INTEGRAL EFFECTS OF DEEP CONVECTION.
(SEND, U and MARSHALL, J), JOURNAL OF PHYSICAL OCEANOGRAPHY, vol. 25, no. 5, pp. pages, 1995.
The large-scale, integral effect of convective elements (plumes) constituting an open-ocean chimney is investigated both theoretically and with a plume-resolving numerical model. The authors consider an initially homogeneous “patch” of ocean of depth H, with Coriolis parameter f, in which buoyancy is lost from the surface at a rate B. Both vorticity constraints on the convection patch and model analyses imply that, irrespective of the details of the plumes themselves, the mean vertical transport resulting from their action must be vanishingly small. Plumes are best thought of as mixing agents, which efficiently homogenize properties of the chimney. Scaling laws are derived from dynamical arguments and tested against the model. Using an expression for the vertical mixing timescale, they relate the chimney properties, the strength of the geostrophic rim-current setup around it, and its breakup timescale by baroclinic instability to the external parameters B,f, and H. After breakup, the instability eddies may merge to form larger “cones” of convected water, which offset the buoyancy loss at the surface by laterally incorporating stratified fluid. Properties of the plumes only enter the scaling results by setting the vertical mixing timescale. The authors argue that the plume scale may be parameterized by a mixing scheme if this implies the appropriate mixing timescale. Finally, the authors suggest that for the estimation of deep-water formation rates the volume of convectively modified fluid processed by a chimney should be computed rather than the mean vertical transport during the convection phase.