Science & goals
CONTEXT
The distribution of heat, freshwater, biogeochemical parameters (e.g. carbon), and pollutants (e.g. plastic) in the ocean is at the heart of current environmental concerns. The observation and modelling of ocean currents, which largely control their distributions from the basin scale to the dissipation scale, is essential to describe the anthropogenic climate transition and reduce uncertainties on its long-term projection and impacts. An essential vector for the propagation of climate-relevant quantities within the global ocean interior is the North Atlantic Deep Water (NADW), a dense watermass that regularly connects with the atmosphere at high northern latitudes before sinking and spreading southward as far as the Indian and Pacific Oceans . The NADW, which occupies the deep limb of the so-called Atlantic Meridional Overturning Circulation (AMOC – Fig. 1), is one of the two water masses filling the deep global ocean – the other one being the Antarctic Bottom Water. The key role of NADW on heat uptake, on the spreading of Arctic-origin freshwater anomalies, and on the penetration of anthropogenic carbon into the abyss, is now widely recognized. Such a role has placed the NADW and hence the AMOC at the heart of the oceanographic community interests, with an ultimate goal of foreseeing their responses to future climate changes at high latitudes, such as, for instance, the expected weakening of deep convection throughout increased meltwater discharge from the Greenland ice sheet.
The southward export of NADW and its two components – the Upper NADW (UNADW) formed in subpolar seas and the Lower NADW (LNADW) formed in the Nordic Seas – was for a long time simplified as a confined laminar flow, heading continuously southward from high latitudes along the western Atlantic margins: the so-called Deep Western Boundary current (DWBC). This paradigm has now been revised, as improved observational and modelling capabilities revealed the NADW circulation as a three-dimensional system with complex patterns of vertical and meridional connectivity. Such complexity is largely determined by local ocean dynamics around a Transition Zone (TZ) that separates subpolar and subtropical basins: Flemish Cap and the Grand Banks of Newfoundland. This wide underwater plateau represents an important topographic obstacle for the southward propagation of NADW, and encompasses several processes that can significantly affect its intrinsic properties. Those include topography-controlled leakiness of boundary flows towards eddy-driven recirculating gyres and dispersive interior pathways , near-shore recirculation associated with the meandering North Atlantic Current (NAC), intense eddy-driven vertical motions, or isopycnal and diapycnal mixing near the steep continental slopes of the region. The spatial and temporal sparseness of existing observations, however, only enables a partial description of typical circulation and mixing patterns and cannot support a comprehensive analysis of NADW dynamics across the western subpolar-subtropical boundary. This requires a dedicated network that will measure concomitantly and continuously oceanic conditions upstream, within, and downstream of the TZ. This integrated observational approach, combined with novel modelling, is a prerequisite for predicting how fast and how far forthcoming high-latitude climate changes could imprint the deep ocean globally.
OBJECTIVES & HYPOTHESIS
The overarching objective of CROSSROAD is to create synergy between novel observational and numerical tools for tracking NADW components and enabling a mechanistic assessment of their pathways and transformation at the western subpolar-subtropical boundary. This challenge requires to implement a network of complementary instrumentations that will target the key components and processes of the system (Fig. 2). It also demands to analyse newly-available high-resolution simulations of the region that can both decipher the dominant physics at play and provide a full basin-scale and interannual to decadal context.
WP0. Produce a sub-kilometric numerical simulation
The first methodological cornerstone of the CROSSROAD project is to design and run a multi-annual sub-kilometric numerical simulation of the TZ region (GIGATZ). This sub-regional zoom will be embedded in a newly- available high-resolution and multi-annual realistic numerical simulation of the Atlantic Ocean, named GIGATL3, which is developed within the framework of the ANR JCJC DEEPER (PI. J. Gula). The simulations use the ROMS (Regional Oceanic Modelling System) model in its CROCO version (Coastal and Regional Ocean COmmunity version, https ://www.croco-ocean.org/). The latter solves the primitive (hydrostatic) or non-hydrostatic equations for an incompressible fluid with free surface. It uses vertical sigma coordinates, which follow the topography, and allows better resolution of processes related to interactions with the topography by increasing the vertical resolution near the bottom. The reference simulation to be used in CROSSROAD was produced in a realistic oceanic context and includes adequate internal wave sources (hourly winds and tides). Initiated and forced at the boundaries by the SODA ocean reanalysis, GIGATL3 currently spans 10 years over the period 2004-2014 with a resolution of 3 km and 100 vertical levels. A 1 km resolution version of the same simulation (GIGATL1) is also vailable over a shorter time span (2 years). We will here use hourly GIGATL3 outputs to generate boundary forcing for a sub-regional zoom over the TZ region (GIGATZ). GIGATZ will have an identical numerical setup, but with increased horizontal (< 1 km) and vertical resolutions (256 vertical levels), to better consider fine scale effects on local NADW pathways and mixing. Finally, sensitivity experiments will be run where sources of internal waves (tidal and near-inertial) will be alternatively switched off.
The GIGATL3 and GIGATZ simulations will be evaluated against available observations, including satellite-based (e.g. altimetry) and in situ (e.g. Argo) data. A preliminary validation of the coarser simulation GIGATL6 (identical to GIGATL3 but at 6 km resolution) with various observational datasets has already confirmed the realistic representation of mean currents (e.g. Gulf Stream separation and NAC) and of surface and interior mesoscale turbulence levels compared to previous simulations with coarser resolution. This validation will be here extended to the TZ region and its western boundary current system (NAC and DWBC) and associated mesoscale field (spatial and temporal scales). If available, the use of sea-level data at kilometric resolution during the fast-sampling phase of the the Surface Water and Ocean Topography (SWOT) satellite mission will be considered to evaluate the representativeness of submesoscale processes in the model (an orbital track of the satellite during the high-resolution calibration phase of the mission indeed falls just within the TZ).
WP1. Implement an observational array and build datasets
See Work at sea
WP2. Leakiness and Pathways of NADW
Once NADW signals (e.g. buoyancy or nutrient concentration anomalies) have entered the DWBC at high latitudes via eddy-driven entrainment or downslope cascading and overflows, how far can they travel ? A unique steady continuous flow along continental margins is very unlikely, and the hypothesis that the TZ effectively disrupts the pathways of NADW is in fact suggested by both observational and modelling studies. A most famous illustration comes from the trajectories of isobaric floats deployed in the UNADW portion of the DWBC at the exit of the Labrador Sea : a majority were not exported to the south and recirculated eastward within the NAC, whilst the few floats reaching subtropical latitudes did so via interior pathways and eddy-driven recirculation gyres instead of the confined DWBC. A recent high-resolution modelling study deciphered this leakiness of the DWBC in the TZ, and suggested that a significant fraction of the cross-shore boundary-to-interior transport of NADW (∼ 15 Sv in total) was typically occurring as inertial separations (i.e. Eulerian-mean detachment of a rapid boundary current due to bathymetric curvature and steepening) at some key hot spots along the TZ margins – namely Flemish Cap and the Grand Banks. The potential occurrence of other mechanistic sources, including NAC-DWBC interactions or instabilities of the DWBC itself, is yet not to be excluded. The objective of WP2 is to provide a novel and comprehensive observation-based description of NADW leakiness and pathways within the TZ. Particular attention will be paid to describing the respective behaviors of UNADW and LNADW along the boundary, and hence to the respective role of these two water masses in maintaining deep meridional connectivity in the North Atlantic.
WP3. Mixing and transformation of NADW
The TZ may not just behave as a passive multi-route gateway for the export of NADW. The latter presumably
undergoes important diapycnal transformation and vertical motions as it flows around and leaks from the continental boundary, breaking further down subpolar-to-subtropical coherence. Realistic numerical simulations show significant mixing-driven diapycnal transformation of NADW to lower isopycnal classes in the TZ that leads to the set-up of an overturning cell internal to the subpolar gyre (4 Sv). Although those results depend on the representativeness of deep mixing in z-level hindcast simulation which can include spurious numerical diapycnal mixing), they are not utterly inconsistent with observations. A global representation of vertical diffusivities derived from the Argo dataset is indeed characterized by generally high values in those regions with rugged topography, energetic mesoscale circulation, and weak stratification, including the TZ. Fine-scale analyses of CTD-LADCP data also show pronounced vertical turbulent mixing around the continental slope of the TZ, with vertical diffusivity reaching up to 10 −3 m 2 s −1 , the very upper range of observed estimates in the region. Moreover, significant deep-reaching upwelling and downwelling and associated tracer transport resulting from both Eulerian mean and eddy-induced bolus flows are likely to take place in the TZ. The objective of WP3 is to provide a quantitative and qualitative description of NADW transformation and vertical motions within the TZ and an assessment of the dominant controlling processes.
WP4. AMOC meridional coherence and ocean heat content budgets
Describing and explaining the horizontal pathways (leakiness and export) and the mixing-driven transformation of NADW in the TZ could eventually shed light on the evolution of basin-scale climate-relevant metrics. As already stated, the local TZ dynamics can influence the meridional coherence of the AMOC through its control on the along-boundary propagation of buoyancy anomalies and the associated intensity of the meridional geostrophic flow. Moreover, the likely importance of diapycnal and vertical motions in the TZ makes it a plausible direct contributor to the sinking branch of the AMOC. Sayol et al. (2019) indeed estimated in a realistic high-resolution model a downward transport of nearly 4 Sv (1 Sv = 10 6 m 3 s −1 ) across 1500 m in the vicinity of Flemish Cap. Secondly, the fate of NADW after its passage in the TZ may partly control the distribution of deep ocean heat and freshwater contents, as already demonstrated for upper ocean water masses. In fact, realistic ocean state estimates show the TZ as a key area for the sequestration and redistribution of ocean heat content (OHC) anomalies into the UNADW layer, and as a dominant contributor to the heat budget of the whole subpolar domain. The objective of WP4 is to evaluate how local NADW dynamics in the TZ can impact the basin-scale hydrography and circulation of the North Atlantic. We will hence attempt a large-scale contextualisation of WP2 and WP3 findings using the AMOC and the OHC as key climate-relevant metrics, and once again deciphering the respective roles of UNADW and LNADW.