IJIS

[gregcharles]

3)  If you want to wager a 6 pack of Deschutes beer (yum)....I would be MORE THAN HAPPY to bet you that the ice extent will be equal to or below 2.5 million square kilometers.  I don't bet on "even money bets".....that is crap shoot.

OK, you're on. Yes, even money bets are a crap shoot ... that's a tautology actually. Even though I'm taking what you estimate to the be the 15% chance, I like the idea of the bet for a few reasons.

1. A little fun will ease the pain inherent in this discussion, like Nero's beautiful violin music eased the pain of Rome's citizens as their city burned.

2. It solves a problem for me of how to communicate this crisis with people I know. If I win the bet, I'll share the beer with my coworkers. (We do weekly beer tastings of various microbrews.) I can segue from details of the bet into the general situation.

3. Deschutes, yum!

Details: I'm betting the minimum arctic sea ice extent reported by IJIS for 2016 will be above 12.5 million square kilometers. You are betting it will be below that. The stakes are a six pack of Deschutes against a six pack of Stone or Ballast Point or other San Diego brewery ... just don't expect me to buy Deschutes and ship it back to Oregon for you! I'll try to come up with a nice selection of stuff you can't just buy yourself up there. I hope you'll do the same, but I'll leave it to your discretion. Sound good?

[Tor Bejnar]

Even the non-betting person I am would bet against "minimum arctic sea ice extent reported by IJIS for 2016 will be above 12.5 million square kilometers", given that it is already way below 12.5.  Oh, that was a typo?   Aw shucks... 

[gregcharles]

Gah! I was trying to be so careful too. Yes, 2.5 million. OK, Buddy? Otherwise, I guess I just owe you a six pack right now.

[Buddy]

Gah! I was trying to be so careful too. Yes, 2.5 million. OK, Buddy? Otherwise, I guess I just owe you a six pack right now.

OK...the bet is on.  We'll find out the winner come September.....if not before

I can taste that lovely Deschutes right now.  And for those who have BEEN TO OREGON....no better place than the Deschutes River in central Oregon.

[gregcharles]

Awesome! Let me re-re-re-iterate though, I'm staking San Diego beer (Stone, Ballast Point, AleSmith, Green Flash/Alpine, etc. I'm open to suggestions) against Oregon beer (yes, Deschutes is my preference). If you win, you get San Diego beer, because I'm in San Diego and so have easy access to it. I really hope you don't win though, and not just because of the free beer.

[Espen]

IJIS:

10,989,954 km²（May 20, 2016）

[Magma.]

Earliest date for IJIS sea ice extent to drop below 11,000,000 km2

1. May 20, 2016
2. May 29, 2015
3. June 3, 2011

Average date
2003-2015: June 9
2000s: June 15
1990s: June 24
1980s: July 3

[jdallen]

Earliest date for IJIS sea ice extent to drop below 11,000,000 km2

1. May 20, 2016
2. May 29, 2015
3. June 3, 2011

Average date
2003-2015: June 9
2000s: June 15
1990s: June 24
1980s: July 3

We are definitely in unknown territory.

[pauldry600]

Unknown indeed

However slow decline for a couple of weeks and damage may not be below 2012. Its touch n go really but not good.

Id guess a final figure of 2.7m

[Meirion]

CCI forecast http://cci-reanalyzer.org/wxmaps/#ARC-LEA (isn't their new Hi-Res product great) does not suggest "slow decline" but biggest melt week of year so far with sub 10.5 by May 27 figs if at all accurate.

[abbottisgone]

I would suggest sub-10.3 is a given by the 27th 

[abbottisgone]

...no, sorry: by the end of the month I would say the numbers are sub-10.3->my apologies.

[DavidR]

Earliest date for IJIS sea ice extent to drop below 11,000,000 km2

1. May 20, 2016
2. May 29, 2015
3. June 3, 2011

Average date
2003-2015: June 9
2000s: June 15
1990s: June 24
1980s: July 3

We are definitely in unknown territory.

At 14 days, thats' the fastest decline from 12 M to 11 M and at 32 days the second fastest  decline from 13M to 11M, behind 2010 at 31 days, despite starting much earlier in the year  on both counts.

[Espen]

IJIS:

10,951,843 km²（May 21, 2016）

[abbottisgone]

I think this is the closest to up we're going to get... 

[Bill Fothergill]

Prior to becoming a fully paid-up member of the growing legion of coffin-dodgers, I used to be involved with performance analysis on communications networks. As well as "short period" data, there was quite a fondness for annual and rolling-period stats.

With that in mind, here are just a few observations on the data from the IJIS csv file...
(Apologies if anyone has done something similar earlier in this thread.)

1) Annual average extents for the following decades...
(NB All extent values in millions of sq kms, unless otherwise stated.)

1980's   11.957
1990's   11.440
2000's   10.773


2) Actual annual average extents from 2003 onwards.

2003   10.988
2004   10.821
2005   10.489   9th
2006   10.349   7th
2007   10.079   3rd
2008   10.585   11th
2009   10.543   10th
2010   10.324   5th=
2011   10.046   2nd
2012     9.971   lowest
2013   10.418   8th
2014   10.324   5th=
2015   10.111   4th

(NB there were 24 days with missing data. Place-holder values for these were calculated using simple linear interpolation from the before/after values.)

3) The rolling 12-month average as of the May 21st data is 10.049 million sq kms - virtually identical to the 2011 average. This value is currently dropping at between 1,300 and 1,500 sq kms per day.

Should the 2016 extent drop at a mere 30,000 sq kms per day until it coincides with 2015, and then track 2015 until the year end, the 2016 annual average will end up at approximately 10.032 million sq kms.

I wonder how close it will get to 10.000

[AbruptSLR]

I wonder how close it will get to 10.000

It is my guess that it will be below 9.8 million sq kms

[jdallen]

I wonder how close it will get to 10.000

It is my guess that it will be below 9.8 million sq kms

I think this is still to uncertain to call. A lot depends on what happens in the Kara and Labrador seas, and to a lesser degree, the Beaufort and Chukchi. We're talking about a steep drop in May over a very short time, when most of the "easy" ice has for the most part fled.

[AbruptSLR]

I wonder how close it will get to 10.000

It is my guess that it will be below 9.8 million sq kms

I think this is still to uncertain to call. A lot depends on what happens in the Kara and Labrador seas, and to a lesser degree, the Beaufort and Chukchi. We're talking about a steep drop in May over a very short time, when most of the "easy" ice has for the most part fled.

jdallen,

When there is no urgency it is always best to collect all the data before formulating a conclusion; however, when time is of the essence it has been mathematically proven that it is optimal to make a decision based on only 37% of the data that you will eventually gather.  As I think the longer decision makers take to realize that we are at a climate tipping point already the worse things are going to get, so I prefer to make my decision based on 37% of the data that we will all know come September (without making any effort at all).

Cheers,
ASLR

[jdallen]

I wonder how close it will get to 10.000

It is my guess that it will be below 9.8 million sq kms

I think this is still to uncertain to call. A lot depends on what happens in the Kara and Labrador seas, and to a lesser degree, the Beaufort and Chukchi. We're talking about a steep drop in May over a very short time, when most of the "easy" ice has for the most part fled.

jdallen,

When there is no urgency it is always best to collect all the data before formulating a conclusion; however, when time is of the essence it has been mathematically proven that it is optimal to make a decision based on only 37% of the data that you will eventually gather.  As I think the longer decision makers take to realize that we are at a climate tipping point already the worse things are going to get, so I prefer to make my decision based on 37% of the data that we will all know come September (without making any effort at all).

Cheers,
ASLR

Here, its about timescale, ASLR;   

The variability in two weeks - very hard to speculate even with the data we have so far, and surprisingly difficult to put together a prediction that doesn't have a fairly wide margin of error. 

The variability in 20 years - I think we have enough data to understand the trend and to act, as the longer duration and comparative history - using your 37%, that being about 7 years - smooth's out the bumps we see at the micro-scale.

[AbruptSLR]

I wonder how close it will get to 10.000

It is my guess that it will be below 9.8 million sq kms

I think this is still to uncertain to call. A lot depends on what happens in the Kara and Labrador seas, and to a lesser degree, the Beaufort and Chukchi. We're talking about a steep drop in May over a very short time, when most of the "easy" ice has for the most part fled.

jdallen,

When there is no urgency it is always best to collect all the data before formulating a conclusion; however, when time is of the essence it has been mathematically proven that it is optimal to make a decision based on only 37% of the data that you will eventually gather.  As I think the longer decision makers take to realize that we are at a climate tipping point already the worse things are going to get, so I prefer to make my decision based on 37% of the data that we will all know come September (without making any effort at all).

Cheers,
ASLR

Here, its about timescale, ASLR;   

The variability in two weeks - very hard to speculate even with the data we have so far, and surprisingly difficult to put together a prediction that doesn't have a fairly wide margin of error. 

The variability in 20 years - I think we have enough data to understand the trend and to act, as the longer duration and comparative history - using your 37%, that being about 7 years - smooth's out the bumps we see at the micro-scale.

It looks like your space is for rent.

[jdallen]

It looks like your space is for rent.

I guess I'm not understanding your point, ASLR.

[Espen]

IJIS:

10,923,963 km²（May 22, 2016）

[abbottisgone]

  Yay, summers over!

We're saved 

[oren]

It looks like your space is for rent.

I guess I'm not understanding your point, ASLR.

Your signature...

[Neven]

Congratulations, Governor Espen. 

[AbruptSLR]

It looks like your space is for rent.

I guess I'm not understanding your point, ASLR.

My scientific point is that the IPCC AR5/CMIP5 model projections of sea ice loss from the Arctic assume gradual change, but in reality it does not occur gradually but rather in a ratcheting fashion from one quasi-static/critical point to another as indicated by the attached image (using chaos theory concepts, in this case with our recent Super El Nino acting as the strange attractor).

This is verified by the information in the linked open access reference that indicates that: "… Arctic sea ice may have critical points beyond which a return to the previous state is less likely."  Current AR5/CMIP5 models do not adequately exhibit this behavior:

Goldstein, M. A., Lynch, A. H., Arbetter, T. E., and Fetterer, F.: Abrupt transitions in Arctic open water area, The Cryosphere Discuss., doi:10.5194/tc-2016-108, in review, 2016.

http://www.the-cryosphere-discuss.net/tc-2016-108/

Abstract. September open water fraction in the Arctic is analyzed using the satellite era record of ice concentration (1979–2014). This analysis suggests that there is a statistically significant breakpoint (shift in the mean) and increase in the variance around 1988 and another breakpoint around 2007 in the Pacific sector. These structural breaks are robust to the choice of algorithm used for deriving sea ice concentration from satellite data, and are also apparent in other measures of open water, such as operational ice charts and the record of navigable days from Barrow to Prudhoe Bay. Breakpoints in the Atlantic sector record of open water are evident in 1988 and 2007 but more weakly significant. The breakpoints appear to be associated with concomitant shifts in average ice age, and tend to lead change in Arctic circulation regimes. These results support the thesis that Arctic sea ice may have critical points beyond which a return to the previous state is less likely.


Regarding my point about the 37% rule see:

Optimal stopping:
https://en.wikipedia.org/wiki/Optimal_stopping
The Secretary Problem:
https://en.wikipedia.org/wiki/Secretary_problem

And as oren points out, see your signature for my comment about your space being for rent.

[pauldry600]

Someone mentioned before but the "easy" ice seems gone and now the more robust is slower to melt.

Coupled with the end of el nino the graph is going to correct itself and we will be tracking with the other lines at the bottom soon.....i hope

Just watching n lurking

[Csnavywx]

Someone mentioned before but the "easy" ice seems gone and now the more robust is slower to melt.

Coupled with the end of el nino the graph is going to correct itself and we will be tracking with the other lines at the bottom soon.....i hope

Just watching n lurking

Changes in atmospheric circulation and temperatures generally lag Nino 3.4 by 2-3 months -- so we'll be living with the effects through most of the summer. A quick and dirty look at the 97-98 Nino from a GISS global temp standpoint reveals this (in ballpark fashion):

1998    60   88   61   63   71   77   70   68   45   46   49   57

The bolded values were August and September. Note that Nino 3.4 turned sharply down in May and June, but temps stayed up through the summer.

[Neven]

Some of the 'melt' so far has actually been wind-driven compaction, opening of polynyas. I think there is plenty of easy ice, and unless weather conditions change drastically, 2016 will stay well ahead for a while to come.

And there is no clear-cut correlation with ENSO, although this winter/year does seem to be in line with global SAT.

[dnem]

2012 hit the "June cliff" on June 5 and in the nine days following reeled off ~1.13 million k^2 of losses.  Between today's date and the cliff, losses in 2012 averaged around 44k/day which happens to be exactly the 5-day running average of JAXA loses right now.  I can't see any reason why 2016 won't keep its head start intact until losses pick up in June (or sooner).  I can't comment on whether or not the conditions that led to the massive losses in June in 2012 will be repeated this year.

[helorime]

Someone mentioned before but the "easy" ice seems gone and now the more robust is slower to melt.

Hardly.  The vast majority of the Hudson Bay, Baffin, the Kara, the ESS are yet to go and the Hudson always melts out entirely.  Lots of easy stuff yet to go.

[Meirion]

Don't be fooled by the "hiatus" in #Arctic melt in last few days we're about to hit a plunge (if CCI is right)

[Andreas T]

if the ice in Kara and other areas is easy, why does it not melt at this time in other years? The areas which were melting in 2012 causing this drop of 44k / day are already cleared. As Neven points out in many cases mostly not by melting but by drifting elsewhere. If elsewhere is the Greenland or Barents sea then that means melt elsewhere. But the low ice cover in Greenland sea indicates export there is not strong. Ice drift from HYCOM and what I see on worldview confirms this I think. What I don't know is how much weaker the ice is due to the warm winter. I do expect water which is open now but wasn't in other years to take up heat but getting this to where it has an effect on the ice takes a while I expect.
The logic "extent it must drop x km² because thats what it did before" is flawed.
In Beaufort I see hefty Ice floes at the ice edge which won't disappear soon. Chuckchi and Kara will see melt but probably not at higher rate than in other years. CAB which is low due to the area north of Svalbard could even grow if pushed from the pacific side.
I think extent above 10.5 mio at the start of June is a possibility which does not mean later large (and continuing) drops are unlikely.

[seaicesailor]

And there is no clear-cut correlation with ENSO ...

If this year does not go below 2012, otherwise scientists will have to find a clear-cut correlation

[Tor Bejnar]

...
 But the low ice cover in Greenland sea indicates export there is not strong. Ice drift from HYCOM and what I see on worldview confirms this I think.
...

I've not look recently, so I don't have an opinion about current export of CAB ice, but a month ago we were watching ice flows approaching Svalbard from the CAB and it was melting 'before our eyes' sometimes as fast as the ice front progressed, and sometimes about half as fast.  I believe that a bunch of CAB-exported ice that in past years would now be counted as Greenland Sea ice is this year just water.

[AbruptSLR]

And there is no clear-cut correlation with ENSO ...

If this year does not go below 2012, otherwise scientists will have to find a clear-cut correlation

In the linked article Robert Scribbler does a nice job of summarizing some of the key physics related issues of our current situation:

http://robertscribbler.com/2016/02/25/as-a-titanic-el-nino-begins-to-fade-what-fresh-trouble-will-a-record-warm-world-bring/

Extract: "But as El Nino weakens and the Equator cools, the Jet Stream would tend to slow even more. Such an atmospheric state would tend to further exaggerate already significant Jet Stream wave patterns — transferring still more low-Latitude heat poleward. In addition, the ocean gyres tend to speed up as El Nino fades or transitions to La Nina. The result is an amplified pulse of warmer waters emerging from southern Latitudes and entering the Arctic.
It’s for these combined reasons — tendency to amplify south to north atmospheric heat transfer into the Arctic post El Nino and tendency to flush warmer waters toward Arctic Ocean zones during the same period that it appears we are entering a high risk time for potential new sea ice melts and possible related Greenland land ice melts during 2016 and 2017.

…

Finally, extreme above average sea surface temperatures are predicted to intensify over the Barents and Greenland seas through to end of Summer 2016. This is an area to watch. The added ocean heat would tend to pull the Jet Stream northward over Eastern Europe and Western Russia — generating risk of heatwaves and drought for this region even as Central Asia fell under risk of floods. Long range CFS precipitation and temperature model runs for Europe have not yet picked up this risk. However, given the intensity of heat predicted for Barents sea surfaces and the related tendency of warmth over oceans and in the far north to influence the formation of blocking patterns, heat domes, and high amplitude troughs, it’s worth keeping a weather eye on the situation.

...

Meanwhile, the shift toward La Nina has tended to enhance a range of global heating related issues including record rainfall events and large injections of heat toward the poles in the drop off from El Nino to La Nina."

Edit: The key consideration is that we are probably now transitioning from an El Nino to a weak La Nina (see the thick yellow curve for the average of the Mid-May dynamic ENSO model forecasts in the attached image).

[E.]

When there is no urgency it is always best to collect all the data before formulating a conclusion; however, when time is of the essence it has been mathematically proven that it is optimal to make a decision based on only 37% of the data that you will eventually gather.

<snip>

ASLR, I don't think you have the Secretary Problem quite right . . . it gets the correct answer 37 percent of the time, which is different from drawing conclusions from only 37 percent of the available data.

[AbruptSLR]

When there is no urgency it is always best to collect all the data before formulating a conclusion; however, when time is of the essence it has been mathematically proven that it is optimal to make a decision based on only 37% of the data that you will eventually gather.

<snip>

ASLR, I don't think you have the Secretary Problem quite right . . . it gets the correct answer 37 percent of the time, which is different from drawing conclusions from only 37 percent of the available data.

Perhaps it is better to think of this question in terms of optimal stopping.  Here it is very clear that the problem at hand is choosing a time to take a particular action, such as imposing carbon pricing and/or carbon regulations (or even when to place a bet/guess on the ASIE):

Per Wikipedia: "In mathematics, the theory of optimal stopping is concerned with the problem of choosing a time to take a particular action, in order to maximise an expected reward or minimise an expected cost. Optimal stopping problems can be found in areas of statistics, economics, and mathematical finance (related to the pricing of American options). A key example of an optimal stopping problem is the secretary problem."

https://en.wikipedia.org/wiki/Optimal_stopping

[Phil.]

Some of the 'melt' so far has actually been wind-driven compaction, opening of polynyas. I think there is plenty of easy ice, and unless weather conditions change drastically, 2016 will stay well ahead for a while to come.

Looking at the Barrow webcam today there's lot's of meltwater on the in-shore ice, and the air temperature is 37ºF, so I expect some melting there too.

[Espen]

IJIS:

10,880,319 km²（May 23, 2016）

[Sleepy]

And there is no clear-cut correlation with ENSO ...

If this year does not go below 2012, otherwise scientists will have to find a clear-cut correlation

Clear cut correlations are rare and/but ENSO is the third largest climate phenomena/variation on earth.

Checking ONI against NSIDC one can observe drops after El Nino in:
09-10
06-07
02-03 is kind of special together with 04-05 since there was only a short dip into negative values.
97-98
94-95
91-92
86-87 and 87-88 is also kind of special and there the drop came later in 1990.
82-83

The 2014-2016 El Nino should generate a new record unless it holds on to neutral for another year. It might do so considering the amount of heat in the pacific.
As I see it that is... I'm only a sleepy obeserver with too little time to spend on this.

[abbottisgone]

Some of the 'melt' so far has actually been wind-driven compaction, opening of polynyas. I think there is plenty of easy ice, and unless weather conditions change drastically, 2016 will stay well ahead for a while to come.

..you just had to say 37 didn't you 
Looking at the Barrow webcam today there's lot's of meltwater on the in-shore ice, and the air temperature is 37ºF, so I expect some melting there too.

[Andreas T]

Some of the 'melt' so far has actually been wind-driven compaction, opening of polynyas. I think there is plenty of easy ice, and unless weather conditions change drastically, 2016 will stay well ahead for a while to come.

Looking at the Barrow webcam today there's lot's of meltwater on the in-shore ice, and the air temperature is 37ºF, so I expect some melting there too.

there definitely is melting going on but it does not feature strongly in the extent figures (yet) if those are representing water surface covered with ice and are not too sensitive to ice covered by water.

[seaicesailor]

Someone mentioned before but the "easy" ice seems gone and now the more robust is slower to melt.

Hardly.  The vast majority of the Hudson Bay, Baffin, the Kara, the ESS are yet to go and the Hudson always melts out entirely.  Lots of easy stuff yet to go.

I would define 'easy'. These seas may be 'easy' for a super warm land NH year like 2016 (in the sense of melt out sooner you mean?),
definitely ESS is not easy

[helorime]

I will give you the ESS, yes, in most years it is not an "easy" melter.  The Hudson certainly is though and the implication in the original post was that there is nowhere for continued fast drops to occur in.

[seaicesailor]

I will give you the ESS, yes, in most years it is not an "easy" melter.  The Hudson certainly is though and the implication in the original post was that there is nowhere for continued fast drops to occur in.

Actually I am agreeing with you (except for ESS) given the extreme warmth of this year.
Temperatures around Hudson were lagging in April, see first plot, but this has changed.

Cannot forget that +2 degC of NH land anomaly represents same temperatures but 1 month in advance in April, compared to the seventies (see monthly temperatures in third plot). Post-2010 years anomalies were quarter to half a degree C below 2016, meaning this year's NH Spring is one week in advance with respect to post 2010 years.

Hope it is not perceived as off-topic, given the surprising trend of IJIS.

[6roucho]

Perhaps it is better to think of this question in terms of optimal stopping.  Here it is very clear that the problem at hand is choosing a time to take a particular action, such as imposing carbon pricing and/or carbon regulations (or even when to place a bet/guess on the ASIE):

Interesting idea. I think the obstacle to using stopping theory to make bets on the ASIE is that the data is ordered: those secretaries keep getting better, until one day they keep getting worse. Stopping theory requires random data, like stock market indices or real secretaries.

[AbruptSLR]

Perhaps it is better to think of this question in terms of optimal stopping.  Here it is very clear that the problem at hand is choosing a time to take a particular action, such as imposing carbon pricing and/or carbon regulations (or even when to place a bet/guess on the ASIE):

Interesting idea. I think the obstacle to using stopping theory to make bets on the ASIE is that the data is ordered: those secretaries keep getting better, until one day they keep getting worse. Stopping theory requires random data, like stock market indices or real secretaries.

This distinction can be addressed by normalizing the betting odds that you can get.  In this regard you can think of this as a salary negotiation between a union & and owner, which is also a well studied/formulated "optimal stopping" case that is different the more simple Secretary Problem (just look at Wikipedia for references).

[AbruptSLR]

And there is no clear-cut correlation with ENSO ...

If this year does not go below 2012, otherwise scientists will have to find a clear-cut correlation

Clear cut correlations are rare and/but ENSO is the third largest climate phenomena/variation on earth.

I would like to note that correlating to past El Nino or to past La Nina events, is different from correlating to El Nino to La Nina transition periods, in the May to July timeframe, as we may be currently experiencing.