Coal transitions—part 2: phase-out dynamics in global long-term mitigation scenarios

A rapid phase-out of unabated coal use is essential to limit global warming to below 2 °C. This review presents a comprehensive assessment of coal transitions in mitigation scenarios consistent with the Paris Agreement, using data from more than 1500 publicly available scenarios generated by more than 30 integrated assessment models. Our ensemble analysis uses clustering techniques to categorize coal transition pathways in models and bridges evidence on technological learning and innovation with historical data of energy systems. Six key findings emerge: First, we identify three archetypal coal transitions within Paris-consistent mitigation pathways. About 38% of scenarios are ‘coal phase out’ trajectories and rapidly reduce coal consumption to near zero. ‘Coal persistence’ pathways (42%) reduce coal consumption much more gradually and incompletely. The remaining 20% follow ‘coal resurgence’ pathways, characterized by increased coal consumption in the second half of the century. Second, coal persistence and resurgence archetypes rely on the widespread availability and rapid scale-up of carbon capture and storage technology (CCS). Third, coal-transition archetypes spread across all levels of climate policy ambition and scenario cycles, reflecting their dependence on model structures and assumptions. Fourth, most baseline scenarios—including the shared socio-economic pathways (SSPs)—show much higher coal dependency compared to historical observations over the last 60 years. Fifth, coal-transition scenarios consistently incorporate very optimistic assumptions about the cost and scalability of CCS technologies, while being pessimistic about the cost and scalability of renewable energy technologies. Sixth, evaluation against coal-dependent baseline scenarios suggests that many mitigation scenarios overestimate the technical difficulty and costs of coal phase-outs. To improve future research, we recommend using up-to-date cost data and evidence about innovation and diffusion dynamics of different groups of zero or low-carbon technologies. Revised SSP quantifications need to incorporate projected technology learning and consistent cost structures, while reflecting recent trends in coal consumption.