Coherent Flow Structures in the Pore Spaces of Permeable Beds underlying a Unidirectional Turbulent Boundary Layer: A Review and some New Experimental Results

Flow over a permeable wall is perhaps one of the most common, yet also one of the most complex, types of boundary condition found in any natural geophysical flow. Our current state of knowledge concerning turbulent boundary layer flow is dominated by studies of impermeable boundaries, with far less research having addressed the influence of bed permeability. This chapter briefly reviews the current understanding of flows above and within a permeable bed by considering experimental and theoretical studies of both the freeflow and transition layers (the Darcian layer is not considered here). Thiswork highlights the fundamental differences between flows above permeable and impermeable beds, and examines the implications of these differences for the mechanisms of mass and momentum exchange. Some of the peculiarities of coherent flow structures formed within the freeflow over permeable beds have now been revealed. For example, permeable beds are characterized by significant injection and suction events that move fluid across the interface, and result in an absence of the low-speed streaks that are ubiquitous over impermeable beds. Moreover, contrary to flow over impermeable surfaces, ejection events dominate over sweep events, and this has significant implications for the mechanisms of energy dissipation that occur within the transition layer. However, these mechanisms are still poorly understood, partly because of a lack of direct observations from this experimentally challenging layer. Using new experimental techniques, namely endoscopic particle imaging velocimetry (EPIV) and refractive index matching (RIM), the present chapter examines the complexity of flow within the pores of the transition layer. Coherent flowstructures, including vortices, jets and bifurcations, are imaged for the first timewithin the pore spaces of a submerged permeable bed. These new data suggest that the manner in which the transition layer is commonly represented in current numerical models may be inappropriate, and that future work must better account for flow across this dynamic interface.