Topographically generated internal waves and boundary layer instabilities

Abstract

Flow over topography has been shown to generate finite amplitude internal waves upstream, over the topography and downstream. Such waves can interact with the viscous bottom boundary layer to produce vigorous instabilities. However, the strength and size of such instabilities depends on whether viscosity significantly modifies the wave generation process, which is usually treated using inviscid theory in the literature. In this work, we contrast cases in which boundary layer separation profoundly alters the wave generation process and cases for which the generated internal waves largely match inviscid theory. All results are generated using a numerical model that simulates stratified flow over topography. Several issues with using a wave-based Reynolds number to describe boundary layer properties are discussed by comparing simulations with modifications to the domain depth, background velocity, and viscosity. For hill-like topography, three-dimensional aspects of the instabilities are also discussed. Decreasing the Reynolds number by a factor of four (by increasing the viscosity), while leaving the primary two-dimensional instabilities largely unchanged, drastically affects their three-dimensionalization. Several cases at the laboratory scale with a depth of 1 m are examined in both two and three dimensions and a subset of the cases is scaled up to a field scale 10-m deep fluid while maintaining similar values for the background current and viscosity. At this scale, increasing the viscosity by an order of magnitude does not significantly change the wave properties but does alter the wave’s interaction with the bottom boundary layer through the bottom shear stress. Finally, two subcritical cases for which disturbances are able to propagate upstream showcase a set of instabilities forming on the upstream slope of the elevated topography. The time scale over which these instabilities develop is related to but distinct from the advective time scale of the waves. At a non-dimensional time when instabilities have formed in the field scale case, no instabilities have yet formed in the lab scale case.