Non-stationarity in the NAO-Gulf Stream interannual relationship

The North Atlantic Oscillation (NAO) is widely recognized as a primary driver of the Gulf Stream (GS) path variability on interannual to decadal timescales. While the canonical view holds that the GS path responds to the NAO in a temporally stationary and approximately linear manner, we show that the sign and strength of the NAO-GS relationship vary markedly through time. Using satellite altimetry, atmospheric reanalysis, and a wind-driven, two-layer ocean model, we investigate the temporal evolution and dynamical origins of this non-stationarity in the relationship between the NAO and the meridional position of the separated GS throughout the altimetry era (1993-2023). Lead-lag analysis of the full record shows a significant, positive correlation between the NAO and GS position when the NAO leads by 3–20 months, with the maximum correlation occurring at a lag of 8 months. However, moving-window correlations demonstrate pronounced modulation of this relationship through time, with transitions from strong positive correlations to weak or reversed correlations. This behavior is reproduced by the wind-driven model, indicating that wind forcing plays a dominant role in driving the observed non-stationarity in the NAO-GS relationship. Analysis of sea level pressure variability reveals substantial shifts in the spatial structure of the NAO. The transition from positive to weakly negative NAO–GS correlation coincides with southwestward shifts of the subtropical high and subpolar low, consistent with changes in the geometry of the associated wind stress curl forcing over the GS region and intergyre boundary. These results suggest that the sensitivity of the GS to atmospheric variability depends strongly on the spatial structure of the forcing, rather than solely its magnitude, and highlight the importance of accounting for non-stationarity when interpreting and predicting GS variability and its impacts on regional climate, sea level, and marine ecosystems.