Bishop Hill Gavin Schmidt slams Skeptical Science
Apr 18, 2013 Climate: sensitivity Gavin Schmidt, a much misunderstood character in the global warming debate, has demonstrated his good faith and honourable intentions by issuing a denunciation of Skeptical Science.
Earlier today Gavin and I exchanged some tweets about the use of means and modes in climate sensitivity studies. Gavin's thoughts were as follows:
@aDissentient @micefearboggis Comparing the mode to previously reported means is a sleight of hand.
I was slightly confused at first, as I was unaware of anyone who had done such a wicked thing. However, having now read Dana Nuccitelli's post about Nic Lewis's paper at Skeptical Science I can now see that Gavin calls out scientific malfeasance whereever he sees it. Here's the relevant excerpt from the Nuccitelli piece:
One significant issue in Lewis' paper (in his abstract, in fact) is that in trying to show that his result is not an outlier, he claims that Aldrin et al. (2012) arrived at the same most likely [i.e. the mode] climate sensitivity estimate of 1.6°C, calling his result "identical to those from Aldrin et al. (2012)." However, this is simply a misrepresentation of their paper.
The authors of Aldrin et al. report a climate sensitivity value of 2.0°C [per the paper, the mean] under certain assumptions that they caution are not directly comparable to climate model-based estimates. When Aldrin et al. include a term for the influences of indirect aerosols and clouds, which they consider to be a more appropriate comparison to estimates such as the IPCC's model-based estimate of ~3°C, they report a sensitivity that increases up to 3.3°C. Their reported value is thus in good agreement with the full body of evidence as detailed in the IPCC report.
A sleight of hand indeed. I will not hear a bad word said about Gavin at my blog. :-)
Reader Comments (49)
I, for one, would like to congratulate Gavin as he wins yet another "Michael Mann Lookalike of the Year" prize - the beard, it's a winner.
Sometimes on looks for inspiration outside the field in question. And sometimes one finds it.
http://www.telegraph.co.uk/news/uknews/crime/10001149/Scientist-jailed-for-faking-medicine-test-results.html
And there I was on Twitter last Saturday:
Maurizio Morabito @omnologos 13 Apr
A sort of Guy of Gisborne character to Judith Curry's Marian.
"A sleight of hand indeed. I will not hear a bad word said about Gavin at my blog. :-) "
That doesn't leave me much room to coment then, because I refuse to discuss that website.
One swallow doth not a summer make.
Oh, dear Gavin. The selective pedant.
Our Gav
Is a very very very fine Gav
Life is a can of beans
When you 'dem modes and means
Now everything is easy
'Cause of you
Andrew
Excellent post Andrew. But as you point out our thanks are due above all to Gavin.
Not sure that Gavin should receive any medals and honours just yet. As Steve M might suggest, It is always a good idea to keep a very careful eye on 'The Pea' when dealing with "The Team".
Real Climate looks much the same as it ever did to me right now (18/4/13) ...
"Wrongness about Antarctic Ice by deniers - Eric Steig"
"GCM's may be s***e but regional models are now verified "
"Thin Ice — the movie (and I thought this was a Science site)"
"Should regional climate models take the blame?" (GCM's may be s***e but regional models are now verified?)
"Movie review: SWITCH"
"Climate Science from Climate Scientists" ... indeed.
???
I believe Gavin is speaking of _Lewis_ in this regard, who apparently did not report the commonly compared mean of his CS estimates:
------
Bishop Hill @aDissentient 16 Apr
Another paper finds that climate sensitivity is low http://www.bishop-hill.net/blog/2013/4/16/another-paper-finds-that-climate-sensitivity-is-low.html …
Gavin Schmidt
@ClimateOfGavin
@aDissentient What is the mean CS from this analysis? Mode is 1.6, but the more commonly compared metric of the mean is not mentioned.
4:53 PM - 17 Apr 13
Bishop Hill @aDissentient 16h
@ClimateOfGavin I'll ask, but why would you use a mean for a skewed distribution?
John Kennedy @micefearboggis 14h
@aDissentient @ClimateOfGavin No one number will capture the shape of the distribution.
Bishop Hill @aDissentient 14h
@micefearboggis @ClimateOfGavin Sure, but if you want to compare the central tendencies median or mode will tell you something. Mean won't.
Gavin Schmidt @ClimateOfGavin 10h
@aDissentient @micefearboggis Mean is the expectation - hardly a novel concept.
Gavin Schmidt @ClimateOfGavin 10h
@aDissentient @micefearboggis Comparing the mode to previously reported means is a sleight of hand.
Bishop Hill @aDissentient 10h
@ClimateOfGavin Who has done such a thing? @micefearboggis
Bishop Hill @aDissentient 10h
@ClimateOfGavin And strongly succeptible to influence of outliers. Not new either. @micefearboggis
Oh Schmidt!
Pity that the true believers won't realize you are joking, Bishop. But maybe you know that.
Nic Lewis did actually comment on Aldrin at WUWT a few weeks ago:
"Yes. Note that the paper quotes the mean estimated sensitivity of 2.0°C, but with a strongly asymmetrical distribution the mean is not a good central estimate. I cited, as stated, their most likely estimate (the peak probability density from their main results sensitivity PDF graph) – it is actually more like 1.55°C than 1.6°C. The extra forcing component mentioned is additional aerosol-cloud interaction forcing, which there is very little reason to think is needed.
This type of study provides what constitutes a standard observationally based (strictly, observationally-constrained) estimate in climate science. So far as the climate model involved goes, it is about as simple as you can get while retaining separate hemispheres (vital to constrain the aerosol forcing estimate). And I think Aldrin et al. did a thorough job. But the Bayesian approach is full of pitfalls. In particular, use of uniform (or expert) priors for climate sensitivity and/or effective ocean diffusivity will typically lead to climate sensitivity being overestimated and having far too long an upper tail. I have been trying to persuade the key IPCC lead authors involved of this, and that it is essential to use a computed noninformative prior (with a view to achieving probabilistic results that reflect objective measures of probability, not the standard Bayesian subjective belief). But I don’t think they really understand the issue properly – maybe they don’t want to either.
I have a paper that uses an objective Bayesian method to estimate climate sensitivity undergoing peer review."
Well spotted, your eminence.
Bit of a chuckle, that.
Going to get crowded under that bus...
See! No bad word, though the plural of word did tempt me.
Schmidt happens.
I would have thought it uncontroversial amongst grown ups that in a probability distribution function with one mode, i.e. one hump, the mode is actually the *one* thing you can understand visually without any further wordage. If you have it shown on a graph, and suitable scale applied, then there it is - waving at you saying "look at me my mode is here!" :) Even Tom Curtis on SKS said the same thing.
I haven’t seen Lewis’s actual paper but from his layman write up on BH he clearly only compares his mode against Aldrin’s mode. No mix up is possible there. So judging by the cute way SKS write it up above, and knowing their tendency for literally assembling others words in new ways to criticise, SKS’s good faith is the last thing assumed by me. I don't know why any sane person uses them as a reference.
The illustrative thing about episodes like this is that it again shows how the great climate science educator’s like to play games with words when it suits them.
Leopard - I cannot imagine anybody actually reaching such a level of misleading of the public. I prefer to think that Dana Nuccitelli simply (and as usual) has no clue on what he's writing about, doesn't know his mean from his mode, can't read a graph and in general is completely blinded and self-fooled by an unbounded desire to save the world, a desire compounded by just as boundless stupidity.
Apr 19, 2013 at 7:52 AM | omnologos
Mmmm yes but I would find it hard to imagine that a site referenced by many quite respectable scientists as source of climate advice could be manned by people who literally can't understand graphs at this level (I speak as a layman), but you may be right ;)
I won't speculate as to why they speak such rubbish (for fear of the ever circling Lewandowsky's pathologising) but however it happens, it happens.
BTW I have to congratulate the Bish for his bone dry wit here :) I was one of those initially led to think maybe that Schmidt had actually shown some good faith and so I started puzzling out the detail by following the back story (not paid much attention so far), the pay off in my investigation made the joke even funnier, thanks.
Schmidts words really are only applicable to what SkS has done :)
Is Gavin Schmidt poor at sums?
The IPCC asserts that all the warming since 1850, about +0.7°C and is wholly due to Man-made CO2 emissions. However there are other published and peer-reviewed opinions that differs on the actual level of the impact of Man-made additions of CO2 to the atmosphere. A well-accepted view was provided by CDIAC, the Carbon Dioxide Information Analysis Centre of the U.S. Department of Energy (DOE). (Their useful FAQ 2 giving percentages making up the Greenhouse effects seems to have been withdrawn within the last year).
The CDIAC figures can be transposed into parts of the 33°C total Greenhouse Effect are as follows:
Water Vapour ~95% of effect ~31.35°C
Greenhouse Gases ~5% ~1.65°C
Carbon Dioxide at 390 ppmv ~75% non H2O GHGs ~1.24°C
Natural CO2 ~86% (assuming 50% emissions since 1850 Man-made) ~1.07°C
Man-made CO2 2010 (50% of the 110 ppmv increase from 280 ppmv of 390 ppmv ) ~14% ~0.17°C
Other non H2O GHGs gases ~25% ~0.41°C
The CDIAC figure of ~0.17°C for Man-made influence since 1850 is roughly a quarter of the measured temperature rise of ~0.7°C: therefore it accords well with the notion that the temperature increase can be allocated roughly 50% solar influence and 50/50% natural / Man-made CO2 emissions. A value of rather less than 0.17°C for the effect of worldwide emissions to date has been accepted in correspondence with Professor David MacKay, the chief scientific advisor of the UK Department of Energy and Climate Change.
On the other hand, Gavin Schmidt of NASA GISS, one of the topmost scientists involved in the Catastrophic Anthropogenic Global Warming Cause, in his 2010 peer reviewed paper clearly lays out the following calculation:
Water Vapour and Clouds ~75% of effect ~24.75°C
Greenhouse Gases ~25% ~8.25°C
Carbon Dioxide at 390 ppmv ~95% non H2O and CO2 GHGs ~7.84°C
Natural CO2 ~72% 280 ppmv (assuming 100% emissions since 1850 are Man-made) ~5.63°C
Man-made CO2 2010 (100% of the 110 ppmv increase from 280 ppmv of 390 ppmv ) ~28% ~2.21°C
Other non H2O GHGs gases as above ~0.41°C
As the reported and acknowledged temperature increase since 1850 is only ~0.7°C in total, Gavin Schmidt’s assertion has to be in error. So this premier scientist supporting the concept of Catastrophic Anthropogenic Global Warming markedly exaggerates the Man-made effect by more than three times than the accepted rise since 1850 and estimates it at ~13 times the CDIAC Man-made temperature estimate.
Clearly neither Gavin Schmidt nor his peer reviewing colleagues carried out these trivial check sums before publication. Had they done so, they would have seen that these figures grossly exaggerate Man-made influence on temperature from CO2 emissions since 1850.
Nonetheless the alarmist Cause expects the Western world to revolutionise its economies based on this type of casual assertion and sloppy calculation. This is the type of trivial due diligence that seems never to be undertaken when radical and vastly expensive policies are formulated to address Catastrophic Man-made Global Warming. Instead inaccurate assertions of this nature are made regularly and have been widely accepted by governments.
This is just another example of the sloppy, deceptive and unscientific work carried out in the name of the “Cause” by NASA GISS, the group directed by Jim Hansen. NASA GISS is also one of the official repositories of global Temperature record. His organisation has been continually rewriting history, (1984 style) , over the last 30 years. Such retrospective and un-auditable revisions of the historic records have occurred particularly in the USA where the records of temperature earlier in the 20th century have been consistently lowered to emphasise warming. This has been a continuing process. NASA GISS also alter records outside the USA. One the most egregious instances recent being the past records in Reykjavik Iceland where NASA GISS have amended past values in spite of protest from the Icelandic Meteorological Office .
But these are the records and the climate experts that World Governments via the UN IPCC depend upon and on which the Western world is basing its self-destructive and costly policy decisions.
Leopard - Gavin is a Brit too. His words as I said can have multiple interpretations and that's why he will never appear a disaster such as Rahmstorf or even Hansen (Mann being in a different league).
I repeat: a careful analysis of Gavin's writing shows that he does utter sensible thoughts.
As for SkS, it reminds me of those web sites were people talk like doctors and put pictures like doctors, but aren't really doctors at all. Many unsuspecting idiots fall for it, and being a scientist doesn't mean one cannot be an idiot as well.
This episode springs to mind.
Apr 19, 2013 at 9:37 AM | omnologos
Yes interesting, I admit I haven't followed Gavin Schmidt's position too closely, I can speak of my sense I have sometimes picked up a more reasonable and rational vibe from him before in his dealings, though I suffer from the fact I may unfairly weight that vibe against associating him with RealClimate and Michael Mann ;)
You are right to remind me, I forget climate science isn't a mature science and nuance is something to look out for more there. I will pay more attention to what he says ;)
Just to add a few comments about what has been written on the "Climate Sensitivity Single Study Syndrome, Nic Lewis Edition" thread at SkS.
Dana Nuticelli suggests that I may be mistaken in describing what is estimated as "equilibrium climate sensitivity" rather than "effective climate sensitivity". This claim is rather dented by the treatment in AR4 WG1 of the sensitivity estimates in Forest 2006 and other observationally-constrained studies. The x-axis of Fig 9.20, where the F06 and other sensitivity PDFs were shown, is labelled "Equilibrium climate sensitivity" and in the text the estimates are described as being of equilibrium climate sensitivity (using the acronym ECS).
Actually there is a reason why I think F06 and my study probably did actually estimate "equilibrium climate sensitivity" rather than "effective climate sensitivity", but it is a bit technical and I won't go into it here. I'm doubtful that there is that much of a difference between effective and equilibrium climate sensitivity in any case. The Armour et al. 2012 paper that Dana Nuticelli cites in that regard seems dubious to me. Armour himself has accepted that the feedback pattern in the CCSM4 GCM he uses may be a bit of an outlier.
Tom Curtis, in a measured and otherwise quite accurate comment at SkS, wrote about Lewis 2013 that "It is misleading in that he does not point out the difference between the modal value and the central estimate commonly quoted for the IPCC, and that they are not directly comparable." But of course Fig. 9.20 in AR4 WG1, didn't provide the mean at all, it just showed the median and mode (the median being marked with a dot in the uncertainty range bar, the mode being visually obvious). And the accompanying text likewise discusses only modes and medians, describing the modes as being the "best estimates of the ECS".
But I thought Nuccitelli was Mann's (new) best friend..!
For the layman like me:
The mode is the value that appears most often in a set of data.
In statistics, mean has three related meanings:
the arithmetic mean of a sample (distinguished from the geometric mean or harmonic mean).
the expected value of a random variable.
the mean of a probability distribution.
Shevva - I think the latter two are the same thing.
Martin A
Not quite the same. For a Bayesian, probability represents our state of knowledge of a parameter. It does not have to be a random variable.
It is a circular definition though ... "the mean is the mean of a probability distribution"
Nic -
Thanks for the rebuttal to the SkS post.
We all assess sources (sites and authors/commenters) as to their reliability. After having perused several posts at SkS, my assessment was "not necessarily wrong." So I can't say that I am surprised at their response to your article, nor at the misrepresentation thereof.
I've got one further comment regarding the SkS thread. Karsten (I believe climate researcher Dr Karsten Haustein, of the Oxford University School of Geography and the Environment) wrote there about Lewis 2013:
"As aerosol forcing seems the most constrained parameter, I wonder what a small shift in the prior to a more plausible range of -1.5 to 0.0 W/m2 would do to the final results. Given that the F06 aerosol forcing mode isn't too far off the one in Lewis 2013, it seems that the method is very sensitive to the prior choice. I might very well be wrong in this regard."
Karsten is indeed wrong. Since the main results 95% upper bound on the estimated aerosol forcing is 0.0 W/m2, reducing the upper limit on the aerosol forcing prior from 0.5 W/m2 to 0.0 W/m2 would make very little difference.
"a much misunderstood character in the global warming debate"
That pretty mischievous for a man of the cloth. Keep it up!
Nic Lewis (1:16 PM) -
Karsten might also consult Libardoni and Forest 2011, which shows the posterior on aerosol forcing to be well-confined between 0 and -1.0 W/m^2, with almost no dependence on the surface temperature data set (unlike posteriors on ECS and diffusion).
I found its constraints on aerosol forcing to be equally interesting as those on ECS, as Hansen (among others) speculates that aerosol forcing is large and negative, and that the observations do not therefore rule out large sensitivity values. Of course, the MIT 2D model doesn't actually *prove* ECS to be ~2 K/doubling nor that aerosol forcing is ~-0.5 W/m2; but from an abstract perspective ("50,000-foot view") it's reasonably credible.
Fascinating. Montford can't seem to get over the fact that he got the basic premise of his post wrong here. Gavin was slamming Lewis, not Dana. And Lewis can't seem to get over the fact that Dana, in very short order of time, was able to identify Lewis' slight of hand. Very humorous to witness. Comments are reminiscent of the Keystone Cops.
Skepticalscience.com is a joke. Nuccitelli in the Nic Lewis critique states: "It's most important not to fall into the trap of thinking that any single study will overturn a vast body of scientific evidence..." Okay, fair enough.
Yet just last month he was gushing about Kevin Trenberth's single study where he claims to have found his missing heat and thus climate sensitivity is indeed high just like the IPCC claims. Yet, in that critique Nuccitelli states: "Those (recent low climate sensitivity) studies are fundamentally flawed because they do not account for the warming of the deep oceans."
Those guys have simply abandoned science and the unfortunate part is that is not easy to spot (well for laymen like me anyways).
Buzz
You should look up the word "humour" in a dictionary. If you need it spelt out for you, no, I don't think Gavin was slamming SkS. But he is obviously correct that comparing means to modes would be sleight of hand. It was unfortunate for Dana that he decided to do just that at the moment Gavin posted his tweet.
Now then Buzz, Gavin has made an unambiguous statement decrying the practice of comparing the mean to the mode as a "sleight of hand", and Dana Nuccitelli is the only person who has done that, see the clear example above.
I wouldn't be hard on Dana and call it "slamming", seems a bit extreme ;)
>I believe Gavin is speaking of _Lewis_ in this regard, who apparently did not report the commonly compared mean of his CS estimates
Chronicles of Narnia AND climate science. The man was a genius!
>Is Gavin Schmidt poor at sums?
Yes he is. See the RC post, the tragedy of climate commons. In it he relies on old data to say
top 5 catchers, catch 20% of the fish, and the top 20 catch almost 75% of the fish. Meanwhile the least efficient 50 catch only 10% of the fish
When I called him on it, that the US isn't catching 20%, and where are China and India in this example, he tried to backtrack by saying that he meant population, not countries. The numbers still don't work out as those countries are almost 20% by themselves, so he then says the countries are split up among the fishermen, which means his whole analogy is worthless.
You guys clearly have not read Dana's post. It is Dana who points out that Aldrin doesn't use the 1.6C figure that Lewis claims. Lewis is, therefore, pulling a sleight of hand by, oh so conveniently (and without stating it, which is what makes it sleight of hand), using a mode figure instead of the actual 2C mean figure used in the paper.
Buzz -
Nic Lewis is comparing modes to modes. From Nic's post,
This is not rocket science. Look at Aldrin's paper. What is the mode of their distribution?
Yes, Dana wants to compare means to means. Nic has stated a reason why the mode is to be preferred -- see Manfred's comment supra. Dana may not agree; that's his prerogative of course. But he is incorrect (as are you) in claiming that there is sleight of hand involved by Nic.
Buzz
We agree that Aldrin doesn't mention the mode. However, it can be seen on his PDF.
For ECS, the IPCC uses modes and medians and eschews means. It refers to modes as "best estimates". You can hardly criticise Lewis for using them too. The fact that Aldrin prefers to talk about means (which will be highly influenced by outliers) is irrelevant.
Strangely, Dana has tweeted that he has not compared mean to mode. When even his own commenters are saying he has done this, it's hard to credit the idea that he'd try to tough it out.
Mmmm, this is weirdly fascinating – I do hope this continues! In my travels around the links following this story I also noticed a guy call "thingsbreak" gleefully making this same point. The words 1.6 degrees don’t appear in Aldrin’s paper!
What are we to make of this? Ignore our lying eyes?
Call me a naive but I would say the thing is the thing, and just because the words don’t appear saying what that thing is, doesn't void the existence of any obvious and relevant important feature that can be seen and needs to be accepted and discussed.
I mean if I saw a paper that talks of experiments on a large mammal, with graphic photos showing the procedure, that nevertheless fails to mention that it is an elephant we see, would I be wrong to point that out? ;)
I mean should we apply a scotoma and not see Aldrin’s graph? Why did he put graphs in his paper giving away the shape of his pdf? Why not just the single mean?
This is a fascinating place to be at when you see the argument reduced to this level in science of all subjects – yeah you see it in politics and so on – but science!?
I can’t help coming to the conclusion that it is either a real cognitive impairment or some application of willful dissembling. There aren't any other descriptions for this behaviour. This is coming from the side who are "pro-science" arguing against "deniers" - I love it, you can't make this up :)
Despite the best care, it's inevitable that occasionally one will be caught out in a mistake. Even Homer nods. [And Dana is no Homer.] The character question is how one reacts. Errare humanum est, perseverare autem diabolicum.
Apr 19, 2013 at 3:53 PM | Buzz Fledderjohn
Ahh. when comedy was simple, how I miss those times (as you can guess, old,white & conspires to be a nut)
"The Keystone Cops (often spelled "Keystone Kops") were fictional incompetent policemen, featured in silent film comedies in the early 20th century. The movies were produced by Mack Sennett for his Keystone Film Company between 1912 and 1917. The idea came from Hank Mann who also played police chief Tehiezel in the first film before being replaced by Ford Sterling."
well Hank has his torch bearers, who still make me chuckle..
Dana has removed the bracketed mode/mean from his post at SKS without a word that he did....
sue - SkS is well known for editing comments/postings without comment.
@Apr 21, 2013 at 4:48 PM | sue
Do you mean the bolded square bracketed words in the quotes above? If so they are provided by the Bish describing what Dana has done. Personally I would be more specific and say Dana chose to say this:
When the Lewis paper says this:
My emphasis above.
Mmmm, now what has Dana dropped there? Is that a sleight of hand? ;)
Leopard,
My mistake then. I thought that when I first read this post that I clicked on the link to SKS and they were there also, but maybe I didn't.