2.

Do you actually read? Or do the ideas in your head scream so loud that you can not hear anything else?

Vergent,

As spotted by Richard, my mistake, that should read EXTENT, not area. Sorry for that. IIRC I think I can edit my posts, if so I will correct now.

I clearly stated misgiving about my interpenetration of what you wrote. In spite of the clear statement you made, yet you give credit for catching your mistake to someone posting after me. Clearly, I did not hold you accountable for what you actually wrote, I gave you a chance to correct your post, but instead, you give someone else credit for catching your misstatement. Why?

Verg

[ChrisReynolds]

Because I didn't re-read your comment and forgot the exact phrasing of your opening sentence, remembering only the detail you went into.

If you want the credit you have it.

[Vergent]

Because I didn't re-read your comment and forgot the exact phrasing of your opening sentence, remembering only the detail you went into.

Let me get this straight, you started a topic where you opened up by defining "ice free" as 1.0 Mkm^2 area and concluding that the definition should be 1.083 Mkm^2 extent, but you did not actually mean what you wrote. I responded with a post where every sentence was a comparison or contrast of an "area" definition of "ice free" with an "extent" definition, plus a graphic representation. But, now you are saying that the problem was that my writing was so bad that it required re-reading to understand what I was saying?

Unless I am misreading your post....

How did I know that I could not rely on the actual words that you wrote?

Verg

[Jim Hunt]

Verg - Pedantic? Moi?

[ChrisReynolds]

Vergent,

A facet of the human brain is that it picks out what it.considers salient information and retains that. This happened to me, I found the bulk of what you were saying interesting and retained it. Now you are claiming I accused you of poor writing, had I considered that to be the issue I would have said so. I did not. I could have pointed out that you actually said: "Unless I am misreading your post", while you read what I said, I just said the wrong thing. I chose not to.

Now I find myself wondering what the subtext is to your belligerent attitude. I suspect such because you are making so much of one incorrect word.

[P-maker]

Vergent,

I think Chris is trying to ”pull your leg”.

He is fully aware that 1 million square km of solid first year ice roughly 0,5 m thick is a lot of ice. Actually, it is – volume-wise – equivalent to the expected annual loss of calving ice from Greenland (500 cubic km) in a few years time, when the sea ice is essentially gone. Thus, no matter how you define it, we will never be able to observe less than 500 cubic km of glacier ice/1 million square km of solid first year sea ice (area)/3 million square km of >15 % multi-year sea ice (extent).

Taking into account that the IPCC in its most recent SYR report has even raised the bar to:

“A nearly ice-free Arctic Ocean (When sea-ice extent is less than one million km2 for at least five consecutive years. ) in the summer sea-ice minimum in September before mid-century is likely for RCP8.5 (medium confidence). {2.2.3, Figure 2.1}”

We will most likely never see five years in a row with less sea ice/calf ice than stated above. Whether Jim or Chris will win the bet is utterly meaningless. The fact is, that we will soon reach close to the 3 million square km extent/1 million square km area/500 cubic km volume in September, if we go on like this. Then we should start counting to five to prove the IPCC utterly wrong in their forecast. I don’t really care if we do not make it in the finer detail before 2020. The essence is, that we are  destroying the Arctic sea ice about one generation before we thought we would.

Moving the goal posts does not make this tragedy more acceptable.

[ktonine]

Area and extent are both essentially meaningless.  Let's cover the arctic in September from end to end with 1 mm thick ice.  Is it completely and fully 'recovered'?

Maslowski couched his definition of "nearly ice-free" in terms of volume - and by that he meant an 80% reduction from the 1979-2000 baseline.  Ice is not two-dimensional and volume makes much more sense to me.

Just because it's easier to 'see' area and extent does not mean they are the best metrics to use.  Our use of them is due to daily data availability - and we like horseraces - but that can often be misleading.

P.S. - Verge, chill out.

[Vergent]

Okay, its's been a devastating summer and there is only left a single contiguous ice flow 300 kkm^2. The arctic is essentially ice free! Along comes a storm and spreads the ice out over 1.083 Mkm^2, without melting or making new ice. The Arctic is no longer ice free! Another storm comes along and spreads it out over 1.9 Mkm^2. We are nowhere near an ice free arctic! Another storm spreads it out to 2.1 Mkm^2. The arctic is totally ice free!

I get it.

Verg

[ChrisReynolds]

Vergent,

I think Chris is trying to ”pull your leg”.

P-Maker,

If I pull someone's leg I let them know with a wink emoticon. I am not pulling anyone's leg.

Calving from Greenland is probably one of the reasons it is best to settle for <1M km^2 as ice free. However 500km^3? Is that from the very north of Greenland alone? I ask because one could easily discount the Greenland Sea from consideration of September extent. Furthermore 500km^3 of glacial ice is not like 500km^3 of sea ice. The glacial ice is much thicker and will add less to overall extent.

EDIT - the only paper I have is Bigg, 1999, whose estimate is 300km^3 for the whole of Greenland.

Vergent,

Yes that is a problem with extent, but there it is. All datasets contain flaws, surely they are still useful if one is aware of such issues.

Ktonine,

I agree, indeed it is volume loss that drives the decline in area and extent at the September minimum. But where do we get volume from if not from PIOMAS, or a similar reanalysis model? PIOMAS assimilates sea ice concentration, from which area and extent are derived.

[P-maker]

Chris,

in that case you must be "taking the piss" on Jim (see http://en.wikipedia.org/wiki/Taking_the_piss ).

Having no less than three candidates to boost your minimum (MYI, 1YI & calf ice) and the IPCC to demand five minmimas in a row, you must be sure to win your bet. It will not make any sense to exclude any of the peripheral seas from the calculations. You don't even need your "force majeure" clause. Even glacial ice breaks down these days, The rubble seen in the Jakobshavn Isfjord now  is as volatile as any sea ice, once it has left the fjord.                 

[Peter Ellis]

What?

I don't see that anything Chris has posted has anything to do with the bet.  The bet is quite simply "there will be less than 1M km^2 ice extent some time before 2022".  That's it.  Chris and Jim seem to have a gentleman's agreement that anything not reasonably foreseeable may be called Force Majeure and void the bet - but that's for them to decide on and the rest of us to watch.

This thread has nothing to do with the bet, it's simply asking the question, "what do we take as the benchmark for seasonally ice free" in the surrounding seas.  The proposed measure of ~15% of 1980s extent = ice-free seems reasonable to me, since that's pretty much the same benchmark widely used for the Arctic as a whole, just re-calculated on a region-by-region basis for the purpose of more detailed analysis.

I find most of the posts in this thread utterly baffling.

[Jim Hunt]

P-maker - Actually I'm extracting the Michael from a variety of "skeptical" fellows, and most certainly not Chris!

Peter - I rather wish I hadn't mentioned it now! However the wager (as I understand it at least) is based on area. As Vergent seems to be suggesting that makes more sense to me in this context. However extent is what is generally referred to in the literature, such as this from the IPCC's AR5 WGI final report:

Nearly ice-free conditions (i.e., when sea ice extent is less than 1 million km2 for at least five consecutive years).

[epiphyte]

What?

I don't see that anything Chris has posted has anything to do with the bet.  The bet is quite simply "there will be less than 1M km^2 ice extent some time before 2022".  That's it.  Chris and Jim seem to have a gentleman's agreement that anything not reasonably foreseeable may be called Force Majeure and void the bet - but that's for them to decide on and the rest of us to watch.

This thread has nothing to do with the bet, it's simply asking the question, "what do we take as the benchmark for seasonally ice free" in the surrounding seas.  The proposed measure of ~15% of 1980s extent = ice-free seems reasonable to me, since that's pretty much the same benchmark widely used for the Arctic as a whole, just re-calculated on a region-by-region basis for the purpose of more detailed analysis.

I find most of the posts in this thread utterly baffling.

I wanted to follow up on my post yesterday when I suggested that the definition of "ice-free" would be settled for us by the oil companies. This really wasn't pure cynicism and I really do think that for human purposes, "ice free" is most meaningfully defined as the point at which human endeavors can reasonably proceed as if there was no ice. There is precedent for this; viz: sea passages are considered ice-free when they become open for shipping. Conversely, the Titanic was sunk in an area of ocean which probably would show up on a chart as <0.01% in area or extent - but was manifestly not ice-free.

[ChrisReynolds]

I find most of the posts in this thread utterly baffling.

Same here.

[ChrisReynolds]

Chris,

in that case you must be "taking the piss" on Jim (see http://en.wikipedia.org/wiki/Taking_the_piss ).

Having no less than three candidates to boost your minimum (MYI, 1YI & calf ice) and the IPCC to demand five minmimas in a row, you must be sure to win your bet. It will not make any sense to exclude any of the peripheral seas from the calculations. You don't even need your "force majeure" clause. Even glacial ice breaks down these days, The rubble seen in the Jakobshavn Isfjord now  is as volatile as any sea ice, once it has left the fjord.               

But the bet with Jim is for one day below 1M km^2 in CT Area. What the IPCC says has nothing to do with it.

Exclusion of the Greenland Sea was for your suggestion that calving Greenland glaciers were important. Note that the definition of the Arctic Ocean I use in my calculations is Beaufort round to Bering, Central, CAA, Greenland. Baffin is not included and if Greenland is left out the calving issue is massively reduced. You didn't answer my question regards your 500km^3 calving referring to the whole of Greenland - I take that as an affirmation that indeed it is. That combined with my point about calved glacial ice being much thicker than sea ice confirms my view - it's a minor factor.

Jakobshavn Isfjord is in Baffin Bay. So I would be happy to discount it from consideration regards my bet with Jim.

I'm doing it again, giving a serious answer to someone who is clearly taking the piss. 

See what I mean about a wink emoticon?   

[P-maker]

Chris,

I think you are still adding to the confusion here:

1)   You originally made a bet with Jim, that CT area would not go below 1 million square km for one day before 2022
2)   You then opened up this thread by randomly selecting the 1980ies average sea ice extent (provided by Wipneus) as your baseline and then again in a rather random fashion decided to reduce this by 85 percent
3)   You then started to pick and choose amongst the peripheral seas in order to get to your own definition of “virtually ice free”.

When several contributors tried to get you back on a secure footing, you either ignored (Vergent and epiphyte) or disputed (me) our contributions.

In order not to end up with the same mess, as we have with the “Global 2 degree target” and the “nearly zero-energy buildings” in the EU, I strongly advice you to help clear up this thread for the benefit of all. The confusion of area versus extent is certainly not helpful, and the IPCC requirement of five consecutive years is also a show stopper.

If you really wish to make it clear to everyone, what you think will happen, you  could state that – according to your calculations - the area covered by sea ice in the Central Arctic Ocean will remain above 1 million square km until at least 2022 no matter how it is measured. This will eliminate a lot of confusion and complexity coming from the peripheral seas, as well as exclude major contributions from glaciers.

[Peter Ellis]

P-maker:  Rubbish

The point is simply that currently there are some regions (Hudson Bay, St Lawrence, Barents and Baffin) that are clearly "seasonal seas", i.e. they are only ice-covered in winter.  Then there are others like the Central Arctic Basin that are currently ice-covered year-round.  Finally, there are a bunch of others (Kara, Laptev, Beaufort, ESS, Chukchi) which are not clearly one or the other.

As the Arctic ice cap continues to shrink, it will be useful to have a benchmark of some kind, so we can discuss whether (e.g.) the ESS is behaving more like a perennially ice-covered sea or more like a seasonally ice-free sea.  That's all.

Accordingly, Chris opened the thread in order to discuss possible definitions of "seasonally ice-free" for these peripheral seas.  No more, no less.  It has nothing to do with the Central Arctic, nothing to do with the IPCC,  nothing to do with specific predictions, and nothing (much) to do with his bet with Jim!

[seaicesailor]

Actually I found the idea of ice-free regions per year brilliant.

I would also count Kara sea as a zero if it wasnt because of the Eastern part.
2012 was the first peripherally seasonally ice-free year then!

[oren]

P-maker:  Rubbish

The point is simply that currently there are some regions (Hudson Bay, St Lawrence, Barents and Baffin) that are clearly "seasonal seas", i.e. they are only ice-covered in winter.  Then there are others like the Central Arctic Basin that are currently ice-covered year-round.  Finally, there are a bunch of others (Kara, Laptev, Beaufort, ESS, Chukchi) which are not clearly one or the other.

As the Arctic ice cap continues to shrink, it will be useful to have a benchmark of some kind, so we can discuss whether (e.g.) the ESS is behaving more like a perennially ice-covered sea or more like a seasonally ice-free sea.  That's all.

Accordingly, Chris opened the thread in order to discuss possible definitions of "seasonally ice-free" for these peripheral seas.  No more, no less.  It has nothing to do with the Central Arctic, nothing to do with the IPCC,  nothing to do with specific predictions, and nothing (much) to do with his bet with Jim!

Thanks for the clarification.  I only now understood what this thread is all about, due to my too-quick reading...

[ChrisReynolds]

P-Maker,

Garbage.

1) I have stopped using CT Area as I have doubts about it and prefer Wipneus's regional extent & area. Jim wanted the bet in terms of CT Area. I was not going to start messing around arguing so accepted on Jim's terms.

2) The 1980s is the earliest full decade of sea ice. As sea ice does not know anything about human decades it seemed a reasonable early baseline period. And it worked out for a justification of Arctic wide 1M km^2 as a virtually ice free state.

3) You keep throwing out canards and running away when challenged, you can do better. If you disagree with the careful justifications I have given region by region - say why, explicitly, region by region. Don't pretend I have arbitrarily picked when the first post of this thread proves you wrong.

I did not ignore Epiphyte, I said, "Many a true word is said in (half) jest."

I made an initial mistake saying area instead of extent. I did not raise the issue of the bet in this thread.

I fully agree with Peter's comments above.

From your further comments I think it is becoming clear what the real problem is, I suspect the same problem with Vergent but have had no further clarification from s/he. Could the subtext be that I have been arguing that a rapid crash is not imminent and that I have broken some taboo round here by taking Jim up on his offer of a bet? Should that be the case I advise those who believe in the imminence of a crash to put up some solid arguments (perhaps in another thread). I am about to wrap up my argument on my blog and after that I really want to move on.

[ChrisReynolds]

Actually I found the idea of ice-free regions per year brilliant.

I would also count Kara sea as a zero if it wasnt because of the Eastern part.
2012 was the first peripherally seasonally ice-free year then!

Thanks Sea Ice Sailor.

Kara is 1's all the way down because that is the way the numbers fall. If it would help I can easily post numbers for the regions.

[jdallen]

Actually I found the idea of ice-free regions per year brilliant.

I would also count Kara sea as a zero if it wasnt because of the Eastern part.
2012 was the first peripherally seasonally ice-free year then!

Thanks Sea Ice Sailor.

Kara is 1's all the way down because that is the way the numbers fall. If it would help I can easily post numbers for the regions.

I'd be curious to see that, Chris.  I've similarly found this thread confusing, but becoming somewhat less murky.  The "ice free regions" concept got lost in the back and forth.

It makes sense, much as how your arguments for a somewhat slower melt-out have.  That doesn't mean a sudden melt out is precluded (looks warily at the current weather and conditions); but does mean that even if one happened, it would not be persistent.  There'd be pull back much as 2013 following 2012.

In understanding and watching the transition, the state of the Peripheral seas will provide a good sense of where we stand.  Documenting how much longer it takes them to re-freeze and the more rapidly they clear in spring could be used to help people who are uncertain on AGW to better understand the impact.  I'd say starting with the Bering and Barents might prove very illustrative.  We are used to them being either completely or mostly ice-free at the end of the melt season, but I don't think this has always been the case.   They both certainly have far less winter extent at max than they did 50 years ago, and that contrast would be understood immediately.
(Note differences near max between years below for the Bering, Okhotsk, and Barent's seas.)

Edit: 
I've listed some questions for entirely rhetorical purposes below.  I'm not asking for answers.  Rather I think they may be useful for focusing thought.

What impact do the observed regional changes have on local climate?
What is the cascading effect of these changes on regional climate?
How do the changes influence weather over the remaining pack?
What measurable influence does the change in extent have on annual minimums?
What impact does the reduced coverage have on climate outside of the arctic?

[P-maker]

Chris,

Following up on the “Rubbish” and “Garbage” statements was not my idea of cleaning up this thread.

Let me quote you Chris:

“A facet of the human brain is that it picks out what it.considers salient information and retains that. This happened to me,…”

You are absolutely right. We pick out words (you have about 45,000 of them in the English language and I have about one third of that in my native language).

Using these words we try to communicate our thoughts.

I do not wish to go down into the details of the various peripheral seas. Time does not allow this pedigree activity. Instead I would like you to stick to the main controversy – will the Central Arctic Ocean be essentially ice free before 2022 or not?

Your most recent blog statement was:

“I've been looking at Beaufort/Chukchi/East Siberian Seas, a region I call the BCE region. For the BCE region, the 2007 to 2012 period has an average September extent 14% of that for the 1980s. For comparison with the whole Northern Hemisphere extent, 14% of the 1980s average would be 0.979M km^2, which fits the commonly accepted definition of below 1M km^2 for which the Arctic be considered virtually ice free in September. Therefore BCE can be considered virtually ice free in September for the period 2007 to 2012.”

Let me challenge that observation by issuing another statement: I have been looking at the Baffin Bay, Davis Strait & Greenland Sea areas, a region I call the BDG region. For the end game, I expect there will be more than 1 million square km (both area and extent) in these waters throughout September all the way to 2022, which means that both Wipneus and CT Area will be above 1 million square km and you will win the bet with Jim. All this ice will have to come through the Nares and Fram straits and will be topped up by calf ice from Greenland. Since you apparently prefer a 5 year average, we will eventually have to wait until 2027 before you are proven wrong.

You have on many occasions claimed that the peripherel seas were of little importance. Now is the time to let go and not waste time discussing whether they are seasonally or perenially ice-free.

[Peter Ellis]

I do not wish to go down into the details of the various peripheral seas.

So go post in a different thread rather than shitting up this one?  This is a thread about the regional seas and their ice-free status (or otherwise).  There are plenty of threads with vague evidence-free prognostications of doom and When All The Ice Will Be Gone (TM), is it too much to ask to have one that looks at more detailed data?

Edit:  Chris - you might want to change the title of the thread to assist the terminally confused.  Maybe something like "Peripheral sea transition to a seasonally ice free state".

[Yuha]

Actually I found the idea of ice-free regions per year brilliant.

I would also count Kara sea as a zero if it wasnt because of the Eastern part.
2012 was the first peripherally seasonally ice-free year then!

Thanks Sea Ice Sailor.

Kara is 1's all the way down because that is the way the numbers fall. If it would help I can easily post numbers for the regions.

There is some problem with the numbers for Kara in 2012 though. According to Wipneus' AMSR2 data September 2012 average extent for Kara is only 0.0025M km2, less than 1% of the 0.30M km2. The NSIDC data must have some coastal ghost ice or something like that.

[ChrisReynolds]

Yuha,

Yes I am using Wipneus's calculation based on NSIDC gridded concentration, easier to use his figures than to run my code. NSIDC data does have such problems but is consistent and goes back to 1979. Kara 2012 Sept average is 0.0598, 15% of the 1980s average is 0.0448, so it just misses out on the ice free state.

[ChrisReynolds]

Let me challenge that observation by issuing another statement: I have been looking at the Baffin Bay, Davis Strait & Greenland Sea areas, a region I call the BDG region. For the end game, I expect there will be more than 1 million square km (both area and extent) in these waters throughout September all the way to 2022, which means that both Wipneus and CT Area will be above 1 million square km and you will win the bet with Jim. All this ice will have to come through the Nares and Fram straits and will be topped up by calf ice from Greenland. Since you apparently prefer a 5 year average, we will eventually have to wait until 2027 before you are proven wrong.

You have on many occasions claimed that the peripherel seas were of little importance. Now is the time to let go and not waste time discussing whether they are seasonally or perenially ice-free.

Let go of what? Hard data? Reality?

Here are the September average extents for Baffin Bay and Greenland Sea (M km^2).

Year   Baffin   Greenland
1979   0.068   0.230
1980   0.047   0.329
1981   0.055   0.331
1982   0.089   0.371
1983   0.150   0.328
1984   0.070   0.241
1985   0.046   0.123
1986   0.059   0.238
1987   0.045   0.352
1988   0.031   0.310
1989   0.074   0.377
1990   0.055   0.149
1991   0.052   0.138
1992   0.091   0.372
1993   0.083   0.348
1994   0.059   0.340
1995   0.071   0.372
1996   0.159   0.159
1997   0.077   0.310
1998   0.059   0.291
1999   0.069   0.286
2000   0.062   0.257
2001   0.063   0.113
2002   0.050   0.070
2003   0.045   0.073
2004   0.049   0.096
2005   0.040   0.271
2006   0.042   0.232
2007   0.042   0.340
2008   0.043   0.157
2009   0.040   0.254
2010   0.055   0.194
2011   0.034   0.284
2012   0.034   0.258
2013   0.061   0.144
2014   0.029   0.189

Taking the average September extent for the post 2000 period:
Baffin Bay 0.046M km^2
Greenland Sea 0.195M km^2
Combined they are already considerably less than 1M km^2, and have been throughout the record. Greenland sea has 325% more post 2000 September average ice extent than Baffin Bay.

Using Google Earth I have calculated the rough lengths from Nares Strait entry in the Central Arctic to the southern tip of Greenland.
Baffin Bay 3019km
Greenland Sea 3778km
The Greenland sea (and Central Arctic) coast of Greenland is 25% longer than the Baffin bay coastline. So the difference is not likely to be due to calved glacial mass.

Baffin Bay has only 0.046M km^2 average September extent since 2000, this presents a indication of maximum extent that might be attributed to glacial calving. While Greenland Sea has an equivalent extent of 0.195M km^2. The Greenland sea extent is due to export from the Arctic Ocean proper. Therefore once the Arctic Ocean is ice free and this export stops extent should be substantially less even with substantially increased calving.

I am not convinced by your counter proposal.

Oh, and by the way, it was you who raised the 5 year average, it has not been a requirement of mine.

Instead I would like you to stick to the main controversy – will the Central Arctic Ocean be essentially ice free before 2022 or not?

No, it will not. The peripheral seas play less of a role in the establishment of such a condition precisely because they are already so close to that state.

Peter,

No problem, thread name changed.

[ktonine]

No, it will not. The peripheral seas play less of a role in the establishment of such a condition precisely because they are already so close to that state.

I would disagree.  It's not enough that the peripheral seas are already essentially ice free at the end of the melt season, for the Central Arctic Basin to be ice-free in September the peripheral seas will have to become ice-free earlier - so that warm seas, wave action, etc can eat away at the CAB for a more prolonged period of time.

In this sense the peripheral seas can play a pivotal role in the march towards < 1M km^2 - by reaching their ice-free state even earlier.

[Gray-Wolf]

So, like Barentsz through the noughties, we'd need to see low winter ice conditions across the peripheral seas so as to give them a chance of becoming ice free by Aug and so give the C.A.B. a good chance of being melted by melt seasons end?

Well we have seen both Barentsz and Baffin have a go a really low winter ice levels and the past winter saw Bering and Okhotsk both on the low side so we have been seeing the pre-requisite for a good attack on C.A.B. in late season ...... just not all in the same year!

[Vergent]

Chris,

The methodology you propose is flawed. The definition of "extent" is where the ice area is 15% or higher. Your methodology says that a region is seasonally ice free when the extent drops to 15% or lower, and falsely equate the two. Lets look at ESS for which you give an '80s average of 0.67. We multiply by 0.15 and then by 0.15 again to ask what is the least area that could preclude calling it ice free? Chris's answer: 15k If there is as little as 2% area in the ESS, Chris is prepared to call it not seasonally ice free.

I am going to apply the WUWT test. Would this idea fly over at WUWT?

Looking at the arctic basin 100k area could preclude you calling it seasonally ice free. In that circumstance, you would be welcomed over at WUWT, you would be a hero for arguing that the Basin wasn't seasonally ice free. But, over here, I think you would have few agreeing with you. Just picture 100k on this graph:



Verg

[ChrisReynolds]

Actually I found the idea of ice-free regions per year brilliant.

I would also count Kara sea as a zero if it wasnt because of the Eastern part.
2012 was the first peripherally seasonally ice-free year then!

Thanks Sea Ice Sailor.

Kara is 1's all the way down because that is the way the numbers fall. If it would help I can easily post numbers for the regions.

I'd be curious to see that, Chris.  I've similarly found this thread confusing, but becoming somewhat less murky.  The "ice free regions" concept got lost in the back and forth.

It makes sense, much as how your arguments for a somewhat slower melt-out have.  That doesn't mean a sudden melt out is precluded (looks warily at the current weather and conditions); but does mean that even if one happened, it would not be persistent.  There'd be pull back much as 2013 following 2012.

In understanding and watching the transition, the state of the Peripheral seas will provide a good sense of where we stand.  Documenting how much longer it takes them to re-freeze and the more rapidly they clear in spring could be used to help people who are uncertain on AGW to better understand the impact.  I'd say starting with the Bering and Barents might prove very illustrative.  We are used to them being either completely or mostly ice-free at the end of the melt season, but I don't think this has always been the case.   They both certainly have far less winter extent at max than they did 50 years ago, and that contrast would be understood immediately.
(Note differences near max between years below for the Bering, Okhotsk, and Barent's seas.)

Edit: 
I've listed some questions for entirely rhetorical purposes below.  I'm not asking for answers.  Rather I think they may be useful for focusing thought.

What impact do the observed regional changes have on local climate?
What is the cascading effect of these changes on regional climate?
How do the changes influence weather over the remaining pack?
What measurable influence does the change in extent have on annual minimums?
What impact does the reduced coverage have on climate outside of the arctic?

Sorry for not replying last night, I got distracted and am determined not to spend every night on sea ice.

I'll post figures after this comment.

The state of the peripheral seas is important because once they are ice free every year the first date of ice free can move earlier in the season on a more regular basis... And then we get to the exciting bit, where the Central Arctic starts getting eaten into on a regular basis.

Barents in particular shows the power of ocean warming, with substantial winter retreat, but both have virtually totally melted out through the satellite era. That raises the question of how representative the early satellite era is of long term summer conditions. Due to lack of data I am not convinced either way on that question.

Your questions are interesting, with one question put in a different place in the list.

What impact do the observed regional changes have on local climate?
What impact do they have on the formation of the Arctic Dipole and what role does that have in loss of summer sea ice off Siberia? I have no answers here.

What is the cascading effect of these changes on regional climate?and
What impact does the reduced coverage have on climate outside of the arctic?
Sea ice loss in the summer drives faster snowline advance in autumn which Cohen argues drives negative AO index winters and cold boreal winters. Liu et al (Judith Curry is a co-author) find the same as Cohen, but note broader meridional meanders of the jet stream than the 'classic' AO. Wu et al find an impact of lower sea ice concentration on summer rainfall in China...

I have 32 papers on the links between sea ice loss and weather - and I do not have them all!

How do the changes influence weather over the remaining pack?
I don't know.

What measurable influence does the change in extent have on annual minimums?
I don't get what this question means, sorry.

[ChrisReynolds]

   Beaufort   Chukchi   ESS   Laptev   Kara   Barents
1979   0.199   0.094   0.838   0.576   0.141   0.011
1980   0.427   0.274   0.738   0.327   0.479   0.055
1981   0.278   0.266   0.587   0.333   0.461   0.044
1982   0.206   0.197   0.683   0.308   0.291   0.193
1983   0.457   0.379   0.721   0.227   0.220   0.010
1984   0.402   0.180   0.726   0.364   0.136   0.006
1985   0.441   0.265   0.786   0.272   0.081   0.022
1986   0.303   0.135   0.784   0.424   0.398   0.054
1987   0.259   0.249   0.706   0.417   0.182   0.069
1988   0.351   0.366   0.646   0.354   0.328   0.101
1989   0.278   0.152   0.281   0.400   0.409   0.161
1990   0.314   0.102   0.133   0.138   0.349   0.034
1991   0.449   0.206   0.354   0.161   0.260   0.026
1992   0.399   0.265   0.488   0.461   0.329   0.038
1993   0.162   0.033   0.236   0.368   0.174   0.149
1994   0.364   0.312   0.695   0.323   0.184   0.014
1995   0.212   0.181   0.403   0.022   0.044   0.004
1996   0.452   0.142   0.784   0.633   0.413   0.017
1997   0.192   0.066   0.630   0.278   0.134   0.017
1998   0.044   0.073   0.701   0.376   0.378   0.086
1999   0.207   0.026   0.533   0.128   0.216   0.087
2000   0.232   0.145   0.525   0.100   0.158   0.005
2001   0.313   0.187   0.666   0.522   0.146   0.003
2002   0.129   0.006   0.319   0.195   0.264   0.038
2003   0.209   0.005   0.141   0.229   0.414   0.108
2004   0.158   0.001   0.264   0.428   0.199   0.011
2005   0.273   0.032   0.148   0.059   0.063   0.003
2006   0.237   0.125   0.226   0.162   0.193   0.008
2007   0.091   0.001   0.006   0.207   0.065   0.003
2008   0.041   0.001   0.017   0.219   0.070   0.031
2009   0.211   0.001   0.132   0.094   0.066   0.023
2010   0.064   0.001   0.216   0.120   0.060   0.010
2011   0.060   0.002   0.166   0.015   0.060   0.002
2012   0.001   0.003   0.005   0.012   0.059   0.001
2013   0.224   0.018   0.347   0.018   0.112   0.001
2014   0.160   0.004   0.251   0.009   0.089   0.103

[ChrisReynolds]

Greenland   Central   CAA   Arctic Ocean   All
0.230   4.237   0.613   6.939   7.066
0.329   4.442   0.532   7.602   7.680
0.331   4.425   0.335   7.059   7.143
0.371   4.449   0.493   7.191   7.316
0.328   4.386   0.401   7.128   7.407
0.241   4.286   0.419   6.759   6.864
0.123   4.291   0.377   6.659   6.732
0.238   4.413   0.557   7.306   7.412
0.352   4.446   0.498   7.179   7.279
0.310   4.440   0.397   7.293   7.347
0.377   4.441   0.410   6.909   7.009
0.149   4.309   0.498   6.026   6.143
0.138   4.328   0.459   6.381   6.474
0.372   4.412   0.540   7.306   7.442
0.348   4.390   0.403   6.262   6.397
0.340   4.395   0.423   7.048   7.138
0.372   4.284   0.441   5.962   6.051
0.159   4.310   0.470   7.380   7.561
0.310   4.408   0.555   6.591   6.687
0.291   4.268   0.239   6.457   6.537
0.286   4.213   0.327   6.024   6.118
0.257   4.379   0.351   6.153   6.234
0.113   4.262   0.440   6.652   6.733
0.070   4.299   0.439   5.759   5.828
0.073   4.349   0.524   6.051   6.117
0.096   4.197   0.544   5.897   5.970
0.271   4.146   0.455   5.450   5.504
0.232   4.294   0.329   5.806   5.863
0.340   3.246   0.252   4.210   4.268
0.157   3.789   0.303   4.626   4.684
0.254   4.007   0.419   5.208   5.262
0.194   3.855   0.277   4.798   4.866
0.284   3.711   0.212   4.512   4.562
0.258   2.980   0.189   3.509   3.556
0.144   3.842   0.424   5.132   5.209
0.189   3.900   0.474   5.179   5.221


'Arctic' is Beaufort round to Greenland, Central, and CAA.
'All' is total NH sea ice including Hudson Bay.
Regions follow Cryosphere Today regions.

All figures are September average extent in million kmsq.

[ChrisReynolds]

No, it will not. The peripheral seas play less of a role in the establishment of such a condition precisely because they are already so close to that state.

I would disagree.  It's not enough that the peripheral seas are already essentially ice free at the end of the melt season, for the Central Arctic Basin to be ice-free in September the peripheral seas will have to become ice-free earlier - so that warm seas, wave action, etc can eat away at the CAB for a more prolonged period of time.

In this sense the peripheral seas can play a pivotal role in the march towards < 1M km^2 - by reaching their ice-free state even earlier.

I'm not sure how fast this earlier date of melt out will proceed though. In the PIOMAS -1m experiment organised by Blanchard Wigglesorth, artificially thinning the pack by 1m on 1 June kills all the peripheral seas by the July data! (not clear if the July data is an average or last day of the month.)

[Peter Ellis]

@Vergent: So, in less hyperbolic terms - you disagree with using an extent based measure to define "ice-free", because of the 15% threshold used to define extent. You would prefer to use an area-based metric.

That's fair enough, however it would be worth bearing in mind there are problems associated with area measurements, particularly in summer, due to melt ponding etc - neither extent nor area is a true reflection of how much water is actually covered by ice.

It's also worth bearing in mind that to people not familiar with Arctic ice, the notion that "ice-free" means anything other than no ice at all is quite counter-intuitive. (Let's leave aside calved icebergs for the moment, since (a) they're easily distinguished from sea ice and (b) they're almost negligible in both area and extent).  If you go for the "ice free means completely free" viewpoint, then area vs extent is irrelevant, since zero extent = zero area.

Why might extent be a defensible measure? Firstly, because it's robustly measureable without the confounding effects of melt ponding.  A replicable measure is useful to reduce arguments.  Secondly, (as I understand it) the 15% threshold is used for extent measurements because that's about the limit for safe navigability in boats other than icebreakers, so it has a real meaning in terms of the impact on Arctic activity.

Chris' proposed threshold is thus "the point at which >85% of a given region is accessible for ordinary shipping", which doesn't seem particularly unreasonable.  More defensible than "zero means zero" at any rate, and arguably more reasonable than an area-based metric which could see a region being called "ice-free" even though little of it is actually save to visit except in an icebreaker.

[ChrisReynolds]

Chris,

The methodology you propose is flawed. The definition of "extent" is where the ice area is 15% or higher. Your methodology says that a region is seasonally ice free when the extent drops to 15% or lower, and falsely equate the two. Lets look at ESS for which you give an '80s average of 0.67. We multiply by 0.15 and then by 0.15 again to ask what is the least area that could preclude calling it ice free? Chris's answer: 15k If there is as little as 2% area in the ESS, Chris is prepared to call it not seasonally ice free.

I am going to apply the WUWT test. Would this idea fly over at WUWT?

Looking at the arctic basin 100k area could preclude you calling it seasonally ice free. In that circumstance, you would be welcomed over at WUWT, you would be a hero for arguing that the Basin wasn't seasonally ice free. But, over here, I think you would have few agreeing with you. Just picture 100k on this graph:



Verg

On my blog I wanted to say "There is a clear increase in September virtually ice free state within the peripheral seas of the Arctic Ocean basin..." I cannot assert such a thing without back up. I see it in the data, but that just isn't good enough. So I wonder, how might I define seasonally ice free for a sea within the Arctic Ocean?

First I ponder the generally accepted definition of a virtually ice free state, less than 1M km^2 sea ice extent for the whole Arctic in the September average. Now I cannot just portion that down to regions, I have no way of doing that fairly.

I ponder some more...

Then while playing around with the figures I realise that using an impartial early baseline, the 1980s average, 1M km^2 is as near as damn it 15% of 1980s September average extent for the whole Arctic. I could have used the 1979 to 1994 average (re Lindsay & Zhang's tipping point), because that gives 1.06M km^2.

OK. So 15% looks good, and it is neat and tidy because it not only matches with the whole Arctic definition of virtually ice free, but it also fits neatly with the 15% used for extent (BTW - I have written code I can use to calculate extent for both NSIDC gridded concentration data and the PIOMAS area.h variable, i.e. I do know what I am doing).

"Looking at the arctic basin 100k area could preclude you calling it seasonally ice free."

Let's try this numerically. You have the numbers for the regions in September, depends what you mean by the Arctic Basin of course.

The proper definition used in the literature is Beaufort round to Laptev and the Central Arctic. The 1980s average extent is 5.997M km^2, 15% of that is 0.90M km^2, or 900k extent, the minimum area that could qualify as 900k extent is 134k km^2 area (15% of 900k).

So, 100k would not be above the seasonally ice free limit, even with implausibly low compactness.

As Peter Ellis points out, there are good reasons why professional scientists use extent. I used to use area, my familiarity with the data has changed my mind - I now mainly use extent, not area.

[ktonine]

I'm not sure how fast this earlier date of melt out will proceed though.

Well, yes, there's the rub

I'm unable to make any predictions myself, though I think in general I have not written off the possibility of a fast collapse - i.e., in the next 3 to 5 years.

What I find interesting in the extent data is the correlations between the Beaufort, Chukchi, and ESS with the CAB.  Especially when they are looked at for the first 20 years of the record (1979 to 1998) and compared with the last 20 years (1995 to 2014).

1979 to 1998
Beaufort - CAB 0.080219436
Chukchi - CAB 0.393517979
East SS - CAB -0.0609996

1995 to 2014
Beaufort - CAB 0.601123336
Chukchi - CAB 0.516668461
East SS - CAB 0.660549527

And then look at the combined Beaufort, Chukchi, ESS extents vs the CAB
1979 to 1998 0.112745638
1995 to 2014 0.708761282
2005 to 2014 0.790956637

I think it's clear from this that the CAB September extent is highly correlated (and I would say dependent) on the combined extent in the Beaufort, Chukchi, and ESS.

Across the NH we see phenomenological data showing a steady northwards progression of warming; spring arriving earlier, plants and animals and insects migrating, inland lakes freezing later and melting out earlier.  Given some of the ridiculous temperature anomalies we've seen this spring in the arctic - temps +20 to +30 C above normal - I'd say that progression is just as relevant to the arctic and arctic amplification is in full swing. So an 'Oh shit" moment isn't that outlandish to envision.