• Adaptation and mitigation are complementary measures.
• Globally, one cannot trade the concept of ‘1 degree of mitigation’ with that of ‘1 degree of adaptation’, owing to the very different potential for adaptation in different sectors and regions, the issue of scale, the dynamics of the process, and incomplete information about the limits to adaptation. The concept of ‘reaching the 2 degree target’ under 3°C temperature rise with ‘1°C of adaptation’ is not realistic.
• Most non-complex integrated models mislead users by attempting to derive optimal tradeoffs between mitigation and adaptation based upon an incomplete and inadequate representation of the climate system, climate change and climate change impacts, and an incomplete representation of the adaptation process. Outcomes are dependent on subjective and sometimes non-transparent judgements. Hence they fail to properly account for the significant uncertainties that pervade climate policy questions.
• The integrated models are perhaps most misleading when considering policy options under extreme climate change.
• The PAGE09 model overcomes most, although not all, of these difficulties through its rigorous uncertainty analysis and other merits, and can yield useful insight.
• Physically based models have limited treatments of adaptation, with the exception of crop models. Representation of adaptation processes in models could be improved, e.g. through the representation of additional water storage and levels of flood protection, or through the representation of reduced demands and improved efficiency in resource use.
• Modelling adaptation requires consideration of non-climate influences as well as climate change, as illustrated in the review of coasts. The DIVA coastal model provides a good approach to begin such analysis for this sector.
• A new conceptual framework for modelling combinations of mitigation and adaptation using a risk assessment approach is put forward.