Enhancing predictive climate models
Mathematical models of any process whose long-term behaviour is of interest must rely on careful formulation of relevant parameters and their inter-relationships while incorporating a statistical basis for extrapolations to the future. In the case of global climate change, the highly dynamic nature of the parameters of interest combined with their relatively recent documentation saddles predictions with unacceptable uncertainty. Scientists initiated the EU-funded project Ampere to compare 19 internationally recognised energy economy and integrated assessment models. Including the latest climate system information and scenarios will enable evaluation of underlying causes of model differences. The intercomparison of models will increase robustness of interpretations and thus effectiveness of associated policies. Such a comparison is quite complex, incorporating not only many models but many different parameters in each model and many outputs of interest. Obtaining interpretable results requires careful definition of a model intercomparison framework. This was established during the first project period and encompassed scenario specifications, harmonised baselines, reporting formats and documentation templates. Given that cost and policy are key components of the analysis, investigators harmonised population and gross domestic product (GDP) trajectories and established a policy benchmark capturing current and planned climate policies. The accompanying model validation framework has received considerable interest within the research community as there are currently no standards for evaluating an integrated assessment model. Ampere has completed two series of model comparisons. In addition, Ampere scientists developed a first-of-its-kind model diagnostics analysis for assessing differences among models in responses to carbon price signals. Scientists also assessed differences in emissions response of climate modules. Results should have important impact on integrated assessment modelling, diagnostics and intercomparison far beyond the scope of Ampere. Ampere results are expected to provide foundations for the preparation of long-term climate policy roadmaps focused on climate stabilisation. The range of models included in the analysis should ensure robust findings, reducing uncertainty and leading to more effective policies appropriately targeted at real-world scenarios.
