EURYI_Call1_Demography, uncertainty, and learning in integrated assessment models of climate change

The overall aims of the research were to improve integrated assessment models (IAMs) of climate change and to develop new methods of applying them to the climate change issue. Integrated assessment models are computer models that link both social science and natural science elements of the climate change issue in order to analyse questions that are relevant to policy. This research project focused on several related aspects of IAMs: advances in the integration of demographics, more systematic accounting for uncertainty, and novel treatments of the potential for learning (i.e., changes in uncertainty over time) within such models. Within the demography focus, research aimed to explicitly model links between the major demographic trends of aging, urbanization, and changes in living arrangements, on the one hand, and energy consumption, land use, and associated carbon emissions on the other hand. This work was organized around a set of country case studies that were then integrated into a global assessment. Within the uncertainty and learning areas, a series of analyses were undertaken to examine the implications of learning (or changes in uncertainty over time) for how decisions could be made over time, adapting to what was learned at each step. Results showed that alternative demographic futures can have substantial effects on future carbon emissions both regionally and globally. In particular, we quantified the extent to which slower population growth would reduce carbon emissions, after accounting for a number of important demographic- economic interactions that are often ignored in such studies. We also demonstrated that aging and urbanization, separate from the effects of population size, can substantially affect emissions as well. Aging, for example, tends to lead to lower emissions than otherwise expected, while urbanization has the opposite effect. Aging has a particularly large effect in industrialized countries (and also in China), and urbanization has a particularly large effect in Asia. These results suggest that the emissions scenario research community needs to better represent these trends in their models. In our work on uncertainty, we applied methods of uncertainty analysis and learning to components of the climate change issue that had not previously been examined, including population growth and the carbon cycle, demonstrating that while large uncertainties in these factors remain, there is a correspondingly large potential for reducing them over the coming decades through additional observations. The project also contributed to the development of a novel concept - negative learning - that identified a commonly experienced but rarely acknowledged pattern of learning, in which research leads the community farther away from what eventually turns out to be the "truth" for a period that is long enough to affect policy decisions. These new developments have helped broaden the treatment of uncertainty in climate analyses and counter a tendency toward overconfidence. We applied these concepts to the problem of how best to achieve long-term climate change policy goals, such as the aim to limit warming to less than 2 degrees Celsius. We concluded that it is crucial to achieve particular outcomes by the middle of the century in order to avoid closing off options for the longer term. Our work was the first to explicitly examine the question of what would need to be achieved by 2050 in order to keep particular long-term climate goals technically feasible. For example, we found that in one scenario of high growth in energy demand, the widely cited 2 degree limit supported by the EU is already impossible to achieve with even odds.