Johan Rockstrom and company have a new paper talking about possible trajectories for the climate through the year 3000. There is a fair amount here, but one thing I find interesting is that while they are not talking about the highest emissions scenarios anymore (SSP3 under the CMIP6 framework), they are looking at a range of equilibrium climate sensitivities (ECS), and this makes a big difference. ECS, if I remember correctly, is the amount of global warming that occurs from a doubling of atmospheric carbon dioxide concentrations. If this value is 4 degrees C rather than 3, our situation gets much worse, with a peak of around 5 degrees C above preindustrial temperatures rather than the 3.5 expected under the lower ECS value. They show the peak sometime between 2200 and 2300. If this sets off catastrophic feedback loops, the system could get to a point well outside human control, even if we found the capacity and geopolitical will to get our act together.

That might happen, for example, in the case of large-scale dying or burning of forests (Reyer et al., 2015), or when carbon release from thawing permafrost (Turetsky et al., 2020) or gas hydrates on the ocean floor (Minshull et al., 2016) increases beyond current expectations. If such “natural” carbon release exceeds a potentially modest threshold (compare residual emissions experiment, Figure 1), warming may continue even if human forcing from GHG emissions or further land-use changes are virtually stopped. These human efforts would then be thwarted by the Earth system itself.

An important point is we don’t know the correct value of ECS yet – that should become clear with more data collection and research in coming decades. But the risk is existential enough that a species and civilization able to act together in a rational manner would be managing it now.