Meeting Abstract
Recent studies on how fish filter food particles out of the water have identified the importance of vortices that form ‘hydrocyclonic nets’ that repel zooplankton away from the filter surface formed by the gill rakers, and concentrates them within the buccal cavity. Furthermore, physical models in flow tanks showed that similar vortices keep part of the branchial sieve void of adhering particles, and thus avoid clogging. One of these models, a simplified representation of the morphology of the ram filter feeding paddlefish, is the cross-step model by Sanderson and co-workers published in Nature Communications in 2016. Apart from flow visualization with dye and local flow magnitude measurements, the hydrodynamics of this cross-step model remain unclear, which limits our insight in the links between the filter’s geometry, flow patterns and filtration performance. I quantified the 3D-hydrodynamics of the cross-step model with ‘ribs’ at 90° and solved particle tracks using the computational fluid dynamics software ANSYS Fluent. The model shows the importance of the resistance by the gill rakers (represented by a nylon mesh in the physical model, and a ‘porous media model’ in the computational model) to establish the vortices. As predicted, the zone on the filter where no particles were retained showed a low pressure and a strongly shearing flow. The vortices consistently showed a helical pattern with a ventral-to-dorsal flow direction. Calculation of the paths of neutrally buoyant spherical particles of a range of sizes did not show the repelling action of the vortex to cause separation from the flow. Therefore, the current computational model predicts that filtration by this model is occurring by cross-flow or dead-end filtration the level of the gill rakers instead of more medial inside the mouth cavity.
