Western Boundary Current Separation: Inferences from a Laboratory Experiment

Abstract

Observations of a laboratory model of a western boundary current, and its separation and subsequent meandering, are described. The current is established by pumping fluid through a rotating channel that contains a topographic ? effect and continental slope topography. The observations are compared with a theoretical model of all three aspects of the current: the structure of the attached current, the process of separation, and the dynamics and path of the meandering jet. This model includes a viscous boundary layer for the attached current, with a thickness of order [?/(dvI/dy)]1/2, where ? is kinematic viscosity and dvI/dy is the velocity gradient of the inviscid (free slip) flow along the boundary. Comparison between the observations and the model show that the attached boundary current is governed by potential vorticity conservation and the Bernoulli equation, and the pressure decreases along its length. The separation of this current from the sidewall is then caused by the minimum pressure level that is set by the downstream conditions in the tank, which forces the current into deeper water. The process is analogous to the separation of a boundary layer from a surface in an adverse pressure gradient in nonrotating flows. This process has implications for the separation of ocean boundary currents, where the details are more complex but clear analogies exist. Meanders in the separated current are qualitatively consistent with an inertial jet model, although detached eddies attributable to instability are also observed.