Author response to reviewer comments

<strong class="journal-contentHeaderColor">Abstract.</strong> We report initial results for G6-1.5K-SAI, a climate model experiment proposed by the Geoengineering Model Intercomparison Project (GeoMIP). G6-1.5K-SAI, which simulates a stratospheric aerosol injection (SAI) to limit global warming to &sim; 1.5 &deg;C above preindustrial in each model, features several design updates relative to previous GeoMIP experiment G6sulfur, such as hemispherically symmetric subtropical injection (30&deg; N and 30&deg; S) instead of equatorial injection. Due to differences in climate sensitivity, models disagree on the amount of warming to be offset, and therefore on the total injection required. While they agree strongly on the rate of cooling per unit rate of injection (&sim; 0.1 &deg;C per Tg SO₂ yr^&minus;1, a similar value to G6sulfur models with interactive SO₂), similarities in aerosol representation and disagreements in aerosol optical depth (AOD) per rate of unit injection and in rate of cooling per unit AOD mean this agreement may not imply accuracy. In all participating models, SAI cools the land surface more than the ocean and offsets mid- and high-latitude precipitation increases under global warming, but models disagree on the magnitude of residual Arctic amplification and changes to tropical precipitation. Relative to G6sulfur, G6-1.5K-SAI cools the Arctic more strongly, and also decreases precipitation less, especially in the tropics and over land. All in all, while the new G6-1.5K-SAI experiment constitutes an update over the older G6sulfur, due to the differences in scenario across these two experiments, any differences in SAI impacts must be evaluated carefully.