The Devil’s in the Details

I am sure I’m not not the first to call Devil’s Kitchen a Fuckwit, and I am sure I’m not going to be the last.

However, I have developed a reputation for a “willingness to engage (politely!) with rightists“, and as DK has “politely” pointed out, I did not engage with the substantive topic of his last post on climate change.

First, a Fuckwit

To bring everyone up to date DK put together a post discussing the views of climate scientists, and the implications of a survey taken of their views on the actuality of climate change and the modelling of climate change.

I picked him up for utterly misunderstanding what the “airborne fraction” meant when referring to carbon in the atmosphere, but did not engage further. DK took the new report by Knorr to argue that, as the “airborne fraction” of carbon had been unchanged by 150 years, the amount of carbon in the atmosphere had not changed over the last 150 years, this is a massive misinterpretation.

A far better discussion of the airborne fraction is available here but briefly, it refers to the percentage of carbon emitted by humans which is absorbed by natural carbon sinks rather than staying in the atmosphere. The below graph illustrates this rather well. DK boobed and has retracted the claim.

Second, a Survey

The bulk of DK’s post however concerns something else entirely, and it is this I wish him to retract as well. He cites Bishop Hill and his reporting of the results of survey of climate scientists from 2008.

Most of the scientists responsible for creating the delusion still believe global warming is man-made and will be a crisis. We know this from an international survey conducted in 2008 by Dennis Bray and Hans von Storch. They surveyed 373 scientists who work for climate research institutes and appear in the climate journals that are controlled by the now-notorious Climategate gang. [LO: Objective reportiong PJM’s raison d’être]

Thirty-five percent responded “very much” when asked the following question: “How convinced are you that most of recent or near future climate change is, or will be, a result of anthropogenic causes?” On a scale from 1 to 7, with 1 being “not at all” and 7 being “very much,” 83 percent answered 5, 6, or 7. Only 1 percent said “not at all” and only 11 percent answered 1, 2, or 3. Answers to the question “How convinced are you that climate change poses a very serious and dangerous threat to humanity?” were similar.

However, the Bray and von Storch survey also reveals that very few of these scientists trust climate models — which form the basis of claims that human activity could have a dangerous effect on the global climate. Fewer than 3 or 4 percent said they “strongly agree” that computer models produce reliable predictions of future temperatures, precipitation, or other weather events. More scientists rated climate models “very poor” than “very good” on a long list of important matters, including the ability to model temperatures, precipitation, sea level, and extreme weather events.

The quote above, refers to Question 21 “How convinced are you that most of recent or near future climate change is, or will be, a result of anthropogenic causes?” and Question 22 on “How convinced are you that climate change poses a very serious and dangerous threat to humanity?” These questions clearly shows an overwhelming majority of scientific opinion in favour of the idea of  anthropogenic causes for climate change.

This is then contrasted with their views on the climate models from which their conclusions are drawn. There appears to be an inexplicable leap in the number of scientists who think modelling is adequate to the number who think that climate change is happening, is man made and is a threat to humanity.

This is something which has been described by Bishop Hill, PJM and Devil’s Kitchen as a “gotcha” moment. A proof, if you will, that the difference between the two figures is a palpable measure of the amount of “faith” used over science to reach their opinions.

In their own words, previous surveys conducted by Dennis Bray and Hans von Storch have before been badly misrepresented by climate change deniers. [1] The misrepresentation of their previous survey has led them to be more careful with their latest survey [pdf available here], however we can see that it has still thrown up interesting results.

There are a number of reasons to not jump to conclusions quickly. First of all is that the above survey is comparing apples with oranges. The information on modelling actually refers to a vast array of questions which for ease of reading I have included in a footnote [2] below.

Good, now you’ve scrolled down you will have realised that the above may not be the fairest of comparisons.

An example may help further elaborate why I don’t think this comparison is fair. It is mentioned that “more scientists rated climate models “very poor” than “very good”” in the above “gotcha.”

However, this wrongly contrasts what could be a specific and widespread dissatisfaction with, say, the ability of global models to predict precipitation over the next 10 years with a general and widespread mild dissatisfaction with scientific modelling leading few to rate it “very good.”

As you can see the above comparison is vague and is full of holes, I wasn’t able to find the full results from the Dennis Bray and Hans von Storch but the lack of specifics in the above quote should start ringing alarm bells. If the results are so unflattering why report so selectively?

Third, a Theory which doesn’t hold water

There is a further reason to doubt the above interpretation as well.

Opposite is the logo of the Cochrane Collaboration. It may seem odd to include the logo of a an organization which is dedicated to conducting meta-analyses of clinical trials in a post about climate change but it is relevant. From Ben Goldacre’s Bad Science:

The logo of the Cochrane Collaboration features a simplified “blogogram”, a graph of the results from a landmark meta-analysis which looked at an intervention given to pregnant mothers. When people give birth prematurely, as you might expect, the babies are  more likely to suffer and die. Some doctors in New Zealand had the idea that giving a short, cheap course of steroid might help improve outcomes, and seven trials testing this idea were done between 1972 and 1981. Two of them showed some benefit from the steroids, but the remaining five failed to detect any benefit, and because of this, the idea didn’t catch on.

Eight years later, in 1989 a meta-analysis was done by pooling all this trial data. If you look at the blobbogram in the logo you can see what happened. Each horizontal line represents a single study: if the line is over the to the left, it means the steroids were better than placebo, and if it is over to the right, it means the steroids were worse. If the horizontal line touches the big vertical “nil effect” line going down the middle, then the trail showed no clear indication either way. One last thing: the longer the horizontal line is, the less certain the outcome of the study was….

The diamond at the bottom shows the pooled answer: that there is, in fact, very strong evidence indeed for steroid reducing the risk – by 30 to 50 per cent – of babies dying of complications of immaturity.

I hope the comparison and implication is clear. Asking the doctors to look at the various mentioned above would not have elicited a response that the treatment was “good” or “very good,” the responses would, in fact, be somewhat similar to the attitude to various climate models in our survey.

But the combination of these results – or our climate change models – is greater than the sum of their results. This is what DK, Bishop’s Hill and PJM have all missed. This is the fatal flaw in their logic.

The gap between climate scientist’s confidence in the various climate models and their professed belief in anthropogenic climate change does not mean they have a “faith” in climate change in excess of the evidence.

The multitude of climate models used by the IPCC all contain flaws, some are more flawed than others. But dissatisfaction with climate modelling – and with so many questions  on it in our survey, it seems little wonder some areas of disatisfaction were identified – does not mean that you cannot be convinced by a large number of complimentary studies.

Forth, a Fuckwit again

The above information and methods are not difficult to come by. A brief stint on Google gave me a copy of  the survey questions of Dennis Bray and Hans von Storch, and a cursory glance showed me that the above comparison is not fair.

Those active in the blogosphere and blogging on science are surly aware of concept of meat-analyses. Although our climate models have not and cannot be subject to a meta-analysis, the concept should not be alien that a large number of small studies – or models – can produce a result greater than its components.

As Tamino has illustrated there really are divergences between models used by the IPCC which may lead to divergences of scientific opionion. But to use the above survey to pour scorn on the whole scientific community is dishonest.

After the leaking of the CRU, Devil’s Kitchen lamented the standards of investigation and quality control of the scientific community. In this posting on climate change he has shown none of the rigour he demands of others.

I try not to propagandise here. I believe there is a time to propagandise but blogging on science is not that time – especially not when you have swearily demanded higher standards from others.

So to conclude, it is you Devil’s Kitchen who is the “Disingenuous Fuckwit of the Day” and the last 1500 words explain exactly why. I make that two Fuckwit Awards for the same post – with 5,376 still to look through I’m sure you’ll have more than a few in the post.

[1] “Deniers” is the prefered term when outright misrepresentation rather than sceptical analysis is the method used, I think this distinction is fair.


  • 12 on atmospheric modelling
  • 13 on oceanic modelling
  • 14 on the combination of atmospheric and oceanic models
  • 15 on the current state of scientific knowledge of various mechanisms involved in climate modelling
  • 16 on the ability of global models to:
  1. reproduce temperature observations
  2. reproduce precipitation observations
  3. model temperature values for the next 10 years
  4. model temperature values for the next 50 years
  5. model precipitation values for the next 10 years
  6. model precipitation values for the next 50 years
  7. model sea level rise for the next 10 years
  8. model sea level rise for the next 50 years
  9. model extreme events for the next 10 years
  10. model extreme events for the next 50 years
  • 17 on the ability of regional models to:
  1. reproduce temperature observations
  2. reproduce precipitation observations
  3. model temperature values for the next 10 years
  4. model temperature values for the next 50 years
  5. model precipitation values for the next 10 years
  6. model precipitation values for the next 50 years
  7. model sea level rise for the next 10 years
  8. model sea level rise for the next 50 years
  9. model extreme events for the next 10 years
  10. model extreme events for the next 50 years
  • 18 on How relevant is the study of paleoclimatology to the understanding of both climate sensitivity and anthropogenic induced climate change.
  • 19 on How would you rate the ability of paleo models to reproduce proxy temperature observations or proxy precipitation observations.

7 thoughts on “The Devil’s in the Details

  1. The other fact about climate science is that it is no way only driven by climate modeling, there are a wide array of observational disciplines which people dedicate their lives to. these fields give invaluable evidence of the presently occurring changes and can compellingly relate them to past climates which we have a better understanding of. modeling is just one way of looking at the problem. And scientists who aren’t involved are often rightly, considering all the assumptions which must be made, cautious of its results.

    Even if a scientist isn’t convinced by the modeling of future or past climates the physics behind climate change is simple enough to be comprehensively understood by a bright A level student and this is what, in my experience, even the most skeptical scientist won’t argue with. modeling is attempting to quantify how such a vastly complex system will cope with the inevitable change in the balance of that system.

    Models aren’t perfect but if you read just a few chapters of the IPCC report its striking how so many institutions with so many different priorities and methods come to widely the same conclusion.

    1. Thanks for the contribution.

      Its often said that modelling is the only “proof” proper that global warming is happening. “If you can’t model it, it ain’t science,” some would say. What would you say to that?

  2. I imagine it depends on how comfortable you are with the simplifications made in models. It is very easy to make a simple or even moderately complex numerical model which demonstrates changes in the global temperature in response to changes in radiative forcing. But as with all complex systems I can think of to have any chance of modelling it you must parameterise certain aspects of the system, this is where a bit of art works its way into the science-knowing what you can safely and intelligently parameterise!

    I would disagree with the premise that if you can’t model it aint science. Just think of the attempts to model the human brain, this is a far more constrained system but any attempts to produce a comprehensive model of the brain fail but I doubt you would argue that a specialist couldn’t identify and predict possible imbalances in the brain dangerous to the patient.

  3. There might be global warming or cooling but the important issue is whether we, as a human race, can do anything about it.

    There are a host of porkies and not very much truth barraging us everyday so its difficult to know what to believe.

    I think I have simplified the issue in an entertaining way on my blog which includes discussion on the CO2 issue.

    Please feel welcome to visit and leave a comment.



    PS In my country a porky is not a fat person but refers to a statement or assertion of gross falsehood or extreme exaggeration.

  4. Left Outside,

    A very pretty analysis that is missing only one thing—proof.

    Where is the response data that proves all of your assertions?

    You are arguing on precedent and by analogy (and I know that you must have been desperate to crowbar Ben’s stuff in, but it isn’t really analogous): you actually have no proof that any such manipulation has happened with the 2008 survey.

    I linked to an article, giving an analysis of the response data (which I have also not seen): as always with such things, my comment was on the assumption that the article was substantially correct.

    It may not be.

    But that doesn’t alter the fact that your above analysis contains no facts about the survey results under discussion.


    1. Its called occam’s razor, the above explanation is simpler than your own.

      I also seem to think you’re shifting the burden of proof somewhat as you were the one who approvingly reported what seems to be a somewhat dubious interpretation of the above survey.

      This may not contain details of the survey results, but it does contain information on the survey structure, which casts serious doubts on the interpretation offered by you PJM and Bishop Hill.

      By bringing in Ben’s work I just hope to illustrate how exactly a combination of less than perfect models can produce something greater than its parts in the eyes of climate scientists. As Dan above says, there are other observational disciplines which compliment the models as well.

      Neither of us at this moment have all the evidence necessary to discuss this further, but I’ll get involved. You should get your retractions written up in advance, yeah?

Comments are closed.