Anthropogenic aerosol forcing of the Atlantic meridional overturning circulation and the associated mechanisms in CMIP6 models

Abstract. By regulating the global transport of heat, freshwater, and carbon, Atlantic meridional overturning circulation (AMOC) serves as an important component climate system. During late 20th early 21st centuries, indirect observations models suggest a weakening AMOC. Direct AMOC also during century but with substantial interannual variability. Long-term has been associated increasing greenhouse gases (GHGs), some modeling studies build up anthropogenic aerosols (AAs) may have offset part GHG-induced weakening. Here, we quantify 1900–2020 variations assess driving mechanisms in state-of-the-art from Coupled Model Intercomparison Project phase 6 (CMIP6). The CMIP6 forcing (GHGs, volcanic aerosols, solar variability, land use change) multi-model mean shows negligible changes to ∼ 1950, followed by robust strengthening second half (∼ 1950–1990) afterwards (1990–2020). These multi-decadal are related North atmospheric circulation, including altered sea level pressure gradient, storm track activity, surface winds, heat fluxes, which drive subpolar density flux. To further investigate these relationships, perform regression analysis decompose responses into aerosol-forced subsequent AMOC-related feedback. Similar previous studies, GHG simulations yield weakening, particularly century. Changes natural forcings, variability changes. In contrast, AA (weakening) response (decreasing) aerosols. Moreover, all-forcing occur simulations, suggests largely driven aerosol emissions. More specifically, our results that is initiated optical thickness perturbations net shortwave radiation temperature (and hence density), turn affect gradient wind – via latent sensible fluxes flux through its thermal component. feedbacks act reinforce this response, due salinity temperature-related cloud-related) acting mute initial response. Although aspects resembles observations, notable differences exist. This includes 1950 1990, when estimates underestimates observed increase ocean content, although overall patterns similar simulated weaker than those observed, implying they only partially externally forced. possible causes include internal observational uncertainties, model shortcomings, excessive forcing. A handful realizations evolution since 1900 inferred 1990 even 1930 1990) significant contribution (i.e., unforced) Nonetheless, robust, forced changes, bulk