# Turbulent flows over dense filament canopies

## TL;DR

This study uses DNS to analyze turbulent flows over dense filament canopies, revealing how element spacing and height influence flow structures, turbulence characteristics, and the formation of Kelvin-Helmholtz-like rollers.

## Contribution

It provides new insights into the flow dynamics over dense canopies, especially regarding the role of spacing and height in turbulence and roller formation, supported by DNS data.

## Key findings

- Kelvin--Helmholtz-like rollers dominate within dense canopies.
- Spacing influences dispersive velocity fluctuations and roller wavelength.
- Shallow and very small spacings inhibit roller formation.

## Abstract

Turbulent flows over dense canopies of rigid filaments of small size are investigated for different element heights and spacings using DNS. The flow can be decomposed into the element-coherent, dispersive flow, the Kelvin--Helmholtz-like rollers typically reported over dense canopies, and the background, incoherent turbulence. The canopies studied have spacings $s^+ = 3$--$50$, which essentially preclude the background turbulence from penetrating within. The dispersive velocity fluctuations are also mainly determined by the spacing, and are small deep within the canopy, where the footprint of the Kelvin--Helmholtz-like rollers dominates. Their typical streamwise wavelength is determined by the mixing length, which is essentially the sum of its height above and below the canopy tips. For the present dense canopies, the former remains roughly the same in wall-units, and the latter, which scales with the drag length, depends linearly on the spacing. This is the result of the drag being essentially viscous and governed by the planar layout of the canopy. In shallow canopies, the proximity of the canopy floor inhibits the formation of Kelvin--Helmholtz-like rollers, with essentially no signature for height-to-spacing ratios $h/s \approx 1$, and no further inhibition beyond $h/s \approx 6$. Very small spacings also inhibit the rollers, due to their obstruction by the canopy elements. The obstruction decreases with increasing spacing and the signature of the instability intensifies, even if for canopies sparser than those studied here the instability eventually breaks down. Simple models based on linear stability can capture some of the above effects.