October 05, 2008 | Graham

How good are climate models?

The only way General Circulation Models can produce catastrophic CO2-induced warming is to introduce positive forcings from other agents, such as water vapour. Without these forcings temperature increases are relatively benign. What most don’t understand is that the values attributed to these forcings are largely imaginary.
I’ve just come across two pieces of information which demonstrate the precariousness of the modelling assumptions of forcings. The first is courtesy of a post on Jennifer Marohasy’s blog. In a soon to be published paper Douglass and Christy eliminate all of the known forcings on climate from the El Nino Southern Oscillation (ENSO) and aerosols (produced by volcanoes) and demonstrate that the underlying increase in global temperature without these independent forcings is 0.062±0.010ºK/decade, which they argue is consistent with the IPCC’s formula for unforced CO2 which gives a value of 0.07ºC/decade.
Of course correlation doesn’t equal causation, but Douglass and Christy point out that if the IPCC assumptions of positive forcings are correct, then there must be other, as yet undiscovered, negative forcings which have acted over the last 30 years.
The other piece of information is these notes posted by Roger Pielke Senior. They are by Ann Henderson-Sellers, Professor in the Department of Physical Geography at Macquarie University, and until 2007, the Director of the World Climate Research Programme http://wcrp.wmo.int (WCRP) based in Geneva at the headquarters of the World Meteorological Organisation. She is one of the most frequently cited researchers in the world.
Professor Henderson-Sellers notes are of feedback from the IPCC lead authors on the AR4 IPCC report. I have reproduced the bullet-points below that refer to climate models.

Serious inadequacies in climate change prediction that are of real concern

  • The rush to emphasize regional climate does not have a scientifically sound basis.
  • Prioritize the models so that weaker ones do not confuse/dilute the signals.
  • Until and unless major oscillations in the Earth System (El Nino-Southern Oscillation (ENSO), Pacific Decadal Oscillation (PDO), North Atlantic Oscillation (NAO) and Atlantic Multidecadal Oscillation (AMO) etc.) can be predicted to the extent that they are predictable, regional climate is not a well defined problem. It may never be. If that is the case then we should say so. It is not just the forecast but the confidence and uncertainty that are just as much a key.
  • Climate models need to be exercised for weather prediction; there are necessary but not sufficient things that can best be tested in this framework, which is just beginning to be exploited.
  • Energy budget is really worrisome; we should have had 20 years of ERBE [Earth Radiation Budget Experiment] type data by now- this would have told us about cloud feedback and climate sensitivity. I’m worried that we’ll never have a reliable long-term measurement. This combined with accurate ocean heat uptake data would really help constrain the big-picture climate change outcome, and then we can work on the details.
  • [Analyse] the response of models to a single transient 20th century forcing construction. The factors leading to the spread in the responses of models over the 20th century can then be better ascertained, with forcing separated out thus from the mix of the uncertainty factors. The Fourth Assessment Report missed doing this owing essentially to the timelines that were arranged.
  • Adding complexity to models, when some basic elements are not working right (e.g. the hydrological cycle) is not sound science. A hierarchy of models can help in this regard.

This should give readers some idea as to just how uncertain the outputs from these models are. On another note, given the reservations about regional climate modelling, it makes one wonder what our government agencies think they are doing. This Google search on “gcm climate” turns up reference after reference to the CSIRO and various Australian governments making regional predictions about climate under enhanced CO2 conditions. According to Henderson-Sellers’ notes, they are wasting their time and our money.

Posted by Graham at 8:27 am | Comments (10) |
Filed under: Environment


  1. Three observations:
    1. If the weather forecast is for rain I believe it prudent to take an umbrella. Similarly if there is a possibility that by changing our behaviour we can reduce the risk of harmful climate change then we should do so.
    2. If we take action and there is no harmful climate change then we will have proved nothing. The sceptic will still be able to argue that our action made no difference. If we dont take action and the predictions are 100% fullfilled we will be too busy coping with the mess to waste our time worrying about who was right or wrong.
    3.Climate operates on geological time not human time – our observations are a blink in terms of geological time – our research about climate from ice cores and tree rings is at least consistent with geological time but we are still a long way from having an unambiguous profile of climate that is consistent with a human time frame.

    Comment by John Tons — October 6, 2008 @ 9:56 am

  2. “Energy budget is really worrisome; we should have had 20 years of ERBE [Earth Radiation Budget Experiment”
    If observation and radiation research does exist to the detriment of the Global Warmlings, why would acknowledged scientists release it at this time? while disproving the current fads are yet to play out in full. Hell!! why send in the cavalry now? there are so many still lost minds to captured and souls yet to be saved.

    Comment by Dallas Beaufort — October 6, 2008 @ 12:04 pm

  3. To me the take away message is as I’ve said before. Our understanding of complex relational science is at best still incomplete and consequently our models should be regarded as indicative rather than predictive let alone absolute.
    There are examples in biology, compute programming and physics where the whole appears more than the sum of the parts i.e. there are unknown or unconsidered (unknown) extraneous factors (eg Newtonian Physics etc.). Logic dictates that we can’t predict what we don’t yet know or understand. The bell weather indicator here is clearly the more complex the science/model are the more judgement, less ‘precise’ and therefore the less absolute the result. One could summarize by postulating “that the more complexity the less predictability” finding fault/weaknesses therefore is no surprise.
    At the basis ‘scientific’ reasoned debate both sides are based elements of truth, unknowns, and educated judgement hence we the ‘Great Unwashed’ rely on education, experience, and expertise of ‘experts’. Experts are human too with limitation, egos, biases and self interests.
    Notwithstanding the above and that I am at the thoughtful end of “GU”, I suggest that the real issue isn’t as myopically binary contest about which side is absolute. RATHER, what do we do?
    Playing academic/political one-upmanship strikes as akin to the Capt. of the Titanic while still in port engaging in statistical probability analysis of hitting an iceberg.
    Given the observable and logical evidence as to the degradation of the environment, an increasing population, the limitation of natural resources ‘business as usual’ isn’t a prudent or I believe defensible option. Particularly if we add the overtly complex and unpredictable dynamics of trade, financial system, economics and world wide political imperatives and their consequences.
    We don’t have a Hari Selton and his psychohistory skills (Isaac Asimov’s “Foundation” series). In the mean time I would put it to Climate change sceptics to offer alternative strategies.

    Comment by examinator — October 6, 2008 @ 12:24 pm

  4. Examinator, my intention in critically analysing how accurate the models might be, and whether we are over-relying on them is just so that we can get the most realistic estimates of risk possible. Many of these models are just “good stories” expressed in mathematical language, and unlikely to represent reality. One would be foolish to rely on them to make one’s decisions.
    One also has to take account of other potential catastrophes and ensure that the noise of one doesn’t drown out conversation about the others.
    So we should take much more account of the actual history of the world, and use a prudent adaptive approach to deal with the problem.
    If I find the time I’ll set out such an approach soon in a blog post.

    Comment by Graham Young — October 6, 2008 @ 10:30 pm

  5. Graham
    I agree your comment its one of the underlying concerns I have with modelling.
    • Like Tor Hunloe I have real sadness over the loss of the polymath (generalist) way of addressing issues.
    • To paraphrase John Donne,
    No ‘discipline’ is an island unto its self.
    Every ‘item/issue’ is a piece of the continent, a part of the whole.
    • I think it is too easy for specialists to get lost in their jargon and minutiae of statistical aggregations.
    • Aggregation is the end result not the driving force.
    • Many of the ‘criticisms’ particularly in the ‘lame street’ media are simply sensationism criticism for criticism sake and perhaps tendentious. Even those who that have ‘pure’ motives seem to be based on the false assumption that the majority of the readers are able or willing to comprehend the context i.e. ‘Global warming’ as such is a moot point but;
    a. The real issue is our impact on our world and the inevitable consequences if we continue ‘business as usual’.
    b. Can we afford to wait until the science is unequivocal?
    c. Why does the debate need to polarize (on either of two extremes)? There is at least a third way.
    • Concepts like ‘science will save us’ so all we need to do is develop “clean coal” for example is both tenuous and myopic the ‘conservation of energy’. All our ‘commercial’ science tends to do is ‘relocate’ the problem e.g. sequestration doesn’t solve the problem it simply ‘buries it’.
    It is this aggressive polarization into two opposing camps and therefore the need to therefore elevate aggregation to an immutable defining force that tends to bother me. To me balance is an overused term I would prefer objectivity any day.:-)

    Comment by examinator — October 8, 2008 @ 9:22 am

  6. For the sake of balance, here is another view:
    Or for those less inclined to the technical stuff, some questions and answers here:

    Comment by Q&A — October 18, 2008 @ 10:55 am

  7. How can they provide balance when they don’t even address the same question Q&A? What is it drives people like yourself to tour the Internet trying to side-track debate?
    The papers you have cited are to do with discrepancies between the observations and the models in terms of the warming of the troposphere compared to the warming of the surface. The authors claim to have solved this discrepancy. But none of the information I posted even deals with this issue.

    Comment by Graham Young — October 18, 2008 @ 11:22 am

  8. Graham
    How good are climate models?
    Many so called “deniers” lampoon the GCMs and often take them out of context to denigrate the theory of AGW.
    I agree totally with Professor Henderson-Sellers, it’s unfortunate that many others (from both sides) don’t really understand what she is saying.
    For what its worth, those links I provided came via email from a colleague researching coupled ocean/atmospheric systems.
    It seems again you want to pass aspersions on my character – more akin to playing the man and not the ball.

    Comment by Q&A — October 18, 2008 @ 2:23 pm

  9. Professor Henderson-Sellers also had this to say;
    I hope this puts her position in perspective to what Graham has posted above.

    Comment by Q&A — October 25, 2008 @ 12:34 pm

  10. Another red herring Q&A. The post you link to is about whether she believes global warming is accelerating. The notes above are of comments by IPCC authors and don’t necessarily reflect her views. And they are about the shortcomings of climate models. There is nothing in what you have linked to that rebuts any of the concerns listed.

    Comment by Graham Young — October 25, 2008 @ 2:27 pm

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.