We have investigated the role of stratospheric dynamics for Northern Hemisphere regional climate changes under continuous and steadily increasing stratospheric sulfate injections to meet multiple annual mean surface temperature targets (of global mean temperature and Equator-to-pole and interhemispheric temperature and gradients) under the RCP8.5 scenario in the Geoengineering Large Ensemble (GLENS) simulations. This geoengineering approach avoids many of the large surface climate impacts of greenhouse gas forcing under RCP8.5, for example in temperature and precipitation, but there are residual impacts on the high latitude, wintertime NH which we have studied here. Sulfate-aerosol-driven warming of the tropical lower stratosphere and the consequent strengthening of the stratospheric polar vortex is a key difference in the climate response under geoengineering compared to a non-geoengineered climate in this model and adds to the robustness of this finding in previous studies. The strengthening NH polar vortex, as reflected in the stratospheric Northern Annular Mode, correlates well in time with surface climate responses; regression analysis suggests that an increasing NAM at 50 hPa leads to a positive trend in the wintertime North Atlantic Oscillation and is consequently the main cause of Eurasian continental winter warming and a dipole response in precipitation. Experiments forced with just stratospheric heating from aerosols further cement the major role of dynamical stratosphere–troposphere coupling over other effects such as seasonal changes in insolation, as suggested by Kravitz et al. (2017), in leading to the NH wintertime surface climate changes (see also Jiang et al., 2019, and Simpson et al., 2019). Trends in North Atlantic sea level pressure congruent with the stratospheric NAM are, however, offset by a negatively signed component that also dominates in the summer season; the causes of this need to be investigated further. Since these are results from a single model and a unique geoengineering strategy, the robustness of the dynamically driven winter warming simulated here needs to be ascertained from other models and other sulfate injection strategies. A recent study also finds winter warming, and argues similarly for stratosphere–troposphere coupling, under the sulfur injection scenario (G6sulfur) of the GeoMIP experiments with the UKESM1 and CESM2-WACCM6 models (Jones et al., 2021).

The role of volcanic forcing in causing the observed wintertime warming following the large Pinatubo and Krakatau eruptions has recently been questioned in the face of large internal variability (Polvani et al., 2019; Polvani and Camargo, 2020). The findings of this paper add further evidence to those studies. The forced warming in the GLENS simulations by the end of the century, when sulfate emissions have reached around 50 Tg (SO2) yr−1, equivalent to several eruptions like Mount Pinatubo each year and sustained for many decades, is found to be within the 1σ spread of unforced extratropical wintertime interannual variability, suggesting that a single large eruption is very unlikely to be detectable. Nonetheless, the forced response in the GLENS simulations is robustly demonstrated with the ensemble approach. The side effects of smaller and perhaps more plausible sulfate injections on the Eurasian continent would be smaller than shown here, with longer timescales needed for detection, underscoring the need for mitigation of greenhouse gas emissions.