HighlightsWe attribute variation across global climate mitigation scenarios to three factors The three factors are climate ambition, scenario background and model choice Many indicators are well-explained by the average effects of one or two factors We also calculate the residual not explained by these average effects This shows which indicators give outliers for some specific input combinations AbstractWe attribute variations in key energy sector indicators across global climate mitigation scenarios to climate ambition, assumptions in background socioeconomic scenarios, differences between models and an unattributed portion that depends on the interaction between these. The scenarios assessed have been generated by Integrated Assessment Models (IAMs) as part of a model intercomparison project exploring the Shared Socio-economic Pathways (SSPs) used by the climate science community. Climate ambition plays the most significant role in explaining many energy-related indicators, particularly those relevant to overall energy supply, the use of fossil fuels, final energy carriers and emissions. The role of socioeconomic background scenarios is more prominent for indicators influenced by population and GDP growth, such as those relating to final energy demand and nuclear energy. Variations across some indicators, including hydro, solar and wind generation, are largely attributable to inter-model differences. Our Shapley-Owen decomposition gives an unexplained residual not due to the average effects of the other factors, highlighting some (such as the use of carbon capture and storage (CCS) for fossil fuels, or adopting hydrogen as an energy carrier) with outlier results for particular ambition-scenario-model combinations. This suggests guidance to policymakers on these indicators is the least robust.Graphical AbstractKeywordsEnergy transition, climate change mitigation, Integrated Assessment Models, Shapley-Owen decomposition