Role of Eddies in Shaping the Vertical Structure of the Beaufort Gyre Halocline and its Geostrophic Circulation

The Beaufort Gyre is driven by the prevailing anticyclonic wind stress curl over the western Arctic Ocean, which pumps relatively low-salinity water into the gyre interior and deepens the halocline. Storage/release of freshwater occurs during periods of anomalous wind stress curl, which has considerable implications for the Atlantic Meridional Overturning Circulation and thus the global climate. Current theories portray the halocline as a buoyant layer with a thickness that is governed by a balance (or lack thereof) between the Ekman downwelling and mesoscale eddy transport. In equilibrium, the isopycnal slopes are predicted to be proportional to the strength of the surface stress and inversely proportional to the eddy diffusivity. However, eddy diffusivity is often parameterized to be proportional to the isopycnal slope. Climatological observations of the gyre suggest that its isopycnal slope significantly increases with depth, while the mooring-derived isopycnal eddy diffusivities decrease with depth, suggesting that the current theory is insufficient. Here, we develop a simple new theory that reconciles these seemingly contradictory statements.
We extend the single-layer halocline paradigm to a multi-layer model in which the isopycnal interfaces are coupled with each other. The strength and vertical profile of the eddy diffusivities are determined from the baroclinic instability characteristics of the geostrophic currents, which are linked to isopycnal slopes via thermal wind balance. Using this basic model of eddy-mean flow interactions, we identify the critical processes that set the vertical structure of the gyre. Specifically, we prove that potential vorticity sources at the gyre boundaries (over continental slopes) are essential for a realistic vertical distribution of the isopycnal slopes. In the absence of the boundary fluxes, the gyre attracts to a state with depth-invariant isopycnal slope, regardless of the initial conditions. Our results further justify the need for observational constraints on eddy diffusivity and boundary fluxes.