Drivers of regional variation in the de-emergence of climate change under negative emissions

A central question of climate change impact assessments is identifying when the signal of change emerges from the noise of natural variability, though to date, this framework has been almost exclusively applied under warming conditions. In this work, we investigate the phenomenon of de-emergence, where temperatures rise beyond a given threshold and subsequently drop back below that threshold as forcing reverses. Large-scale carbon dioxide removal (CDR), also referred to as negative emissions, has been proposed to lower global average temperatures and to mitigate damage to the Earth system, though the regional effects of this have been understudied to date. Here, we analyze the results of eight Earth system models (ESMs) that participated in the CDR Model Intercomparison Project (CDRMIP). Decades after CO ₂ concentrations return to preindustrial levels following a gradual quadrupling, the only regions that most models agree to return to preindustrial temperatures are an area across northwestern Eurasia and a small area off the coast of West Antarctica. Around half of the models project de-emergence for areas of North America and East Asia. We investigate potential causes for these patterns and find that the land areas that cool more also exhibit enhanced latent heat flux, net carbon uptake, and precipitation, implying enhanced evapotranspiration. The region off West Antarctica appears to experience a cooling influence from changes to sea ice and meltwater. Both regions experience changes to cloud radiative effects that cause net cooling. The distribution of de-emergence patterns raises questions about the equity and fairness of overshoot scenarios.