Also known as Qinghua Ding et al., Nature Climate Change (2017) doi:10.1038/nclimate3241; published online 13 March 2017 (PDF as submitted). And the abstract:
The Arctic has seen rapid sea-ice decline in the past three decades, whilst warming at about twice the global average rate. Yet the relationship between Arctic warming and sea-ice loss is not well understood. Here, we present evidence that trends in summertime atmospheric circulation may have contributed as much as 60% to the September sea-ice extent decline since 1979. A tendency towards a stronger anticyclonic circulation over Greenland and the Arctic Ocean with a barotropic structure blah blah you're not reading this bit are you in the troposphere increased the downwelling longwave radiation above the ice by warming and moistening the lower troposphere. Model experiments, with reanalysis data constraining atmospheric circulation, replicate the observed thermodynamic response and indicate that the near-surface changes are dominated by circulation changes rather than feedbacks from the changing sea-ice cover. 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.
In case you're completely asleep, I've bolded the bit of general interest. Can it really be true that our beloved Arctic sea ice decline is no more than natural variability? Well of course they aren't saying that; but they are saying that a substantial fraction of it might be. Are we really going to sit still for that? Of course not, I'm going to quibble, but don't get too excited; this isn't a Curry paper, it isn't fatally flawed. Should you believe it? If you're a septic no you certainly shouldn't, because this paper is totally reliant on models for it's conclusions, and we all know that a good septic doesn't trust models, only hard data. Carbon Brief has a go at spinning it as Humans causing up to two-thirds of Arctic summer sea ice loss, study confirms but I'm not sure that's entirely satisfactory.
But enough of that. What does it say? Weeell an awful lot of it is about the connection between September sea-ice extent and the preceding summer (June–July–August, JJA) atmospheric circulation. We choose thisp receding 3-month window because sea-ice extent anomalies have a ∼3-month decorrelation timescale and the reference for that is Blanchard-Wrigglesworth et al., one of the authors. Again, without being able to pick any obvious holes I feel somewhat uncomfortable with that; the idea that September ice depends just on JJA circulation doesn't feel at all right. Having decided that, though, they then run a variety of model experiments, for example "nudging" the circulation back to re-analysis, with and without an ocean-ice model underneath. And the result seems to be that it is mostly the circulation forcing the sea ice, rather than the sea ice changes forcing the atmosphere. This kinda-fits the "information flow" meme from way back so I should be prepared to accept that mostly. having done that they then convince themselves that most of the circulation changes that matter to the ice are not GW forced, and so must be natural variability; and hence the conclusion. If you took all of this at face value then they'd have solved one of the puzzles, that on the whole models show much less ice decline that reality. But of course if the decline is substantially a freak of variation, not forced, that would fit.
The flaw in this overall, without looking at the details, is that it's hard to see a near-40-year trend and being so much natural variability. That seems to be asking for an awful lot of one-way variation.
- Log in to post comments
Seems to be a study that misses the debate over connection between arctic amplification and global weather circulation patterns.
http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-16-0762.1
Attributing to the dog the direction of the walkers stroll .
The usual suspect will focus on this one study and ignore the implications of all others .
Qinghua Din lead author, has signed off on other papers with Willie "oilcan" Song, collector of 1+ million dollars from oil, energy & business PR propaganda poopers. "wuwt" loves G.D., who has devoted many of his years to finding "natural variation" warming constructs. Hopefully, Willie shared more than a few bucks with his co-hort.
[I think you mean, Soon, not Song. And Ding not Din. But you're right, he does seem to have "form" with WS, e.g. http://www.soest.hawaii.edu/MET/Faculty/bwang/bw/paper/177_Jian_Liu_et_…. That is interesting -W]
"blah blah you’re not reading this bit are you"
Why, yes.
...but I'm not a meteorologist so much of it does end up sounding like blah blah. OTOH not being a skeptical type I don't automatically assume anything I don't understand is wrong.
[I just thought I'd throw that bit in to test you :-) -W]
Obvious question: Are the changes in atmospheric circulation themselves natural or a consequence of warming? Otherwise we are just seeing enhanced heat transport to the poles melting more ice, which is expected.
[They do try to answer that. I'm afraid I slightly glazed over that part of the analysis, which I suspect is somewhat flawed, but they do <something> which convinces them that the circulation changes are largely just variability -W]
"blah blah you’re not reading this bit are you"
Of course I am, but you're using my name out of context...
Found it by searching on 'natural'. Apparently this doesn't happen in the models, but it's a result of pacific tropical variability.
It's possible that there is a PDO / Arctic sea ice connection, and we just haven't noticed it before because all it did was reduce sea ice thickness a bit; add AGW and suddenly the extent drops a lot. But's it's all quite vague. Plus the focus on one 3-month period dosen't help.
Since I would consider volume to be a more important metric than extent, I'd like to see the same analysis on volume minima.
Since all of these Arctic trends (area, extent, volume) are the inverse of CO2 emissions and Arctic temperature it *is* hard to believe in a 40 year run of natural variation being a major cause of losses.
In my own initial analysis of the "skeptical" brouhaha I've been focusing on the reporting of the paper, rather than the substance of it.
I shall add a link to The Stoat forthwith of course, but here's the first question that sprang into my ice befuddled brain:
http://GreatWhiteCon.info/2017/03/is-arctic-ice-loss-driven-by-natural-swings/#Q1
"All this excitement in the Twittosphere and elsewhere leads one to wonder whether Ding, Schweiger et al. saw (or should have seen?) all this coming, and if so what might have been done differently?"
[For that, I think you should see #2, which points out an intriguing connection to Soon. Which, if valid, would lead you to change the question somewhat -W]
William - Further to your #9, surely you're not suggesting that the likes of Axel Schweiger (of PIOMAS fame) and Eric Steig would somehow allow their strings to be pulled by Big Willie?
[Certainly not directly. But Eric is last author, you never know quite how involved he might be in all the analysis -W]
There's a general issue with this type of study, which is that formal attribution in a changing system about which there is incomplete information will be wrong to some degree. That applies in spades to the Arctic.
Seems surprising to claim that much variability and yet not observe it in recent centuries.
Sometimes we can take guilt-by-association suspicions a little too far. Qinghua and I recently were co-PIs on an NSF proposal and I have no reason to suspect that he is anything but the careful and imaginative large-scale dynamicist that he appears to be. He has previously worked on teleconnections to the Arctic and Antarctic from lower latitudes. This seems like a logical progression from that research.
Scientific criticism of the results and methodology are of course fair game as always. But I'd be astonished if there was any sort of untoward motive behind this work.
[Damn, and the conspiracy was going so well -W]
The Met Office climate prediction model: HadGEM1 climate model has projected Arctic ice loss with reasonable accuracy up to the present day. The model projects that sea ice decline will continue over the 21st Century; however it also shows the possibility of a period of up to 20 years when the ice loss is temporarily slowed. This period is caused by reductions in ice export from north of Greenland, and by a weakening of the North Atlantic overturning circulation (which results in a reduction of warm water transported into the Arctic).
http://www.metoffice.gov.uk/research/climate/cryosphere-oceans/ice-pred…
>"[Damn, and the conspiracy was going so well -W]"
I think you missed out the word theory. ;)
Steve #15 - Here's a long video of the first ever public event at the Met Office's new High Performance Computing Complex near Exeter:
https://youtu.be/lE7fDp3kxqo?t=14m30s
I was in attendance, and if you skip to the "Questions" section at the end you can even see me asking Vicky Pope a tricky question! After that I had a long chat with David Underwood, since nobody else seemed terribly interested in what 16 petaFLOPS and counting might ultimately be capable of. Arctic sea ice inevitably entered the conversation. David said he was proud of the UKMO's "unified" model's performance in that regard. In ~3 yrs the ocean and atmospheric components of their model are due to become "close coupled", at which point their sea ice predictions should improve as well.
The proof of that pudding will be in the eating of course!
Crandles #16 - I think so too!
More learned comments via a head's up from Neven:
http://www.sciencemediacentre.org/expert-reaction-to-research-on-natural-climate-variability-and-arctic-summer-sea-ice/
Thanks, Jim!
Hopefully the first summer ice-free period won't precede that of the model by more than a couple decades. ;)
Seriously, the circulation changes we're seeing, especially this last winter, seem guaranteed to have lots of interesting consequences as they strengthen.
I look at all of this through the lens of the inability of (earth) models to transition to a Pliocene-like climate state. Something is missing (maybe several somethings) that moves additional heat toward the poles, and IMO we're seeing its early effects now.
Well I'm pretty offended, William. This kind of guilt by association and insinuation is below you. Qinhgua -- a former postdoc with me, and a close collaborator with whom I've written a number of important papers, is a great scientist and has unassailable integrity. Same goes for all the other authors on the paper (if you don't know who Battisti is you need to catch up on the last, oh, 30 years of climate dynamics research).
Apology accepted. Just don't do it again, eh?
Oh, and yes, I was very involved in the work. That's why my name is on the paper. I don't coauthor papers I don't understand and/or don't agree with.
[What are you offended by: QD's connection to Soon being pointed out and speculated on? That does seem slightly odd. Or are you suggesting that despite all the fuss, WS is actually perfectly respectable? -W]
please ELIF "have form with"
https://www.google.com/search?q=define+"have+form+with"
suggests you're referring to having goodness?
https://books.google.com/books?id=tTsqCgAAQBAJ&pg=PT30&lpg=PT30&dq=defi…
Hank #21 - This is a Britishism which doesn't seem to have made it into UD!
"Has form" = "Has a criminal record" or more loosely "has a track record"
Does "has a track record with" compute?
[Yes indeed. Here for example. Or, indeed, here -W]
Eric #20 - Since you seem to be here, what do you make of my question in #9?
I have a supplementary one too:
http://GreatWhiteCon.info/2017/03/is-arctic-ice-loss-driven-by-natural-…
Is “a simple ocean–sea-ice model” up to the task in hand?;
Jim hunt: your question being whether we "saw this coming"?
I certainly did. I written or been involved with that a number of papers that more or less say' "hang on folks, not every bump and wiggle is forced change; much of it is intrinsic variability (or at least, more work is needed to demonstrate otherwise." The response is always as expected from the no-nothings, which is some combination of quoting us as showing that natural variability dominates (no, we didn't), and claiming that the scientific mainstream has previously ignored natural variability. I'm more surprised when I get criticism from my colleagues who've usually not yet read the paper carefully, though I'm beginning to get used to that. Should we have done something different? I don't know what. Not published the paper, lest we inadvertently help the "skeptics"? Put in more disclaimers (I thought we had, actually)?
In response to William's comment on my comment: I'm finding it a bit hard to tell you apart from Roger Pielke Jr at the moment. You're kidding, right? Or are you high on something? But given my long standing (and continued) respect for you -- and just in case you are just having an episode of temporary cluelessness -- I'll spell it out: 1) Of course I'm not trying to suggest WS deserves any particular respect; but it doesn't follow that every paper he's been a coauthor on isn't any good. 2) The point is that public speculation of the "true motives" of your fellow scientists is juvenile at best. This kind of thing is McIntyrism (or McCarthyism if you prefer), particularly when it's based primarily on guilt by association. Don't engage in it.
[I'm not sure I can entirely agree with that. There has been extensive "speculation" on WS's motives, and that seems quite reasonable: qui cum canibus concumbunt cum pulicibus surgent. I know of no reason why QD should get that level of speculation, and conversely by default deserves a higher level of default respect; but he isn't getting that level of speculation or anything like it.
I'm happy, however, to agree that it is far better to read, understand and evaluate the paper than speculate on the motives of the authors. Perhaps a post at RC would be in order?
Also, do you have any comment on my reluctance to believe in 40 years of natural variation? -W]
Cat fight!
A good piece in the IBT: http://www.ibtimes.co.uk/how-predict-when-arctic-sea-ice-will-reach-its…
"Previous studies have shown that the observed decrease of Arctic sea ice cannot be explained only by internal variability. That means that the decline of Arctic sea ice is one of the strongest indicators of climate change. Having said that, the total decrease cannot be attributed to climate change. Several model studies revealed that 30 to 50% of the decrease could be caused by internal variability. Thus, it is possible that the Arctic summer sea ice extent might not decrease further during the next 10 years...
On the other hand, the Arctic sea ice has become thinner and more vulnerable. An extreme year... could already melt most of the summer sea ice under current climate conditions... Altogether, this makes it impossible to predict the precise year when the Arctic will be ice-free in summer... even for a "perfect" climate model."
Peaceful #27 - Thanks for the heads up, I'd not stumbled upon that one yet. It will be intriguing to discover how well the CPOM forecast works out this year!
Eric #24 - Any thoughts on my Q2?
Regarding my Q1, any thoughts on this recent comment of mine at Neven's Arctic Sea Ice Blog?
http://neven1.typepad.com/blog/2017/03/lowest-maximum-on-record-again.h…
So when they write:
"... warming and moistening the lower troposphere. Model experiments, with reanalysis data constraining atmospheric circulation, replicate the observed thermodynamic response and indicate that the near-surface changes are dominated by circulation changes rather than feedbacks from the changing sea-ice cover."
So are they saying having open water rather than ice does not warm and moisten the Arctic lower troposphere and thereby cause more warming, although moisture is a greenhouse gas (water vapor) -- or are they saying that of the moisture in the atmosphere that's present, more of that moisture comes in from elsewhere than evaporates from the open Arctic ocean water (so moving the storm track around will matter more than opening more ice-free surface water)?
Eric, thank you for responding here, and William, thank you for your criticism.
As far as I can see, the essence of the conclusion (attribution to AGW) of the paper lies in this section :
Now, both these fractions (70% and 60%) are questionable.
First of all, the 60% number refers to the correlation between sea-ice trend and atmospheric circulation over the Arctic.
However, that does NOT say how much atmospheric circulation over the Arctic is influenced by temperature.
Since higher temperature means expanding air mass, geopotential height will always increase with increasing temperature, which their own findings in figure 1e of the paper clearly shows (best correlation of geopotential height at 200 mb is with temperature).
So that 60% influence of atmospheric circulation can very well be simply caused by atmospheric temperature increase, which can easily be AGW in origin. After all, Z200 is high up in the atmosphere, which means that even during summer it is not much influenced by melting sea ice below.
And the 70% (natural variability) refers only to the influence of “high-altitude winds”.
Here, again, high altitude winds (such as the jet stream) are caused by geopotential height, which is again caused by temperature changes over the Arctic. If the Arctic warms more than the rest of the planet (due to albedo feedback or increase in moisture or any other reason), the geopotential height over the Arctic will increase more than the rest of the planet, and thus the high altitude winds will be less “cyclonic” than otherwise. That means this “70% (natural variability of the atmospheric circulation over the Arctic)” may very well be completely caused by temperature increase.
So both numbers are highly dependent on temperature increase, and since the paper does NOT investigate the correlation of these variables to temperature increase, even though its own analysis establishes that correlation very clearly (fig. 1) its conclusions are not sustained by the evidence they collected.
So it seems that William had a point when he wrote :
Also, do you have any comment on my reluctance to believe in 40 years of natural variation?
I don't have the same reluctance myself. This is a fairly small region of the planet (about 4% of global surface area, I think?) which is known to be highly variable, and we're also talking about seasonal trends rather than annual average. The 2007-2012 period features a few extreme anomalies within the context of the long-term trend, and its position near the end of the trend period will affect the slope. September extents since 2012 have not really got close to challenging even the 2007 anomaly despite the on-going downward trend.
There's a general view now that variability globally over the past couple of decades has been unusual in a centennial context. It seems reasonable that the Arctic would likewise be affected in some way. Looking elsewhere I think it's plausible the sign of Antarctic sea ice change has been dictated mostly by internal variability since 1979.
(Pressed submit by mistake before [removed - W])
[I find it easy to believe in natural variability. I find it much harder to believe that it has been forcing all one way for 40 years; or something like that. I would expect much larger decadal variability in that case; though I haven't thought about it too hard -W]
I find it much harder to believe that it has been forcing all one way for 40 years; or something like that.
It's not really about a steady influence over 40 years. It's all about the 2007-2012, or perhaps 2007-2014 period. If you just remove those years I suspect the extent trend would be about 30-40% smaller. Because this particularly unusual period is currently near the end of the record it has a strong effect on trend.
They're arguing that this short period is an example of large decadal variability.
[Really? I think you'd have to drop everything from ~2000: https://tamino.wordpress.com/2016/09/11/climate-deniers-embarrass-thems… -W]
Really? I think you’d have to drop everything from ~2000
Using the NSIDC data (ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/monthly/data/N_09_…) I get a 1979-2014 trend of -0.087 million SqKm/yr and a 1979-2006 trend of -0.059 million SqKm/yr. That's more than a 40% difference.
Going beyond the study period...
For the 1979-2016 trend I get -0.087 million SqKm/yr and, after removing 2007-2012 or 2007-2014, I get -0.071 million SqKm/yr. So, just over a 20% difference there.
Paul #33 - What do you get if you use the PIOMAS modelled Arctic sea ice volume instead?
http://psc.apl.uw.edu/research/projects/arctic-sea-ice-volume-anomaly/
We are discussing an Axel Schweiger paper after all!
Eric Steig said "quoting us as showing that natural variability dominates (no, we didn’t)"
Clearly they didn't say they expect a 60% natural variability.
From ftp://sidads.colorado.edu/DATASETS/NOAA/G02135/north/monthly/data/N_09_…
decline from 7.86 to 3.63 is 4.23
The next year was back up to 5.35 a rise of 1.72 and 1.72 is over 40% of the 4.23 reduction. I am sure I am doing these calcs all wrong to say we already know natural variability can account for at least 40% of the decline just by looking at the extent data.
Still if they are not claiming natural variability is as high as 60%, more like 30% and we can do a calc to show it was potentially over 40% in 2012, are people making a fuss about something that is obviously true?
Sorry if this is a stupid question, but might it be sensible to ask how high the authors expect extent to jump back up if the natural variability ceased to have a negative effect?
Jim Hunt,
Can't seem to get access to the PIOMAS data at the moment. The volume trend fit between their model and PIOMAS looks less good than for extent - the PIOMAS loss seems a bit more sustained. Though conclusions should probably be tempered by the fact PIOMAS is an indirect model of ice thicknesses based on environmental measurements, rather than being direct measurement.
I'd be interested in seeing what the paper's method expects up to the present. Globally, many variability patterns seem to have flipped since 2013/14 and I would suspect (perhaps wrongly) that the Arctic atmospheric circulation pattern they identify has likewise subsided. Yet PIOMAS found September 2016 was close to the 2012 record, and 2017 looks like a strong contender for a new record.
In related news, March around Antarctica looks set for another huge chart-busting and trend-busting record low for sea ice.
Crandles #35 - From my perspective at least "the fuss" is over the way the paper has been misrepresented in the "skeptical" media, where something that is obviously untrue (to those in the know) is being asserted.
I certainly don't think that Eric's "Not publish[ing] the paper" is the way forward! What else might have been done differently? "Natural variability" for example, is a "strange attractor" for "skeptical" sorts!
[I'm not bothered by the paper being misrepresented by the "skeptics"; they will misrepresent anything -W]
Paul #36 - I see the problem with the PIOMAS data. Click "Send" and see only an indefinitely spinning hourglass?
Wipneus is an alternative source:
https://sites.google.com/site/arctischepinguin/home/piomas
although that doesn't seem to be quite up to date
PaulS for daily volume data see
http://forum.arctic-sea-ice.net/index.php/topic,119.1500.html
post#1507
>"[I’m not bothered by the paper being misrepresented by the “skeptics”; they will misrepresent anything -W]"
Yes I also am not bothered by such misrepresentation, denialati will do that. But sane people seem to be in incredulity mode not believing the papers conclusions might be correct.
What I am trying to say is, the situation may depend on what is meant. At the other end of scale of possible mathematical interpretations, we could decide that the 40% below for 2012 is really 20% below, then 2013 is 20% above, average these with 8 other years and we would get close to 0% for the 10 year average in which case 30% would be a big deal.
If you can do different maths and one way with a single year you can get 40% so around 30% is obviously true / no big deal and another way using 10 year average, 30% becomes a big deal issue.
Is it possible that the sane commentators need to understand what the authors mean before getting into their incredulity mode?
(or maybe it is just me failing to parse it sensibly.)
[I'm still waiting for RC to tell me what to think -W]
William #37 & Crandles #40 - You are aware that the lunatics are claiming to be in charge of the asylum? Allegedly one of them was even voted into that position:
http://GreatWhiteCon.info/2017/03/that-deaf-dumb-and-blind-kid-sure-pla…
It is not easy to debunk a paper in the comment section of a blog on that paper, but it’s not impossible either.
Typically, if a paper suggests that natural variability is responsible for X % of Arctic sea ice extent reduction, a red flag goes off since we have not seen and trend in variables that are “natural” in nature other than temperature, which goes up with AGW.
In Ding et al 2017, the key figure to keep in mind is figure 1c, which shows development of Arctic “temperature”, “specific humidity”, “long wave downwelling radiation” and how they relate to “sea ice extent” from 1979 to 1014.
The key is that “temperature”, “specific humidity”, “long wave downwelling radiation” all very closely correlated AND that they all have a trend and some natural variability imposed on top of that.
Now, below, I’ll show that Ding et al 2017 convoluted that trend with natural variability, and how they did it.
So without further adieu, here we go :
From the abstract :
So where does this 60% (caused by summertime atmospheric circulation) come from ?
Here is the section :
So it is the result of the difference between Experiment 5 and 6 that lead them to this conclusion.
What IS the difference between experiment 5 and 6 ?
For that, we need to go to the “Methods” section :
That is a very convoluted and complex way of saying that in Experiment 6 then kept the summer-time climate constant, while in Experiment 5 they let if follow the variables from Fig. 1 (Which we know have a trend).
If you keep the summer-time climate constant, then obviously the trend in sea ice will be reduced. And they find that it is 60% reduced. So the conclusion from this (Exp 5 versus Exp 6) is that 60% of Arctic sea ice reduction is caused by summer-time climate change, while 40% is caused by climate change over the remaining 9 months.
That would have been a fine conclusion, but is completely different from their conclusion that "summertime atmospheric circulation” may have contributed as much as 60% to theSeptember sea-ice extent decline since 1979.
It’s not “summertime atmospheric circulation” that caused it; it is “summertime climate change (with ALL the variables, including temperature) being the cause. And with AGW being real, we know that temperature for one is NOT a natural variability variable.
So Ding et al 2017 made the ASSUMPTION that the temperature TREND in the Arctic is part of natural variability and they base their conclusions on that.
Needless to say that that assumption (in a warming world) is quite preposterous, and I wonder what the authors were thinking when they drew their conclusions.
[That was terribly confusing, because in your quotes (from, I presume, the published paper) the submitted Exp-4 has become 5, and Exp-5 has become 6. Anyway, going on from that:
Having looked again, you appear to be correct. They've detrended on Z220, apparently on the assumption (which I can't find stated) that Z200 must be "variability" but I can't see any obvious reason to believe that. So, yes; by removing a pile of the forcing they've removed a pile of the result, i.e. less sea ice loss. Quite why this is supposed to be valid is, as you say, something of a mystery.
You'd expect the referees to pick this up -W]
Please note that I quoted from the published paper :
http://www.nature.com/articles/nclimate3241.epdf?referrer_access_token=…
and not from the "submitted" paper that William refs above
https://www.atmos.washington.edu/~david/Ding_etal_2016_submitted.pdf
The two versions differ in details, but not in substance.
http://www.timeslive.co.za/scitech/2017/03/23/Sea-ice-hits-new-record-l…
[I liked CIP's http://capitalistimperialistpig.blogspot.co.uk/2017/03/minimax.html -W]
OK. Kind of silent after my post.
Not sure if people understood the main problem with Ding et al 2017, so let me try to summarize it as brief as possible :
From the abstract we read "trends in summertime atmospheric circulation may have contributed as much as 60% to the September sea-ice extent decline since 1979".
However, that is NOT what the paper's methods show.
For starters, in the methods (model experiments) "atmospheric circulation" variable is represented by "geopotential height" over the Arctic summer, specifically Z200 (geopotential height at 200 mb). So the correct conclusion would be "trends in summertime geopotential height may have contributed as much as 60% to the September sea-ice extent decline since 1979".
With that conclusion, you can feel the correlation-versus-causation issue coming up : Arctic temperature increase (due to AGW) may be the cause of or geopotential height increase.
And consequently you can no longer claim that the (up) trend in both variables is a source of "internal variability".
And thus the final conclusion (that "Internal variability dominates the Arctic summer circulation trend and may beresponsible for about 30–50% of the overall decline in September sea ice since 1979") no longer holds.
Now that we know that, the question is if Eric Steig and Alex Schweiger (both of whom I respect very much) knew about this.
Eric, I know you are reading this, so could you please comment on if you agree that the paper's methods don't support the conclusions ?
And if you do, then did you know about this ?
And if you don't, can you please start a technical discussion on where I misinterpreted the methods used in Ding et al 2017 ?
Let me put this differently (in case Eric does not respond) :
I am ready to start a technical discussion on if the paper's conclusions are sustained by the methods (and experiments) or not. I claim they are not. At all.
This paper has unleashed the alarmist commentariat versions of WUWT commenters at Neven's in the Arctic sea ice changes: Natural variation vs human influence Forum topic.
Sad to see.
Also sad to see so few of the regulars even taking these nutters to task for their accusations against climate scientists.
Kevin #47 - I'm too busy taking on the nutters at David's, Judy's, Anthony's, Tony's and Lamar's:
https://twitter.com/GreatWhiteCon/status/845756197353476097
"In which we idly speculate whether the David & Judy show will be performed tomorrow and/or on Wednesday"
Rob #46 - Please feel free:
"Is Arctic Ice Loss Driven by Natural Swings?"
I am very disappointed to see #2 comment on our paper. But I really appreciate that Raymond said something for me and Eric fought back for me before I step in.
Here I only want to quickly respond to that suspicion.
The primary funding source of the paper is from NOAA.
Please check this if you want to know more
http://cpo.noaa.gov/ClimatePrograms/EarthSystemScience/ClimateVariabili…
I don't know Willie Song. The main reason I was on that paper was that the paper focused on the global monsoon, a concept proposed by one of my early paper ( Wang and Ding 2005) with my Ph.D advisor ( Pro. Bin Wang). I helped the first author doing some analyses and so she put my name in the paper.I probably won't check this blog again. If you have any further question, please reach me at my email address qinghua@ucsb.edu
[Thanks for stopping by. It is good to hear what you have to say about WS. If you have any comments regarding RD's analysis of an apparent flaw in your paper, that too would be welcome -W]
I gave up wasting my time with debates in the comments section of blogs a long time ago. One could spend all one's time on it, leaving no time for anything serious. See here: http://www.realclimate.org/index.php/archives/2008/10/greenspan-einstei…
I will just note that yes, I know what we did in the paper on which I am a coauthor, and no, I don't agree with the statement that the "results don't support the conclusions". Most of the ideas RD has written here and elsewhere don't make sense. For example, no, PV = nRT does not tell you how geopotential height responds to surface temperature! There is one aspect that might be worth discussing, which is that we assume, in effect, that most of the the trend in z200 is "natural variability". (We don't actually assume it -- that's a result of the analysis, but in the end it amounts to the same thing, pretty much). But this doesn't come out of nowhere! it comes largely from our previous work published in 2014, showing that the trend in z200 is related to tropical forcing. To claim it is not natural variability you would have to show that the way that the spatial pattern of convection tropics has evolved in the last 30+ years is not natural variability. It has certainly argued that it is not. Trenberth claimed to have to demonstrated an anthropogenic signal in tropical variability in 1997, during the massice 1997/1998 ENSO event, but eventss since then have largely negated that analysis (though it was a good analysis for the time). An honest assesement of the literature (see IPCC fore example) would say that the jury remains out on that. To the extent that the decadal tropical changes ARE a forced response, then our estimate for the natural variability is an upper bound, which I think we were clear about in the paper.
Homework: Ding et al., 2014: http://www.nature.com/nature/journal/v509/n7499/full/nature13260.html
I also highly recommend Perlwitz et al., 2014.
http://journals.ametsoc.org/doi/pdf/10.1175/JCLI-D-14-00095.1
[Thanks for the reply. The point of the trend in z200 is "natural variability" is indeed the point that RD is questioning; and your statement that your estimate is "an upper bound" probably helps; I'm not at all sure that was clear in the paper, though you can certainly read that once you know it is there -W]
Eric,
when you say, "
To the extent that the decadal tropical changes ARE a forced response, then our estimate for the natural variability is an upper bound, which I think we were clear about in the paper. "
We have only this year realized a definitive fingerprint of anthropogenic (SE Asian) aerosol impacts on tropical variability. In fact it now appears that the ~30% cut in Chinese hi-temp industrial SO2 emissions over the last 2.5 years is 100% responsible for the expansion of WV, standing rossby wave functions, abnormal winter arctic heating and portends a near-permanent +IPO condition as SO2 emissions are phased out fully over the next decade.
With this then is a false 'natural' variability signal that is more agnostic, relying on imperfect modeling, than careful science. In view of these missing drivers in the model, any declaration of upper bounds within the paper is plain hubris.
These two papers in 2016 reveal significantly missed tropical variability. I really wish you had been more careful publishing the paper, especially since the Arctic will likely reach ~sept. ice free condition (<1X10^6 Km^2 SIE) in the next 3 years and most assuredly within the next 6.
Role of volcanic and anthropogenic aerosols in the recent global surface warming slowdown
http://www.nature.com/nclimate/journal/v6/n10/full/nclimate3058.html?WT…
Observational Evidence for Aerosols Increasing Upper Tropospheric Humidity
http://www.atmos-chem-phys.net/16/14331/2016/acp-16-14331-2016.pdf
jai - did you read the reference supplied? The abstract alone makes your comment nonsensical: Tropical forcing of the recent rapid Arctic warming in northeastern Canada and Greenland, Ding et all, 2014
"We find that the most prominent annual mean surface and tropospheric warming in the Arctic since 1979 has occurred in northeastern Canada and Greenland. In this region, much of the year-to-year temperature variability is associated with the leading mode of large-scale circulation variability in the North Atlantic, namely, the North Atlantic Oscillation. Here we show that the recent warming in this region is strongly associated with a negative trend in the North Atlantic Oscillation, which is a response to anomalous Rossby wave-train activity originating in the tropical Pacific. Atmospheric model experiments forced by prescribed tropical sea surface temperatures simulate the observed circulation changes and associated tropospheric and surface warming over northeastern Canada and Greenland. Experiments from the Coupled Model Intercomparison Project Phase 5 models with prescribed anthropogenic forcing show no similar circulation changes related to the North Atlantic Oscillation or associated tropospheric warming. This suggests that a substantial portion of recent warming in the northeastern Canada and Greenland sector of the Arctic arises from unforced natural variability."
Kevin,
Yes, I have and my analysis still stands. The lack of secondary cloud effects, tropical upper troposphere anthropogenic aerosol drivers and impacts on larger MJO/+IPO surface winds are not included in their model and so they have provided an overconfident assessment of 'natural variability'. In addition,
Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability
http://centaur.reading.ac.uk/30590/1/30590booth_et_al_nature_2012-accep…
p.s. When asked what amount of currently observed global mean surface temperature warming above pre-industrial is human-caused, the correct answer is ~110%. This is due to the cooling effect of aerosols on the current climate regime, producing additional cooling effects. Similarly, and quite separate from my earlier posts, a very reasonable argument can be made that, in the total absence of human aerosols, the Arctic would be significantly warmer than today. This does not include the recently observed changes in atmospheric circulation but is exclusively looking at cloud effects. combining these two effects, we will see an additional >+3C warming from today. This human-induced cooling has suppressed sea ice loss and so anthropogenic forcing has produced significantly MORE sea ice loss than those currently observed.
Indirect Aerosol Effect Increases CMIP5 Models’ Projected Arctic Warming
http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-15-0362.1
Attribution of Arctic temperature change to greenhouse-gas and aerosol influences (note: here they state 1.8C of arctic cooling due to anthro aerosols)
www.homepages.ed.ac.uk/shs/Climatechange/Arctic%20ice/Arctic%20temp%20G…
Jai - "The atmospheric circulation changes this year (I believe as a result of China shutting down the 30 domestic coal mines with the highest sulfur content - significantly reducing SO2 emissions) have destroyed the arctic sea ice - this has terrified the arctic sea ice community"
Of course Ding et al (2014) looked at data from 1979 to 2012. Not *this year* - duh. And they showed that just by nudging two variables they could replicate the arctic patterns. Hint: aerosols weren't one of the variables.
Besides your snide comments towards scientists is what's really offending. You were one of the WUWT-like commenters over in the forum at Neven's. Here are a few excerpts from what you wrote over there:
"This effort is an attempt at "Cover Your Ass" or CYA because they are attempting to deflect the coming criticism of their work..."
"... this is an obvious attribution overconfidence in the model's capabilities - the attribution is glaring and not good science."</i
"...this paper, with its selfishness and unscientific assertions will be used to allay the global response to the climate emergency..."
Comment on March 16th
Frankly, if I were one of the authors I wouldn't give you the time of day.
Eric, thank you for your reply.
I'm sorry that my comments "don't make sense". Let me clarify my points in better detail here.
William mentioned that "The point of the trend in z200 is “natural variability” is indeed the point that RD is questioning" and that is certainly a point of concern. After all, we KNOW that at least a part of the atmospheric temperature increase is due to AGW, which MUST have put an imprint on geopotential height over the Arctic. Jennifer Francis theory of wavy jetstreams due to reduced geopotential height difference between the tropics and the Arctic is based on that. Are you now saying that that theory is wrong ?
If not, then at least a part of trend in geopotential height increase (Z200) since 1979 must be anthropogenic.
But in fact the core of my argument goes one step further :
I claim that your "regression coefficient" in Experiment 6 eliminates ALL trends, natural or anthropogenic. And thus, that you essentially kept the climate constant in Experiment 6.
To see that, check the regression formula you use for Exp-6 :
B(x, y, t) = β(x, y) × Z200 GL (t) (1)
Here, β(x, y) is the regression coefficient between Z200-GL and the variable under consideration, and B(z,y,t) is the "forcing" that you later subtract from the ERA data during summer.
Now, the regression coefficient is very good in finding trends.
So even if the correlation of Z200-GL and the variable (say "temperature") is not very good (like further away from Greenland) the regression coefficient will still find a 'trend', and thus the "forcing" you subtract for Exp-6 will still be pretty darn close to the linear relation we expect to see from the ideal gas law.
Qinghua has all the data available, so could you please ask him to reproduce the "adjusted forcing" for Exp-6 for, say the temperature variable north of 70deg, and check if it still has a trend since 1979 ? I claim it doesn't, since the regression method eliminated it.
And that means that all you have calculated (between Exp-5 and Exp-6) is the influence of climate change over the JJA period (60%) which means the remaining 40% was caused by climate change over the other 9 months.
Which is an interesting conclusion, but positively doesn't have anything to do with "natural" or "anthropogenic" causes of the trend in Z200-GL.
Kevin,
Don't forget my follow up comment:
[i}it is impossible to say if this is a CYA paper from entrenched arctic scientists who have dedicated their lives to learning and teaching about these things or if their unsubstantiated interpretations of global atmospheric circulations being 'natural' even though recent studies indicate a significant anthropogenic component (due to many feedbacks, ENSO impacts and aerosol impacts) are rather based on institutional thinking, bias toward confidence in their previous work (models), failure on the SLD (Side of Least Drama) and trying to find a reason an explanation why their models are so pathetically, horribly, dis-servingly off base.[/i]
Kevin,
within context, you will (all) soon realize that at 495 ppm CO2e we have crossed a critical threshold where ECS is rapidly approaching 6.0C in its non-linear push toward early Pliocene climate regimes. At this point, only a WWII-scale climate mobilization will prevent +4C of warming by 2065 and we have already locked in >+3.0C at today's levels (and the absence of aerosols). in this view I think you may understand why this paper made me so upset.
I received the third email from Rob and then test his idea tonight. Actually, we did all these calculations before we implemented Exp. 6. The test shows we still have 35% of trend retained after that Z200-GL influence is removed. I have sent a figure to Rob. The original trend of LW ( the purple curve in Fig. 4 of the paper) from 1979 to 2014 is 2w/m2/decade. The modified one still owns a 0.7W/m2/decade trend.
Rob,
You miss the point of the Paper, they make their claim by unsing an ensemble of CMIP5-Models, see Section "Estimation of internal and anthropogenic contributions"
You can also see their used models in the Supps Table 1 and that is what makes there conclusion not so confident, because of very strong vary of state of climate in each model, Q. Shu (2015) gives a good impression about the models and there cyrosphere states (1979-2005)
Observed:
Annual Extent: 12.02 Mio km^2
Extent-Aplitude: 8.08 Mio km^2
Trend-Dekade: -4.35 Mio km^´2
Models used in the Paper (on 95 Conf)
Annual Extent: 12.82 +-0.59 Mio km^2
Extent-Aplitude: 10.55 +- 1.27 Mio km^2
Trend-Dekade: -3.73+- 0.98 Mio km^´2
That mean, the Model they used, show a little bit to much ice, have a to strong aplitude and a little bit to low trend over the historical period 1970-2005, by 2005 the models are in there projection range. So the Question is, what happen since 2005 with the models, if there is a not so strong trend like in Observation, it not have to mean, that the rest is unforced, it could also mean, that models not able(or less able) to give correct atmosphere answer to the loos of sea ice extent.
So in the End i do have doubts about the conclusion, first of the models they used and that blocking events are vice/versa between near surface and 200mb
Soory:
"So in the End i do have doubts about the conclusion, first of the models they used and that blocking events are vice/versa between near surface and 200mb"
Have to mean:
"So in the End i do have doubts about the conclusion, first of the models they used and because of that blocking events are vice/versa between near surface and 200mb"
within context, you will (all) soon realize that at 495 ppm CO2e we have crossed a critical threshold where ECS is rapidly approaching 6.0C in its non-linear
I don't see how this is possible (other than as having a small probability based on the standard ECS distribution). Even the ESS (which includes all the slow feedbacks) is probably less than 6C in our current climate state (relatively small ice sheets).
Another Thing to point out is, that arctic Summer-Saison (JJA) z200 is extrem strong correlate with Sea-Ice-Concentration(SIC) in Models, here i used the mean von CMIP5 RCP4.5
If we take z200 and SIC in 70-90N for JJA for 1979-2016
R^2: 0.925
This means, if Models in the Paper (as shown before) had a less Sea-Ice-Loos, the increase of z200 have to be lower then in Observation. So therefore, that the Rest of increase in z200 is internal can not be driven, it could also be the less decrease of Sea-Ice in the Models itself.
Soon connection will do fine.
Ding et al is climate revisionism. Dump.
To the point of the more frequent blocking (which is good the see in z200): http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-15-0391.1
". Reduced summer sea ice favors stronger and more frequent occurrences of blocking-high pressure events over Greenland. "
Under the line:
The Paper (in my optinion) make a weak conclusion about internal effects (long term) on sea ice, weak because its not clear yet how much the circulation changes are caused by arctic sea ice loos or internal variability, for both good arguments, but no clear evidence
ATTP,
wrt ESS,6C
you are working from old data,, the models used in AR4 and most in AR5 held that we would not breach effective ice free arctic in September under 2X CO2. We now know that our first effective ice free minimum will be by 2024 at the very latest.
under 2XCO2 (if we ever go there) we will experience an ice free Arctic by June 21. Under this scenario the difference in albedo alone is equivalent to ~2.3 W/m^2 (globally averaged TOA) however it will ALL be received above the arctic circle).
Not to mention a semi-permanent positive IPO, consequential loss of Amazon forest/peat system, collapse of the boreal forest system due to heat stress, algae growth in the arctic further decreasing albedo and the acidification-driven loss of phytoplankton that produce a significant volume of Dimethyl Sulfide which is working to further cool the earth.
Booth et al. (2012)
Aerosols implicated as a prime driver of twentieth-century North Atlantic climate variability
http://centaur.reading.ac.uk/30590/1/30590booth_et_al_nature_2012-accep…
"Individually, volcanoes and aerosols explain 23% and 66% of the temporal multidecadal variability (10 year smoothed) of the detrended NASST (Figure S5). Combining both contributions, 76% (80% after inclusion of mineral dust aerosols) of the simulated variance is explained."
Miettinen et al. (2011)
North Atlantic sea surface temperatures and their relation to the North Atlantic Oscillation during the last 230 years
https://www.researchgate.net/publication/227298859_North_Atlantic_sea_s…
"The aSST record and the August North Atlantic Oscillation (aNAO) index show similar multidecadal-scale variability indicating a close coupling between the oceanic and atmospheric patterns. The aSST record shows a negative correlation with the aNAO indicating cold aSST during the positive aNAO trend and vice versa. Results suggest that the wind driven variation in volume fluxes of the North Atlantic surface waters could be the major mechanism behind the observed relationship. North Atlantic sea surface temperatures and their relation to the North Atlantic Oscillation during the last 230 years."
RE: "our first effective ice free minimum will be by 2024 at the very latest"
I don't see how this prediction can be made with this accuracy.
Variation in weather drives variation in ice loss. Variation puts about a two decade uncertainty on first ice out predictions. If we are at 1% odds this year, we will not be to 99% odds until 2037 or so... Yes, we might be at 1% odds for ice out this summer. But this is the first year that this was a realistic possibility.
Also the Pacific Decadal Oscillation or PDO has been favorable until recently. The recent switch to the positive phase may slow Arctic warming.
http://www.nature.com/nclimate/journal/v6/n9/full/nclimate3011.html
If this prediction is based on recent trends in sea ice volume or extent, these trends extrapolated may be an overestimate of ice loss.
Jai,
Even we do see an ice free Arctic (summer) I'm still not convinced that that implies an ECS of 6K. That's well above any best estimate that I've seen. It's not ruled, but it still seems unlikely.
ATTP,
That for me climate alarmism,its the same as someone is telling ECS is much smaller then 1.5K
Christian,
Yes, that's one of my issues. It's hard to see how claiming that the ECS will be 6K is different to claiming that it will be 1.5K.
ATTP,
These are not included in the models:
under 2XCO2 (if we ever go there) we will experience an ice free Arctic by June 21. Under this scenario the difference in albedo alone is equivalent to ~2.3 W/m^2 (globally averaged TOA) however it will ALL be received above the arctic circle).
Not to mention a semi-permanent positive IPO, consequential loss of Amazon forest/peat system, collapse of the boreal forest system due to heat stress, algae growth in the arctic further decreasing albedo and the acidification-driven loss of phytoplankton that produce a significant volume of Dimethyl Sulfide which is working to further cool the earth.
(note: these responses are identified under 2X CO2 conditions and in the absence of anthropogenic aerosol emissions)
Jai,
Please? They are fully coupled Models, i think you disunderstand what forcing is and there are much more effects also cloud effects..
Belive it or not, its include the model, CMIP5 see -2W/m^2 from 2000-2100 in Surface-Upwelling-Shortwave-Irradiance, on Global Scale, in arctic (70-90N) its -20W/m^2
https://climexp.knmi.nl/data/icmip5_rsus_Amon_modmean_rcp85_0-360E_-90-…
https://climexp.knmi.nl/data/icmip5_rsus_Amon_modmean_rcp85_0-360E_70-9…
So please, in future, first learn about models and talk later
ATTP,
Thats why i use now Denial for all of them, its the same just invert, its the denial of the possibility that ECS lower then the upper limit which cant be ruled out yet
[You can, but using words in a way totally different to everyone else is just bound to confuse people. If you're interested in intelligent discussion rather than a confused morass, it is a mistake. "Denial" in the accepted global-warming-wars sense means "people who deny that GW is happening" or similar -W]
Back to the Paper here discussed,
As shown before, the Models they use have different states of arctic sea ice, if you pick the best model to the parameters Annual Extent, Trend and Amplitude ( MPI-ESM-MR) its nearly perfect reproduce the observed SIC-Minimum (September)
http://www.directupload.net/file/d/4674/5kode2k9_png.htm
It would support the idea that the response of z200 could be a feedback or by a coincidence the internal natural variability is nearly the same in this model and in observation (which seem unliky, because of different up and downs)
Data to get here:
https://climexp.knmi.nl/data/icmip5_sic_OImon_MPI-ESM-MR_rcp26_0-360E_7…
https://climexp.knmi.nl/data/iice_index_n_0-360E_70-90N_n.dat
(NSIDC * 100 to get %)
In other Words:
The most realistic Model in State of Arctic Ice until 2005, have no problem the forcast the strong ice loos after 2005
William,
> You can, but using words in a way totally different to everyone else is just bound to confuse people.
Thats why i told here.
> If you’re interested in intelligent discussion rather than a confused morass, it is a mistake.
You now what, since years i am activ on blogs, i learned if you want a good discussion (which also increase your level of understanding) you have to ignor such people who pick sensitivities and tell that the rigth one, or people who claim, there is no warming by co2, therefore its for me the same, they are trapped in there own world.
> "“Denial” in the accepted global-warming-wars sense means “people who deny that GW is happening” or similar "
Yes, that correct, but it is hard to call the others "low ECS-Denials" or something else, because they have both the same bias, just the opposite direction
Christian,
2.3 W/m^2 globally averaged and applied to 70'N-90'N is ~ -70 W/m^2 shortwave emission anomaly. This will happen about 2065 (under rcp 8.5 and zero anthropogenic aerosols)
Even the most aggressive models (here CanESM2) under RCP 8.5 only show +8C TXX anomaly (annual highest daily temp) (75'N-90'N) https://climexp.knmi.nl/data/icmip5_txx_yr_CanESM2_rcp85_0-360E_75-90N_…
see more here: http://file.scirp.org/Html/17-4700471_66158.htm especially fig 7.
Jai,
Y, and what the problem? Its just all about balance, if sea ice is gone it would not mean, that full Insolation is going into the local climate system, clouds also would increase,increase latent heat excess and and and.
So your 2.3W/m^2 for Global Scale is just useless, because its only a calculation on one side. Also your talking about Models, nearly ice free arctic minimum is forecast to come arround 2060-80 in RCP8.5
And as i hope you know, the climate sensivites are all the same in RCP4.5, 8.5 ....the difference is the forcing projection, a 6K ECS (in this case would also mean higher TCR) would imply for RCP8.5 nearly 8K-10K increase of temperature untill 2100
So its totally nonesens to say ECS becomes higher of faster ice melt, its more that TCR would become closer to ECS and in a ECS-State (after 100s of Years) its no more a matter since when the sea ice is gone the only difference is, that the early warming would be stronger and the later warming would be weaker or less intens
Christian,
Using CERES NASA scientists have already observed (in 2014) an average increase in shortwave absorption in the Beaufort Sea at 50 watts per meter squared. They used cloud effects in their analysis.
https://www.nasa.gov/press/goddard/2014/december/nasa-satellites-measur…
read more on the process here: https://www.nasa.gov/press/goddard/2014/december/nasa-satellites-measur…
with regard to regional impacts to boreal forest and permafrost please see regional warming impacts under total loss of sea ice here: http://www.colorado.edu/geography/class_homepages/geog_4271_f10/reading…
ECS is not fixed. It rises as the temperature rises, especially if one (more than appropriately) incorporates feedbacks that are operating within the projected temperature regime.
please see: http://advances.sciencemag.org/content/2/11/e1501923.full for more information.
note: this last paper is a follow up to the PALEOSENS (2012) paper here: http://www.nature.com/nature/journal/v491/n7426/full/nature11574.html
After reading this paper, I quickly saved it onto my database under the title, “holy s**t the S blows up when temperature moves above neg 1point3 K and arctic ice is gone “
Jai,
Thats the point why i talk about selectiv use of science Paper, i can give u also are multi-Studie about that ECS is near 3K in Paleoclimate. And as said before, you always doing it in 1x Way, what about deep water production and weakening of MOC, what about increase Albedo in the greening of much places arround the arctic cycle (green has a higher albed then brown ground), what about more intens wild fires, which a oBCs and a good shorterm cooling, what about less stabile atmosphere stratification in the arctic, Clouds....
and so on..
So i dont want so say its impossible to have 6K ECS but just to select scientific Papers only of there massage to this ECS its some kind of denial for me
And jes ECS is not fixed, but if the same feedback works it not so important for ECS if this feedback is fully integrate in climate after 50 years or 200 years, its only would bring TCR close to ECS
Jai,
Yes, I'm aware of those papers. However, the Science Advances paper had a best estimate for ECS of 4.9K (not 6K) and even this doesn't really justify claiming that it will be this high, simply that it could be. My own view is that the Science Advances estimate for the TCR (2.7K) is unrealistically high and so also brings into question their ECS estimate. You can also read James Annan's post.
ATTP,
Yes, I know you know since it was your site that brought it to my attention, thank you!
however, you must realize that the CMIP3 projections of sea ice were shortened by ~30 years under CMIP5 and that even the CMIP5 projections (used in SA) are dramatically underestimating sea ice loss as observed (and modeled by second author of the nat var paper in question).
Therefore, the regional warming effect I have described, and not present in the models will engage, and summer sea ice free conditions much earlier than model projections. This works to move Transient Climate Sensitivity (talking about SA paper now) on a higher slope.
Finally, from the paper,
"The polynomial fit is shown in Fig. 3. The slope of the fit is subject to substantial change, steadily increasing for warmer background climate. Interpreting the (local) slope of the fit in terms of a specific equilibrium climate sensitivity is not straightforward. We referred to the detailed discussion by Köhler et al. (50). Here, we are mainly interested in deriving an average of S that is representative of warm climates and can be used for a global warming projection for the 21st century."
Note that they conservatively fit a linear projection on a binomial curve fit and when they did this they took the AVERAGE value. Another very significant conservatism.
This doubly conservative value produced a GST for RCP 8.5 projection to 2100 that is roughly DOUBLE the IPCC estimate and its value is (as you said) 4.88 K (likely range, 4.29 to 5.44 K)
If you then include ocean acidification effects on the production of DMS, algae growth in the arctic circle and then include warming soils and decomposition of frozen soils, you will find that these feedback alone dramatically change the Earth System response in very near (lifetime scale) projections. These are not included in the CMIP5 ensemble.
Well, here is an interesting development.
After Qinghua posted his findings that the adjusted forcing for experiment 6 still had a trend (in post 61; thank you Qinghua) I decided respectfully take the conversation private (polite email with the authors).
Yet now Eric Steig decided to take his opinion about me and the process of expressing scientific criticism on his paper in this fine blog public in a rather unprofessional rant of 9 twitter posts :
https://twitter.com/ericsteig
[Hmm. He does seem rather offended. I think that sequence of tweets will be hard to understand for those not "in the know". I've responded, briefly -W]
Now, since Eric decided to take the discussion public again, why not take the scientific discussion public again also :
The main problem with Ding et al 2017 is that the regression method they used for experiment 6 :
B(x, y, t) = β(x, y) × Z200 GL (t) (1)
eliminates ANY trend in ANY of the variables, regardless of if these trends are "anthropogenic" or "natural variability" in nature.
And thus they can no longer claim to attribute any part to "natural variability" alone.
Their method essentially makes the climate constant during JJA in experiment 6. When confronted about that (by me in post 58).
Qinghua presented a graph (in post 61) which still has a (35%) residual trend.
The problem is that this graph represents the annual average climate used in experiment 6, and decidedly NOT the summertime (JJA) average, which shows no trend.
The residual 35% trend is the contribution from the other 9 months, where they let the climate follow the original ERA data (which still shows an uptrend since 1979).
We KNOW that Qinghua's graph does not show JJA climate in experiment 6, since when I asked Eric to look at that data he replied (on twitter) that he had to "re-do a bunch of calculations" to obtain that data :
https://twitter.com/ericsteig/status/847494244524736517
The flaw in Ding et al 2017 is there for everyone to see, if you just follow the science.
I'm disappointed that instead of embracing the criticism and engaging in a technical discussion, Eric Steig instead resorted to an argument of authority on twitter.
[ES doesn't seem to want to respond in any detail here - I think blog comments are a far better venue than twitter - so I hope he takes up my suggestion of writing the paper up for RC -W]
i got a new email from rob again so I have to say something here.
[Well, you don't have to. But it is good that you've responded, thank you -W]
I think I made it very clear in my plot and email to rob that the curve is for jja only .
It is Not annual mean. it doesn't include any non summer month. It is June-July-august.
The regression we used can only partially remove the signals in jja since correlations between the z200 index with those jja variables are around 0.7 to 0.8. Please see our fig 1 .
There is still 30to 40% stuff left in exp6 that includes both residual trends and year to year changes.
Qinghua,
I know that correlation between the z200 index with those jja variables are around 0.7 to 0.8, but correlation is does not determine the beta factor in linear regression.
Even with a poor correlation, linear regression will still obtain a beta factor that is 'average' for the relation of the variables.
The point is that linear regression is designed to find a trend, even in noisy data with poor correlation.
Also, if the graph you show is really "It is June-July-august" then why does Eric mentions that he needs to "“re-do a bunch of calculations” to obtain that very same data ?
Hi Rob,
I know the problem. Probably you thought we used some fancy methods to do the regression. But we just used a simple linear regression that only depends on the correlation coefficient. This is the most widely used method in our field to do this sort of thing.
As I said, the curve I sent to you is for JJA only and there is still 30 to 40% of the warming trends left after we applied that regression. Thus, we didn't make summer climate constant in exp6. Thanks for your comments on our paper.
Best
Qinghua
Qinghua.
Thank you for your reply.
If you used simple linear regression, then your method will eliminate any trend in the JJA record.
The graph you showed still shows a 35% residual trend, which suggest it is NOT the JJA climate for experiment 6.
After all, where would that residual 35% trend come from ?
the residual trend could be due to Co2 or cloudiness changes that are not related to the Z200 index.
the Z200 index has a strong interdecadal jump around 2003-2005. In our 2014 paper, we have argued that this interdecadal jump is not totally due to the Co2 rise.
So our regression can only remove this interdecadal like jump in the forcing fields. Some slowing increasing trends, as those due to Co2 forcing, cannot be removed by this regression.
the residual trend could be due to Co2 or cloudiness changes that are not related to the Z200 index.
the Z200 index has a strong interdecadal jump around 2003-2005. In our 2014 paper, we have argued that this interdecadal jump is not totally due to the Co2 rise.
So our regression can only remove this interdecadal like jump in the forcing fields. Some slowing increasing trends, as those due to Co2 forcing, cannot be removed by this regression.
the residual trend could be due to Co2 or cloudiness changes that are not related to the Z200 index.
the Z200 index has a strong interdecadal jump around 2003-2005. In our 2014 paper, we have argued that this interdecadal jump is not totally due to the Co2 rise.
So our regression can only remove this interdecadal like jump in the forcing fields. Some slowing increasing trends, as those due to Co2 forcing, cannot be removed by this regression.
Qinghua,
I appreciate your response, but the regression method you used in Ding et al 2017 uses a beta factor against Z200 for ALL variables. Thus, ANY trend in ANY of the variables is reduced by the trend in Z200, and since the trends in all variables (see fig.1c in your paper) is very similar, you knock out ALL trends in ALL variables for the JJA period.
There is no indication in your paper that any specific variability other than CO2 increases (like "cloudiness changes") have any long term trend that could support your conclusions.
Anyway, it is fine that you don't accept my explanation.
But my calculation shows that there is still 30-40% of the trend left in JJA.
This is my last post.
Qinghua,
Maybe we can finalize this conversation in a RC post as William suggested.
Qinghua Ding
2017/03/31
the residual trend could be due to Co2 or cloudiness changes that are not related to the Z200 index.
the Z200 index has a strong interdecadal jump around 2003-2005. In our 2014 paper, we have argued that this interdecadal jump is not totally due to the Co2 rise.
So our regression can only remove this interdecadal like jump in the forcing fields. Some slowing increasing trends, as those due to Co2 forcing, cannot be removed by this regression.
-----------
Within the uncertainty of indirect aerosol effects and the suppression of tropical water vapor expansions, the interdecadal switch from long-trend decreases in SO2 emissions to a global jump in emissions correlates well to this transition period as shown in the image below.
http://iopscience.iop.org/1748-9326/8/1/014003/downloadFigure/figure/er…
soruce: The last decade of global anthropogenic sulfur dioxide: 2000–2011 emissions
Klimont et al. (2013)
http://iopscience.iop.org/article/10.1088/1748-9326/8/1/014003
note: this is an addendum to my post directly above.
It should also be noted that the increase in emissions on a global scale during this time can be entirely attributed to a rapid expansion of high-sulfur coal as a primary fuel for Chinese electric power generation. These hi-temp upper Tropospheric SO2 loading are not treated differently in the models, but observational evidence points toward a much greater indirect/cloud effect, upper tropospheric cooling and GPH impact in the pacific tropics as compared to low-temp coal combustion.
it should be noted that subsequent to the SO2 inventory study posted in #98 satellite analysis by NASA identified 39 additional emission sources that were not previously known, This lifts the overall SO2 emission rate for this period by 12% above that in the study cited above.
http://in.reuters.com/article/nasa-pollution-idINKCN0YO1XJ
This paper went to publish yesterday.
http://onlinelibrary.wiley.com/doi/10.1002/2017GL073138/abstract
North Pacific 20th century decadal-scale variability is unique for the past 342 years
Williams et al. (2017)
quote: "Furthermore, what has been assumed as natural variability in the North Pacific, based on 20th century instrumental data, is not consistent with the long-term natural variability evident in reconstructed SSTs pre-dating the anthropogenic influence."
> qinghua ding
> 2017/03/31
>
> Anyway, it is fine that you don't accept my explanation.
> But my calculation shows that there is still 30-40% of the
> trend left in JJA.
> This is my last post.
> finalize this conversation in a RC post as William suggested
Please do follow up.
Peer review before publication is a low bar to accomplish.
Reviews after publication are the real test of the work.
Thanks to 'Steven' at the ASIF, I realized that I don't need to do a full ERA data download and a gridded regression analysis to test how much of the trend in summertime temperature and LW is taken out by Z200 changes if we apply the Ding et al 2017 regression method.
We can get a pretty good estimate by using the data from fig1c and run the numbers.
I did that today, and indeed it shows (just like Qinghua already told me) a 35% residual trend for LW. Temperature shows a similar (35%) residual trend.
So, even though I find it difficult to accept that 65% of the trend in summertime Arctic temperatures would be caused by 'atmospheric circulation' (natural variability), leaving only 35% for AGW, I do understand the mistake I make and I (humbly) apologize to Qinghua and Eric for my arrogant stand in the issue.
Hi Rob,
I really appreciate that you said that. I also learned something from you. The paper is very condensed because we submitted it as a letter. So we couldn't cover everything in the main text. But I think we should include more description of Exp6 in the method part given its important role in illustrating our main conclusion.
In addition, we argued in the paper that 30% of the Z200 change over Greenland could still be due to anthropogenic forcing. So combining 60% and 70% (100% minus 30%) together (60% *70%) eventually gives us that about 40% of the sea ice trend since 1979 could be owing to a natural source. As I said, this is a very condensed letter-style article, which leaves almost no room for us to discuss this natural variability in detail. I think we will do better in the future if we would publish anything on a similar topic.
.
Again I really appreciate your interests and comments on our paper.
Have a nice weekend!
Best
Qinghua
Steven, I really appreciate your help in sorting this out. Many thanks!!
@ Rob Dekker
I told you before, see #62
"So, even though I find it difficult to accept that 65% of the trend in summertime Arctic temperatures would be caused by ‘atmospheric circulation’ (natural variability),"
Because it isnt, the Paper cant claim that in confidence because their Models they used, have to much ice, a to strong Aplitude and to low trend in the hindcast. As i shown before (#77) the model with the most realistic ice-state until 2005, also can explain what we have observed since 2005. If anyone claim, that Models would represent the forced response, it would mean, near all ice loos is due forced response
This entire discussion has reminded me of my dealings with Steve McIntyre and his followers, and in multiple ways. It's really worth asking ourselves, as a community, how we're any different when it comes to insinuation about motives and strongly-worded claims about facts. I readily admit I've gone too far myself on some occasions. Looking in the mirror is never a bad idea.
One key difference here of course: Rob Dekker apologized to Qinghua and me. Well done. Classy. Thank you.
It's been quite interesting to follow the back-and-forth among Rob Dekker, Jai MItchell, Qinghua Ding and Eric Steig. Eric most recently said
I presume I've missed portions of "This entire discussion", because many of the comments here have seemed both substantive and remarkably civil. There are a few of the random-DK-afflicted-party-on-the-Interwebz sort, but don't seasoned people like Eric shrug those off?
[I can't say I appreciate being compared to McI; unless perhaps the connection goes through QD and Soon -W]
Quoth Hank:
Not being a professional scientist, I'm unclear on where post-publication review is expected to take place, although I'm pretty sure it's not climateaudit.org. If someone could clue me in I'd be grateful.
Mal - you need to go scan some of the comments over at Neven's - where this all originated.
jai mitchell - "This is the most egregious example of placing the entire world at risk through protecting personal self-interests that I have ever seen from reputable scientists (read: not paid climate deniers) and shows that, just because you are a scientist, you are not necessarily a good or even a moral person."
epiphyte - "Brilliant. Perhaps they should take out the atmosphere altogether. Then they could observe that daytime would be hot enough to boil water, and at night it would be cold enough to freeze CO2.
...and humans would plainly have nothing to do with it!"
Jim Williams - "My vague sense in scanning it that it is crap, but the only thing dangerous about crap are the worms which might wait within it."
AndrewB - "And Rob, since you asked this question "I wonder what the authors were thinking when they drew their conclusions(?)", my answer is that they already had the preposterous claim that the disappearance of sea ice is mostly due to natural variability in mind, and just found the models and performed the simulations (which they call "experiments") that would somehow justify their a priori conclusion."
and many, many more .... sigh.
Eric #105 - I'm currently attempting to extract an apology from Steve McIntyre on Mike Mann's behalf. I don't anticipate announcing success any time soon!
https://twitter.com/jim_hunt/status/848782005068324864
Kevin #107 - Qinghua Ding has recently been commenting at length on the ASIF. A summary at:
http://GreatWhiteCon.info/2017/03/is-arctic-ice-loss-driven-by-natural-…
Jim - yes, I obviously follow the topic thread. Interesting to note the two-faced reception he received from those that were calling him and his co-authors everything name they could think of up until his actual appearance in the thread.
This little incident has caused me to lose all respect for most of the commenters in the forums at Neven's.
Does two-faced reception followed by refusal to see any need to apologise and also repeat initial denigration make it three faced or four faced?
Sigh!
[Pardon? -W]
I am talking about the thread at Nevens and expressing exasperation that it is impossible to get some people to play nice.
[Oh, sorry. That's all right then -W]
Crandles - I would have +1 ed your comment in the forum suggesting apologies were needed. But by the time I read it Neven had already called for an end to the apology requests.
And I was not surprised that jai mitchell and AndrewB basically doubled down on their earlier stupid statements. You don't make the kind of comments they made in the first place if you have a shred of decency or common-sense.
I would have said something/more earlier in thread but for Neven effectively saying stop it and move on.
Not sure whether to think Neven is being sensible about there being no chance of getting apologies so better to move on or be irked that he wants to brush it under the carpet rather than deal with it.
I know that I've burned up much of my credibility during the recent exchange with Qinghua Ding and Eric Steig.
So I don't expect anyone to take the following notes seriously.
[Don't be too hard on yourself. It *should* be hard to poke holes in stuff like this; after all, they're pros and have spent months / years working on it. By comparison your resources are limited -W]
However, the subject is important.
If Arctic sea ice declines to record levels, this paper suggests that 30-50% is due to natural variability, and thus that it could revert if the "weather" changes.
Next, I can't deny that I believe Ding et al 2017 still may have some serious flaws which were summarized best by Michael Mann :
https://twitter.com/MichaelEMann/status/841362467603255298
"study doesn't support that conclusion. Atmospheric circulation changes may have anthropogenic component."
[This feel like my objection; that it doesn't feel right. I notice that Mann is only tweeting. And still no RC post on the paper -W]
Let's take this one step at a time.
Ding et al 2017 has a claim that :
This claim consists of two parts :
1) Cause :
Here, we present evidence that trends in summertime atmospheric circulation may have contributed as much as 60% to the September sea-ice extent decline since 1979.
Now, this claim is based on the Exp.5. and Exp.6. differences.
Here, Ding et al 2017 runs a model between two 'climate' regimes : One with the ERA climate data, and one with ERA data adjusted to exclude the Z200 geopotential height trend since 1979, and also adjusted 'temperature' and 'LW downwelling radiation' and 'humidity' variables, which are adjusted to the extent that Z200 changed. Their regression method causes these variables to knock out 65% of the long term trend. This all suggests that Z200 long term changes are the "cause" of the trend in temperature, humidity and LW downwelling radiation. But it may very well be that the "cause" is the long term trend in 'temperature' which may very well be (entirely or partially) anthropogenic (see Mann's note).
Which brings us to 'attribution' :
2) Attribution :
Ding et al 2017 suggests that 70% of the 60% ice loss from (1) is caused by natural forcing, leading to their final conclusion of some 40% natural cause for ice loss in the Arctic. [Edited per request -W]
This (70%) is based on their experiment 7 and 8.
There, they took out the long term trend in 'high latitude wind', assuming that was anthropogenic.
However, the 'high latitude winds' may very well be caused by increase in Z200GL geopotential height increase. After all, increase in geopotential height over the Arctic leads to reduced cyclonic winds. If you assume that Z200GL increase is largely (or half, according to Ding et al 2014) caused by natural variability, then these long term changes in 'high latitude winds' that cause 30% ice loss may be caused by 'natural variability' and the remaining 70% may be caused by anthropogenic forcings.
Which turns the tables around since now 'natural variability' only causes 30% of the 60% (from point (1)) and thus only 18 % could be caused by 'natural variability'.
Needless to say that there is much to argue about this paper's findings.
[It would be interesting to see some "proper" scientific responses. But that may take a while. And if the response is, essentially, to ignore it that would take even longer -W]
Zach Labe points to a comment in NATURE CLIMATE CHANGE by Neil Swart on the Ding et al letter. doi:10.1038/nclimate3254
The heart of the comment is in these paragraphs:
"The challenge, until now, is that there has been no clear understanding of the relative contributions of human-induced warming versus internal variability to the observed long-term decline in Arctic sea ice.
Ding et al. make a significant advance in this area by estimating the contribution of internal variability to the observed long-term sea-ice decline. Beginning with a statistical analysis, they show that observed September near-surface warming and seaice loss are strongly correlated with changes in the upper-level atmospheric circulation centered over Greenland. To determine the causality of these changes, the authors conduct a novel series of model simulations that show that the circulation change is indeed a driver of, not a response to, the sea-ice loss. Remarkably, their simulations suggest that the large-scale atmospheric circulation changes could be responsible for up to 60% of observed summer-time Arctic sea-ice loss since 1979.
If the circulation changes are caused by anthropogenic greenhouse warming (or other human or natural external forcings such as ozone depletion, aerosol emissions, or solar activity) this pattern of atmospheric change should emerge as a clear signature when averaging together many climate model simulations of this period. Averaging together many simulations effectively cancels out all random internally generated fluctuations seen in individual climate simulations, leaving behind the model response to external forcings, such as increasing greenhouse gases. Using this approach, Ding et al. find that external forcing accounts for very little of the observed circulation changes, and therefore attribute the changes predominantly to internally generated variability. The result is surprising, in that it attributes a multidecadal atmospheric circulation anomaly to internal variability. Typically, internal variability is most prominent at shorter timescales — from months to years — and accounts for less of the observed variability as one extends to longer and longer timescales. The simulation-averaging approach they use to make this attribution is common, but it relies on the assumptions that the models have been supplied with the correct forcing, and are faithfully replicating the real world response to that forcing — assumptions that are always open to question12. Nonetheless,multi-decadal-scale internal variability does exist, most often relying on the longerterm memory of the ocean. For example, tropical Pacific sea surface temperature(SST) variability has been shown to have a strong connection with atmospheric circulation anomalies over Greenland and Arctic sea ice . Thus, there is a plausible link between tropical SST variability and Arctic sea-ice decline, but firmly establishing this relationship would require further work. "
[Thanks. It is interesting that's really just what's been said here already. It isn't really the detailed analysis I was hoping for -W]
William said "It would be interesting to see some “proper” scientific responses."
Neil Swart wrote an article in Nature Climate Change on Ding et al 2017 :
http://www.nature.com/nclimate/journal/v7/n4/full/nclimate3254.html
It is being discussed in Neven's ASIF.
Interesting is Swart's graph from the Canadian Earth System Model version 2 model runs, which pretty nicely match observations, and thus do NOT have that discrepancy of the CMIP5 that Ding et al 2017 uses to argue that models mismatch observations and thus could be natural variability.
Swart also adds :
Can't argue with that...:o)
Qinghua Ding was kind enough to reply to my post regarding the 'attribution' experiment (Exp 7/8), in a private email, explaining this experiment in more detail :
It seems to me that if you want to tease out the 'anthropogenic' part of the warming trend, that you should not just eliminate the "CO2 forced wind changes", but you should also include the much more obvious "CO2 forced" factor : Temperature !
This is my main criticism of the Ding et al 2017 paper : At no point (neither in the 'causation' (Exp.5/6) nor in the 'attribution' part (Exp 7/8) did they even attempt to address the influence of AGW, the KEY 'anthropogenic' factor in this game.
I had cited the Swart comment - but my comment never made it out of moderation.
[For some reason WP decided you're a new person -W]
And Rob is misreading Exp-6.
Jim Hunt:
Heh. Wouldn't that yank the rug out from under Mark Steyn and his co-defendants?
Kevin, welcome.
Would you care to elaborate on where I'm "misreading Exp-6" ?
Oh, OK. Stoat is one of those post-publication peer review venues. PPPR is in progress as I lurk. Cool.
Discussion of this paper is on-going at Neven's ASIF, here :
http://forum.arctic-sea-ice.net/index.php/topic,1920.msg109608.html#msg…
Rob Dekker:
Thank you, Mr. Dekker. ASIF is a blog I visit only occasionally due to its relatively narrow brief. In this instance, I gather from above that the "tone" of the discussion there was unsatisfactory for some of the professional participants. On ASIF and even (or especially ;^D) on Stoat that might get called a tone troll, at least if it came from a random party on the Internet. IMO however, actual scientific peer review work better when norms of collegial courtesy are observed pre- and post-publication. Indeed I'm pretty sure that's how those norms arose.
To clarify, collegial courtesy isn't all that distinguishes scientific peer review from the ordinary Internet shouting match. There have to be defensible arguments, too.
Mal Adapted:
Quinghua Ding posted numerous times at length in that ASIF thread, with remarks like "So many interesting discussions!" and "Thanks for your comment." and "Many thanks for all your interesting/burning/critical/constructive comments and questions on our paper and my thoughts." and "Since most scientific discussions about our paper have been happening on ASIF..."
I guess there were some people that didn't like the 'tone' in the ASIF thread, but the only professional around did not seem to share that opinion.
Yes, Q.D. just ignored the shit being thrown around - such as this recent comment:
JaiMitchell: "There is no excuse for being [the lead authors] this far removed from reality. Either they are completely incompetent, unaware, overconfident in their supposed knowledge, don't care or are a fifth column element working against the greater good intentionally."
Note Rob D. has never even chastised or suggested morons posting stuff like this should apologize - in fact he's defended them. Sad.
Rob Dekker:
One supposes there may have been other professionals 'lurking' on that ASIF discussion, who were perhaps too disgusted to participate, or some who may have exited early. Too, Dr. Ding seems unusually diplomatic for a mid-career scientist, at least by the standards I'm acquainted with. Perhaps it's a cultural thing, or maybe it's just the company I've kept 8^}.
I lurk on ASIF, but I am both not a professional and don't like the tone of the discussion, so don't participate.
Kevin, I don't think I am in a position to "chastise" anyone anywhere.
Mal Adapted, yes, there always could be more professionals lurking.
Phil, yet you decided to participate here.
All three of you : If you don't like a particular comment on a particular web site, you can always choose to ignore it.
Also, I find it interesting that on the ASIF Qinghua proposed a "70 year" cycle in the Arctic climate, although there is very little evidence to support that.
And that in the Walsh reconstruction "The sea ice extent in 1938/sep is even larger than that in 1938/aug. "
Which is false.
[It is also a strange comment: way back then the values are unlikely to be accurate -W]
I don't participate here for the bad bee keeping. That's a bonus. Or the rowing, or the bad economics, or watching Brexit make Great Britain into Little England.
I've followed discussions on sea ice with Dr Connolley for decades. I value and respect his thoughts on sea ice. Even when I do not agree.
[It must be about time for another post... -W]
Rob Dekker:
Perfectly correct, Mr. Dekker. We can even ignore the whole blog, for as long as we like. Ain't freedom grand?
The same goes for the authors of a scientific article under post-publication peer review. Their profession offers a protocol that helps advance the goals of Science. Their preferred venues all limit "free speech" in some fashion, not from fear of public scrutiny or any totalitarian impulse, but simply because it's hard to advance the goals of Science in an Internet slapfight.
"It must be about time for another post… -W"
But not on sea ice. Sigh.
["Science is hard" said Barbie. But politics I can knock out in a moment -W]