HR: 08:15h
AN: U51B-02 INVITED    [Abstracts]
TI: Probabilistic inference for future climate change
AU: * Annan, J D
EM: jdannan@jamstec.go.jp
AF: FRCGC/JAMSTEC, 3172-25 Showamachi, Yokohama, 236-0001, Japan
AU: Hargreaves, J C
EM: jules@jamstec.go.jp
AF: FRCGC/JAMSTEC, 3172-25 Showamachi, Yokohama, 236-0001, Japan
AB: The experiment we are currently performing on the Earth system is intrinsically non-repeatable, so the concept of "reliable probabilities" over a large ensemble of predictions cannot apply. Therefore, prediction of future climate change is strongly Bayesian in a way that numerical weather prediction is not: probabilistic predictions are necessarily a matter of informed belief conditioned on our interpretations of the data and models around us. The climate sensitivity (S) in response to anthropogenic and other forcings has long been one of the dominant uncertainties in predicting future climate change. Many observationally-based estimates have been presented in recent years, with many of them disagreeing with the original Charney/IPCC estimates, for instance by assigning substantial probability to extremely high sensitivity, such as P (S > 6C ) > 5%. However, there is now an abundance of observational evidence available to us covering a wide range of time scales and processes. The broad consistency of these data (and success of explicit predictions) suggests that we should rather have increased confidence in our scientific understanding. We will show that explicit probabilistic analysis supports this standpoint, and a much greater confidence in a moderate value for S is easily justified, with climate sensitivity very unlikely to be as high as 4.5C.
DE: 1616 Climate variability (1635, 3305, 3309, 4215, 4513)
DE: 1626 Global climate models (3337, 4928)
SC: Union [U]
MN: 2007 Fall Meeting