Absent mitigation, we’ll be forced to geoengineer

The most likely levels of global warming are very unlikely to cause human extinction.15 The existential risks of climate change instead stem from tail risk climate change – the low probability of extreme levels of warming – and interaction with other sources of risk. It is impossible to say with confidence at what point global warming would become severe enough to pose an existential threat. Research has suggested that warming of 11-12°C would render most of the planet uninhabitable,16 and would completely devastate agriculture.17 This would pose an extreme threat to human civilisation as we know it.18 Warming of around 7°C or more could potentially produce conflict and instability on such a scale that the indirect effects could be an existential risk, although it is extremely uncertain how likely such scenarios are.19 Moreover, the timescales over which such changes might happen could mean that humanity is able to adapt enough to avoid extinction in even very extreme scenarios.

The probability of these levels of warming depends on eventual greenhouse gas concentrations. According to some experts, unless strong action is taken soon by major emitters, it is likely that we will pursue a medium-high emissions pathway.20 If we do, the chance of extreme warming is highly uncertain but appears non-negligible. Current concentrations of greenhouse gases are higher than they have been for hundreds of thousands of years,21 which means that there are significant unknown unknowns about how the climate system will respond. Particularly concerning is the risk of positive feedback loops, such as the release of vast amounts of methane from melting of the arctic permafrost, which would cause rapid and disastrous warming.22 The economists Gernot Wagner and Martin Weitzman have used IPCC figures (which do not include modelling of feedback loops such as those from melting permafrost) to estimate that if we continue to pursue a medium-high emissions pathway, the probability of eventual warming of 6°C is around 10%,23 and of 10°C is around 3%.24 These estimates are of course highly uncertain. It is likely that the world will take action against climate change once it begins to impose large costs on human society, long before there is warming of 10°C. Unfortunately, there is significant inertia in the climate system: there is a 25 to 50 year lag between CO2 emissions and eventual warming,25 and it is expected that 40% of the peak concentration of CO2 will remain in the atmosphere 1,000 years after the peak is reached.26 Consequently, it is impossible to reduce temperatures quickly by reducing CO2 emissions. If the world does start to face costly warming, the international community will therefore face strong incentives to find other ways to reduce global temperatures. The only known way to reduce global temperatures quickly and cheaply is a form of climate engineering called Solar Radiation Management (SRM), which involves cooling the Earth by reflecting sunlight back into space.27 The most researched form of SRM involves injecting aerosols into the stratosphere.28 Most of the evidence so far suggests that ideal SRM deployment programmes would reduce overall damages relative to an un-engineered greenhouse world.29 However, SRM brings its own risks. Of the currently known potential negative direct effects of SRM, only abrupt termination could plausibly bring about an existential catastrophe.30 If a very thick stratospheric veil were deployed and SRM was suddenly terminated and not resumed within a buffer period of a few months, then there would be very rapid and damaging warming. There is some reason to think that an SRM system could be made very resilient to this termination shock risk, especially if the knowledge and capability for SRM deployment was widely shared, but termination shock could occur as a result of another global catastrophe, such as a global war or devastating pandemic.31 Aside from the known risks of SRM, current climate models are imperfect, and SRM could have currently unforeseen catastrophic effects. SRM also creates some indirect risks. Firstly, SRM could be unilaterally deployed32 and it will be difficult to agree on a level of SRM acceptable to all regions.33 Some nations could counter-geoengineer to reverse the effects of SRM, creating tensions and accident risk. Natural regional weather events could be incorrectly attributed to SRM, potentially leading to opposition from some parts of the world. SRM therefore poses serious international governance problems and, if mishandled, could increase the risk of political conflict.34