Redrawing the Iceland–Scotland Overflow Water pathways in the North Atlantic

The Atlantic Meridional Overturning Circulation (AMOC) plays a crucial role in global climate, transporting heat and climate signals (like anthropogenic carbon) from high latitudes. A key component of the AMOC's lower limb is the transport of deep waters, including Iceland-Scotland Overflow Water (ISOW), Denmark Strait Overflow Water, and Labrador Sea Water, away from their formation regions. While the pathways of Denmark Strait Overflow Water and Labrador Sea Water are relatively well-understood, the spreading pathways of ISOW remain less certain. ISOW originates in the Norwegian Sea, flowing between Iceland and Scotland into the North Atlantic. Its salinity contrasts with fresher Denmark Strait Overflow Water and Labrador Sea Water due to entrainment of saltier Subpolar Mode Water during its descent. Historically, simplified circulation diagrams portrayed ISOW as a single, counter-clockwise deep boundary current flowing around the subpolar North Atlantic, passing through the Charlie-Gibbs Fracture Zone (CGFZ) – a major gateway between the eastern and western North Atlantic. West of the CGFZ, ISOW was typically shown turning northward along the Reykjanes Ridge and continuing around the Irminger Sea. However, recent decades have seen evidence challenging this simplified view.

Over the past three decades, research has indicated a more intricate picture of ISOW pathways than the single boundary current model suggests. Studies utilizing Lagrangian observations, hydrographic sections, and model simulations have demonstrated that shallower ISOW flows westward through deep gaps in the Reykjanes Ridge before reaching the CGFZ. Furthermore, observations show ISOW bypassing the westward passage through the CGFZ and instead flowing southward along the Mid-Atlantic Ridge (MAR) into the Western European Basin. Several studies have challenged the notion that all ISOW turns northward after passing through the CGFZ. Hydrographic data revealed westward-spreading tongues of higher-salinity water, indicating alternative westward pathways. Modeling studies further detailed this westward flow, suggesting a significant portion of ISOW follows a deep west-northwestward current extending from the CGFZ almost to Greenland's southern tip, with less flowing northward into the Irminger Sea. Recently, direct observations using Deep-Argo floats have confirmed direct westward flow from the CGFZ to the Flemish Cap, joining the Deep Western Boundary Current. Shipboard LADCP measurements and moored observations within the CGFZ also showed ISOW transport reversals linked to the position of the eastward-flowing North Atlantic Current (NAC). However, the NAC's impact on ISOW spreading pathways downstream of the CGFZ remained unclear.

This study uses a combination of Lagrangian observations and numerical model simulations to improve understanding of ISOW pathways. **Lagrangian Observations:** Twenty-one neutrally buoyant RAFOS floats, deployed at ISOW depths (1800–2800 m) as part of the Overturning in the Subpolar North Atlantic Program (OSNAP), were tracked. Nine floats were released in the CGFZ's northern valley, and twelve others drifted through or near the CGFZ. Float trajectories were analyzed to identify main pathways. Mooring observations provided context on mean velocity and salinity fields across the CGFZ. **Numerical Model Simulations:** The high-resolution Family of Linked Atlantic Models Experiment (FLAME) was used to simulate ISOW pathways. 3593 synthetic floats were released monthly from 1990 to 2004 at ISOW levels in the CGFZ's northern valley, integrated forward in time using 3D velocity fields. The probability distribution of simulated trajectories was analyzed, quantifying the proportion of floats following different pathways. A comparison with another eddy-resolving model (HYCOM) was also conducted. The impact of the North Atlantic Current's (NAC) eddies and meanders was assessed by calculating eddy kinetic energy (EKE) west of the CGFZ and comparing the simulated pathways during high and low EKE periods. Finally, 10-year simulated trajectories were examined to determine long-term spreading patterns and export timescales to subtropical latitudes.

The Lagrangian float observations revealed three main ISOW pathways from the CGFZ: a west-northwestward pathway into the central Labrador and Irminger Seas (eight floats); a surprising southward pathway along the western flank of the MAR (eight floats); and an eastward recirculation pathway (four floats). The southward pathway, previously undocumented, represents a significant fraction of ISOW transport. Analysis of float trajectories showed that the interaction with meanders and eddies associated with the NAC influenced pathway selection, particularly diverting floats southward. The FLAME model simulations corroborated the observed pathways, identifying similar dominant branches: a west-northwestward, southward, eastward, and a locally recirculating branch near the CGFZ. The model quantitatively estimated the proportion of floats following each branch: approximately 59% west-northwestward, 19% southward, 6% eastward and 9% locally circulating near the CGFZ. A small percentage (5%) followed the historically recognized boundary current along the western flank of the Reykjanes Ridge. While the model generally agreed with observations, discrepancies might be due to the limited number of observed floats and differences in NAC activity during the observation and model periods. Analysis of high and low EKE periods in the model confirmed the influence of the NAC's eddy/meandering field on pathway partitioning: high EKE resulted in reduced westward and increased southward spreading. The 10-year simulations showed widespread ISOW distribution throughout the subpolar North Atlantic, with some reaching subtropical latitudes. The southward pathway from the CGFZ appeared to provide a faster track for ISOW export to the subtropical regions.

This study's findings significantly revise the traditional understanding of ISOW pathways. The dominant pathways are not solely confined to the deep boundary current but involve substantial interior spreading. The previously unrecognized southward pathway along the western flank of the MAR adds complexity to the picture. This diffusive export pattern of ISOW has implications for understanding the transport of heat, carbon, and other tracers from the high-latitude North Atlantic to other ocean basins. The influence of the NAC's mesoscale dynamics on deep ISOW pathways highlights an unexpected connection between upper and lower ocean processes. The results suggest that a significant part of the interannual variability in the overturning circulation is linked to upper-ocean mesoscale dynamics.

This research presents a revised understanding of Iceland-Scotland Overflow Water (ISOW) pathways in the North Atlantic, demonstrating significant interior spreading, and a previously unknown southward branch. The influence of the North Atlantic Current's eddies and meanders on ISOW pathway partitioning underscores the need to incorporate upper-ocean mesoscale processes into models of the overturning circulation. Future research should focus on refining the quantification of these pathways and further investigating the interaction between upper-ocean processes and deep-water transport.

The limited number of RAFOS floats in the observational dataset restricts statistical power and might lead to uncertainties in pathway quantification. Differences in the timing of the observed float trajectories and the model simulation period could also impact the comparisons. While FLAME and HYCOM model simulations support the observations, these models have their own limitations, including inherent uncertainties in model resolution and forcing data.