Transition paths of North Atlantic Deep Water

TL;DR

This study applies Transition Path Theory to analyze the equatorward routes of North Atlantic Deep Water using float data, revealing organized boundary currents and pathway durations in the subpolar North Atlantic.

Contribution

It introduces a robust TPT-based framework to quantify NADW pathways, providing new insights into their organization and timescales compared to previous analyses.

Findings

UNADW follows well-organized boundary currents along the southern edge.

LNADW pathways are more diffusive and less organized.

Pathway durations vary from 2-3 years for UNADW to up to 8 years for LNADW.

Abstract

We use Transition Path Theory (TPT) to frame, in a statistically more robust fashion than earlier analyses, equatorward routes of North Atlantic Deep Water (NADW) in the subpolar North Atlantic. TPT is applied on all available RAFOS and Argo floats in the area by means of a discretization of the Lagrangian dynamics described by their trajectories. By considering floats at different depths, we investigate transition paths of NADW in its upper (UNADW) and lower (LNADW) layers. We find that the majority of UNADW transition paths sourced in the Labrador and southwestern Irminger Seas reach the western side of a target arranged zonally along the southern edge of the subpolar North Atlantic domain visited by the floats. This is accomplished in the form of a well-organized deep boundary current (DBC). LNADW transition paths sourced west of the Reykjanes Ridge reveal a similar pattern, while those sourced east of the ridge are found to hit the western side of the target via a DBC and also several other places along it in a less organized fashion, indicating southward flow along the eastern and western flanks of the Mid-Atlantic Ridge. Naked-eye inspection of trajectories suggest generally much more diffusive equatorward NADW routes. A source-independent dynamical decomposition of the flow domain into analogous backward-time basins of attraction, beyond the reach of direct inspection of trajectories, reveals a much wider influence of the western side of the target for UNADW than for LNADW. For UNADW, the average expected duration of the pathways from the Labrador and Irminger Seas was found to be of 2 to 3 years. For LNADW, the duration was found to be influenced by the Reykjanes Ridge, being as long as 8 years from the western side of the ridge and of about 3 years on average from its eastern side.