Presumed Parameters - Climate Change: Identification and Projections - page 57

assumptions. The final evaluations will thus only carry over an informal
degree of confidence for a given evaluation.

5.2. Climate sensitivity S clim

Looking once again at the SPM: “Equilibrium climate sensitivity is the
change in global equilibrium temperature caused by a doubling in the
concentration of CO 2 in the atmosphere. It is likely to be in the range of
1.5°C to 4.5°C, ( high confidence ), extremely unlikely to be under 1°C, ( high
confidence ), and very unlikely to be above 6°C ( medium confidence )”.

There is nothing to indicate that the confidence level of likely (at 66%)
would be symmetrical, especially given that the upper and lower limits are
unsymmetrical (respectively 95-100% and 90-100%), and are weighted
respectively to either high or medium confidence levels.

The AR4 (2007) gave 3°C as the most likely value, which was not
centered in the middle of the given range [2 to 4.5°C]. The second order
draft of the AR5 (Chapter 12, p 7, line 18) again stated that, “the most likely
value remains near 3°C”. Due to a lack of consensus, this was removed from
the final version of the AR5. It would therefore be unsafe to infer from all of
the information above that there is a hypothetical value which equates to the
most likely . We are clearly in a situation where probabilities which are given
do not arise from actual statistical calculations, but from negotiations
between experts.

If we only retain the extreme values:

, 1 °< CS <° 6 C

clim

after dividing by log(2), it follows:

1 . 4427 ° C < S < 8.6562 ° C

1

When comparing these ranges to our own confidence level (at 90%), we
will bear in mind that those of the IPCC are hindered by a certain level of
fuzziness.