Climate change is increasing the global mean surface air temperature and is likely to trigger irreversible tipping points in the Earth system within the next decades. To avoid some of the adverse effects of global warming, solar radiation management (SRM) has been suggested as a potential temporary countermeasure. The only form of SRM that has a natural analogue and is thus in principle proven to work is stratospheric aerosol injection (SAI) of sulfate. It works similar to a large volcanic eruption, which releases sulfuric gas into the stratosphere, where sulfate aerosol particles form and remain for months to years and reflect sunlight. SAI may have dangerous side effects, however, which are currently poorly constrained. Two major concerns are depletion of stratospheric ozone due to an artificially enhanced aerosol burden and the heating of the lower stratosphere from the absorptive sulfate aerosol. These two effects also influence each other through chemical-dynamical links and feedbacks.
To examine these side effects and their associated uncertainty we conducted a multi-model analysis of stratospheric chemistry and dynamics in the CCMI-2022 sensitivity experiments, where SAI is employed to offset all the surface warming after 2025 (over 1 K by 2100) in a moderate greenhouse gas emission scenario. Whereas previous analyses of this type have been conducted with models that have widely different treatments of aerosol microphysics and chemistry, this novel experiment is designed to minimize this source of uncertainty by prescribing one stratospheric aerosol forcing in all participating models. To implement the same forcing in different models, we developed a Mie scattering code and used it to calculate the optical properties of the stratospheric aerosol for this scenario on the different spectral grids required for each model.
We present the results from five models, with a focus on the role of ozone-related processes, such as chemical catalytic destruction cycles and changes in large-scale transport. We find that models agree reasonably well in global mean total column ozone changes, indicating that SAI delays the ozone recovery between three and five decades in this scenario. However, we find substantial inter-model differences in specific regions of the stratosphere, such as the Antarctic polar stratosphere and the tropical lower stratosphere, where heterogeneous chemistry on aerosol surfaces destroys ozone. In this experiment we further separate the chemical and dynamical contributions to the ozone changes und SAI for the first time and show that dynamical changes from the stratospheric heating play the dominant role in this scenario.