Deep ocean circulation in the subpolar North Atlantic observed by acoustically-tracked floats

Abstract

The deep circulation in the subpolar North Atlantic determines the spread and mixing of high latitude climate signals to lower latitudes. However, our current understanding of the subpolar deep circulation has been limited due to relatively sparse observational data. To improve that understanding, we construct gridded fields of mean velocity and eddy kinetic energy (EKE) in the deep (1800–2800 dbar) subpolar North Atlantic using direct velocity measurements from 122 subsurface acoustically-tracked floats that drifted during June 2014–January 2019. The mean velocity field reveals a relatively strong deep boundary current around Greenland and in the Labrador Sea, with a weaker deep boundary current over the eastern flank of the Reykjanes Ridge, and near-zero mean flow over the western flank, implying a discontinuous deep boundary current across the subpolar basin. The deep EKE, albeit with smaller magnitudes, generally resembles the EKE pattern at the ocean surface, including relatively high values along pathways of the North Atlantic Current and west of Greenland where the Irminger Rings are formed. A surprising finding about deep EKE is an elevated band east of Greenland that parallels the coast and is not present in the surface EKE field. This high EKE band is possibly attributed to the combined influence from propagating Denmark Strait Overflow Cyclones, variability of the wind-driven recirculation offshore of southeast Greenland, and/or topographic waves. The float-based flow fields constructed in this study provide an unprecedented quantitative view of the kinematic properties of the large-scale deep circulation in the subpolar North Atlantic. Combined with cross-basin Eulerian measurements, we believe these recent observations provide a benchmark for testing and improving numerical simulations of deep ocean circulation and the Meridional Overturning Circulation, which are urgently needed for climate change predictions.