Examining the Impacts of Aerosol Forcing on Atlantic Multidecadal Variability.

Atlantic Multidecadal Variability (AMV) impacts the Earth system in myriad ways, such as drought and precipitation variability in the Sahel and Americas, sea ice thickness and extent in the North Atlantic, hurricane frequency and intensity in the Atlantic, and the Indian monsoon. Therefore, understanding AMV is crucial for making accurate predictions about the evolution of the Atlantic and its broader effects. However, AMV is notoriously difficult to identify from sea surface temperature records, often being aliased by global warming or other forcing signals. Originally, AMV was believed to be an internal oscillation. Over the years, this view has been scrutinized, with many researchers suggesting that AMV arises from aerosol forcing—specifically, volcanic aerosols in preindustrial and last-millennium simulations and, more recently, anthropogenic aerosols. To examine the role of aerosol forcing on AMV, a novel methodology, low-frequency component analysis, is employed alongside state-of-the-art climate models. Low-frequency component analysis is analogous to EOF or Principal Component Analysis; however, the decomposed signal is organized into modes according to the ratio of low-frequency to total variance rather than the highest to lowest percent variance explained. This distinction helps identify patterns in the Earth system on decadal to multidecadal timescales, which provides more meaningful insights into the long-term role of aerosols on AMV. Applying this technique to observational data and various model simulations—including historical, historical with aerosols fixed at preindustrial levels (hist-piAer), and single-forcing simulations—helps determine the overall nature of AMV. If AMV is an internal oscillation, the phasing should be unique for each simulation, model, and corresponding ensemble member. Conversely, if aerosols externally force AMV, consistent phasing should be observed between historical models, observational data, and aerosol-only forcing models, with hist-piAer having random phasing.