Multidecadal SST Variability Assessed as Primarily Forced in the Atlantic and Internal in the Pacific Using Rotated Low-Frequency Component Analysis

Whether multidecadal sea surface temperature (SST) variability in the Atlantic and Pacific basins originates internally or is externally forced has long been debated. Here, we apply a novel Rotated Low-Frequency Component Analysis to large-ensemble climate model simulations and observational SST datasets to disentangle the forced and unforced components of Atlantic Multidecadal Variability (AMV) and the Pacific Decadal Oscillation (PDO). By rotating low-frequency modes from individual ensemble members toward externally forced reference modes derived from their corresponding ensemble means, we separate external and internal SST anomalies. Across the CESM1, CESM2, GFDL-CM3, and E3SMv2 ensembles, we find that AMV is predominantly forced, while the PDO is largely internal. Application to observations agrees with these findings: the observed AMV is almost entirely explained by external forcing, whereas the PDO retains a substantial internal component. These results have important implications for the attribution, prediction, and understanding of historical and future multidecadal SST variability.