Climate modelling: Achievements and prospects

Abstract The significant progress made with coupled (atmosphere‐ocean) general‐circulation models (CGCMs) over the past decade or so is discussed in the context of the reports of the Intergovernmental Panel on Climate Change (IPCC). In their first report in 1990, most climate‐change GCMs were atmospheric GCMs coupled to ‘slab’ ocean models and being used only to determine the ‘equilibrium’ climate response to specified increases in carbon dioxide. the wide range of estimates of their ‘climate sensitivity’ is discussed, especially in the context of the uncertainty introduced by model sensitivity to representations of cloud. the modelling focus then moved to transient‐response, climate‐change experiments. These and the related modelling problems are noted; as are the simultaneous growing awareness and improved quantification of a range of natural and human‐induced radiative forcings. the consequent inclusion of this information in CGCMs to study the detection and attribution of climate change, leading to the IPCC statement in their 1995 report that ‘the balance of evidence suggests that there is a discernible human influence on global climate’, is also discussed. A significant modelling breakthrough since the 1995 report is the capability of a few CGCMs to produce stable and realistic climate (control) simulations without using artificial ‘flux adjustments’. This is illustrated with results from such a model, the Hadley Centre HadCM3. In conclusion, a brief perspective is given of high‐priority modelling developments and applications to be addressed over the next decade.