Comment on “Abyssal Upwelling and Downwelling Driven by Near-Boundary Mixing”

McDougall and Ferrari have estimated the global deep upward diapycnal flow in the boundary layer overlying continental slopes that must balance both downward diapycnal flow in the deep interior and the formation of bottom water around Antarctica. The decrease of perimeter of isopycnal surfaces with depth and the observed decay with height above bottom of turbulent dissipation in the deep ocean play a key role in their estimate. They argue that because the perimeter of seamounts increases with depth, the net effect of mixing around seamounts is to produce net downward diapycnal flow. While this is true along much of a seamount, it is shown here that diapycnal flow of the densest water around the seamount is upward, with buoyancy being transferred from water just above. The same is true for midocean ridges, whose perimeter is constant with depth. It is argued that mixing around seamounts and especially midocean ridges contributes positively to the global deep overturning circulation, reducing the amount of turbulence demanded over the continental slopes to balance the buoyancy budget for the bottom and deep water.