Lagrangian clustering detection of internal wave boluses

The shoaling of internal waves on a continental slope or shelf produces boluses that travel up the slope with the wave. The boluses are regions of trapped fluid that are transported along with the wave, unlike fluid in the bulk that is temporarily pertubed by a passing wave. Boluses have been observed to transport oxygen-depleted water and induce rapid changes in temperature (Walter et al, JGR, 2012), both of which have potential ramifications for marine biology. Several previous studies have investigated boluses in systems with two layers of different density (e.g., Helfrich, JFM, 1992, and Sutherland et al., JGR, 2013). We conduct laboratory and computational studies of bolus generation and material transport in continuously stratified fluids with a pycnocline, as in the oceans. Our laboratory experiments in a 4 m long tank are complemented by 2-dimensional direct numerical simulations of the Navier-Stokes equations. Efforts have been made to identify boluses with Eularian measures in the past, but a Lagrangian perspective is necessary to objectively identify the bolus over its lifespan. Here we use a Lagrangian based coherent structure method relying on trajectory clustering using the fuzzy c-means approach (Froyland and Padberg-Gehle, Chaos, 2015). The objective detection of a bolus enables examination of the volume, distance traveled, and increased available potential energy of a bolus, as a function of the stratification, wave properties, and the angle of the sloping topography. The decay of a bolus through turbulent mixing is investigated by locating where the Richardson number drops below ¼, where velocity shear overcomes the tendency of a stratified fluid to remain stratified. (supported by ONR MURI grant N000141110701)