Untangling the Multi-model Spread in 21st Century AMOC Projections

The Atlantic Meridional Overturning Circulation (AMOC) plays an important role in the global climate by transporting heat northward. According to the latest IPCC report (AR6) the strength of the AMOC is very likely to weaken by 2100 (Fox-Kemperer et al. 2021). A weaker AMOC would significantly impact local and global climate. However, there is large model spread in the magnitude of the projected reduction in AMOC strength (Weijer et al. 2020) so it is unclear to what extent the AMOC will weaken by the end of the 21^st century.

This study investigates the spread in AMOC response among CMIP6 models. As an initial step we investigated the model correlations of AMOC weakening across different ScenarioMIP experiments. Preliminary results show that the decline for similarly forced scenarios, such as ssp370 and ssp585, have stronger correlations than for scenarios with significantly different forcing, such as ssp126 and ssp585.

Further analyses into the relationship between the projected weakening and model biases in ocean temperature,  salinity and meridional density gradients are performed. In addition, we investigate how the weakening correlates with possible drivers. A better understanding of how model biases influence AMOC changes will allow for more accurate projections of future AMOC changes and their impacts, as well as improved understanding of what the driving processes of the weakening are in various models.