Bayesian climate change assessment using multi-AOGCM ensembles: global and regional surface temperatures

A Bayesian approach is applied to the observed surface air temperature (SAT) changes using multi-AOGCM ensembles (MMEs) of the IPCC AR4 simulations and single-AOGCM ensembles (SMEs) with the ECHO-G model. A Bayesian decision method is used as a tool for classifying observations into scenarios (or hypotheses) concerned. The Bayes factor (or likelihood ratio) provides an observational evidence for the each scenario against a predefined reference scenario. Four scenarios are used to explain observed SAT changes as 'CTL' (control or no change), 'Nat' (natural forcing induced change), 'GHG' (greenhouse-gas induced change), and 'All' (natural plus anthropogenic forcing induced change). Parameters needed to define the four scenarios are estimated from SMEs or MMEs. Taking the 'CTL' scenario as a reference one, application results for global mean SAT changes for the whole 20th century (1900-1999) show 'decisive' evidences (logarithm of Bayes factor > 5) for the 'All' scenario only. While 'strong' evidences (log of Bayes factor > 2.5) for both 'All' and 'Nat' scenarios are found in SAT changes for the first half (1900-1949), there are 'decisive' evidences for the 'All' scenario for SAT changes in the second half (1950-1999), supporting previous results. It is demonstrated that our Bayesian decision results for global mean SATs are largely insensitive to both inter-model uncertainties and prior probabilities. Extension to regional and seasonal SATs will be also presented considering more scenarios such as 'S' (sulfate aerosol forcing only) and 'ANTHRO' ('GHG' plus 'S').