The dependence of the northern extratropical climate response to external forcing on the phase of Atlantic Multidecadal Variability

Abstract

The extent to which decadal-to-multidecadal climate trends forced by carbon dioxide ($\text{CO}_2$) and anthropogenic aerosol (AER) emissions depend on the phase and amplitude of internal modes of climate variability, such as Atlantic Multidecadal Variability (AMV), is an open question. This study uses a fully coupled CMIP5-era general circulation model (GCM) to investigate climate trends forced by increasing $\text{CO}_2$ and AER in the presence of opposite decaying phases of the AMV. Ensembles of simulations are initialized from extreme warm ($\text{AMV}_{(+)}$) and cold ($\text{AMV}_{(-)}$) phases of AMV and, while the global mean temperature trends are similar, significant regional differences are found over the Arctic and northern extratropics. Specifically, the response to $\text{CO}_2$ forcing is hemispherically asymmetric, with western Eurasia warming 20–30% more, and North America and the extratropical North Pacific warming 20–30% less, in $\text{AMV}_{(+)}$ than in $\text{AMV}_{(-)}$. This asymmetry is explained by the atmospheric response to differences in the initial concentration of sea ice in the Atlantic Arctic sector, and by a large-scale atmospheric teleconnection pattern originating in the tropical Indo-Pacific. A decomposition of the temperature trends reveals that the AMV influence occurs mostly through atmospheric dynamics; however, thermodynamic processes are important in regions of sea ice change, western Eurasia, and eastern North America. The difference in the responses to $\text{CO}_2$ and AER forcing between AMV phases reveals that some aspects of the AMV modulation of the response, such as a more positive (negative) temperature trend in $\text{AMV}_{(+)}$ than in $\text{AMV}_{(-)}$ in western Eurasia (the extratropical North Pacific), are largely independent of the sign, magnitude and spatial pattern of external forcing.