In most emissions scenarios consistent with the temperature targets of the Paris Agreement, carbon dioxide removal (CDR) would be required to achieve net negative emissions. The efficiency of CDR depends on the behavior of the natural carbon reservoirs, land and ocean, that regulate atmospheric CO2 concentrations, but their change in response to negative emissions is highly uncertain. Here, we investigate the response of the terrestrial and oceanic carbon cycle to negative emissions based on an idealized emission-driven simulation using a state-of-the-art Earth system model. The terrestrial and oceanic carbon sinks become carbon sources ~30 years after the onset of negative emissions. Thereafter, although the atmospheric CO2 concentration returns to its initial level, the terrestrial ecosystem and the ocean continue to release carbon. The ocean recovers as a carbon sink within a few decades, while the terrestrial ecosystem remains a carbon source until the end of the simulation. The prolonged carbon emissions from the land are due to the delayed response of respiration to the earlier increase in terrestrial carbon uptake. As a result, the total carbon stock on land gradually decreases but still remains higher than its initial state. The ocean carbon inventory shows an irreversible change within a few centuries due to the accumulation of anthropogenic CO2 in the deep ocean, which would take more than centuries to be removed naturally.

Climate extremes, such as hot temperature and heavy precipitation events, have devastating effects on human societies. As the planet gets warmer, they have become more intense and more frequent. To avoid irreversible damages from climate extremes, many countries have committed to achieving net-zero anthropogenic carbon emissions, or carbon neutrality, by the 2050s. Here, we quantify the impact of carbon neutrality on population exposure to climate extremes using multi-model projections from the Coupled Model Intercomparison Project Phase 6 (CMIP6) Shared Socioeconomic Pathway (SSP)1-1.9 and SSP3-7.0 scenarios. It is found that the increasing population exposure to hot-temperature and heavy-precipitation extremes under SSP3-7.0 scenario can be substantially reduced by 87–98% in the late 21st century by achieving the carbon neutrality based on SSP1-1.9 scenario. The benefits of carbon neutrality are particularly pronounced in Africa and Asia. The potential benefits of carbon neutrality are also significant in North America, Europe, and Oceania, where a reduction in climate extremes is more than twice as important as population decline in reducing population exposure to climate extremes. These results provide important scientific support for ongoing efforts to achieve net-zero carbon emissions by the 2050s to reduce potential climate risk and its inequity across continents.

Global warming will soon reach the Paris Agreement targets of 1.5C°/2°C temperature increase. Under a business-as-usual scenario, the time to reach these targets varies widely among climate models. Using Coupled Model Intercomparison Project phase 5 and 6, we show that a 2°C near-future global warming rate is determined by Southern Ocean (SO) state closely tied with a low-level cloud (LLC) amount feedback strength during the reference (1861-1900) period; climate models with cold SO tend to accompany more low-level cloudiness and Antarctic sea ice concentration due to a strong LLC amount feedback. Consequently, initially cold SO models tend to simulate a fast near-future warming rate by absorbing more downward shortwave radiation compared to initially warm SO models because more LLC disappears due to a strong LLC amount feedback during the 2°C rise. Our results demonstrate that climate models that correctly simulate initial SO state can improve near-future projections with reduced uncertainties.