Accounting for transience in the baseline climate state changes the surface climate response attributed to stratospheric aerosol injection

Abstract Stratospheric aerosol injection (SAI) is a proposed means of climate intervention that could halt global temperature rise, though it would imperfectly offset climate change. To estimate this imperfection, it is common to compare the simulated climate under SAI against that of a baseline state at the same global mean temperature without SAI. Here, we combine a recent set of SAI simulations (ARISE-SAI-1.5) in the earth system model UKESM1, with simulations of idealised abrupt and transient warming scenarios, to assess the impact of transient warming through this baseline state on surface climate changes attributed to SAI. We quantify the effect of temperature stabilisation as the expected change in surface climate between a climate state under warming and one in quasi-equilibrium at the same global mean temperature. We estimate that accounting for temperature stabilisation eliminates the land-sea warming ratio change attributed to SAI. However, relative to the hypothetical scenario with lower CO 2 concentrations that would achieve a stabilised climate at the same temperature, SAI produces a 69% larger reduction in global precipitation. Accounting for stabilisation can also meaningfully change the spatial pattern of surface temperature response attributable to SAI. We repeat our analysis for the GeoMIP G6sulfur scenario, to show that effects qualitatively consistent with these findings are seen when comparing the SAI state against the faster and slower warming baselines of the SSP5-8.5 and SSP2-4.5 scenarios. The changes in climate state attributable to temperature stabilisation are generally small compared to changes due to warming since pre-industrial. However, these differences can be significant in the context of assessing residual changes under SAI because these residuals are themselves roughly an order of magnitude smaller than the effects of warming. Our findings have implications for the design and assessment of future SAI simulations, and for the attribution of changes in surface climate to SAI.