The Third Referee Waits In The Wings
ERL has now published the comments of the second referee for Bengtsson's rejected paper. After pointing out that the authors used the wrong equation, calculating the smaller transient climate sensitivity when they meant equilibrium climate sensitivity and used the wrong, well inappropriate, units for the climate sensitivity, K/(W/m2) instead of K for same, the second referee piles on, concluding
The study would be much more valuable if it attempted to also begin to address the four questions posed in the conclusions. I suspect the answers are really quite mundane, although the tone of the discussion implies otherwise.And, oh yes, the ERL chief editor, Dan Kammen, is mad as hell. Graham Readfearn interviewed him, eliciting
He said the recent news coverage appeared to be an attempt to publish research “via the media” after it had been rejected through the academic peer review process.
He pointed out that even though Bengtsson’s paper had been rejected by ERL, “they are free to submit the paper elsewhere”.We at Rabett Run, await the review of the third referee
Posts
50 comments:
The "K/Wm2" in your first paragraph should be K/(W/m2).
I don't understand the referee's first comment, about this unit. K/(W/m2) is commonly used in climate science, including many of James Hansen's papers. Perhaps the referee thinks it's not appropriate for ERL, but the difference (between this unit and just K) starts to matter when the temperaature difference delta-T is large, so that the climate sensitivity isn't a constant across its range. The climate sensitivity also depends on the temperature you start at, too, e.g. climate sensitivity at this point in time probably isn't the same as it was during, say, the PETM when surface temperatures were signficantly higher, or what it would be for a snowball Earth. (Or what it might be in 200 years.)
Thanks David, corrected. ECS is commonly the temperature rise if CO2 is doubled. You could define a sensitivity as Bengtsson did but it would be a slightly different thing.
This reminds Eli of his organic colleagues who don't understand spectroscopists using cm-1 as an energy unit
"It's always the third reviewer!" - except when it's also the first and the second...
But the change in CO2's radiative forcing for a doubling of CO2 depends on CO2's and temperature's starting value. Climate sensitivity is defined as
DT/DA = -(DG/DA)/(DG/DT)
= -S (DG/DA)
where "D" is a partial derivative, G is the net TOA flux, A is whatever parameter you're talking about (here CO2), and S is the climate sensitivity for A.
To get the ECS you'd integrate this equation, so S as a function of A and T matters if the range is large enough.
Anyway, perhaps unnecessary for ERL, if all you're talking about is the first doubling of CO2 from 1850. But there'd be some error for the second doubling.
Anyway.
David Appell: But K/(W/m2) is also defined for a single point (the climate sensitivity at current concentrations): climate sensitivity as usually stated for a doubling of CO2 is just going to be larger than K(/W/m2) by a factor equal to the number of watts for the CO2 doubling.
But in any case, the key issue is that you need to compare like to like: if you are discussing results that are measured in CO2 doublings, you shouldn't compare to K/(W/m2) and vice versa.
-MMM
MMM: I'm making a mountain out of a molehill, I'll admit.
You're right, in that climate sensitivity (call it S) itself is different at every radiative forcing point. It has units of K/(W/m2), as Bengsston and other scientists use.
So S for T=14 C and CO2=300 ppm is probably not the same as for T=20 C and CO2=900 ppm. (Which raises questions about using S as determined from paleoclimate information; it's not necessarily the same today as it was then, if the periods didn't have similar climates.)
It could even change significantly across a tipping point.
ECS is the integral of S*dF (F=forcing), from some value CO2=C0 to CO2=2*C0.
If S is nearly a constant from CO to 2*CO you can take it outside the integral so delta(T)=ECS=S*delta(F), and that's what we commonly think of when we say "climate sensitivity."
But over wide ranges of T or CO2 you can't take it outside the integral, especially if there's a tipping point in those ranges, and it makes sense to distinguish between S and ECS.
Perhaps that was Bengtsson's point. Or perhaps not, if the rest of the paper had so many flaws. Or perhaps I'm full of it.
I thought the referee's point about the units was essentially trivial. If you use the term ECS, then that refers - I think - to the Equilibrium temperature change when CO2 has doubled, hence it has units of K. On, the other hand, climate sensitivity could be presented as K/(W/m^2) and then that can be used to determine the ECS (by multiplying by 3.7 W/m^2).
I haven't seen the paper, but I assumed that it had simply confused ECS and climate sensitivity - or maybe I'm full of it too :-)
Maybe the authors will kindly post the paper somewhere,
"Maybe the authors will kindly post the paper somewhere,"
Post it at least, if not kindly.
Finally I realized what I really wanted to say:
Climate sensitivity (S), and so ECS, depend on the level of CO2 that you're at, and that you're going to.
So units of "K" can be simplistic.
Seems that the New Climate Scam is moderating:
Din kommentar inväntar granskning.
Se även:
http://rabett.blogspot.ca/2014/05/the-third-referee-waits-in-wings.html
Var kan vi hitta det papper?
Tack!
We'll see if Herr Bengtsson will make the paper available.
ATTP and David. Yes, and the referee really enjoyed sticking the knife in.
Climate Dialog has a post on exactly David's point
http://www.climatedialogue.org/climate-sensitivity-and-transient-climate-response/
Hmmm,
Seems my comment did not pass moderation:
http://www.klimatupplysningen.se/2014/05/18/nagra-tankar-om-klimatet-och-var-mojliga-framtid/#comment-372840
Ja, men RC måtta.
David Appell,
The referee was right. The AR4 glossary says:
"In IPCC reports, equilibrium climate sensitivity refers to the equilibrium change in the annual mean global surface temperature following a doubling of the atmospheric equivalent carbon dioxide concentration."
That's a standard, and the referee should require adherence.
Nick,
That sentence from the AR4 glossary is a definition specifically for "IPCC reports" and the very reason they included it is that there is NOT a single convention for expressing climate sensitivity.
Without seeing the paper in question, it's simply not possible to conclude that the referee was justified in the criticism.
If the authors were actually comparing ECS in K/(W/m2) to ECS given by IPCC and others in K without taking into account the 3.7 factor for CO2, then that would be wrong (worthy of criticism).
But if they were making no direct comparison to other results that expressed ECS in "degrees K" (not making an apples to oranges comparison), the units they used are not only fine, but are actually the most general(as pointed out by David above)
Great commentary by Graham Wayne.
"Lennart Bengtsson was doomed, no matter how his appointment was received, or what he did after he discovered the alarm with which his foolish indulgence was greeted. It is not unreasonable for the scientific community to feel betrayed by such a prominent scientist, since he allied himself with a group whose actions are, at every turn, antithetical to good science or responsible reactions to what it tells us. The utterly predictable way Bengtsson was used by the GWPF, the media and on the Internet was callous, manipulative and tawdry, yet it was his own caprice that brought this about, and I doubt if many have much sympathy for him or his self-inflicted wounds. Out of his ivory tower, Bengtsson became cannon fodder in a war he seems barely to have understood, and the collateral damage is widespread."
"the collateral damage is widespread"
outside climate blogs, I'd bet that the public has never even even heard about Bengts&son so the idea that the "collateral damage is widespread" seems to be a bit of a stretch.
But bloggers like Wayne are free to dream.
Is it not a little ironic to see Judy's tribe of pseudo-skeptics fall over each other to validate Lewandowsky? Why, a pseudo-skeptic might even speculate if this was part of Lewandowsky's evil plan all along, and if Bengtsson was either his willing accomplice or unwitting victim, sacrificed for "the cause."
Taylor B
captcha: malndsmo nature
'Anonymous said...
"the collateral damage is widespread"
outside climate blogs, I'd bet that the public has never even even heard about Bengts&son so the idea that the "collateral damage is widespread" seems to be a bit of a stretch.'
Another example of those 'careless cut [1] & paste fingers' so beloved of the anymouse, let us see what came next:
'The real damage, however, is the way this story has provided a handy smoke-screen, covering up several crucial issues that were worthy of far more discussion than this sorry mess. While Bengtsson’s resignation and a follow-up story about his work being ‘suppressed’ (that even Bengtsson was quick to refute) dominated even the front page of the UK’s Times newspaper, two unfolding stories of enormous import were drowned out, if you’ll forgive the pun.'
The collateral damage indicated is the continuance of the populace at large being subjected to bafflegab. Baffleqab at a level which can confuse even really clever people. I am not sure if this particular anymouse qualifies with that latter. Bafflegab that ensure adequate measures for mitigation are delayed further.
[1] copy & paste if being pedantic, but see 'Propositions and Dots' for the quote.
"While Bengtsson’s resignation and a follow-up story about his work being ‘suppressed’ (that even Bengtsson was quick to refute) dominated even the front page of the UK’s Times newspaper, two unfolding stories of enormous import were drowned out,"
Oh, my. How earth-shattering.
I could feel the reverberations all the way across the Atlantic.
To say nothing of the tsunami that followed.
FWIW, in the future, I promise to cut and paste stupid quotes in their entirety rather than just the stupidest parts.
Interesting recaptcha: "BRITISH gythose"
For some backstory, see Did Lennart Bengtsson Know GWPF And Heartland Institute?
"I don't understand the referee's first comment, about this unit. K/(W/m2) is commonly used in climate science, including many of James Hansen's papers. Perhaps the referee thinks it's not appropriate for ERL, but the difference (between this unit and just "
I have to agree with Appel.
Moreover it is not the kind of mistake that is damning. In fact none of the mistakes are damning. All can be corrected.
The two reviewers appear to agree on one thing.
The important questions didnt get answered.
Maybe they cant be.
Steven:
"Moreover it is not the kind of mistake that is damning. In fact none of the mistakes are damning. All can be corrected."
In which case, of course, the authors could've corrected the mistakes and resubmitted, rather than one of them raising holy hell over it.
D'oh.
Steven, the proposed revisions that both reviewers indicate would make the paper potentially publishable would have required a complete rewrite and a lot of new analyses.
As reviewer #2 makes clear, the interesting part would be if the authors actually tried the answer the 4 questions they had in the conclusion.
"But if they were making no direct comparison to other results that expressed ECS in "degrees K" (not making an apples to oranges comparison), the units they used are not only fine, but are actually the most general(as pointed out by David above)"
No, it's not a matter of converting units. ECS, properly expressed, includes information about the forcing implied by doubled CO2. 3.7 isn't a SI conversion factor. It's a modelled result, which doesn't have a standard value.
Mention of the third reviewer (still in the wings) reminds me of the "Peer Review Circa 1945" YouTube video. Bunnies who have never seen it may, or may not, find it hilarious.
"trurifi This"
Ach! I habe bemerkt -- ur, I just noticed -- that our host linked the same video to the word "third" in his original post. Ich bin ein Dummkopf!
Nick,
You are arguing in circles.
The IPCC itself has adopted "an RF of +3.7 W m–2 for a doubling in the CO2 mixing ratio. "
Nick Stokes said...
The referee was right. The AR4 glossary says...
Yes, ECS has units of temperature. But it isn't a constant; it depends on both temperture, CO2, and potentially other climate factors.
So citing a single number for it is misleading; it leaves out information that could matter. That's why scientists usually cite the derivative version.
Anonymous said...
The IPCC itself has adopted "an RF of +3.7 W m–2 for a doubling in the CO2 mixing ratio."
But the 3.7 W/m2 is a result from modeling.
A recent paper by Byrne and Goldblatt finds slightly different formulas (that imply this number should be 3.8 W/m2), that could be used in models. I gave their new equations here:
http://davidappell.blogspot.com/2014/03/what-if-co2-gets-really-high-say-1000.html
At CO2=1000 ppm, the new forcing is about 10% higher than the old one. That's about the uncertainty cited in the IPCC 4AR link you gave.
"That's why scientists usually cite the derivative version."
But the referee was objecting to LB citing ECS in K/(W/m**2). He was right.
But the 3.7 W/m2 is a result from modeling.
..which actually argues for expressing climate sensitivity in K/(W/m^2) as Bengtsson et al did rather than simply as K (as IPCC does)
The values given by "IPCC" actually assume this 3.7 factor, albeit explicitly.
But any scientist reading the Bengtsson paper should certainly be capable of taking a climate sensitivity to CO2 doubling given in K/(W/m^2) and multiplying by the 3.7 factor quoted by IPCC and then comparing that value to the values in the IPCC report.
It's really not that hard nor particularly confusing as the referee stated.
I should have stated above thus:
"any scientist reading the Bengtsson paper should certainly be capable of taking a climate sensitivity given in K/(W/m^2) and multiplying by the 3.7 factor quoted by IPCC and then comparing that value to the values in the IPCC report"
The values given by "IPCC" actually assume this 3.7 factor, albeit explicitly.
They don't assume anything -- it's a result from modeling. Each model can decide on its own what they will use for radiative forcing.
They might not even use any such equation, but instead solve the Schwarzschild equations at each grid level in the atmosphere.
But the referee was objecting to LB citing ECS in K/(W/m**2). He was right.
ECS has units of K, but citing a specific numerical value for it in units of K is not right.
I agree with Anon; anyone should be able to go back and forth from K/(W/m2) to K, at least in the simple case where climate sensitivity is assumed to be a constant.
David,
I am agreeing with you on your point about sensitivity.
By "assume", I meant they include the value in their calculations.
When they do so, they are effectively "assuming" that it is the "best" value to use (by whatever criteria they have used to come to that conclusion. Indeed, there are assumptions that go into the models used to come up with the forcing factor)
The fact that there are assumptions about the factor is the best argument for not simply expressing climate sensitivity in degrees K as IPCC and many others do.
Anon: Yes, I realize we agree. Sorry if I seem to imply otherwise.
Cheers.
http://retractionwatch.com/2014/05/21/gremlins-caused-errors-in-climate-change-paper-showing-gains-from-global-warming/
Yeah retractionwatch is interesting, and developing.
---Brief excerpt follows:---
... a correction was quietly posted on the journal’s website last week. The correction points out that the original paper concluded that “there were net benefits of climate change associated with warming below about 2°C”, but the updated analysis shows “impacts are always negative”.
This is significant because a more recent paper by Professor Tol ... contained many of the same errors ... was used ... in the report of the IPCC on ‘Impacts, Adaptation and Vulnerability’.
...
I have now drawn the attention of the IPCC to Professor Tol’s correction and suggested that it needs to amend the text of Chapter 10 before the final version is published later this year.
...
Tol, for his part, tells us that the errors don’t undermine his conclusions:
'Although the numbers have changed, the conclusions have not. The difference between the new and old results is not statistically significant. There is no qualitative change either.'
Update, 12 p.m. Eastern, 5/21/14: Tol added more comments in response to the LSE post at retractionwatch.
---end excerpt----
As Eli pointed out on Tol's Demon (a bag of wind if there were ever one), a major problem with the errors in Tol 2009 is that they echo down into many other studies.
It is another fine mess.
Tol:
"Although the numbers have changed, the conclusions have not. The difference between the new and old results is not statistically significant."
What Tol means is that the difference, while not *statistically* different, is *politically* different, therefore should not be changed.
The "it is good for us for a bit" argument was never *statistically* robust, but was good politics for the delayers/deniers like Tol.
Obviously, he does not want to give this up, and is using "statistical significance" as cover for political spin.
As far as I can understand the argument Tol is making on retraction watch (and I would point it out there, but I really can't be bothered and someone else essential has) is that when he used 14 or so results, it suggested a positive benefit for up to 2 degrees of warming, but the uncertainty intervals were large. Now that he's added 7 new results, the analysis suggests that benefits are negative for all future rises in temperature. However, the confidence intervals overlap those of the older analysis, therefore the two result are statistically consistent and therefore the result doesn't change! Now that is a remarkable use of statistical significance testing. That would appear to be embarrassingly nonsensical. Of course, if you're immune to embarrassment, maybe it's not.
I imagine that one could push this argument to the limit. Even if we get more and more results that suggest that benefits will become more and more negative, as long as the confidence interval of the new analysis overlaps with those of the previous (which overlaps those of the one before, etc) we can continue to assert that the evidence suggest positive benefits for up to 2 degrees of warming. Now that is clever.
Bengtsson guest posting at our place... http://uppsalainitiativet.blogspot.se/2014/05/guest-post-by-lennart-bengtsson-my-view.html
"imagine that one could push this argument to the limit."
Zero studies (no results) has effectively infinite uncertainty, which means warming *might* have a positive effect and because of the infinite "confidence interval", this is consistent with every other possible result.
TolED
Tol's logic seems to be fundamentally quantum mechanical*, for which two (or more) different possibilities can exist simultaneously until the climatefunction collapses (quite literally)
*which would make him a "quantum economist"
ATTP, tell him to toss his garbage and the really negative outlier on the other side and then see what is "signficant".
Note: Bengtsson's post has been translated to English at the Uppsalainitiativet and the comments are in English
Oh boy. From the Uppsala link.
"Because of chaos theory it is practically impossible to make climate forecasts, since weather cannot be predicted more than one or several weeks."
This strikes me as just plain dumb. What subject is he supposed to be an expert in?
Sounds like he may have Climatesheimers.
Comments here are too gripping for my handheld. It refuses to quit coming to this comment box.
Post a Comment