Mainstream IPCC climate models do not predict human extinction. Although many individual lives will be at risk due to the most likely climate change, temperature rise and sea level rise in itself cannot threaten humanity in its entirety, for there will always be some land area with modest temperatures left.

However, there are some unlikely extreme cases where extinction risk looks slightly more possible:

  • We could dramatically fail to curb our greenhouse gas emissions. In such a case, burning a lower bound of all available fossil fuels (5000 Gt) could lead to 9 to 13 degrees warming in 2300 according to established IPCC models.
  • It could be that the positive feedback loops of melting permafrost and/or methane clathrates emit large amounts of extra greenhouse gas on top of our direct emissions, increasing warming.
  • The climate could be more sensitive to the amount of greenhouse gas that we emit, which would also increase warming. Current climate models are not really that certain about the amount of warming a given amount of greenhouse gas causes.
Ord estimates the combined possibility at 0.1% as an estimate. It does not appear to be impossible, but it is much less likely than other man-made existential risks.

Because of these reasons, climate change could become a lot more severe than our best guess suggests. When accounting for all the uncertainties, it is hard to give an upper limit. We could plausibly end up with 13 degrees warming by 2300.

If global warming would become that bad (the chance of this is small, but not negligible), would that cause a risk for human extinction?

Heat stress and flooding itself cannot make humanity go extinct: there are always somes places left where humans can live. However, there are three potential ways which could:

  1. Moist greenhouse effect
  2. Extreme biodiversity loss
  3. An unknown climate-caused mechanism

None of these effects has occurred in the earth’s history, but greenhouse gas emission has also never occurred as rapidly as it currently does. It is therefore difficult to know what the chance is that one of these will happen. Ord estimates the combined possibility at 0.1% as an estimate. It does not appear to be impossible, but it is much less likely than other man-made existential risks.