In recent years international climate agreements, such as the Paris Agreement of 2015, have called for global warming to be limited based on specific global temperature targets. Subsequently there has been a large amount of research aimed at understanding the implications of global warming at these targets for the mean climate and extremes. Different model experiments have been used to analyse the implications of climates at specific levels of global warming, but there are still many unanswered questions, such as on the difference between transient and equilibrium climates under these levels of global warming, which our model experiments are not designed to answer.

Our study demonstrates that CMIP5 simulations performed for the extended climate projections to the year 2300 under RCP4.5 (mid-range global warming) and the RCP8.5 projections to 2100 (high-end global warming) may be used to generate samples of years under quasi-equilibrium and transient global warming respectively. The resulting ensembles may be compared directly with differences between the ensemble independent of model climate sensitivity. We use these model ensembles to make quantitative comparisons between equivalent transient and quasi-equilibrium warmer worlds. We find that most land areas are substantially warmer in transient climates than in equivalent quasi-equilibrium climates. This difference extends to extremes with a significantly greater probability of hot seasons in the transient than the quasi-equilibrium warmer worlds in many highly-populated regions.

Our analysis is the first to use a multi-model ensemble to assess differences between transient and quasi-equilibrium climates. Our findings provide important context for international climate agreements which implicitly aim for equilibrium global warming targets, but for which most analyses are based on transient climate projections.