Arctic sea ice changes: Natural variation vs human influence

[Michael Hauber]

If you can show me an ensemble mean ( more than 10 to 20+ realizations) of multiple models forced by anthropogenic forcing ( whatever you like to add in the models, Co2, aerosol, land surface use and ozone etc ) that can well capture the observed circulation change in the past 40 years. I think this would be a good evidence that my argument is wrong. If not, we have to say that a portion of observed circulation change is due to a natural source and this part of natural source can melt sea ice through a dynamical impact rather than the greenhouse effect.
 
Please read our papers in 2014 (fig. 4) and 2017(fig.4) . We checked all available models to do this analysis and we couldn't find a similarity.

(I added emphasis to quotation)

As far as I can understand should the statement be:

The circulation change is due to an unknown cause.  Modelling cannot find a link between Co2 and this circulation change, which is evidence that it is not Co2.  But until this circulation change is better understood it is not really known whether it is purely random natural variation, an alternative external forcing such as aerosols etc, or whether there is an important influence of Co2 on the climate that models cannot yet capture.

[oren]

If you can show me an ensemble mean ( more than 10 to 20+ realizations) of multiple models forced by anthropogenic forcing ( whatever you like to add in the models, Co2, aerosol, land surface use and ozone etc ) that can well capture the observed circulation change in the past 40 years. I think this would be a good evidence that my argument is wrong. If not, we have to say that a portion of observed circulation change is due to a natural source and this part of natural source can melt sea ice through a dynamical impact rather than the greenhouse effect.
 
Please read our papers in 2014 (fig. 4) and 2017(fig.4) . We checked all available models to do this analysis and we couldn't find a similarity.

(I added emphasis to quotation)

As far as I can understand should the statement be:

The circulation change is due to an unknown cause.  Modelling cannot find a link between Co2 and this circulation change, which is evidence that it is not Co2.  But until this circulation change is better understood it is not really known whether it is purely random natural variation, an alternative external forcing such as aerosols etc, or whether there is an important influence of Co2 on the climate that models cannot yet capture.

Thanks Michael, good summary of my thinking in better terms than I could articulate.

[ktonine]

ktonine, you misunderstand Exp-5/6.
These experiments do NOT try to determine the effects of AGW.
They try to determine the effect of 'atmospheric circulation' and specifically the effect of Z200GL (geopotential height over Greenland). Ding et al 2017 uses a regression method to find out how much the other variables (including the most important one : temperature) may have been affected by Z200GL. As a result, the method knocks out 2/3rd of the Arctic temperature trend, effectively assigning it to 'atmospheric circulation' as the cause.

Rob one set has the AGW trend (or most of it) removed - the other does not.  No?  Looking at the difference is a comparison or an attempt to judge the influence of AGW.  Characteriizing it otherwise is incorrect.  Yes, they're looking at the influence on the atmospheric pattern because that's what they judge as significant.  Again, you're just chasing down the same rabbit hole.

[AndrewB]

If you can show me an ensemble mean ( more than 10 to 20+ realizations) of multiple models forced by anthropogenic forcing ( whatever you like to add in the models, Co2, aerosol, land surface use and ozone etc ) that can well capture the observed circulation change in the past 40 years. I think this would be a good evidence that my argument is wrong. If not, we have to say that a portion of observed circulation change is due to a natural source and this part of natural source can melt sea ice through a dynamical impact rather than the greenhouse effect.
 
Please read our papers in 2014 (fig. 4) and 2017(fig.4) . We checked all available models to do this analysis and we couldn't find a similarity.

(I added emphasis to quotation)

As far as I can understand should the statement be:

The circulation change is due to an unknown cause.  Modelling cannot find a link between Co2 and this circulation change, which is evidence that it is not Co2.  But until this circulation change is better understood it is not really known whether it is purely random natural variation, an alternative external forcing such as aerosols etc, or whether there is an important influence of Co2 on the climate that models cannot yet capture.

Hi Michael, thanks for your comment, however it slightly distorts what Dr Ding wrote. Note that Dr Ding does mention various anthropogenic forcings: "CO2, aerosol, land surface use and ozone etc", not just CO2.

If you can show me an ensemble mean ( more than 10 to 20+ realizations) of multiple models forced by anthropogenic forcing ( whatever you like to add in the models, Co2, aerosol, land surface use and ozone etc ) that can well capture the observed circulation change in the past 40 years. I think this would be a good evidence that my argument is wrong.

(emphasis mine)

But this amounts to blaming the CMIP5 model ensemble for not correctly capturing the observed circulation change over the Arctic, something they have not been specifically designed for.
Again, don't blame the models, they are just tools.

And again, there are no "good" or "bad", "right" or "wrong" models. Models can be useful to evidence mechanisms or other phenomena in the chaotic global climate system, and they are an essential tool not only in climate science but in many other areas of human knowledge.
So (imo) it is not a question of finding "better" models here.

Last but not least, I would like to thank Dr. Ding, for taking note of the second part of my question no.1 (*) in one of my previous comments, and for continuing to engage in an open and vigorous discussion of his paper here on ASIF.

* (the full question was:
"1. In the abstract of your 2017 paper, one can read: "Internal variability dominates the Arctic summer circulation trend and may be responsible for about 30-50% of the overall decline in September sea ice since 1979." My question (in two parts) is: how confident are you and your co-authors about these numbers, and what scientific experiment could be made, or physical evidence would be required, if any, to invalidate this claim?")

Dr Ding's most interesting answer is food for thought, and I'll probably spend a few many hours this weekend thinking about it.

[Cid_Yama]

But this amounts to blaming the CMIP5 model ensemble for not correctly capturing the observed circulation change over the Arctic, something they have not been specifically designed for.
Again, don't blame the models, they are just tools.

+1  Although my complaint is the assigning of 'natural variability' to unaccounted for melting.  We know that atmospheric circulation and composition only account for a portion of the melting and that ocean circulation, i.e. warm water incursions, and general warming, as well as ocean (and ice) atmospheric interactions have a major role.

These are all influenced by anthropogenic factors.  And also global circulation changes, as a result of atmospheric chemistry changes and warming.

We are dealing with a badly designed research paradigm.  One that provides no useful information and makes claims not supported but merely conjectured.

   

[AndrewB]

...
We are dealing with a badly designed research paradigm.  One that provides no useful information and makes claims not supported but merely conjectured.

(emphasis mine)
I couldn't agree more.

[Neven]

So, if you would want to tease out the contributions of AGW and natural variation to Arctic sea ice loss, what would your "designed research paradigm" be? How would you go about it?

[Cid_Yama]

Why would I want to do that and what does that mean?  It is clear that human activities have caused both atmospheric chemical and circulation changes that threaten our very existence.

What IS natural variation after a century and a half of our dramatically changing the chemistry of the atmosphere?

Your question, to me seems meaningless.  There is no longer anything natural about what's going on.

Paintings from the 16th to the 19th century show a sky that was robin egg blue.  It's not anymore.

CO2 levels are beyond anything we have experienced throughout our evolution and may already be doing physiological harm.

What would I be parsing out and why?  Would it provide us with anything useful?

All this paper shows is unaccounted for melting in the various climate models.  Then provides unwarranted conclusions.

We already know a slew of factors affecting the ice that is not included in the climate models.

Explain to me how this paper provides me with any information of value.   

[Neven]

You're mixing up two issues here, namely the science and the societal debate. We've talked about how this research had a high likelihood of being spun, and how that perhaps could've been spun prevented (edit N.). And we've talked about the science. I understand that people don't like how it was spun (neither do I) and that they then try to find issue with the science. That's all understandable, and I'm not even taking issue with that.

But the way you criticize the research, implies that you know how it could've been done better. If not, you are saying that the research shouldn't have been done in the first place. But don't we want to know the respective contributions of AGW and natural variation to Arctic sea ice loss? That's a legitimate and interesting scientific question, right?

[DrTskoul]

...
We are dealing with a badly designed research paradigm.  One that provides no useful information and makes claims not supported but merely conjectured.

(emphasis mine)
I couldn't agree more.

Some times science moves forward by posing mere conjectures, for others to prove or disprove....research is not always clean and perfect and accurate....

[FredBear]

I feared the word "Natural" in the Dr Ding paper would become a focus for BAU merchants selling carbon for energy but has this happened? It may be a good attempt find features that are causing ice loss but sadly the "N" word in this particular paper has generated much more heat than light in this forum. There may be factors that can now be watched more closely  .  .   .

I have felt like:-

as arguments beyond me whistled by!

Surely the arguments are going to become more & more academic as the  40%+ of anthropogenic factors increasingly tilts the balance against the ice? The coming melt season may well prove to be more interesting  .   .   .   (and provide more useful data?)

[crandles]

You're mixing up two issues here, namely the science and the societal debate. We've talked about how this research had a high likelihood of being spun, and how that perhaps could've been spun. And we've talked about the science. I understand that people don't like how it was spun (neither do I) and that they then try to find issue with the science. That's all understandable, and I'm not even taking issue with that.

But the way you criticize the research, implies that you know how it could've been done better. If not, ou are saying that the research shouldn't have been done in the first place. But don't we want to know the respective contributions of AGW and natural variation to Arctic sea ice loss? That's a legitimate and interesting scientific question, right?

+1

[Jim Williams]

But the way you criticize the research, implies that you know how it could've been done better. If not, you are saying that the research shouldn't have been done in the first place. But don't we want to know the respective contributions of AGW and natural variation to Arctic sea ice loss? That's a legitimate and interesting scientific question, right?

I can think of better ways to spend our soon to be very limited research dollars.  More satellites would be nice.

[AndrewB]

Why would I want to do that and what does that mean?  It is clear that human activities have caused both atmospheric chemical and circulation changes that threaten our very existence.

What IS natural variation after a century and a half of our dramatically changing the chemistry of the atmosphere?

Your question, to me seems meaningless.  There is no longer anything natural about what's going on.

Paintings from the 16th to the 19th century show a sky that was robin egg blue.  It's not anymore.

CO2 levels are beyond anything we have experienced throughout our evolution and may already be doing physiological harm.

What would I be parsing out and why?  Would it provide us with anything useful?

All this paper shows is unaccounted for melting in the various climate models.  Then provides unwarranted conclusions.

We already know a slew of factors affecting the ice that is not included in the climate models.

Explain to me how this paper provides me with any information of value.

Best post in this entire thread.

Richard hesitated and looked at Rieux:

- Sincerely, tell me your thoughts, do you have the certainty that this is the plague?

- You are not asking the right question. It's not a matter of vocabulary, it's a matter of time.

"Your opinion on the matter," said the Prefect, "would be that, even if it were not plague, the prophylactic measures indicated in cases of plague should, however, be applied.

- If it is absolutely necessary that I have an opinion on the matter, it is indeed this one.

The doctors consulted and Richard finally said,

- So we have to take responsibility for acting as if the disease was the plague.

The formula was warmly approved:

"Is that your opinion, my dear colleague?" asked Richard.

"The wording is indifferent to me," said Rieux. "Let's just say that we should not act as if half the city would not be killed because then it would be."

Albert Camus, The Plague. (1947)

[Neven]

That's a great novel by Camus, but the analogy is false. No one is saying there isn't a plague.

Does anyone wonder why climate risk deniers got so little traction out of this? Because they indirectly admit there is a plague.  That's why Arctic sea ice is such an incredible PITA for them. It's the whole reason the ASIB and then this Forum was set up. So, not too much of a worry there.

Which brings us safely back to the science. First of all, no one has ever claimed that all of Arctic sea ice loss was because of AGW (that's the strawman climate risk deniers use). For scientists and those in the know, it's clear that there may be a natural component causing part of the loss. If there is a natural component, I think it would be extremely useful to know about it. We're not finding out if no one is researching it.

Secondly, it's not just that we want to know this, so we can pinpoint when the Arctic will go ice-free. This is irrelevant, as it is almost certain that is going to happen in 0-15 years from now. We need to increase our understanding of how it all works, and what the contributions from AGW and natural variation (if any) are, if we want to have a good understanding while monitoring the comeback of Arctic sea ice. Either because of some geoengineering, or because we manage to quickly reduce CO2 emissions and then take it out of the atmosphere.

So, attack the science because there are flaws. That's how science advances.

But don't attack the science because you perceive it as some social threat. That energy is better spent elsewhere. In your garden, for instance. And in efforts to change the system.

[jai mitchell]

So, if you would want to tease out the contributions of AGW and natural variation to Arctic sea ice loss, what would your "designed research paradigm" be? How would you go about it?

why would you intentionally attempt to assign a value of 'natural variability' to a dynamic system that is in the process of catastrophic collapse?  Is it because the collapse is not what you expected?

This is hubris. 

The aerosol cooling component in the Arctic is extremely well documented, as is the poor representation by current models of the full effects of aerosols on both global scales and in the Arctic.  With the clear understanding of the CMIP5 ensemble understating these impacts to the Arctic.  We should expect a more rapid loss of sea ice than the models predict.

http://www.nature.com/ngeo/journal/v9/n4/full/ngeo2673.html

Amplification of Arctic warming by past air pollution reductions in Europe

Here we present simulations with an Earth system model with comprehensive aerosol physics and chemistry that show that the sulfate aerosol reductions in Europe since 1980 can potentially explain a significant fraction of Arctic warming over that period. Specifically, the Arctic region receives an additional 0.3 W m−2 of energy, and warms by 0.5 °C on annual average in simulations with declining European sulfur

[DrTskoul]

So, if you would want to tease out the contributions of AGW and natural variation to Arctic sea ice loss, what would your "designed research paradigm" be? How would you go about it?

why would you intentionally attempt to assign a value of 'natural variability' to a dynamic system that is in the process of catastrophic collapse?  Is it because the collapse is not what you expected?

The atmospheric and ocean dynamic system is in no threat of collapse. Biosphere as we know it is. Understanding the internal variability (sans forcing) of the system, helps you improve your understanding to its detailed response to forcings. Then you can more accurately model scenarios and responses. And if it comes to geoengineering, I would like to be able to predict what it might do as accurately as possible.

[AndrewB]

... if we want to have a good understanding while monitoring the comeback of Arctic sea ice. ...

You are joking, right?
We are already committed to the complete and irreversible (on a human time scale) disappearance year round of Arctic sea ice.

[Neven]

... if we want to have a good understanding while monitoring the comeback of Arctic sea ice. ...

You are joking, right?
We are already committed to the complete and irreversible (on a human time scale) disappearance year round of Arctic sea ice.

That kind of makes me want to shut down the blog and forum, and build a big bunker. And so the narrative I tell myself (and hopefully others once I get back to blogging) is that what we need to fight for, is getting Arctic sea ice back after we lose it. Preferably prevent it from going ice-free all year round.

why would you intentionally attempt to assign a value of 'natural variability' to a dynamic system that is in the process of catastrophic collapse?  Is it because the collapse is not what you expected?

If it is all collapsing, there's not much sense in thinking about what the best way to do research is. Or post on a forum. And what I've said in reply to AndrewB.

[AndrewB]

But the way you criticize the research, implies that you know how it could've been done better. If not, you are saying that the research shouldn't have been done in the first place. But don't we want to know the respective contributions of AGW and natural variation to Arctic sea ice loss? That's a legitimate and interesting scientific question, right?

I can think of better ways to spend our soon to be very limited research dollars.  More satellites would be nice.

Satellites are expensive. Modelturbation is cheap.

[AndrewB]

That kind of makes me want to shut down the blog and forum, and build a big bunker. And so the narrative I tell myself (and hopefully others once I get back to blogging) is that what we need to fight for, is getting Arctic sea ice back after we lose it. Preferably prevent it from going ice-free all year round.

Neven,
I would suggest then you re-read The Plague, just to see how various characters react to the situation.

We can't "fight [to get] the Arctic sea ice back". There is no technology to do so, and the scale of the problem vastly exceeds the capabilities of what humankind can achieve, even with a global wartime effort.
What we can do is reduce GHG emissions as fast as possible, and to do so we need to push for public policies that reduce our reliance on fossil fuels. There is no extinction threat here, but for sure hundreds of millions of human beings will suffer and possibly tens of millions will die, if we don't act as quickly as possible. It is a question of time, as Docteur Rieux says in The Plague.

Now, the question of what exact percentage (if any) of atmospheric circulation changes over the Arctic in June July August over the period 1979-2014 can be attributed to "natural climate variability" is purely academic, and at this point in the unfolding global warming catastrophe, a distraction, and a waste of time.
Similarly, calling the disappearance of 75% of sea ice volume a "collapse" or "sea ice changes due to natural variability intrinsic in atmospheric circulation" is just a question of words, and changes absolutely nothing to the reality of what is happening.

[nukefix]

Satellites are expensive. Modelturbation is cheap.

All models are wrong, but some are useful. Meanwhile satellites...are needed to constrain the models. And even with well constrained models extrapolations into the future are problematic.

[jdallen]

... if we want to have a good understanding while monitoring the comeback of Arctic sea ice. ...

You are joking, right?
We are already committed to the complete and irreversible (on a human time scale) disappearance year round of Arctic sea ice.

That kind of makes me want to shut down the blog and forum, and build a big bunker. And so the narrative I tell myself (and hopefully others once I get back to blogging) is that what we need to fight for, is getting Arctic sea ice back after we lose it. Preferably prevent it from going ice-free all year round.

why would you intentionally attempt to assign a value of 'natural variability' to a dynamic system that is in the process of catastrophic collapse?  Is it because the collapse is not what you expected?

If it is all collapsing, there's not much sense in thinking about what the best way to do research is. Or post on a forum. And what I've said in reply to AndrewB.

Bunkers are attractive but in the long run ineffective.

I appreciate greatly Dr. Ding jumping into our discussion, even as I struggle to fully grasp the nuances of argument. His doing so is the quintessence of good scholarship.

While I generally understand his teams methodology, which in context of process seems sound, I am wrestling with the label "natural variation" to describe "unforced" departure from average.

That's a big bucket, and I'm wondering how much of that is anthropomorphic change outside of the Arctic translating into it. Perhaps, Dr. Ding, you could address that thought?

Last thought to Neven - even if we cannot recover the Arctic as it was,  we still must consider the potential loss of billions of lives due to climate change. That deserves our continued discussion.

[Lennart van der Linde]

If natural variability does play such a large role as Ding et al estimate, would it be possible to detect comparably large sea ice losses in the past? Or do we not have good enough data and/or models for such detection?

[jai mitchell]

sorry I was unclear, in talking about collapsing systems, I was specifically talking about summer sea ice.

[Random_Weather]

If natural variability does play such a large role as Ding et al estimate, would it be possible to detect comparably large sea ice losses in the past? Or do we not have good enough data and/or models for such detection?

I dont think its not that easy, as i make clear in this thread, i have doubts about sperate nat. variability and forcing in this way (because in my opinion the choose of models to use must have a near real world cyrosphere state to make conclusion confident), on the other hand, in real climate, external forcing and internal variability interact with each other. What i want to say is, that thinner ice also could amplifier weather effects or internal inter to decadel variability.

Therefore, in past if ice was thick enough, the impact on extent could be less effective as now, we can see in observation some evidence if we looking at the time and the first Sigma on a running 5y

Here 5y running Sigma1 in Exent for September 1979-2016
http://www.directupload.net/file/d/4690/tk4qsvjl_png.htm

We can also watch the weather Pattern, here i use the 500mb geopot for Summer Saison (JJA)
http://www.directupload.net/file/d/4690/cxwxrpya_png.htm

So well we see that variability in 500mb is not increasing(but correlates) but the variability of extent and if we take them together, we see near the end, there is a disconnect also to see is, that ice is probably leads 500mb geopot untill the disconnect
left: 5y running Sigma 1 of 500mb gepot for JJA on 70-90N)
right: 5y running Sigma 1 of sea ice extent in September

http://www.directupload.net/file/d/4690/gu76bgh3_png.htm

Under the line:
The very high Sigma1 of the very most past years can not be explain to variability of the weather, so as some Paper has claim it seem likly that thinner ice acting like a amplifier. So long text, but come to a conclusion to your question, i think it have not to be detect in the past.


greets

[AndrewB]

If natural variability does play such a large role as Ding et al estimate, would it be possible to detect comparably large sea ice losses in the past? Or do we not have good enough data and/or models for such detection?

Datasets for sea ice available here: http://nsidc.org/data/search/#keywords=sea+ice/sortKeys=score,,desc/facetFilters=%257B%257D/pageNumber=1/itemsPerPage=25

I think the earliest data from historical records goes back to 1850. But I doubt very much there is any record, historical or reconstructed from sediments or whatever, of any Arctic sea ice loss comparable to what we have experienced in the last 40 years or so (75% loss in September sea ice volume).

The Eskimo people have lived around the Arctic for the last 3,000 years or so, and their dependence on sea ice is such that we can predict their culture will disappear together with Arctic sea ice, before the end of this century.

I also mentioned a recent paper about the Barnes Ice Cap. It has survived > 2,000 years (ergo, > 2,000 summers) with practically a constant average volume. But now it's bound to disappear within decades. If anything, that is physical evidence that the present warming of the Arctic is totally unprecedented.

[Lennart van der Linde]

How about this paper Piron & Pasalodos 2016 (Neven has blogged about it):
https://diablobanquisa.wordpress.com/2016/01/14/new-time-series-september-arctic-sea-ice-extent-1935-2014/

Their fig5 below shows an average year to year variation of a little over 0.5 million km2 (eyeballing their red line) from the decadal mean/trends from 1935-2014. On decadal scale the variation could be about 1 million km2 from the mean over about 70 years of about 7 million km2. Total sea ice extent loss is about 3 million km2 over 1935-2014, assuming 7 million km2 as the baseline. So if 1 millon km2 of that loss would be due to natural variability, then 2 million km2 should be due to global warming. So that would be about 33%. But is 7 million km2 the correct baseline and 70 years the right period? And have cooling aerosols maybe prevented even more melting? And how about volume loss over that period?

I suppose the period should be at least 90 years since the biggest loss happened from 2000-2014, so the data should be extended at least to 1925 to estimate the right baseline, right? And how about natural variability over centuries?

[Lennart van der Linde]

Also have a look at fig 8 of Walsh et al 2016:
http://onlinelibrary.wiley.com/doi/10.1111/j.1931-0846.2016.12195.x/full

It shows maybe 0.5 million km2 decadal variability from 1850-1925 and about 1 million km2 from 1925-2000. So over 150 years the variability is limited and then ice loss grows fast in the last 10-15 years. If the natural ccyle should have gone up then all of the loss should be due to global warming. If part of that loss is due to natural variability, then the cycle should be even longer than 150 years, so we should look back even further. Ding et al didn't do that, so what can they say about natural variability? or maybe I'm missing something?

The natural long term global (and Arctic?) temperature trend was cooling, as far as I know, so sea ice extent should have been increasing, if not for global warming, I would think.

[StopTheApocalypse]

It seems to me that the assumption is that if there is not already a well understood mechanism causing demonstrating that SIE loss is caused by AGW, it's assumed to be internal (natural?) variability. Is that more or less correct? If so, it's just another example of scientists being conservative. This is fine, but should be emphasized by the media when being reported.

[Lennart van der Linde]

Or we can look at Polyak et al 2010:
https://www.geo.umass.edu/faculty/jbg/Pubs/Polyak%20etal%20seaice%20QSR10%20inpress.pdf

See their fig 2a and fig 12 below.

Or fig 3a in Kinnard et al 2011:
http://projects.iq.harvard.edu/files/climate/files/kinnardetal2011.pdf

That last one shows a variability over many centuries of maybe 0.5-1 million km2 around a slightly growing trend a little above 10 million km2. And then a steep drop since 1950 or so of maybe 5 million km2 by now. So at least 80% should be due to AGW, I would guess, if not more or all of it.

Of course we don't know how reliable these data are, but neither do we know this about the models until they can be tested against more reliable data. So this seems the best we have so far. Or not?

[magnamentis]

... if we want to have a good understanding while monitoring the comeback of Arctic sea ice. ...

You are joking, right?
We are already committed to the complete and irreversible (on a human time scale) disappearance year round of Arctic sea ice.

That kind of makes me want to shut down the blog and forum, and build a big bunker. And so the narrative I tell myself (and hopefully others once I get back to blogging) is that what we need to fight for, is getting Arctic sea ice back after we lose it. Preferably prevent it from going ice-free all year round.

why would you intentionally attempt to assign a value of 'natural variability' to a dynamic system that is in the process of catastrophic collapse?  Is it because the collapse is not what you expected?

If it is all collapsing, there's not much sense in thinking about what the best way to do research is. Or post on a forum. And what I've said in reply to AndrewB.

neven in absoluter hochform ( neven in high performance mode LOL ) this thread made my day, keep going.

[Neven]

I also mentioned a recent paper about the Barnes Ice Cap. It has survived > 2,000 years (ergo, > 2,000 summers) with practically a constant average volume. But now it's bound to disappear within decades. If anything, that is physical evidence that the present warming of the Arctic is totally unprecedented.

And don't forget the Ellesmere Island ice shelves (hat-tip to Judith Curry).

[Archimid]

Figure 2a. wow thanks Lennart van der Linde I've never seen that before. That paper looks like a must read.

[Rob Dekker]

ktonine, you misunderstand Exp-5/6.
These experiments do NOT try to determine the effects of AGW.
They try to determine the effect of 'atmospheric circulation' and specifically the effect of Z200GL (geopotential height over Greenland). Ding et al 2017 uses a regression method to find out how much the other variables (including the most important one : temperature) may have been affected by Z200GL. As a result, the method knocks out 2/3rd of the Arctic temperature trend, effectively assigning it to 'atmospheric circulation' as the cause.

Rob one set has the AGW trend (or most of it) removed - the other does not.  No?

No, ktonine.
You seem to be pretty hard to convince, but take a look at your own quote from Ding et al 2017 :

Given a strong correlation between circulation and surface winds, temperature, specific humidity, sea-level pressure, and downwelling long wave radiation in the Arctic,variability and trends in these six variables that are associated with Z200GL are processed and removed from the forcing.

So they teased out the influence of Z200GL, and NOT "the AGW trend".

If you still believe that Exp.5/6 were about AGW trend, please state it as a question. Qinghua Ding mentioned that he has some time this weekend, and he may be able to answer you better than I did.

[Rob Dekker]

So, if you would want to tease out the contributions of AGW and natural variation to Arctic sea ice loss, what would your "designed research paradigm" be? How would you go about it?

I love it. Thank you Neven.

In general, it is very hard to tease out the contributions of AGW and natural variability to Arctic sea ice loss. As my teacher told me a long time ago : If it were easy, I would have done it myself.

However, Ding et al 2017 DOES provide a tool that we can use to 'tease out' particular variables out of a set of correlated variables.
They did it with this formula, where they 'teased out' the influence of Z200GL (the geopotential height over Greenland) :

B(x, y, t) = β(x, y) × Z200 GL (t)

Here, note that Z200 GL (t) is a scalar time-dependent variable.
Instead of Z200 GL (t) we may be able to 'tease out' the AGW influence using this very same method, but instead of Z200 GL we can insert another scalar time-dependent variable, like the global warming variable LOTI :
https://data.giss.nasa.gov/gistemp/graphs/

We would just need to calculate :

B(x, y, t) = β(x, y) × LOTI (t)

and of course, need to re-calculate β(x, y) for each grid point and each variable against LOTI.
Fairly trivial to do if you have the data set up.

After that, we could re-run Exp.6. and see how sea ice is affected (by LOTI).

What we would have done in that case, would be to 'tease out' AGW out of the Arctic variables, and that would be a start (to see what AGW influence is on Arctic sea ice).

Of course, this method assumes that there is no "Arctic amplification" during the summer months.
But it would be a start...

[AndrewB]

So, if you would want to tease out the contributions of AGW and natural variation to Arctic sea ice loss, what would your "designed research paradigm" be? How would you go about it?

...
However, Ding et al 2017 DOES provide a tool that we can use to 'tease out' particular variables out of a set of correlated variables.
...

After that, we could re-run Exp.6. and see how sea ice is affected (by LOTI).

What we would have done in that case, would be to 'tease out' AGW out of the Arctic variables, and that would be a start (to see what AGW influence is on Arctic sea ice).
...

Hi Rob,
The problem is, you are using the same method as Ding et al 2017, with the same uncertainties associated with the model used (POP2+CICE4) and with similar uncertainties with respect to the data and the basic assumptions.
In the end, you would almost certainly be able to make a similar claim as the one that is found in the Ding et al 2017 paper, which if we really account for the uncertainties in the entire "experiment", would read like:
“Global warming may be responsible (with 95% confidence) for about -40 to +120% of the overall decline in September sea ice since 1979.”

Totally meaningless...

A methodology is inherently flawed when the result can mean something and exactly its opposite, which is the case with the methodology used in the Ding et al 2017 paper.

There is a much "cleaner" method to determine the influence of cumulative CO2 emissions on average Arctic sea ice volume, it's the simple, mathematically correct and obvious linear correlation found in the Notz and Stroeve 2016 paper. Which attributes 100% of the decline in Arctic sea ice volume to cumulative CO2 emissions, with a simple causality chain, and with high statistical confidence.

Ding et al 2017 went around with a very complicated methodology that hides its inherent flaws because they wanted to reach a predetermined conclusion, extremely similar to the conclusion they had already reached in the Ding et al 2014 paper, about natural variability.

But all this is absolutely irrelevant and a waste of time, and time now is of the essence. What really matters these days is what each climate science paper brings to the table for policy makers:
- The Notz and Stroeve 2016 paper clearly underlines the imperative to reduce CO2 emissions from the burning of fossil fuels as quickly as possible. Dr. Notz makes exactly the same point in his COP21 (2015) presentation (see the YouTube video I linked to in one of my previous comments).
- The Ding et al 2017 paper is useless for policy makers concerned with AGW. That is the essence of what Cid_Yama wrote in what I consider is (by far) the best comment in this entire thread.

[AndrewB]

Satellites are expensive. Modelturbation is cheap.

All models are wrong, but some are useful. Meanwhile satellites...are needed to constrain the models. And even with well constrained models extrapolations into the future are problematic.

Sorry, my phrase was a poor attempt at humor (a "boutade" in French). Satellites are indeed expensive but they are essential to provide us with "hard data" to advance climate science (see for example the Jason 1, 2 and 3 series of satellites that measure SLR with high precision). As for climate models, they are indeed (relatively) inexpensive to develop and use, and they can be extremely useful as tools for climate scientists.

But that is a discussion for another thread.

[Lennart van der Linde]

I was just reading this:
https://insideclimatenews.org/news/14032017/arctic-ice-melt-climate-change-science

Ding said his findings in no way minimize the role of anthropogenic warming in melting sea ice. "I don't want people to take the wrong message in our study, that we're not to blame for Arctic warming," said Ding. "The message is that it's more complex than we expected. In the long term, maybe 100 years from now, the Arctic will become ice free in summer because eventually this internal variability will be overwhelmed by anthropogenic forcing."

Dr. Ding, if you're reading this: were you quoted correctly, and if so, what is your assertion based on, or how should it be interpreted?

As far as I know the Arctic could become ice free at the end of summer much earlier than 100 years from now, and ice free during summer probably (or maybe) also quite a bit earlier than 100 years from now.

So what do you mean by "ice free in summer"? Three months of ice free Arctic from June 21st - Sept 21st? Or is one day without ice in those months also ice free in summer?

And what do you mean by "maybe": would that be your best estimate (so should we read "probably"? Or do you mean " maybe as early as 100 years from now"? And does that imply say a 33% chance, or 25%, or 17% or 10%, or even 5%?

In short, what message exactly do you want to give us, the public?

[gerontocrat]

Is the phrase "natural variation" itself a cop-out for a blog such as ASIF ? As an over-curious kid I am sure I asked my parents "Why are summers warm and winters cold ?", and did not accept the answer - "Natural variation, my dear". To learn for the first time about the tilt in the axis of the earth, and from that so much else,  was amazing.

So the question I never asked this thread but hoped for the answer is "What is this natural variation of which Dr Ding speaks?". Or should I be posting this in Stupid Questions ?

[Lennart van der Linde]

Kaufman et al 2009 showed that the Arctic was cooling over the past 2000 years, until global warming came along:
https://www.skepticalscience.com/past-Arctic-sea-ice-extent.htm

The figure below shows the 10-year running mean and seems to be adapted from and based on figure 3c of the paper itself:
https://www.geo.umass.edu/climate/papers2/Kaufman2009a.pdf

[Lennart van der Linde]

I'm also wondering what a thinning trend means for variation in extent: the thinner the ice gets because of AGW, the larger the potential year-to-year variations in extent, I would guess. So the effects of natural variability could get multiplied by AGW? And how should we seperate those factors from each other?

Also this could imply crossing a treshold/tipping point beyond which the ice cannot easily recover from a large decrease due to natural variability, but this has not been convincingly argued/shown (yet?), if I remember correctly.

[DrTskoul]

Is the phrase "natural variation" itself a cop-out for a blog such as ASIF ? As an over-curious kid I am sure I asked my parents "Why are summers warm and winters cold ?", and did not accept the answer - "Natural variation, my dear". To learn for the first time about the tilt in the axis of the earth, and from that so much else,  was amazing.

So the question I never asked this thread but hoped for the answer is "What is this natural variation of which Dr Ding speaks?". Or should I be posting this in Stupid Questions ?

Simple, natural variation is the internal variability of the Earth system ( mostly atmospheric and ocean dynamics ) absent anthropogenic and feedback greenhouse gas emissions....

It includes the Sun, the moon, volcanoes, orbital parameters, etc. which result into patterns and cycles and variability of the trajectory of the atmospheric, ocean and biosphere dynamics, and albedo.

[AndrewB]

I was just reading this:
https://insideclimatenews.org/news/14032017/arctic-ice-melt-climate-change-science

Ding said his findings in no way minimize the role of anthropogenic warming in melting sea ice. "I don't want people to take the wrong message in our study, that we're not to blame for Arctic warming," said Ding. "The message is that it's more complex than we expected. In the long term, maybe 100 years from now, the Arctic will become ice free in summer because eventually this internal variability will be overwhelmed by anthropogenic forcing."

Dr. Ding, if you're reading this: were you quoted correctly, and if so, what is your assertion based on, or how should it be interpreted?

As far as I know the Arctic could become ice free at the end of summer much earlier than 100 years from now, and ice free during summer probably (or maybe) also quite a bit earlier than 100 years from now.

So what do you mean by "ice free in summer"? Three months of ice free Arctic from June 21st - Sept 21st? Or is one day without ice in those months also ice free in summer?

And what do you mean by "maybe": would that be your best estimate (so should we read "probably"? Or do you mean " maybe as early as 100 years from now"? And does that imply say a 33% chance, or 25%, or 17% or 10%, or even 5%?

In short, what message exactly do you want to give us, the public?

Despite Dr. Ding's statement above, the Arctic will be essentially sea ice free at its September minimum probably at some time between 2020 and 2024; and continuously essentially sea ice free for the three months of August, September and October, probably at some time before 2035. And essentially continuously sea ice free year round probably at some time before 2100.

That's what the PIOMAS exponential trendlines in the charts below show.

What I would really like to see Dr. Ding stating is that we need to radically reduce GHG emissions from the burning of fossil fuels ASAP. Then as far as I am concerned he could go back to writing his "natural variability" attribution papers based on model experiments and reanalysis data, and he would never hear one word of criticism from me again. 

[Random_Weather]

AndrewB,

Wow, curve fitting at best, why not using physical stuff, ah would change mind a little bit.

[AndrewB]

AndrewB,

Wow, curve fitting at best, why not using physical stuff, ah would change mind a little bit.

The extremely well designed charts (and exponential trendlines) are from Wipneus, to whom we should all be grateful for making them available and periodically updating them, on his website.

https://sites.google.com/site/arctischepinguin/home

[DrTskoul]

AndrewB,

Wow, curve fitting at best, why not using physical stuff, ah would change mind a little bit.

Eh...hmmm... what "physical" "stuff"??

[Neven]

I was just reading this:
https://insideclimatenews.org/news/14032017/arctic-ice-melt-climate-change-science

Ding said his findings in no way minimize the role of anthropogenic warming in melting sea ice. "I don't want people to take the wrong message in our study, that we're not to blame for Arctic warming," said Ding. "The message is that it's more complex than we expected. In the long term, maybe 100 years from now, the Arctic will become ice free in summer because eventually this internal variability will be overwhelmed by anthropogenic forcing."

Dr. Ding, if you're reading this: were you quoted correctly, and if so, what is your assertion based on, or how should it be interpreted?

As far as I know the Arctic could become ice free at the end of summer much earlier than 100 years from now, and ice free during summer probably (or maybe) also quite a bit earlier than 100 years from now.

So what do you mean by "ice free in summer"? Three months of ice free Arctic from June 21st - Sept 21st? Or is one day without ice in those months also ice free in summer?

And what do you mean by "maybe": would that be your best estimate (so should we read "probably"? Or do you mean " maybe as early as 100 years from now"? And does that imply say a 33% chance, or 25%, or 17% or 10%, or even 5%?

In short, what message exactly do you want to give us, the public?

This is a good and simple question, much better than the rehash of points already made.

[Random_Weather]

Eh...hmmm... what "physical" "stuff"??

There a lot of variables you can use as a proxie for surface melting, ocean melting and so on, means fitting on physical base

[DrTskoul]

Eh...hmmm... what "physical" "stuff"??

There a lot of variables you can use as a proxie for surface melting, ocean melting and so on, means fitting on physical base

That is the beauty of simple elegant fits...The Gompetz equation was developed to model the death rate of Humans. Ironic that so far it fits the death rate of ice...