The Creation and Evolution of Coherent Structures in Canopy Flows

Turbulent structures have been an important focus of investigators seeking to understand transport in turbulent shear flows. One of the most important discoveries in canopy turbulence is that the transport is dominated by canopy-scale coherent structures. As with most turbulent flows, the details of the origin and evolution of structures in canopy flows are still uncertian. This debate is complicated, in part, by the difficulty of visualizing and quantifying coherent structures. Because structures vary in time and space, traditional ensemble averaging techniques are not useful. At relatively low Reynolds number, the range of turbulent length scales is low enough that structures can be visualized instantaneously. However, this leads to much uncertainty as to how these processes translate to flows with large Reynolds numbers, which are of greater interest in environmental applications. This leads to the need for a conditional sampling technique that instead visualizes a structure as a composite or ensemble of individual structure realizations.

In this numerical simulation study, we use an approach that combines instantaneous structure visualization and composite averaging to determine how coherent structures are created and evolve in canopy flows. Of primary interest is to determine how coherent structures are associated with the most striking characteristics of canopy turbulence such as the so-called ‘scalar microfront'. We first examine start-up flow, where the range of turbulent length scales is small enough to allow clear visualization of instantaneous structures. This is then translated to the fully-developed flow using a novel composite-averaging technique. The composite averaged structure is based on a trigger of coherent vertical motions of Lagrangian fluid parcels.

Results reveal the hairpin-pair structures that have been reported in previous studies, which we conclude results from the translative instability combined with vortex ‘tearing'. However, we find that this particular structure formation is not necessarily responsible for scalar microfronts. The hairpin structures are preferentially detected in composite averaging methods that use a trigger seeking the strongest events. Using our method which minimizes this bias, we find that scalar microfronts result from the underlying quasi-two-dimensional spanwise roller structures originally postulated by the mixing-layer analogy.