Support the Arctic Sea Ice Forum and Blog

Poll

When will the Arctic Extent dip below 1,000,000 Km^2

2018-2019
12 (17.9%)
2020-2025
21 (31.3%)
2026-2030
13 (19.4%)
2031-2040
15 (22.4%)
2041-2060
2 (3%)
2061-2080
0 (0%)
2081-2099
1 (1.5%)
2100-beyond
3 (4.5%)

Total Members Voted: 61

Voting closed: July 27, 2018, 07:46:32 AM

Author Topic: When will the Arctic Go Ice Free?  (Read 494667 times)

petm

  • Grease ice
  • Posts: 675
    • View Profile
  • Liked: 318
  • Likes Given: 27
Re: When will the Arctic Go Ice Free?
« Reply #950 on: August 12, 2019, 03:55:13 AM »
Actually that is not so much a heat input as a reduction in heat output.

Sure, but same difference, literally. (When discussing a model that requires a certain change of net "input" to achieve a certain threshold, wrt a factor that wasn't considered in the model. If you catch my drift...)

Klondike Kat

  • Grease ice
  • Posts: 842
    • View Profile
  • Liked: 40
  • Likes Given: 56
Re: When will the Arctic Go Ice Free?
« Reply #951 on: August 12, 2019, 04:16:50 AM »
Quote
Actually that is not so much a heat input as a reduction in heat output.

It is both and input and a reduction of heat output, depending on context.

But really, when you have more heat, and a reduction in heat output, looking at it as input is probably a good thing.

That depends on whether it reduces the heat input also.  It may be a wash.

Archimid

  • Young ice
  • Posts: 3511
    • View Profile
  • Liked: 899
  • Likes Given: 206
Re: When will the Arctic Go Ice Free?
« Reply #952 on: August 12, 2019, 04:34:31 AM »
It doesn't look like it. This year will be an excellent test. Will it be like 2012-2013, a record high ice formation, or be like 2016 followed by anemic growth that lead to the lowest max on record?

The next large confluence of oceanic cycles will also be a great test.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

sja45uk

  • New ice
  • Posts: 21
    • View Profile
  • Liked: 3
  • Likes Given: 17
Re: When will the Arctic Go Ice Free?
« Reply #953 on: August 12, 2019, 05:01:29 AM »
That pattern shows accelerating loss of Sept ice volume. I don't believe that pattern is appropriate.

Following graph is not aggressive enough, but is a better fit and doesn't have R2 anywhere near 99.99%

Based on the increase in CO2, I can see no good reason for you to fit a curve that comes to an asymptote rather than continues to accelerate downwards. Your curve fit seems too dependant on the latest value and seems (to me) to make an unjustified assumption about future values. If you have a physical mechanism that justifies not fitting a linear or exponential, I would be interested to know what it is. Ultimately of course, I suspect it will have to bottom out at close to zero ice in the area, and the refreeze will be delayed longer and longer. How long that will take I am not qualified to guess, but we do not seem to be acting to prevent it happening.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #954 on: August 12, 2019, 12:30:32 PM »
That pattern shows accelerating loss of Sept ice volume. I don't believe that pattern is appropriate.

Following graph is not aggressive enough, but is a better fit and doesn't have R2 anywhere near 99.99%

Based on the increase in CO2, I can see no good reason for you to fit a curve that comes to an asymptote rather than continues to accelerate downwards. Your curve fit seems too dependant on the latest value and seems (to me) to make an unjustified assumption about future values. If you have a physical mechanism that justifies not fitting a linear or exponential, I would be interested to know what it is. Ultimately of course, I suspect it will have to bottom out at close to zero ice in the area, and the refreeze will be delayed longer and longer. How long that will take I am not qualified to guess, but we do not seem to be acting to prevent it happening.

I don't believe the horizontal extrapolation, I believe it will continue downwards while ghgs are rising.

1. There is discussion in the literature of Multi year Ice (MYI) ceasing to make it around the Beaufort Gyre allowing it to last for a decade and have time to thicken. The quantity of thick MYI fell dramatically over ~2000-2012.

2. When first year ice (FYI) melts out, then each winter it pretty much grows back to normal thickness. When thick MYI reduces to FYI thickness, you don't really get any regrowth as the ice thickness is already at the equilibrium thickness where upwelling heat balances outgoing heat. This means that getting rid of ice volume that is thick MYI can happen relatively fast, but reducing FYI is a long slow process.

3. Winter ice thickness negative feedback: Reduce the ice thickness and more heat flows out of the ocean. This is a fast process, see delays to start up of winter refreeze only being a little later now than 30 years ago. Also even when all the sea ice is removed from models allowing lots of heat to build up in the ocean over the summer, the refreeze still starts by November. Basically there is plenty of time over the winter to vent any heat build up that is likely to occur. Freeze up will become a little later, but thinner ice allows faster heat loss so while there is a deficit in ice thickness, ice forms at a faster rate so tend to catch up in thickness. Reducing the thickness in winter is a long process (once down to FYI thicknesses).

4. Models. Models are good at some things but bad at others. Levels of ice can be all over the place. So one model might says 2070 but another 2040, so on this, it is a bad idea to trust models. However they pretty much all agree that as ice declines towards 0 ice in September, the rate of ice loss slows down. This seems a pretty reliable result to me. Not impossible that all the models are wrong, but you need some pretty strong reasoning to discount what we see from models.

5. Since proposed has the data supported or debunked which ideas? Such sigmoid/gompertz curves have been discussed here since at least 2011 and probably earlier. How have the alternative ideas of accelerating decline versus gompertz shape fits proceeded? See a pair of graphs, 5th row of
https://sites.google.com/site/arctischepinguin/home/piomas

The 4 parameter gompertz I used is debateably better than the 3 parameter gompertz shown there, but what is not in doubt is that either gompertz fit is better than the exponential or any other accelerating downwards fit.

6. If the gompertz fits were just fitting noise and given lack of correlation of noise between maximum and minimum, I think it likely that fitting a curve to September and April volume would be likely to have to have differences in timings to fit the noise. The time of the point of inflection point (around 2005) is within a year between April and September fits. This suggests to me that there is a real pattern going on not just noise.


Not quite sure I can genuinely count all that as 6 different lines of evidence. Nevertheless, how many do I need?


>"not seem to be acting"
Perhaps a little overstated, but I agree we are not doing enough / not acting fast enough on action to reduce AGW.


Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #955 on: August 13, 2019, 02:25:06 AM »
It doesn't look like it. This year will be an excellent test. Will it be like 2012-2013, a record high ice formation, or be like 2016 followed by anemic growth that lead to the lowest max on record?

The next large confluence of oceanic cycles will also be a great test.

2016 had insanely high temps in the Fall and early Winter. I remember it happening at the time. 2012 had positive anomalies (heck every winter does now) but not as high.

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #956 on: August 13, 2019, 02:31:20 AM »
Thank you crandles...

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #957 on: August 24, 2019, 05:07:31 PM »
Gone quiet here. Was it my post that caused that?  ::) ;D

If so, was it because it was well argued, or so bad it wasn't worth arguing with?  ::)

Klondike Kat

  • Grease ice
  • Posts: 842
    • View Profile
  • Liked: 40
  • Likes Given: 56
Re: When will the Arctic Go Ice Free?
« Reply #958 on: August 24, 2019, 05:58:16 PM »
Gone quiet here. Was it my post that caused that?  ::) ;D

If so, was it because it was well argued, or so bad it wasn't worth arguing with?  ::)

Probably  the former.  I do not think anyone believes the horizontal curve.  However, until the ice starts taking a downward turn, it best fits the data.  Perhaps this year will change things.

Michael Hauber

  • Nilas ice
  • Posts: 1114
    • View Profile
  • Liked: 168
  • Likes Given: 16
Re: When will the Arctic Go Ice Free?
« Reply #959 on: August 25, 2019, 12:15:11 AM »
Gone quiet here. Was it my post that caused that?  ::) ;D

If so, was it because it was well argued, or so bad it wasn't worth arguing with?  ::)

Perhaps also the recent stall in melt, meaning this season is suddenly highly unlikely to break 2012's record. 

So now we are looking at minimum 8 years to break the record.  The previous longest time to break the record for minimum extent according to JAXA was 5 years.
Climate change:  Prepare for the worst, hope for the best, expect the middle.

gandul

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #960 on: September 02, 2019, 12:17:52 AM »
Based on what I’ve watched since 2013, I am confident that we can see an Arctic ice free really soon, talking before 2030.
The ingredient however is not simply more heat. We’ll have that, unfortunately, but we’ll need something else. If something 2019 has proven is that a very warm season has its limits. 2019 will go as the warmest May-June-July-Half of August with plenty of ridges disrupting the Arctic weather, with sun, WAA, and Pacific ocean heat. It will end nonetheless in a discreet fourth, or perhaps second.

When Friv starts to list the things needed for a warm season to blast records, it is like a Swiss watch, and if a timing or a mechanism fails the record is gone. I’m disappointed at warm season, have to be perfect or the Arctic survives quite well especially the core CAB over deep waters.

There was a season, 2016, way way less “clean”, meaning little surface melting in May and June, mixed July with storms and a big GAC in August. The weather was in the 2013 and 2014 mode of cloudiness. However a few factors helped make the season very destructive even when feeling absent of heat. First there was early open water, then snow cover was gone pretty soon, and NH temperatures were all time record allowing for frequent heat pulling even when sun was relatively absent. Then a very destructive GAC.
 A sort of cloudy but warm Arctic constantly broken by storms.

We had in 2016 the spectacular end of season with that wrecked pack almost to the Pole. The fast refreeze of the CAB and very slow refreeze of the rest. And the humidity prevented during the Winter that pack increased in volume, leading to record low extent and volume starting 2017.

Seasons like that, combined with the AGW effect, is what will bring us a very low volume summer soon. These wet summers are less appealing for people, but I think they’ll come more frequently.
A weather mode that leads us first to a humid Fall and Winter, then very low volume, then clouded (but warmer) summer, then again and the ice is gone.

That is my opinion, I am sure many here have expressed similar views, not looking to be original, just observing that a fair, warm season as we got in 2019 doesn’t quite look the way to a ice free Arctic.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #961 on: September 02, 2019, 09:43:20 PM »


Quote
To know whether or not those indicated changes are real or just the impression given by random noise, a good way is to apply changepoint analysis. That confirms that these changes are indeed real, and gives the following “continuous piecewise linear” model of the anomalies.

https://tamino.wordpress.com/2016/07/17/arctic-heat/

Seems like Tamino confirmed slow down in rate at around 2005 was real/ statistically significant back in 2016?

Paul

  • Grease ice
  • Posts: 575
    • View Profile
  • Liked: 136
  • Likes Given: 9
Re: When will the Arctic Go Ice Free?
« Reply #962 on: September 03, 2019, 01:29:48 AM »
I think we got to remember some years since 2007 has had ice extent which really flattered to deceive when you looked at the ice condition, years come to mind are 2010, 2013, 2016 & 2017 which all had the CAB looking more than sorry for itself it would be fair to say.

A year like 2013 when you look at Worldview really showed how shocking things were down the spine of the basin in August with lots of open water within the ice pack and a true polyna developing near the pole. Admittedly ice in other parts of the basin did look healthier and the ice edge on the Pacific side was quite far South(which may of played a part in the 2014 rebound) but if that summer was not so stormy then no doubt we would of seen a much smaller ice pack although at the same time a more compact one like this year.

The Arctic has changed for sure, I have no idea when the first ice free event will occur. Parts of me do question is a blue ocean event even possible just because he basin melt season is only 3 months or so long, can we realistically melt so much ice within that time no matter how thin it is especially if its an average melt season?

nanning

  • Nilas ice
  • Posts: 2487
  • 0Kg CO₂, 37 KWh/wk,125L H₂O/wk, No offspring
    • View Profile
  • Liked: 273
  • Likes Given: 23170
Re: When will the Arctic Go Ice Free?
« Reply #963 on: September 03, 2019, 06:00:56 AM »
<snip>
Parts of me do question is a blue ocean event even possible just because he basin melt season is only 3 months or so long, can we realistically melt so much ice within that time no matter how thin it is especially if its an average melt season?

Yes, by a step-wise disappearance of the Polar Vortex and a destratification of the toplayers of the Arctic Ocean.

It is underway; the jetstream's amplitude is increasing and often 'disconnecting' and passing over the 80N area.
The destratification (and becoming more saline) is already happening in Barentz and Baffin if I remember correct.
"It is preoccupation with possessions, more than anything else, that prevents us from living freely and nobly" - Bertrand Russell
"It is preoccupation with what other people from your groups think of you, that prevents you from living freely and nobly" - Nanning
Why do you keep accumulating stuff?

binntho

  • Nilas ice
  • Posts: 2193
    • View Profile
  • Liked: 878
  • Likes Given: 235
Re: When will the Arctic Go Ice Free?
« Reply #964 on: September 03, 2019, 06:09:01 AM »


Quote
To know whether or not those indicated changes are real or just the impression given by random noise, a good way is to apply changepoint analysis. That confirms that these changes are indeed real, and gives the following “continuous piecewise linear” model of the anomalies.

https://tamino.wordpress.com/2016/07/17/arctic-heat/

Seems like Tamino confirmed slow down in rate at around 2005 was real/ statistically significant back in 2016?

Good find nanning! Seems like I have to start making apologies!
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

wallen

  • Frazil ice
  • Posts: 115
    • View Profile
  • Liked: 52
  • Likes Given: 2
Re: When will the Arctic Go Ice Free?
« Reply #965 on: September 03, 2019, 06:12:46 AM »
Think it is going to need the Bering Strait and Svalbard and possibly FJL to be ice free at the start of the melt season. Which isn't that far fetched.

binntho

  • Nilas ice
  • Posts: 2193
    • View Profile
  • Liked: 878
  • Likes Given: 235
Re: When will the Arctic Go Ice Free?
« Reply #966 on: September 03, 2019, 08:04:45 AM »
As nanning crandles has pointed out in another post (https://forum.arctic-sea-ice.net/index.php/topic,2348.msg226396.html#msg226396), Tamino has indeed done change-point analysis on Arctic Sea Ice and it turns out that there was a change in trends, and that it is statistically valid. So much of what I've said above turns out to be a bit off ...

Tamino's article can be found here https://tamino.wordpress.com/2016/07/17/arctic-heat/

What he does is take the raw data and calculate anomalies to hide the seasonal fluctuations, as shown here:



The weird behaviour towards the end is apparently caused by the seasonal cycle getting bigger, so to correct for that Tamino computes an "adaptive anomaly" to get this:



And finally for some smoothing and change-point analyisis to get this continous piecewise linear model.



So what Tamino does find is that ice loss generally follows a linear downward trend, and that this trend has been mostly unchanged, but is interrupted by a short speed-up between 2002 and through 2006.

This does not fit very well with the statements that Klondike Kat made earlier,

The minimum was decreasing slowly, until it took a bigger dive in the 90s.  That lasted about 15 years, depending on which curve fitting one chooses.  Since then, the minimum has flatlined.

although admittedly KK was talking about the minima and not the full data which is what Tamino uses. So I made a very crude comparison of the two graphs, my own graph of NSIDC minima and Tamino's changepoint graph from above. And it turns out that the changepoints that Tamino finds can not be seen in the minima graph at all!

If Tamino is right (and he usually is!) then something changed to cause rapid melt in the years 2002 - 2006, and then changed back to normal. But what? Global temperatures were basically flatlining, MYI was still going strong and the average (and minimum) SIE was at least a million km2 bigger than it is now.

And Tamino certainly doesn't find any slowdown since 2007, so any putative explanations of slowdowns since then can be put to rest.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #967 on: September 03, 2019, 01:13:59 PM »

So what Tamino does find is that ice loss generally follows a linear downward trend, and that this trend has been mostly unchanged, but is interrupted by a short speed-up between 2002 and through 2006.

Slight nit pick:

There isn't enough data to distinguish between piecewise linear and something like the gompertz curves that have been/could be fitted. We can say the rate has slowed down in a statistically significant fashion since about 2006 but no more about the shape.

binntho

  • Nilas ice
  • Posts: 2193
    • View Profile
  • Liked: 878
  • Likes Given: 235
Re: When will the Arctic Go Ice Free?
« Reply #968 on: September 03, 2019, 01:50:59 PM »

So what Tamino does find is that ice loss generally follows a linear downward trend, and that this trend has been mostly unchanged, but is interrupted by a short speed-up between 2002 and through 2006.

Slight nit pick:

There isn't enough data to distinguish between piecewise linear and something like the gompertz curves that have been/could be fitted. We can say the rate has slowed down in a statistically significant fashion since about 2006 but no more about the shape.

Are you sure abut this? That's not what I gather from reading Tamino's blog, he uses Lowess smooth and Gaussian smooth and both show something very similar to the stepwise linear model.

So I am unwilling to accept your claim at face value - I do not think that the only thing we can say is that the rate has slowed down in a statiscally significant fashion since 2006. However, if you are able to show me otherwise, I'd be grateful.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #969 on: September 03, 2019, 02:35:10 PM »


I agree that Lowess smooth and Gaussian smooth and both show something very similar.

However I would say Lowess smooth and perhaps more the Gaussian smooth both show something very similar to a gompertz fit. You can see both show curves rather than straight lines meeting at an angle which is what the gompertz fit would do. The Gaussian also starts low and ends high.


All models are wrong, some are useful.
Tamino  says

Quote
They’re in broad agreement, giving a visual impression of steady decline until about 2002, followed by faster decline until about 2008, followed by a return to the previous rate of decline.

To know whether or not those indicated changes are real or just the impression given by random noise, a good way is to apply changepoint analysis.

I think this is saying this analysis is useful in order to tell if the change in rate is statistically significant. This doesn't mean the model is right. All models are wrong.


However, I do admit I should fit a gompertz to the data before claiming too much about the Gaussian smooth looking like a gompertz fit. Other work like comparing residuals may also be of benefit. However, at the moment,  I am not sure how, or if I can work out how, to calculate an “adaptive anomaly” in order to process the data.


The piecewise linear with 3 sections has 6 parameters. The gompertz only 4, so it wouldn't be a surprise if the 3 piece piecewise linear fit had lower error residuals. Maybe 6 parameters is justified if the fit is a lot better, however overfitting is a potential issue and I would argue that it is far better to use the shapes that most models suggest and keep the number of parameters down rather than going for a better fit with more parameters.

aslan

  • Frazil ice
  • Posts: 226
    • View Profile
  • Liked: 175
  • Likes Given: 0
Re: When will the Arctic Go Ice Free?
« Reply #970 on: September 03, 2019, 02:40:21 PM »
Without wanting to look harsh, but I think the discussion is taking a bad way. I am not a specialist of this question, but a big missing piece of data is the problem of land versus ocean distribution in Northern Hemisphere. I mean, in winter, ice edge is constrained more by geography (coast of Russia and Canada) than by physics (warming / cooling atmosphere, ocean currents, and so on). Which is no longer the case in Summer, where ice edge is, but for Greenland, in open Ocean and can retreat as the thermodynamics warrants. This is probably why seasonal cycle changed dramatically after 2005. Until 2005, ice was still mostly bounded by the coast, even in Summer. Now, ice has all the possibility to expand and retreat without being imped by the coasts. One possibility is to look instead at sea ice edge latitude. Doing this, difference between Summer and Winter is lessened and the downward trend is more even.
On top of that, captain obvious helping, it should be noted that sea ice area can't go negative. But this has the implication also that the downward trend should go to zero sooner rather than later, as there is no more ice to melt in summer. Losing 2 million squared km when you have like 20 million at hand is almost nothing, losing 2 million squared km when you have like 1 million at hand is impossible... This is probably the first big answer to the question "why trend is so irregular and the seasonal cycle has gone mad ?".

It should be noted also that from this point of view, the situation is “worsening” on the Pacific side. When the edge was in the Bering Sea, there was still some width to “yielded” big loses. Now that the edge is in the bottleneck of the strait, even a retreat of 1° North –which is quite significant– would amount to almost zero ice area loss. Zero time a thousand km is still zero :p We will have to wait for ice edge in winter to reached Beaufort / Chukchi / ESS for seeing anew some significant ice loss for the Pacific side. Same on the Atlantic side, as Barents and Kara Sea are zeroing, going further North will become complicated in the short term. It will happened in the decades to come of course, but over the course of the coming years, variability of winter sea ice as seen from the metric of extent will probably still be dampened, even if the warming speed up bigly.

https://www.tandfonline.com/doi/full/10.3402/polar.v33.21249

P.S. And on top of that "simple" geographical explanation, geaography also lead to "energy yield" vastly different from Winter to Summer. Land and Ocean distribution are also responsible for bigly different thermal answer in Summer and in Winter under the same forcing. We will also probably have to wait that the Ocean becomes an "heat accumulator" in winter for seing warming speeding up in Arctic.
« Last Edit: September 03, 2019, 02:50:58 PM by aslan »

binntho

  • Nilas ice
  • Posts: 2193
    • View Profile
  • Liked: 878
  • Likes Given: 235
Re: When will the Arctic Go Ice Free?
« Reply #971 on: September 03, 2019, 03:08:26 PM »
All models are wrong, some are useful.
Tamino  says

Quote
They’re in broad agreement, giving a visual impression of steady decline until about 2002, followed by faster decline until about 2008, followed by a return to the previous rate of decline.

To know whether or not those indicated changes are real or just the impression given by random noise, a good way is to apply changepoint analysis.

I think this is saying this analysis is useful in order to tell if the change in rate is statistically significant. This doesn't mean the model is right. All models are wrong.


True, all models are wrong. And the more we talk about these graphs and the various attempts to get at underlying trends, the less I feel I know what is actually going on in the arctic.

Quote
The piecewise linear with 3 sections has 6 parameters. The gompertz only 4, so it wouldn't be a surprise if the 3 piece piecewise linear fit had lower error residuals. Maybe 6 parameters is justified if the fit is a lot better, however overfitting is a potential issue and I would argue that it is far better to use the shapes that most models suggest and keep the number of parameters down rather than going for a better fit with more parameters.

I agree that the fewer the parameters the better.

Which reminds me of something discussed earlier: What physical reasoning can be used to justify using one model rather than another?

And in my view, the only thing we can be absolutely sure about is that it keeps getting warmer, with global temperatures increasing at a more-or-less linear rate over the last 40 years. So assuming a linear reduction in sea ice extent is perhaps the most reasonable.

But as aslan points out above, the geography has a large and, significantly, a changing effect. And something similar might be said about the changes in ice-age composition. Gerontocrat has shown us how the ratio of open water to ice is changing rapidly, that must also have an effect that on the trends. The increased storminess, the changing of oceans from Arctic to Atlantic, etc. etc. - things are certainly changing up there! But how to translate that into parameters that will fit one model more than another? I've absolutely no idea.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

Klondike Kat

  • Grease ice
  • Posts: 842
    • View Profile
  • Liked: 40
  • Likes Given: 56
Re: When will the Arctic Go Ice Free?
« Reply #972 on: September 03, 2019, 03:29:53 PM »
As nanning crandles has pointed out in another post (https://forum.arctic-sea-ice.net/index.php/topic,2348.msg226396.html#msg226396), Tamino has indeed done change-point analysis on Arctic Sea Ice and it turns out that there was a change in trends, and that it is statistically valid. So much of what I've said above turns out to be a bit off ...

Tamino's article can be found here https://tamino.wordpress.com/2016/07/17/arctic-heat/

What he does is take the raw data and calculate anomalies to hide the seasonal fluctuations, as shown here:



The weird behaviour towards the end is apparently caused by the seasonal cycle getting bigger, so to correct for that Tamino computes an "adaptive anomaly" to get this:



And finally for some smoothing and change-point analyisis to get this continous piecewise linear model.



So what Tamino does find is that ice loss generally follows a linear downward trend, and that this trend has been mostly unchanged, but is interrupted by a short speed-up between 2002 and through 2006.

This does not fit very well with the statements that Klondike Kat made earlier,

The minimum was decreasing slowly, until it took a bigger dive in the 90s.  That lasted about 15 years, depending on which curve fitting one chooses.  Since then, the minimum has flatlined.

although admittedly KK was talking about the minima and not the full data which is what Tamino uses. So I made a very crude comparison of the two graphs, my own graph of NSIDC minima and Tamino's changepoint graph from above. And it turns out that the changepoints that Tamino finds can not be seen in the minima graph at all!

If Tamino is right (and he usually is!) then something changed to cause rapid melt in the years 2002 - 2006, and then changed back to normal. But what? Global temperatures were basically flatlining, MYI was still going strong and the average (and minimum) SIE was at least a million km2 bigger than it is now.

And Tamino certainly doesn't find any slowdown since 2007, so any putative explanations of slowdowns since then can be put to rest.

I guess that depends on whether you compare the recent (post 2007 trend) to the sharper decline from the previous decade or the trend from prior decades.  Whether focusing on minima, maxima, or average sea ice, the data shows a steeper decline during the first decade of the 21st century compared to previous and subsequent decades.  Should the trend that Tamino calculated continue unabated, a sustained ice-free Arctic will not occur for at least half a century (although seasonal BOEs could occur well before then).

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #973 on: September 03, 2019, 03:33:42 PM »
Conversation is taking different path but since I have done it may as well post.

This is just fitting the gompertz to a piecewise linear. I thought it might do a little better than this and yes the Lowess and Gaussian smooths look more like the piecewise linear than the gompertz.

Gompertz fitted to data rather than to piecewise linear may do better. While probably not as good a fit, there is less parameters and I still doubt there is enough evidence to decide on which shape.


crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #974 on: September 03, 2019, 03:37:32 PM »
Without wanting to look harsh, but I think the discussion is taking a bad way. I am not a specialist of this question, but a big missing piece of data is the problem of land versus ocean distribution in Northern Hemisphere.

I think it is useful to go in several different directions. Extract what we can from trends but also consider physical reasoning.

I am not sure I agree on the importance of land restricting ice extent. It certainly does in winter but it is constrained at same places each year. So I am not convinced it has much effect on trends.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #975 on: September 03, 2019, 03:53:16 PM »

I guess that depends on whether you compare the recent (post 2007 trend) to the sharper decline from the previous decade or the trend from prior decades.  Whether focusing on minima, maxima, or average sea ice, the data shows a steeper decline during the first decade of the 21st century compared to previous and subsequent decades.  Should the trend that Tamino calculated continue unabated, a sustained ice-free Arctic will not occur for at least half a century (although seasonal BOEs could occur well before then).

I estimated slopes of three sections at -.035 -.188 and -.025. I doubt we can say slope of third section is statistically significantly less than slope of first section with data Tamino used in July 2016, only that it is less than slope of steep section. In part this is because there is uncertainty in timing of change points and that noticeably affects slopes of the sections.

Klondike Kat

  • Grease ice
  • Posts: 842
    • View Profile
  • Liked: 40
  • Likes Given: 56
Re: When will the Arctic Go Ice Free?
« Reply #976 on: September 03, 2019, 04:03:08 PM »

I guess that depends on whether you compare the recent (post 2007 trend) to the sharper decline from the previous decade or the trend from prior decades.  Whether focusing on minima, maxima, or average sea ice, the data shows a steeper decline during the first decade of the 21st century compared to previous and subsequent decades.  Should the trend that Tamino calculated continue unabated, a sustained ice-free Arctic will not occur for at least half a century (although seasonal BOEs could occur well before then).

I estimated slopes of three sections at -.035 -.188 and -.025. I doubt we can say slope of third section is statistically significantly less than slope of first section with data Tamino used in July 2016, only that it is less than slope of steep section. In part this is because there is uncertainty in timing of change points and that noticeably affects slopes of the sections.

I would agree.  It will be rather interesting to see if it continues.

binntho

  • Nilas ice
  • Posts: 2193
    • View Profile
  • Liked: 878
  • Likes Given: 235
Re: When will the Arctic Go Ice Free?
« Reply #977 on: September 03, 2019, 04:07:33 PM »

I guess that depends on whether you compare the recent (post 2007 trend) to the sharper decline from the previous decade or the trend from prior decades.  Whether focusing on minima, maxima, or average sea ice, the data shows a steeper decline during the first decade of the 21st century compared to previous and subsequent decades.  Should the trend that Tamino calculated continue unabated, a sustained ice-free Arctic will not occur for at least half a century (although seasonal BOEs could occur well before then).

I estimated slopes of three sections at -.035 -.188 and -.025. I doubt we can say slope of third section is statistically significantly less than slope of first section with data Tamino used in July 2016, only that it is less than slope of steep section. In part this is because there is uncertainty in timing of change points and that noticeably affects slopes of the sections.

So basically there is no reason to think that there has been a change in trend in the last 10 years, or that this trend is any different from the last 40 years EXCEPT for the 4-6 year period 2002 - 2006/8 (both years show up in Tamino's blog).

So that's really the only thing we have to go after as far as the trend is concerned, and we have absolutely no idea why these changes happened - why the trend was different 2004 from 2000 or 2010.

As for the future, I do not expect it to be linear! Rather, I expect the trend to increase significantly after a decade or two. But let's see what happens!
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

Archimid

  • Young ice
  • Posts: 3511
    • View Profile
  • Liked: 899
  • Likes Given: 206
Re: When will the Arctic Go Ice Free?
« Reply #978 on: September 03, 2019, 04:18:35 PM »
Nice talk about a 2 dimensional object. Important, valid talk.

 I want to talk volume, but I want to wait for that last datapoint, 2019 minimum volume, preferably at the CAB. But I may join the conversation with the August results the difference might just be noise. Or it may be relevant.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

aslan

  • Frazil ice
  • Posts: 226
    • View Profile
  • Liked: 175
  • Likes Given: 0
Re: When will the Arctic Go Ice Free?
« Reply #979 on: September 03, 2019, 04:22:09 PM »
Without wanting to look harsh, but I think the discussion is taking a bad way. I am not a specialist of this question, but a big missing piece of data is the problem of land versus ocean distribution in Northern Hemisphere.

I think it is useful to go in several different directions. Extract what we can from trends but also consider physical reasoning.

I am not sure I agree on the importance of land restricting ice extent. It certainly does in winter but it is constrained at same places each year. So I am not convinced it has much effect on trends.


If I may insist. Considering for example an uniform retreat of 1° northward per decade, Summer as Winter.
If the ice edge is 1 000 km long, the 1° will lead to 100 time 1 000 km², i.e. 100 000 km² ice loss over a decade. Or, said otherwise, peanuts.
If the ice edge is 10 000 km long, this will lead to 10 000 time 100 km², i.e. 1 000 000 km² ice loss over a decade.
This is really what is ongoing. In winter, the "free" ice edge is not really wide, so even if the ice is marching toward the North at the same speed as in Summer, ice loss will be small. And as said, in the Pacific we are now in the bottleneck of the strait, and in the Atlantic it is not really better. So even a retreat of 2 or 3° over a few years will lead to almost no sea ice extent loss in Winter, while in Summer a retreat of 2° or 3° will be almost the end of the sea ice. This is simple but important math’s. Of course, it makes the evaluation of trends in ice extent more complex. But I really do think brain storming over breaks of the linear slope without acknowledging this fact will lead to a sterile discussion. The changes in slope are way more better explained by changing ice edge conditions than by "hard" physics.

https://tamino.wordpress.com/2010/08/22/on-ice-with-a-twist/

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #980 on: September 03, 2019, 04:23:18 PM »
So that's really the only thing we have to go after as far as the trend is concerned, and we have absolutely no idea why these changes happened


As for the future, I do not expect it to be linear! Rather, I expect the trend to increase significantly after a decade or two. But let's see what happens!

Indeed we we see what happens. With more data 2002-2006 may look like a tiny bump, practically insignificant compared to what happens next. Or continue to look like linear sections or look like a continuing negative sigmoid shaped curve.... (I certainly don't believe the completely flat extrapolation shown by 4 parameter gompertz curve but there are many sigmoid shaped curves.)

I tend towards some form of sigmoid shaped curve as our best extrapolation attempt but there could easily be surprises.

>absolutely no idea
Well we can look to literature and see explanations of MYI no longer making it around Beaufort Gyre. see also my post #954 in this thread.

Archimid

  • Young ice
  • Posts: 3511
    • View Profile
  • Liked: 899
  • Likes Given: 206
Re: When will the Arctic Go Ice Free?
« Reply #981 on: September 03, 2019, 04:25:51 PM »
let's say extent loses cease. Yay.

Where do the extra heat goes? If the sides are not melting but the sun is still shining, where all that heat goes?
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #982 on: September 03, 2019, 04:56:10 PM »

If I may insist. Considering for example an uniform retreat of 1° northward per decade, Summer as Winter.
If the ice edge is 1 000 km long, the 1° will lead to 100 time 1 000 km², i.e. 100 000 km² ice loss over a decade. Or, said otherwise, peanuts.
If the ice edge is 10 000 km long, this will lead to 10 000 time 100 km², i.e. 1 000 000 km² ice loss over a decade.
This is really what is ongoing. In winter, the "free" ice edge is not really wide, so even if the ice is marching toward the North at the same speed as in Summer, ice loss will be small. And as said, in the Pacific we are now in the bottleneck of the strait, and in the Atlantic it is not really better. So even a retreat of 2 or 3° over a few years will lead to almost no sea ice extent loss in Winter, while in Summer a retreat of 2° or 3° will be almost the end of the sea ice. This is simple but important math’s. Of course, it makes the evaluation of trends in ice extent more complex. But I really do think brain storming over breaks of the linear slope without acknowledging this fact will lead to a sterile discussion. The changes in slope are way more better explained by changing ice edge conditions than by "hard" physics.

https://tamino.wordpress.com/2010/08/22/on-ice-with-a-twist/

You are certainly entitled to your opinion (as I am to stick to mine, I hope)

If "This is really what is ongoing" then we should be able to model the area at minimum as a circle with linearly declining radius (r) and as the are is proportional to r^2 the declining area should decrease in a sequence that looks proportional to a 36 25 16 9 4 1 sequence. You may be able to fit such a pattern to the data since 2005 and maybe be able to explain away the failure of the model before that as being related to land restrictions. However I think you will struggle at some point.

This indicates a changing seasonal pattern:



this makes it more complex than summer and winter marching at the same speed, which I don't think you wanted to rule out. However, I believe the main explanation is in albedo feedback and it just continues to get more complex until you need a global climate model. And well we already have them.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #983 on: September 03, 2019, 04:59:40 PM »
let's say extent loses cease. Yay.

Where do the extra heat goes? If the sides are not melting but the sun is still shining, where all that heat goes?

Straw man.

I am saying clearly I don't believe the completely flat extrapolation. So we all agree the slope remains downwards. So how does your argument help in any way as to whether slope of decline increases or decreases?

Archimid

  • Young ice
  • Posts: 3511
    • View Profile
  • Liked: 899
  • Likes Given: 206
Re: When will the Arctic Go Ice Free?
« Reply #984 on: September 03, 2019, 05:11:32 PM »
Quote
Straw man.

It's a question. One that you choose not to answer.

Quote
So how does your argument help in any way as to whether slope of decline increases or decreases?

The arctic is a 3 dimensional object, not a 2 dimensional object. If the area stops declining but the temperature keeps increasing then something got to give. That something is the thickness.


A very interesting development happened this year. The CAB was under attack from all 4 directions. The connection with Greenland was broken. The sun is starting to hit places that it has not touched for millennia. All that extra energy accumulates in the form of melt during melting season and warmth during freezing season.


Quote
However, I believe the main explanation is in albedo feedback and it just continues to get more complex until you need a global climate model. And well we already have them.

And what do they say? The latest runs, not runs from 2011.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #985 on: September 03, 2019, 05:13:30 PM »


I find it interesting that the maximums and minimums seem to be showing a sigmoid like shape. If sigmoid there, why not sigmoid for fit of anomaly data?

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #986 on: September 03, 2019, 05:22:42 PM »




The weird behaviour towards the end is apparently caused by the seasonal cycle getting bigger, so to correct for that Tamino computes an "adaptive anomaly" to get this:

This "weird behavior" becomes less weird when we recognize the destruction of the thick MYI that occurred in 2007, 2010 and, to a lesser degree, 2012. The thick MYI served as a constraint on the minimums that any single melt season could achieve. Once this MYI was destroyed, this constraint disappeared and the minimums became driven more by the characteristics of individual melt seasons.

The destruction of the MYI can be seen in the volume graph where we see the strong melt result in new minimums in volume.



crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #987 on: September 03, 2019, 05:26:48 PM »
Quote
Straw man.

It's a question. One that you choose not to answer.

Strawman means that nobody believes the premise in the first place. Consequently any answers are irrelevant. So I believe I did answer it, just maybe not in the way you wanted.


>A very interesting development, ... connection with Greenland
If you believe interesting things are going to happen, then it is easy to allow such beliefs to cause you to believe something slightly odd is significant and means the interesting things are going to happen soon thereby justifying your beliefs. I don't see your evidence that this is something significant rather than just something slightly unusual. Didn't someone find a lesser disconnection but in the middle of winter some years prior?

binntho

  • Nilas ice
  • Posts: 2193
    • View Profile
  • Liked: 878
  • Likes Given: 235
Re: When will the Arctic Go Ice Free?
« Reply #988 on: September 03, 2019, 05:37:26 PM »
So that's really the only thing we have to go after as far as the trend is concerned, and we have absolutely no idea why these changes happened


As for the future, I do not expect it to be linear! Rather, I expect the trend to increase significantly after a decade or two. But let's see what happens!

Indeed we we see what happens. With more data 2002-2006 may look like a tiny bump, practically insignificant compared to what happens next. Or continue to look like linear sections or look like a continuing negative sigmoid shaped curve.... (I certainly don't believe the completely flat extrapolation shown by 4 parameter gompertz curve but there are many sigmoid shaped curves.)

I tend towards some form of sigmoid shaped curve as our best extrapolation attempt but there could easily be surprises.

>absolutely no idea
Well we can look to literature and see explanations of MYI no longer making it around Beaufort Gyre. see also my post #954 in this thread.

A link would have been nice!

But unfortunately, your post #954 gives no explanation for a steep increase in extent loss between 2002 and 2008. The drastic loss of MYI happens later, after 2007 and particularly 2012, so again no cigar.

In other words, we (plural) have no idea why the trend may have changed in 2002 and again in 2006/8.
because a thing is eloquently expressed it should not be taken to be as necessarily true
St. Augustine, Confessions V, 6

crandles

  • Young ice
  • Posts: 3379
    • View Profile
  • Liked: 239
  • Likes Given: 81
Re: When will the Arctic Go Ice Free?
« Reply #989 on: September 03, 2019, 05:48:39 PM »

image not showing, try this link:
http://nsidc.org/arcticseaicenews/2012/10/

Yes looks like area went down after. I do though wonder about volume in excess of say 1.8m thickness. Wonder if I can find some dosbat graphs.

Perhaps not there either a much longer period:

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #990 on: September 03, 2019, 05:51:31 PM »
I don't think we will see a BOE any time soon and certainly not a perennially BOE and that volume chart is the key. The destruction of the MYI ice in 2007 and 2010 has resulted in a new floor for volume and consequently extent and area minimums. As has been mentioned by many here, it is difficult to get the ice north of 80 degrees to melt out entirely for a number of reasons. Despite the relatively stable volume minimums over the past decade, the ice at minimum does look far different than it did in 2007, more fragmented and more mobile.

So, how might we see a BOE in the coming decade? We are seeing increased dispersion at minimum and warmer winters in general. For us to see a BOE, we will need to have record high dispersion at minimum with  much of the remaining ice pushed into the warm waters in the peripheral seas. This dispersion would need to continue well into a very warm, cloudy but not stormy fall and winter where the Arctic struggles to release heat to space due to a blanket of clouds. This would need to be followed by a near perfect melt season.

Possible? Yes. Likely in the next decade? Not in my opinion.

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #991 on: September 03, 2019, 06:07:18 PM »
A very interesting development happened this year. The CAB was under attack from all 4 directions. The connection with Greenland was broken. The sun is starting to hit places that it has not touched for millennia. All that extra energy accumulates in the form of melt during melting season and warmth during freezing season.

I think the ice moving away from Greenland and the CAA is simply the result of a trend that has been in place for much of the 20th century and is due to the general warming of the Arctic. Over the last century, the Ellesmere Ice Shelf has been decimated as has most of the large rafts of MYI adjacent to it. The resulting ice pack has become more fragmented and mobile and, as we have seen, persistent winds can cause the ice to lift off from Greenland and the CAA.

Sam

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #992 on: September 03, 2019, 06:17:55 PM »
SH,

The destruction of the MYI ice in 2007 and 2010 has resulted in a new floor for volume and consequently extent and area minimums.

I see no evidence of a "floor", only continuing decline.

Sam

Archimid

  • Young ice
  • Posts: 3511
    • View Profile
  • Liked: 899
  • Likes Given: 206
Re: When will the Arctic Go Ice Free?
« Reply #993 on: September 03, 2019, 06:30:42 PM »
Crandles, there are two questions pending. very important ones.

1. If extent stalls but global warming continues, where does that extra energy goes? That is a valid question as we know energy doesn't just disappear. It has to go somewhere.

2. You mentioned the models. What do they say about extent? Recent models please.
I am an energy reservoir seemingly intent on lowering entropy for self preservation.

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #994 on: September 03, 2019, 07:11:33 PM »
SH,

The destruction of the MYI ice in 2007 and 2010 has resulted in a new floor for volume and consequently extent and area minimums.

I see no evidence of a "floor", only continuing decline.

Sam

Did you even look at the chart? Here is another.

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #995 on: September 03, 2019, 07:22:46 PM »
Another look at what I would describe as a new floor which supports a minimum threshold that will prove difficult to breach but looking at ice age instead of volume.

Shared Humanity

  • Guest
Re: When will the Arctic Go Ice Free?
« Reply #996 on: September 03, 2019, 07:28:38 PM »
And while it may very well be that I am seeing what I choose to see in these graphs, I feel there was a change in the behavior of extent minimums at the same time that we saw the MYI devastated by the 2007 and 2010 melt seasons.
« Last Edit: September 03, 2019, 09:46:27 PM by Shared Humanity »

gerontocrat

  • Multi-year ice
  • Posts: 20376
    • View Profile
  • Liked: 5289
  • Likes Given: 69
Re: When will the Arctic Go Ice Free?
« Reply #997 on: September 03, 2019, 08:55:37 PM »
I had another look at my 365 Day trailing Averages, i.e. the continuous trend in average annual sea ice over the years.

I don't see how anyone can say that sea ice is not declining. And this concentration on the minimum - one day in the 365 day year, seems designed to ensure lack of light.

ps: When the area & volume lines cross, average thickness for the whole year will be less than 1 metre.
« Last Edit: September 03, 2019, 09:29:39 PM by gerontocrat »
"Para a Causa do Povo a Luta Continua!"
"And that's all I'm going to say about that". Forrest Gump
"Damn, I wanted to see what happened next" (Epitaph)

aslan

  • Frazil ice
  • Posts: 226
    • View Profile
  • Liked: 175
  • Likes Given: 0
Re: When will the Arctic Go Ice Free?
« Reply #998 on: September 03, 2019, 09:22:10 PM »
Ok, I might be stubborn,  ;D but I will rephrase more abruptly. What if the change in sea ice extent does not reflect a change from an energy point of view?

Glen Koehler

  • Grease ice
  • Posts: 931
    • View Profile
  • Liked: 738
  • Likes Given: 1413
Re: When will the Arctic Go Ice Free?
« Reply #999 on: September 03, 2019, 09:33:24 PM »
Another look at what I would describe as a new floor which supports a minimum threshold that will prove difficult to breach but looking at ice age instead of volume.

Good discussion.  And nice chart of early March Arctic sea ice portions by age of ice SH.
The chart shows 1st-year ice percentage increased from ca. 38.3% in 1984 to 65.8% in 2018, which summarizes to an average gain of 0.8% per year.  If that rate continued for another 48 years then 1st-year ice would account for nearly 100% of early March Arctic sea ice. 

    My understanding is that in one melt season, about 2 meters of ice thickness is lost.  And that the average thickness of 1st-year ice is also about 2 meters. 

   So... once we get to ca. 100% 1st-year ice at start of the melt season in March, then that ice would be thin enough to all melt out by the end of what is now an average melt season.
 
    Of course, with current and accelerating warming of global average surface temperature, by 2067, i.e. 48 years from now, the typical melt season will be much warmer so the scenario i'm creating here may arrive by 2050.

  There would still be refreezing in winter.  But progressively less on average each year.  So that's another trend to throw into the blender.  With a lower starting volume each year, the year when all, or nearly all, the Arctic sea ice melts out by end of melt season gets even earlier.

   And I end up back with what I consider the simplest and most convincing predictor of when Arctic sea ice annual minimum.  That being the Wipneus graph showing linear trend of annual minimum volume reaching zero around 2030. 
https://14adebb0-a-62cb3a1a-s-sites.googlegroups.com/site/arctischepinguin/home/piomas/grf/piomas-trnd7.png?attachauth=ANoY7co-JT1jjQGvpGxNEvrl0OaSjsV--ZIsZkJxM9mAn3liTvPonGye-SIwYIoMrdz3WhCjzTF-7eO1y4xjVMqo4Mqe6Py4I4KqkaHO97Qm_F7_j-0wn7hmI_ipG2OPbsk-eONQXLct1Ze62owhGmvWbQvdKVz21eBmKeJaoqz_EeC_3JYndbm6NHkkBDxQHAohuZUi9NrYqBUZ105uBQF7giHXXIyLZiC9G_kb1PqxmVgYs8y-q-WkPdR-VoZitgO_CypQuuTS&attredirects=0

     So no need for my layers of speculation and assumptions, just look at the numbers as graphed by Wipneus!  Also interesting to note that 2019 September minimum volume is going to land just about exactly on the linear trend in the Wipneus graph. 
“What is at stake.... Everything, I would say." ~ Julienne Stroeve