lördag 22 augusti 2009

Penguin Logic of IPCC


Vincent Gray summaries his experience as expert reviewer for the Intergovernmental Panel on Climate Change IPCC in The Triumph of Double-Speak as follows:
  • Despite over 20 years’ of effort and four major Reports, the IPCC has not succeeded in providing any evidence that increases in  greenhouse gases are having a measurable effect on the climate. Why is it, then, that so many people believe that they have done so. The answer lies in their subtle use of  doublespeak, the technique of creating confusion by manipulation of language. This newsletter shows how they have confused and twisted the meanings of words in such a way as to create triumph out of failure.  
If what Gray claims is true, then the Copenhagen Climate Council based on the IPCC reports does not have to open, and the leaders of the world can foucs on solving real problems instead of creating real problems by inventing imaginary problems. 

Let's see if Gray's analysis is correct by going to the documents and then focus on the Fourth Assessment Report AR4 from 2007. In particular let's check if it respresents a form of  Science of Penguin Logic or pseudo-science. AR4 states in Technical Summary:
  • While this report provides new and important policy-relevant information on the scientifi c understanding of  climate change, the complexity of the climate system  and the multiple interactions that determine its behaviour impose limitations on our ability to understand fully the  future course of Earth’s global climate. 
  • There is still an incomplete physical understanding of many components of the climate system and their role in climate change. Key uncertainties include aspects of the roles played by  clouds, the cryosphere, the oceans, land use and couplings  between climate and biogeochemical cycles. 
  • The areas of  science covered in this report continue to undergo rapid  progress and it should be recognised that the present  assessment reflects scientifi c understanding based on the  peer-reviewed literature available in mid-2006.  
  • Models differ considerably in their estimates of the  strength of different feedbacks in the climate system. 
  • Equilibrium climate sensitivity is likely to be in the range  2°C to 4.5°C with a most likely value of about 3°C, based upon multiple observational and modelling constraints. There is a good understanding of the origin of differences in equilibrium climate sensitivity found in different  models. Cloud feedbacks are the primary source of inter-model differences in equilibrium climate sensitivity.
  • The overall response of global climate to radiative  forcing is complex due to a number of positive and negative feedbacks that can have a strong influence on the climate  system radiative balance. 
The key quantity is climate sensitivity measuring global warming vs doubling of the CO2 level in the atmosphere: IPCC states that it is likely to be in the range  2° to 4.5° C, with according to the IPCC Uncertainty Guidance likely = probability > 66%. To help interpretation of this statement IPCC informs us 
  • Finally we come to the most difficult question of when the detection and attribution of human­induced climate change is likely to occur. The answer to this question must be subjective, particularly in the light of the very large signal and noise uncertainties discussed in this chapter. Some scientists maintain that these uncertainties currently preclude any answer to the question posed above. Other scientists would and have claimed...that confident detection of a significant anthropogenic climate change has already occurred...
This can be interpreted as a reservation that convincing scientific support of the IPCC climate sensitivity estimate is lacking. But using doublespeak it is also interpreted by IPCC as something close to a truth:
  • Most of the observed increase in globally averaged temperature since the mid 20th century is very likely due to the observed increase in anthropogenic greenhouse gas   concentrations’.   
We see that IPCC oscillates between not-knowingthe most difficult question is if human induced climate change is likely to occur? and knowing: is very likely due to...anthropogenic greenhouse gas.  This is an extreme form of doublespeak, which is also practiced by modern theoretical physicists in search of a Theory Of Everything saying nothing about the physics of the world we live in. Knowing everything and nothing at the same time! 

 
Let us analyze the logic of the key statement of IPCC:
  • Climate sensitivity between 2° and 4.5° C with probability > 66% = likely.
Suppose we compare with the following possible statement by IPCC: 
  • Climate sensitivity between 1° and 10° C with probability > 95% = extremely likely.
This statement could seem more alarming by threatening with an extreme of 10° C combined with extremely likely. IPCC could take one step further to 
  • Climate sensitivity between -10°  and  +20° C with probability > 99% = virtually certain.
which could seem even more alarming. We seem to be led to the conclusion that IPCC uses Penguin Logic. What do you think? Compare also Sheep Herd Accuracy.

6 kommentarer:

  1. You find a contradiction in the following.

    * Finally we come to the most difficult question of when the detection and attribution of human­induced climate change is likely to occur. The answer to this question must be subjective, particularly in the light of the very large signal and noise uncertainties discussed in this chapter. Some scientists maintain that these uncertainties currently preclude any answer to the question posed above. Other scientists would and have claimed...that confident detection of a significant anthropogenic climate change has already occurred...

    This can be interpreted as a reservation that convincing scientific support of the IPCC climate sensitivity estimate is lacking. But using doublespeak it is also interpreted by IPCC as something close to a truth:

    * Most of the observed increase in globally averaged temperature since the mid 20th century is very likely due to the observed increase in anthropogenic greenhouse gas concentrations’.

    ===

    These are not contradictory. The attribution problem is defined as statistical detection in the absence of theory. There is debate whether this has occurred, i.e., whether this statistical threshhold has been crossed. Does the data already show, independent of any prior knowledge, that something unusual is happening.

    This is rather a fake question, but it is a traditional one among statisticians. Fortunately, we can do much better than modeling the great greenhouse experiment as a single-subject protocol.

    The transition from the first to the second assessment is approximately a transition from a frequentist to a Bayesian statement. Given the nature of the experiment at hand, the second statement is simply more useful. They are not, however, contradictory.

    SvaraRadera
  2. Maybe not contradictory, but to combine a reservation of not-knowing with an assessment of very likely knowing, is doublespeak and doublespeak is not science, only politics. Global warming is about a real physical phenomenon but it is viewed as politics.

    There is a huge difference between non-knowing and knowing, in science, right?

    SvaraRadera
  3. The nature of scientific knowledge is a very philosophical question. I do think about such matters and I'm not sure things are as simple as you portray. Still I'd rather not take that up here.

    I will concede this. There is no doubt that there is politics involved in the choice of words in the IPCC documents which after all must be approved by a committee. There is also no doubt that the influence upon politics at a larger scale is considered as a part of the editorial process.

    But this is exactly the purpose for which the IPCC is convened! How else should the scientific community summarize policy-relevant science to the policy community?

    SvaraRadera
  4. I don't think IPCC represents a good transfer of knowledge from science to politics.

    SvaraRadera
  5. On what basis? That is a very broad statement.

    In my experience, the IPCC first working group reports have been an excellent representation of the state of human understanding of the relevant questions.

    I can't vouch for groups 2 and 3 in the same way, but I am confident that the state of physical climatology is sufficiently rigorous and mature that a perfectly sound summary is feasible and has been effectively achieved four times.

    I have never seen a single statement of fact in an IPCC WG I report that was not justifiable on contemporaneous evidence. Indeed one would not expect less from a panel of eminent scientists in a less controversial field. Why should climate physics be different?

    Given that you collaborate with people I greatly admire, (I was very impressed by a FEniCS and DOLFIN at various presentations at U CHicago) I am certain that you are a man of great intelligence and competence.

    I am consequently dismayed that you have been steered so far wrong that you take the advice of the egregious and unsophisticated Mr. Watts over the careful and moderate conclusions of the IPCC. I urge you to reconsider.

    SvaraRadera
  6. I am always open to reconsideration if there are sound reasons.
    In any case there is no consensus on AGW, so whatever you believe it is wrong according to some scientists.

    SvaraRadera