The following HTML text is provided to enhance online
readability. Many aspects of typography translate only awkwardly to HTML.
Please use the page image
as the authoritative form to ensure accuracy.
Climate Stabilization Targets: Emissions, Concentrations, and Impacts over Decades to Millennia
more complete simulations on longer time scales, once the ocean stops taking up heat.
For the general circulation models listed in Table 8.2 of IPCC, Working Group I, The Physical Science Basis (IPCC, 2007a), the equilibrium Δ T2x has a minimum of 2.1ºC, a maximum of 4.4ºC, and a median of 3.2ºC. Even the least sensitive model has a higher climate sensitivity than the idealized calculations yield for basic clear-sky water vapor and lapse-rate feedback. This is largely because the presence of clouds increases the basic black-body plus water vapor feedback sensitivity to about 1.8ºC even if clouds do not change as the climate warms. This form of cloud effect is not conventionally counted as a cloud feedback. It is more robust than feedbacks due to changing clouds, because it is based on cloud properties that can be verified against today’s climate. This value, too agrees well among models and is considered to be highly certain. Thus, the least sensitive IPCC models correspond very nearly to cloud properties remaining fixed while warming is amplified by water vapor feedbacks alone. The more sensitive models are more sensitive primarily by virtue of having positive cloud feedback. The spread in equilibrium climate sensitivity within the IPCC ensemble of models is primarily due to differences in cloud feedback, and in particular to the feedback of low clouds (Bony et al., 2006). Note that climate sensitivity could only be lower than about 1.8ºC if there are negative feedbacks very different from those of any of the models, such as changes in upper tropospheric or lower stratospheric water vapor, or changes in clouds that are opposite to those expected.
It has long been recognized that a symmetric distribution of the uncertainty in the strength of the feedbacks affecting climate sensitivity results in a skewed distribution in the climate sensitivity itself, with a high probability of large values (e.g., Schlesinger, 1986).2 Roe and Baker (2007) attempt to use this property to argue that it will be extremely difficult to eliminate the significant possibility of very high climate sensitivities. However, there is no a priori reason to expect the uncertainty in the strength of the feedback
This can be understood by noting that the climate sensitivity is proportional to 1/(1-f), where f is the strength of the feedbacks, and is positive if the feedbacks are positive. If one starts with the value f = 0.5, then increasing f by 0.25, say, increases the sensitivity by 100%, or a a factor of two. Decreasing f by the same amount decreases the sensitivity by only 67%. One can also use Figure 3.1 to understand this result pictorially. Climate sensitivity is proportional to the reciprocal of the slope shown in the figure, with the magnitude of the slope determined by the strength of the feedbacks. A symmetric distribution of slopes does not result in a symmetric distribution of climate sensitivity. A symmetric distribution might include zero slope with a finite probability, resulting in infinite climate sensitivity, which, of course, is ruled out by the observed stability of the climate system.