Simulated Climate Sensitivity Uncertainty: Control Bias, Perturbed Physics and the Tuning of Models

Thursday, 18 December 2014
Dietmar Dommenget, Monash University, Melbourne, VIC, Australia
The simulated climate sensitivity on regional and global scale to a doubling of CO2 has large uncertainties that result from model errors. In this study the relationship between climate model biases in the control climate and the simulated climate sensitivity are discussed on the basis of perturbed physics ensemble simulations with a globally resolved energy balance (GREB) model. It is illustrated that the uncertainties in the simulated climate sensitivity can be conceptually split into two parts: a direct effect of the perturbed physics on the climate sensitivity independent of the control mean climate and an indirect effect of the perturbed physics by changing the control mean climate, which in turn changes the climate sensitivity, as the climate sensitivity itself is depending on the control climate. It is discussed whether correcting the models by tuning or by artificially correcting the control mean state (flux correcting) would be a better way of improving the climate change predictions.