When will the Arctic Go Ice Free?

[Klondike Kat]

Shared Humanity:  2111, based on the decline in MYI over the past decade

Could you point out where I said 2111 when I actually said not before 2030? Everyone here has been trying to use charts to determine when we will see our first BOE. I have been providing reasons why I believe the declines in SIE at minimum has slowed. I really don’t give a flying fuck what extrapolating a chart says.

You never said that.  Rather, it was an extrapolation from the chart of sea ice volume at minimum.  Nothing more.  As Oren stated, this are averages, and volatility would likely result in an earlier occurrence, and reverting back to higher minima.  You may not care much about extrapolations, but multiple posters here and been using such to make their predictions.

[Shared Humanity]

Apparently you can only attach 4 items to any single comment so my reference to Chart 5 in the above comment does not have a Chart 5 attached. Here is the SIE chart that I had intended to attach and I do believe it shows the same flattening effect in SIE minimums over the past decade as can be found on the volume chart.

[binntho]

Apparently you can only attach 4 items to any single comment so my reference to Chart 5 in the above comment does not have a Chart 5 attached. Here is the SIE chart that I had intended to attach and I do believe it shows the same flattening effect in SIE minimums over the past decade as can be found on the volume chart.

It does not show any flattening (strike effect). And neither does the volume chart.

EDIT: These charts do not show flattening any more than charts showing annual temperatures showed any "hiatus" in the 2000's. Lots and lots of denialists kept claiming that global warming had stopped, and they thought they could see it in the charts, just as you claim that you see a flattening in your charts.

But there was no hiatus in global warming, just natural random variability combined with an outlier (1998). And the same is happening with the SI graphs - natural random variability combined with an outlier (2012)

[Shared Humanity]

Shared Humanity:  2111, based on the decline in MYI over the past decade

Could you point out where I said 2111 when I actually said not before 2030? Everyone here has been trying to use charts to determine when we will see our first BOE. I have been providing reasons why I believe the declines in SIE at minimum has slowed. I really don’t give a flying fuck what extrapolating a chart says.

You never said that.  Rather, it was an extrapolation from the chart of sea ice volume at minimum.  Nothing more.  As Oren stated, this are averages, and volatility would likely result in an earlier occurrence, and reverting back to higher minima.  You may not care much about extrapolations, but multiple posters here and been using such to make their predictions.

You should exercise more care when speaking for others. I did not say 2111. I said not before 2030 and stating otherwise is not appreciated. I never speak for others here and when I do want to respond to something that another person has said, I quote their exact words. You should have done this instead of putting words in my mouth.

[Shared Humanity]

SharedHumanity, I salute your efforts!

And I mostly agree with you, except (as could be expected)  for where you say "... and the flattening of the graph of SIE at minimum over the past decade ...". I don't think that there is a flattening, but a steady (if highly variable) decline.

And this presumption actually changes everything. If we presume that the ice is flatlining around 4M then what you are saying is quite reasonable.

But if we assume a steady decline, as per the trendline in this image of -0.811 MKm2 per decade then we should be seeing values hovering betwen 3 and 4 in the 20's and 2 and 3 in the 40's.

I agree that this is where our opinions diverge and I think depending on and extrapolating a trend line on any chart is a mistake. I have not been trying to do this. All of the charts are measuring processes that are occurring in the Arctic. I think the appropriate focus should be on what is actually happening in the Arctic and allowing the charts to focus our inquiry on these processes and changes that are occurring in the Arctic.

What do I mean by this?

Example: The Beaufort use to be where MYI would age and thicken and much of the thickest ice could be found on the Pacific side of the CAB. The brutal melt seasons in 2007, 2010 and 2012 have altered this behavior permanently, I fear. The Beaufort now melts out very early due to it being primarily FYI, warms due to increased insolation and has now become a killing ground for MYI. I do not expect 5 year old ice to ever rebound for this reason.

[SteveMDFP]

And I mostly agree with you, except (as could be expected)  for where you say "... and the flattening of the graph of SIE at minimum over the past decade ...". I don't think that there is a flattening, but a steady (if highly variable) decline.

And this presumption actually changes everything. If we presume that the ice is flatlining around 4M then what you are saying is quite reasonable.

But if we assume a steady decline, as per the trendline in this image of -0.811 MKm2 per decade then we should be seeing values hovering betwen 3 and 4 in the 20's and 2 and 3 in the 40's.



But these values to me indicate an increasingly unstable ice pack. If we had half the extent we have now, I think it would more or less automatically be blown all over the place and melt out more or less constantly.

So my attempt at an answer: In the 20's a combination of extremes may give us our first BOE. In the 30's a BOE will start to happen regularly, in the 40's it will be more or less an annual event.

I would tend to agree that a simple linear trend line makes most sense for projections.  Mostly based on Occam's Razor.  A linear trend line includes the smallest number of variables--just one point and a slope define it.  Any curved projection suffers the objection "entities should not be multiplied without necessity."

In particular, polynomial curves to fit the data are almost certainly a poor alternative choice.  Given enough terms, curves can be created to nicely fit any data set.  But in all relevant polynomials, the extreme left and right ends curve to nonsense.

[Klondike Kat]

Shared Humanity:  2111, based on the decline in MYI over the past decade

Could you point out where I said 2111 when I actually said not before 2030? Everyone here has been trying to use charts to determine when we will see our first BOE. I have been providing reasons why I believe the declines in SIE at minimum has slowed. I really don’t give a flying fuck what extrapolating a chart says.

You never said that.  Rather, it was an extrapolation from the chart of sea ice volume at minimum.  Nothing more.  As Oren stated, this are averages, and volatility would likely result in an earlier occurrence, and reverting back to higher minima.  You may not care much about extrapolations, but multiple posters here and been using such to make their predictions.

You should exercise more care when speaking for others. I did not say 2111. I said not before 2030 and stating otherwise is not appreciated. I never speak for others here and when I do want to respond to something that another person has said, I quote their exact words. You should have done this instead of putting words in my mouth.

I never spoke for others, and did not put any words in your mouth.  I clearly stated at the start of my post, "Extrapolating these trends yields forward, the following estimates for the first onset of an ice-free Arctic (< 1M km2) are derived:"

[crandles]

What do the posters know that the experts do not?

Shouldn't we assume the experts know more than the posters here?
Shouldn't it be:
 What do the experts know that the posters here do not?

I certainly don't count myself as an expert.

[crandles]

It does not show any flattening (strike effect). And neither does the volume chart.

EDIT: These charts do not show flattening any more than charts showing annual temperatures showed any "hiatus" in the 2000's. Lots and lots of denialists kept claiming that global warming had stopped, and they thought they could see it in the charts, just as you claim that you see a flattening in your charts.

But there was no hiatus in global warming, just natural random variability combined with an outlier (1998). And the same is happening with the SI graphs - natural random variability combined with an outlier (2012)

I am not sure but think the temperature trend did not reach statistically significant levels if you accounted for cherry picking 2012 as the start point. Tamino is saying slowing rate of arctic sea ice loss is statistically significant. At what point do you start to believe and how fully should you believe?

Rather than the 95% which you don't seem to believe, maybe you want 99% significance or 99.5% or 99.9%? Where to draw the line?

In reality, when something odd/unexpected/different shows up, it could be entirely a new trend or it could be entirely noise and no change in trend, but the likeliest scenario is it is a bit of noise and a bit of trend change.

7 times faster rate suddenly disappearing is, according to my subjective priors not very likely at all, til we see the data. It is now a possibility but a more likely possibility is the trend rate didn't change by as much as Tamino's piecewise linear trend lines show, instead it is likely a bit of noise and some decline in the rate but not as much as 7 times and to fit it would need to be that rate for more than 4 years.

All noise is possible, but a bit of noise and a bit of trend change is more likely.

[Shared Humanity]

Shared Humanity:  2111, based on the decline in MYI over the past decade

Could you point out where I said 2111 when I actually said not before 2030? Everyone here has been trying to use charts to determine when we will see our first BOE. I have been providing reasons why I believe the declines in SIE at minimum has slowed. I really don’t give a flying fuck what extrapolating a chart says.

You never said that.  Rather, it was an extrapolation from the chart of sea ice volume at minimum.  Nothing more.  As Oren stated, this are averages, and volatility would likely result in an earlier occurrence, and reverting back to higher minima.  You may not care much about extrapolations, but multiple posters here and been using such to make their predictions.

You should exercise more care when speaking for others. I did not say 2111. I said not before 2030 and stating otherwise is not appreciated. I never speak for others here and when I do want to respond to something that another person has said, I quote their exact words. You should have done this instead of putting words in my mouth.

I never spoke for others, and did not put any words in your mouth.  I clearly stated at the start of my post, "Extrapolating these trends yields forward, the following estimates for the first onset of an ice-free Arctic (< 1M km2) are derived:"

You attached my nym to a forecast of the first BOE at 2111. How is that not implying that I believe in this ridiculous projection?

Given BAU for the next 2 decades which is likely, I think we will be perennially ice free by 2050. I also think we will see our 1st BOE between 2030 and 2040.

[Tom_Mazanec]

I just searched for /2111/ and all the posts that came up are today's, not an old prediction.

[Klondike Kat]

Shared Humanity:  2111, based on the decline in MYI over the past decade

Could you point out where I said 2111 when I actually said not before 2030? Everyone here has been trying to use charts to determine when we will see our first BOE. I have been providing reasons why I believe the declines in SIE at minimum has slowed. I really don’t give a flying fuck what extrapolating a chart says.

You never said that.  Rather, it was an extrapolation from the chart of sea ice volume at minimum.  Nothing more.  As Oren stated, this are averages, and volatility would likely result in an earlier occurrence, and reverting back to higher minima.  You may not care much about extrapolations, but multiple posters here and been using such to make their predictions.

You should exercise more care when speaking for others. I did not say 2111. I said not before 2030 and stating otherwise is not appreciated. I never speak for others here and when I do want to respond to something that another person has said, I quote their exact words. You should have done this instead of putting words in my mouth.

I never spoke for others, and did not put any words in your mouth.  I clearly stated at the start of my post, "Extrapolating these trends yields forward, the following estimates for the first onset of an ice-free Arctic (< 1M km2) are derived:"

You attached my nym to a forecast of the first BOE at 2111. How is that not implying that I believe in this ridiculous projection?

Given BAU for the next 2 decades which is likely, I think we will be perennially ice free by 2050. I also think we will see our 1st BOE between 2030 and 2040.

I attached your name so that others could find the graph easily from your post.  That is all.  No implications involved.  Others have acknowledged that they posted such graphics, stated that they do not agree, but do not feel implicated by such findings.

[Shared Humanity]

My request stands that you exercise more care when using my name

[Klondike Kat]

My request stands that you exercise more care when using my name

Acknowledged.

[binntho]

It does not show any flattening (strike effect). And neither does the volume chart.

EDIT: These charts do not show flattening any more than charts showing annual temperatures showed any "hiatus" in the 2000's. Lots and lots of denialists kept claiming that global warming had stopped, and they thought they could see it in the charts, just as you claim that you see a flattening in your charts.

But there was no hiatus in global warming, just natural random variability combined with an outlier (1998). And the same is happening with the SI graphs - natural random variability combined with an outlier (2012)

I am not sure but think the temperature trend did not reach statistically significant levels if you accounted for cherry picking 2012 as the start point. Tamino is saying slowing rate of arctic sea ice loss is statistically significant. At what point do you start to believe and how fully should you believe?

We are talking about two different things here. SharedHumanity is claiming that extent flatlined. Tamino shows that extent loss has decelerated. A decelerated extent loss is still an ongoing extent loss, and cannot be counted as flatlining.

Rather than the 95% which you don't seem to believe, maybe you want 99% significance or 99.5% or 99.9%? Where to draw the line?

I don't really believe in anything, let alone statistics! What a strange question. But I'm 100% certain that a value with a 95% threshold has a 95% change of being correct.

In reality, when something odd/unexpected/different shows up, it could be entirely a new trend or it could be entirely noise and no change in trend, but the likeliest scenario is it is a bit of noise and a bit of trend change.

Perhaps. Has anybody done any research to validate the claim that if somebody thinks they see a pattern in a graph, then the most likely explanation is that there is a pattern, and not just random variability?

7 times faster rate suddenly disappearing is, according to my subjective priors not very likely at all, til we see the data. It is now a possibility but a more likely possibility is the trend rate didn't change by as much as Tamino's piecewise linear trend lines show, instead it is likely a bit of noise and some decline in the rate but not as much as 7 times and to fit it would need to be that rate for more than 4 years.

All noise is possible, but a bit of noise and a bit of trend change is more likely.

Where did that 7 times faster rate come from? And of course, it can have any likelyhood it wants, but once the data is in it's either true or false.

[binntho]

Having repeatedly claimed that the only truly quantifiable and reliable underlying physical force behind the ongoing sea ice loss is global warming,  I decided to do a couple of rough-and-tumble graphs comparing the annual minimum with the annual average temperature north of 67 degrees (which is at times surprisingly out of sync with the global temps).

Being a very lazy bugger, I just read the temp figures off Zach Labe's graph from here (graph nr. 5, yellow line) https://sites.uci.edu/zlabe/arctic-temperatures/

I also assigned 2019 to a temp value of 2.5. I could have gone higher, since based on the last graph on Zach's page (the monthly ranking one), 2019 currently ranks higher than 2016, and that year had a temp of 2.95.

For the SIE I used NSIDC, with todays value of 4.2 as a proxy for 2019 minimum. I then made two graphs, one  with the numbers and one with the temps as a running 2 year average. The red lines show SIE numbers in the millions of km2, the blue lines show average temps north of 67N, from a 1981-2010 baseline.

And what surprises me is how good the correspondance is, particularly in the last 20 years. When the 2 year temps go up, SIE goes down and vice versa. Which really goes to show that it is indeed the heat that is running things up there (as expected). And presumably, as the globe continues to warm, so will the Arctic (at double the rate) and SIE will continue to fall.

[binntho]

Some people will no doubt jump at the fact that 2012 is so much lower SIE wise than the other low years. But as we all know, weather has a big effect and 2012 was exceptional when it comes to the various factors such as pre-conditioning, melting momentum and GAC.

On the other hand, 2019 has not had the same favourable factors at play - melting momentum got a late start, august was a damp squib and no GAC. But because of the much higher temps, it's still heading towards the prize platform.

EDIT: Which leads one to the conclusion that if 2019 had had the same weather at the same times as 2012, 2019 would easily have broken the 2012 record. What was missing this year was the very early conditioning (too clear skies early spring), and in August 2019 the weather has been dispersing ice in all directions rather than compacting after one huge storm as in 2012.

[binntho]

Finally I'd like to point out another graph where a linear trend is clearly the way to go. Stephan posted over on the 2019 sea ice area and extent data  thread a graph showing August anomalies, here it is:

[El Cid]

The major problem with this linear (or any other) projections is: why do you start your projection from 1979 why not 1969 or 1989? And the only reason is that that is when the dataset starts. And that is a very very poor reason...

at the minimum you should start your projection when AGW started in earnest...if you can nail it

[binntho]

The major problem with this linear (or any other) projections is: why do you start your projection from 1979 why not 1969 or 1989? And the only reason is that that is when the dataset starts. And that is a very very poor reason...

at the minimum you should start your projection when AGW started in earnest...if you can nail it

Well, that's a non-sequitur. You can only work with the data you have, and the data you have is the data you have and is what you have to work with. Complaining about not working with non-data is a no no.

[kassy]

Acceleration and more recently deceleration appears to have been shown to be statistically significant.

Indeed it is, based on the full daily dataset that Tamino uses. The number of datapoints here are well sufficient to make such claims.

The problem is that we have no explanation for neither the acceleration nor the deceleration, which leaves open the possibility that it didn't happen!

When discussing melt in certain areas it is often mention that some sea areas always melt out and that others don´t. Seas change in behaviour over time due to increased heat input and seas further south are hit first.

Could it be that the acceleration is seas at at the edges melting out after which they disappear from the ice cover data. All the ice left now is the more protected ice so melt for now is slower.

[binntho]

Acceleration and more recently deceleration appears to have been shown to be statistically significant.

Indeed it is, based on the full daily dataset that Tamino uses. The number of datapoints here are well sufficient to make such claims.

The problem is that we have no explanation for neither the acceleration nor the deceleration, which leaves open the possibility that it didn't happen!

When discussing melt in certain areas it is often mention that some sea areas always melt out and that others don´t. Seas change in behaviour over time due to increased heat input and seas further south are hit first.

Could it be that the acceleration is seas at at the edges melting out after which they disappear from the ice cover data. All the ice left now is the more protected ice so melt for now is slower.

This would be quite reasonable if it wasn't for the fact that the acceleration starts in 2002 and is finished in 2007. The deceleration may be caused by loss of MYI and/or periferal-central dynamics as you mention kassy. But what caused the acceleration?

[Klondike Kat]

Acceleration and more recently deceleration appears to have been shown to be statistically significant.

Indeed it is, based on the full daily dataset that Tamino uses. The number of datapoints here are well sufficient to make such claims.

The problem is that we have no explanation for neither the acceleration nor the deceleration, which leaves open the possibility that it didn't happen!

When discussing melt in certain areas it is often mention that some sea areas always melt out and that others don´t. Seas change in behaviour over time due to increased heat input and seas further south are hit first.

Could it be that the acceleration is seas at at the edges melting out after which they disappear from the ice cover data. All the ice left now is the more protected ice so melt for now is slower.

This would be quite reasonable if it wasn't for the fact that the acceleration starts in 2002 and is finished in 2007. The deceleration may be caused by loss of MYI and/or periferal-central dynamics as you mention kassy. But what caused the acceleration?

You may to check out this site, which shows the rapid decline of the ice in the Siberian and Laptev Seas.  The ice in those areas declined rapidly at minimum over those years.  After 2006, the minimum extent is based largely on the ice remaining in the CAB

https://climate.nasa.gov/vital-signs/arctic-sea-ice/

[binntho]

This would be quite reasonable if it wasn't for the fact that the acceleration starts in 2002 and is finished in 2007. The deceleration may be caused by loss of MYI and/or periferal-central dynamics as you mention kassy. But what caused the acceleration?

You may to check out this site, which shows the rapid decline of the ice in the Siberian and Laptev Seas.  The ice in those areas declined rapidly at minimum over those years.  After 2006, the minimum extent is based largely on the ice remaining in the CAB

https://climate.nasa.gov/vital-signs/arctic-sea-ice/

I'm not seeing this, if it is the animation you mean. Both seas melt out occasionally before 2000, but after 2001 they seem to melt out most years (but not all).

As for the extent at minimum being largely based on the ice remaining in the CAB, this seems (from the animation) to be almost true 2007 and particularly 2012 but not so much the other years, with some millions of km2 being outside the CAB (in various directions at various times) but it is extremely difficult to tell from the animation.

If the explanation is that those two seas melted rapidly in those 4 years then we would need a much better dataset, and perhaps most important, WHY would those two seas melt so fast in exactly those 4 years?

[binntho]

But all that aside Klondike Kat, as I have conclusively shown above, the extent follows the temperatures very closely. Add a bit of weather randomness and you get the 2012 outlier, but other than that there is absolutely no reason to think that there is a stall, since the temperatures are not stalling and the extent at minimum clearly follows the temps.

[El Cid]

binntho,

I believe you are 100% wrong about this. What if you had sea ice extent data from 1900 and it showed (likely) that there was not much change in sea ice between 1900 and 1979. I give you one such example and a linear fit. I also give you another one where we have the same data, but only from 1990, not 1979. You can see that the two lines (equation showed) are absolutely different although we used the same dataset but with different timeframes. And both linear fits are different than the one you are using which uses data from 1979.

In fact, the only reason that yout linear fit works (somewhat) is that coincidentally (pure luck) AGW started in earnest at the same time that they started measuring ice extent. What is they started measuring in 1500? How would your fit work? Would you say that we need to start the fit from 1500, "because it is a non sequitur that we use the data we have"

[Klondike Kat]

But all that aside Klondike Kat, as I have conclusively shown above, the extent follows the temperatures very closely. Add a bit of weather randomness and you get the 2012 outlier, but other than that there is absolutely no reason to think that there is a stall, since the temperatures are not stalling and the extent at minimum clearly follows the temps.

Conclusively?  Your temperature plot continues to increase over the past decade, but the minimum does not.  Plotting each variable against a third (time) does not show correlation.  Try plotting minimum against temperature.  The graph shows linearity between -0.5 and +1.5, but digress below and above those temperatures.

[binntho]

But all that aside Klondike Kat, as I have conclusively shown above, the extent follows the temperatures very closely. Add a bit of weather randomness and you get the 2012 outlier, but other than that there is absolutely no reason to think that there is a stall, since the temperatures are not stalling and the extent at minimum clearly follows the temps.

Conclusively?  Your temperature plot continues to increase over the past decade, but the minimum does not.  Plotting each variable against a third (time) does not show correlation.  Try plotting minimum against temperature.  The graph shows linearity between -0.5 and +1.5, but digress below and above those temperatures.

"Conclusively" was provocative, I know, just my Friday teaser! But dragging in data from before the satellite era doesn't really make any difference - besides, everybody would expect there to be changes in trend since 1900 since there have been major changes in temperature trends during the same area, and temperature is after all the driving force.

You suggest I try plotting minimum against temperature. I thought that was what I was doing? If not, perhaps you'd like to do it and show me?

And so that we can both agree on one thing: The graph for SIE minimums does seem to level off after 2010. But what I have beent trying to say all this time is that "seems" is not enough - it has to be statistically valid and there has to be a physical cause.

The data points are too few to make it statistically valid, and the 2012 outlier is just the thing to make a graph "seem" to flatline. So I have learned to distrust exactly this sort of thing. We have a graph that may or may not exhibit "leveling off" - and I am unable to conclude either one thing or another just from looking at the graph.

The possible physical causes of such a "leveling off" have been repeated several times, but not shown to be real. I.e. nobody has demonstrated a causal effect that is stronger than the variability of weather and the relentlessly rising temperatures.

This is exactly demonstrated by Tamino's bomb of an aceleration in 2002 and deceleration in 2006. The best we seem to be able to come up with is that perhaps we can explain the deceleration after 2007 (but NB there is no mention of a "stall" or "levelling off" - Tamino still shows a very decent linear decline after 2007) but we have only vague guesses as to why it accelerated in 2002!

As long as we can't do better to explain something that actually did happen, I for one am not willing to accept at face value that a) there is a stall since 2010 or b) that we have a valid explanation for such a stall.

[Klondike Kat]

But all that aside Klondike Kat, as I have conclusively shown above, the extent follows the temperatures very closely. Add a bit of weather randomness and you get the 2012 outlier, but other than that there is absolutely no reason to think that there is a stall, since the temperatures are not stalling and the extent at minimum clearly follows the temps.

Conclusively?  Your temperature plot continues to increase over the past decade, but the minimum does not.  Plotting each variable against a third (time) does not show correlation.  Try plotting minimum against temperature.  The graph shows linearity between -0.5 and +1.5, but digress below and above those temperatures.

"Conclusively" was provocative, I know, just my Friday teaser! But dragging in data from before the satellite era doesn't really make any difference - besides, everybody would expect there to be changes in trend since 1900 since there have been major changes in temperature trends during the same area, and temperature is after all the driving force.

You suggest I try plotting minimum against temperature. I thought that was what I was doing? If not, perhaps you'd like to do it and show me?

And so that we can both agree on one thing: The graph for SIE minimums does seem to level off after 2010. But what I have beent trying to say all this time is that "seems" is not enough - it has to be statistically valid and there has to be a physical cause.

The data points are too few to make it statistically valid, and the 2012 outlier is just the thing to make a graph "seem" to flatline. So I have learned to distrust exactly this sort of thing. We have a graph that may or may not exhibit "leveling off" - and I am unable to conclude either one thing or another just from looking at the graph.

The possible physical causes of such a "leveling off" have been repeated several times, but not shown to be real. I.e. nobody has demonstrated a causal effect that is stronger than the variability of weather and the relentlessly rising temperatures.

This is exactly demonstrated by Tamino's bomb of an aceleration in 2002 and deceleration in 2006. The best we seem to be able to come up with is that perhaps we can explain the deceleration after 2007 (but NB there is no mention of a "stall" or "levelling off" - Tamino still shows a very decent linear decline after 2007) but we have only vague guesses as to why it accelerated in 2002!

As long as we can't do better to explain something that actually did happen, I for one am not willing to accept at face value that a) there is a stall since 2010 or b) that we have a valid explanation for such a stall.

OK, your teaser dragged me in.  Also, I did not bring in data from before the satellite era, that was another poster.  I presume that you used excel to plot your data.  If so, try switching your X-values for the minimum data from year to temperature, and eliminate the temperature line altogether. 

[dnem]

I find this entire line of argument close to comical. Since 2007 we have 12 years of sea ice data.  The arctic environment has been changing remarkably quickly during this time, resulting in erratic and variable melt seasons with an unmistakable trend toward less ice of poorer quality at the end of the season.  To say we have any idea about how the arctic climate will change over the next 10 or 20 years is also comical.  There are certainly some well-reasoned arguments here about why or why not the system may, or may not, stabilize as it approaches a total melt out.  And I realize it's a lot easier to sit here and throw up my hands and say prediction is pointless, but prediction is ... pointless.

And I predict a shattered and dispersed non-contiguous pack by summer's end before 2030.

[Stephan]

It is time for the monthly update of my extrapolation when the extent [Extent], volume [Volumen], thickness [Dicke] and area [Fläche] (new!) will reach zero. The extrapolation occured linearly and by a logarithmic function; the latter one almost constantly resulting in earlier times (valid for volume, area and thickness, not for extent). August value now includes 2019.
Extent, volume, thickness and area for August 2019 lie below the long-term linear trend lines. These anomalies are significant, therefore the "BOE numbers" decreased by 2 years (extent) and 3 years (volume) compared to last year and slightly increased by one year for thickness.
The order (earlier → later BOE) generally is volume < thickness < area < extent.

Please note that this is not a forecast but a trend!
See attached table.

[crandles]

It does not show any flattening (strike effect). And neither does the volume chart.

EDIT: These charts do not show flattening any more than charts showing annual temperatures showed any "hiatus" in the 2000's. Lots and lots of denialists kept claiming that global warming had stopped, and they thought they could see it in the charts, just as you claim that you see a flattening in your charts.

But there was no hiatus in global warming, just natural random variability combined with an outlier (1998). And the same is happening with the SI graphs - natural random variability combined with an outlier (2012)

I am not sure but think the temperature trend did not reach statistically significant levels if you accounted for cherry picking 2012 as the start point. Tamino is saying slowing rate of arctic sea ice loss is statistically significant. At what point do you start to believe and how fully should you believe?

We are talking about two different things here. SharedHumanity is claiming that extent flatlined. Tamino shows that extent loss has decelerated. A decelerated extent loss is still an ongoing extent loss, and cannot be counted as flatlining.

Ah. I took word 'flattening' to mean 'becoming less steep' rather than completely horizontal so yes if we are talking different things then I have probably gone over the top.

7 times faster rate suddenly disappearing is, according to my subjective priors not very likely at all, til we see the data. It is now a possibility but a more likely possibility is the trend rate didn't change by as much as Tamino's piecewise linear trend lines show, instead it is likely a bit of noise and some decline in the rate but not as much as 7 times and to fit it would need to be that rate for more than 4 years.

All noise is possible, but a bit of noise and a bit of trend change is more likely.

Where did that 7 times faster rate come from? And of course, it can have any likelyhood it wants, but once the data is in it's either true or false.

I eyeballed tamino's piecewise graph and got rates of decline of -.035 -.1875 and -.025.
.025*7.5=.1875 but doubtless someone could pixel count more accurately than my eyeballing of the graph.

The data is the data so once in it is true or false but the appropriate trendline is still uncertain.

[kassy]

When discussing melt in certain areas it is often mention that some sea areas always melt out and that others don´t. Seas change in behaviour over time due to increased heat input and seas further south are hit first.

Could it be that the acceleration is seas at at the edges melting out after which they disappear from the ice cover data. All the ice left now is the more protected ice so melt for now is slower.

This would be quite reasonable if it wasn't for the fact that the acceleration starts in 2002 and is finished in 2007. The deceleration may be caused by loss of MYI and/or periferal-central dynamics as you mention kassy. But what caused the acceleration?

The acceleration is caused by outer areas that used to have ice and parts of it partially survive changing to melting out more and more and then not refreezing. After they remain mostly ice free they cannot contribute to the anomaly data.

In the years when they started to melt out completely these outer areas contributed a lot of change and the areas protected by them also started partially melting.

Lets say if you had extent per subarea so you can make a graph for SIE loss for each year and the amount the separate areas contribute. You should see some areas become more changeable and then disappear.

[binntho]

When discussing melt in certain areas it is often mention that some sea areas always melt out and that others don´t. Seas change in behaviour over time due to increased heat input and seas further south are hit first.

Could it be that the acceleration is seas at at the edges melting out after which they disappear from the ice cover data. All the ice left now is the more protected ice so melt for now is slower.

This would be quite reasonable if it wasn't for the fact that the acceleration starts in 2002 and is finished in 2007. The deceleration may be caused by loss of MYI and/or periferal-central dynamics as you mention kassy. But what caused the acceleration?

The acceleration is caused by outer areas that used to have ice and parts of it partially survive changing to melting out more and more and then not refreezing. After they remain mostly ice free they cannot contribute to the anomaly data.

Yes, that's the explanation that's doing the rounds. But it's just a guess. Nobody has quantified whether this geographical effect (i.e. different seas melting/not melting every summer) has any relevance. I for one am not sure how to even start trying to do that, and I find the suddenness of acceleration and deceleration as found by Tamino doesn't really match the geographical explanation.

So as long as the best we can do is make intuitive guesses about events in the recent past, we should be a bit more careful about making claims about the present or the future.

[crandles]

Is it an intuitive guess if it is in the scientific literature?

http://dosbat.blogspot.com/2012/05/transition-complete.html

discuses a few including
Nghiem et al 2007
Maslanik and Tschudi
Polyakov et al
Barber et al, 2010
Haas et al, 2008,

[binntho]

Is it an intuitive guess if it is in the scientific literature?

http://dosbat.blogspot.com/2012/05/transition-complete.html

discuses a few including
Nghiem et al 2007
Maslanik and Tschudi
Polyakov et al
Barber et al, 2010
Haas et al, 2008,

The "intuitive" guess was the posited explanation for the acceleration in 2002 - nothing in the above linked article even tries to touch on that. In fact, I am totally missing the relevance here!

Besides, isn't it totally counterfactual to claim that the loss of MYI somehow causes a flatlining of the annual minimums? I'd have thought it would be exactly the opposite.

When the slowly vanishing MYI was finally mostly gone in 2012, the minimums should have started to drop ever lower! The difficult-to-melt ice was all gone, only the easy FYI left, and minimums should have plummeted. Instead the annual minimums are (most probably) heading steadily downwards or have (quite possibly) stalled.

It looks as if there is no any causal link between the trends in annual minimums and the loss of MYI.

[Klondike Kat]

Is it an intuitive guess if it is in the scientific literature?

http://dosbat.blogspot.com/2012/05/transition-complete.html

discuses a few including
Nghiem et al 2007
Maslanik and Tschudi
Polyakov et al
Barber et al, 2010
Haas et al, 2008,

The "intuitive" guess was the posited explanation for the acceleration in 2002 - nothing in the above linked article even tries to touch on that. In fact, I am totally missing the relevance here!

Besides, isn't it totally counterfactual to claim that the loss of MYI somehow causes a flatlining of the annual minimums? I'd have thought it would be exactly the opposite.

When the slowly vanishing MYI was finally mostly gone in 2012, the minimums should have started to drop ever lower! The difficult-to-melt ice was all gone, only the easy FYI left, and minimums should have plummeted. Instead the annual minimums are (most probably) heading steadily downwards or have (quite possibly) stalled.

It looks as if there is no any causal link between the trends in annual minimums and the loss of MYI.

I will agree with you on this one.  The decrease in sea ice occurred largely in the peripheral seas, which were predominantly FYI.  I think (guess, if you prefer) that the acceleration was due to the loss of this ice.  The CAB and areas around the Canadian archipelago contain more MYI, and hence more difficult to melt.

[binntho]

Is it an intuitive guess if it is in the scientific literature?

http://dosbat.blogspot.com/2012/05/transition-complete.html

discuses a few including
Nghiem et al 2007
Maslanik and Tschudi
Polyakov et al
Barber et al, 2010
Haas et al, 2008,

The "intuitive" guess was the posited explanation for the acceleration in 2002 - nothing in the above linked article even tries to touch on that. In fact, I am totally missing the relevance here!

Besides, isn't it totally counterfactual to claim that the loss of MYI somehow causes a flatlining of the annual minimums? I'd have thought it would be exactly the opposite.

When the slowly vanishing MYI was finally mostly gone in 2012, the minimums should have started to drop ever lower! The difficult-to-melt ice was all gone, only the easy FYI left, and minimums should have plummeted. Instead the annual minimums are (most probably) heading steadily downwards or have (quite possibly) stalled.

It looks as if there is no any causal link between the trends in annual minimums and the loss of MYI.

I will agree with you on this one.  The decrease in sea ice occurred largely in the peripheral seas, which were predominantly FYI.  I think (guess, if you prefer) that the acceleration was due to the loss of this ice.  The CAB and areas around the Canadian archipelago contain more MYI, and hence more difficult to melt.

I'm not so sure that the peripheral seas were predominantly FYI. Didn't we use to have a big long arm of MYI stretching accross the Beaufort and into the Laptev and ESS?

[Klondike Kat]

Is it an intuitive guess if it is in the scientific literature?

http://dosbat.blogspot.com/2012/05/transition-complete.html

discuses a few including
Nghiem et al 2007
Maslanik and Tschudi
Polyakov et al
Barber et al, 2010
Haas et al, 2008,

The "intuitive" guess was the posited explanation for the acceleration in 2002 - nothing in the above linked article even tries to touch on that. In fact, I am totally missing the relevance here!

Besides, isn't it totally counterfactual to claim that the loss of MYI somehow causes a flatlining of the annual minimums? I'd have thought it would be exactly the opposite.

When the slowly vanishing MYI was finally mostly gone in 2012, the minimums should have started to drop ever lower! The difficult-to-melt ice was all gone, only the easy FYI left, and minimums should have plummeted. Instead the annual minimums are (most probably) heading steadily downwards or have (quite possibly) stalled.

It looks as if there is no any causal link between the trends in annual minimums and the loss of MYI.

I will agree with you on this one.  The decrease in sea ice occurred largely in the peripheral seas, which were predominantly FYI.  I think (guess, if you prefer) that the acceleration was due to the loss of this ice.  The CAB and areas around the Canadian archipelago contain more MYI, and hence more difficult to melt.

I'm not so sure that the peripheral seas were predominantly FYI. Didn't we use to have a big long arm of MYI stretching accross the Beaufort and into the Laptev and ESS?

Yes, but that decades ago.  By the turn of the century, it was mostly FYI.  Since then, there was the occasional summer in which some ice survived, but not much.

[binntho]

Yes, but that decades ago.  By the turn of the century, it was mostly FYI.  Since then, there was the occasional summer in which some ice survived, but not much.

Not so sure about that. I seem to remember last year we had a tongue of MYI in the ESS - and not that many years ago (at least after I joined in 2016) I remember A-Team and others showing the MYI ridges stretching across the Beaufort and Laptev.

[crandles]

The "intuitive" guess was the posited explanation for the acceleration in 2002 - nothing in the above linked article even tries to touch on that. In fact, I am totally missing the relevance here!

Besides, isn't it totally counterfactual to claim that the loss of MYI somehow causes a flatlining of the annual minimums? I'd have thought it would be exactly the opposite.

When the slowly vanishing MYI was finally mostly gone in 2012, the minimums should have started to drop ever lower! The difficult-to-melt ice was all gone, only the easy FYI left, and minimums should have plummeted. Instead the annual minimums are (most probably) heading steadily downwards or have (quite possibly) stalled.

It looks as if there is no any causal link between the trends in annual minimums and the loss of MYI.

The explanation and timeline, as I see it, for acceleration and then later deceleration runs as follows:

Just prior to ~1998-2000 the flywheel of Ice making it around beaufort gyre and having a life of around 10 years in which to thicken meant large MYI extent and thickness. Nghiem et al 2007 has perennial ice extent slowly decreasing towards 4m km^2. Once it reached that level, the ice in Chukchi, East Sirberian and Laptev was melted out such that each year so it was only first year ice in those regions each year so they melted out every year. Some MYI might move into Chukchi, but it didn't make it around East Sirberian and Laptev parts of Beaufort gyre.



This created crash in thick MYI volume, area quickly declined to 3m km^2 and volume even more rapidly. As thick MYI reduces in thickness, it does not grow back during the winter if it is already above the equilibrium thickness. FYI is saltier and more easily melts out so the increase in FYI area and lower thickness allows more open water formation so more albedo feedback. This drives a rapid reduction in extent.



As it took ice ~ a decade to complete the beaufort gyre, this process goes on for at least a decade til the MYI reaches a new much lower level equilibrium basically squashed up against Greenland and CAA. The centre doesn't melt out so extra MYI is added there, being more or less balanced by some MYI getting squashed out into Beaufort or Nares or Greendland sea where it melts out.

There is only so much MYI that can be reduced in the rapid process before the reductions in MYI has to slow down. Once the MYI loss slows down, this then slows down the extent loss which was following the MYI losses.

This explains both the acceleration and the deceleration.

[binntho]

Thanks Crandles, I like this one, the rapid loss of MYI starts around 2002 and is finished around 2007 - so perhaps here we have the smoking gun behind Tamino's rapid patch.

[binntho]

It's a pity this one ends in 2015, but never mind. This is volume at maximum, and the amount of ice up to 3.3 meters stays about the same over the entire period.

So one way to understand this is to say that every year, the same amount of ice is generated (the under 3.3) while every summer more and more of the older ice (over 3.3) melts away. It doesn't really tell us much about the minimum or the average trend, but never mind that.

But my point would be that when the over 3.3 meter ice is finally gone, the under 3.3 should start to show some serious losses and the minimum should take a nosedive.

[Tom_Mazanec]

So from 1979 to 2015 thick ice went from 16 to 6 at an accelerating rate. Should be down to down to 4 or 5 by now. We should hit zero in the next decade or so.

[binntho]

The more I think about it, the less I am able to see why there should be a stall in melt, whether it's the minimum we focus on or the annual average (which I believe is still trending downwards - when did I last see a graph showing the annual average? - Gerontocrat has a running 365 day average which is definitely tending down, but annual calendar year average?)

• Loss of MYI should cause a speed up of melt. MYI is "stored" ice from previous years and takes a hell of a lot of energy to melt out. Once it's gone, the FYI should melt all the faster.
• Melt-out of the peripheral seas also has the potential to speed up melt:
  
  • the periphery is easier to melt (being further south) but on the other hand it used to act as a buffer, protecting the center
  • increased open water means increased summer insolation which means faster Arctic warming
  • more wave action, more storminess with winds etc. cause more melting
  • more clouds can delay ice formation in winter, precondition for melt in spring but perhaps work against melt in summer

So all in all I'd expect melting to accelerate and if I'm right that the current reluctance of the minimums to trend downwards is simply a trick of the randoms, then we might well see some drastic losses in the very near future.

On the other hand, as this current year has shown very well, perhaps we'll get ice that disperses so easily that the good old 15% SIE measurements simply can't be nudged downwards even with rapidly increasing melt.

[crandles]

Yes, but that decades ago.  By the turn of the century, it was mostly FYI.  Since then, there was the occasional summer in which some ice survived, but not much.

Not so sure about that. I seem to remember last year we had a tongue of MYI in the ESS - and not that many years ago (at least after I joined in 2016) I remember A-Team and others showing the MYI ridges stretching across the Beaufort and Laptev.

yes an arm continues in existance but not hugely significant?

an older one

[binntho]

yes an arm continues in existance but not hugely significant?

I guess not.

Crandles, are we nudging towards an FFA?

[gerontocrat]

yes an arm continues in existance but not hugely significant?

I guess not.

Crandles, are we nudging towards an FFA?

Whatever you lot are going to agree about,
I have been disagreeing,
I am disagreeing,
I will disagree,
I will have disagreed.

[crandles]

What is FFA??

Below formation zones and exits for MYI. OK I will accept probably slow decline rather than complete equilibrium in MYI.

[binntho]

yes an arm continues in existance but not hugely significant?

I guess not.

Crandles, are we nudging towards an FFA?

Whatever you lot are going to agree about,
I have been disagreeing,
I am disagreeing,
I will disagree,
I will have disagreed.

A courageous stand Gerontocrat! And I'd tend to agree with you (sorry!), or to quote the famous prophet Brutha, "Let a thousand voices sound".

(FFA = Full Forum Agreement, probably less likely than a BOE this year)