Abstract
The spatiotemporal structure of the recent decadal subsurface cooling trend in the North Atlantic Ocean is analyzed in the context of the phase reversal of Atlantic multidecadal variability. A vertically integrated ocean heat content (HC) Atlantic Multidecadal Oscillation index (AMO-HC) definition is proposed in order to capture the thermal state of the ocean and not just that of the surface as in the canonical AMO SST-based indices. The AMO-HC (5–657 m) index (defined over the area, 0°N–60°N, 5°W–75°W) indicates that: (1) The traditional surface AMO index lags the heat content (and subsurface temperatures) with the leading time being latitude-dependent. (2) The North Atlantic subsurface was in a warming trend since the mid-1980s to the mid-2000s, a feature that was also present at the surface with a lag of 3 years. (3) The North Atlantic subsurface is in a cooling trend since the mid-2000s with significant implications for predicting future North Atlantic climate. The spatial structure of decadal trends in upper-ocean heat content (5–657 m) in the North Atlantic prior to and after 2006 suggests a link with variability of the Gulf Stream–Subpolar Gyre system. The Gulf Stream leads variability in upper-ocean heat content over the Subpolar region by ∼13 years, and the lead increases from east to west from the Iceland Basin to the Irminger Sea to the Labrador Sea. The similarity between the structure of decadal mean anomalies, their change and trends in upper-ocean heat content and salinity in the North Atlantic and anomalies associated with a Gulf Stream index is striking. In this scheme, the displacements of the Gulf Stream–North Atlantic Current systems and their interactions with the Subpolar Gyre, as part of the meridional overturning circulation, have a decisive role for imposing decadal variability in both the Ocean and hydroclimate over the neighbouring continents.
Acknowledgements
The authors are grateful to T. Joyce for making the Gulf Stream index available. Argyro Kavvada performed preliminary analysis of the HC-based AMO indices as part of her doctoral thesis at the University of Maryland in 2014.
Disclosure statement
No potential conflict of interest was reported by the authors.
Supplemental data
Supplemental data for this article can be accessed here: http://10.0.4.56/16000870.2018.1481688.