Forest edge in turbulent inflow: Comparison between full-scale and wind tunnel experiments

Wind and turbulence in the near-surface boundary-layer are strongly modified by forest edges. In order to describe and quantify these effects, a full-scale experiment was performed at a dense and straight beech forest edge in Denmark (Dellwik et al., 2014). The study was focused on the near-edge region (between “the Impact Region” and the “Roughness change” region according to the description by Belcher et al., 2012), where the flow changes rapidly due to horizontal and vertical pressure gradients. Two masts were placed 1.5 canopy heights (hc) upwind and downwind of the edge, with sonic anemometers at five and seven levels, respectively. Three levels of measurement within the canopy were conducted on the mast inside the forest. In 2019, we attempted to reproduce the findings from the full-scale experiment using a forest model in the Wind Engineering, Energy and Environment (WindEEE) Dome facility at Western University in Canada. Forest edges have previously been scaled in wind tunnels by use of, for example, pegs, metal strips, metallic mesh and realistic looking small trees. Because we wanted to match the flow pattern around a specific forest edge, and the connection between model forest and full-scale beech forest could not be deduced before the study, we needed to use a model that could be quickly adapted in terms of porosity. To this purpose, we used 0.4 m tall plastic tubes placed in a regular staggered configuration and iteratively added several layers of netting.

This study concerns an evaluation of how the different iterations of adding netting as well as a change to the forest edge shape affected the similarity between the full-scale and scaled experiments. The forest model is evaluated in terms of resulted first-, second- and third-order wind statistics as well as spectra. We demonstrate that the change from a square forest edge to a tapered edge affected the flow strongly and that several layers of netting were needed to reproduce how the mean wind speed accelerated over the edge in the full-scale experiment. The effect of the netting also strongly increased the similarity between scaled and full-scale flow characteristics taken inside the canopy. A significant change was also observed in the surface pressure observations.

Despite that there were several levels of sonic anemometers on two masts in the full-scale experiment, many benefits were found from the increased vertical and horizontal resolution in the scaled study. In the scaled study in WindEEE Dome, wind measurements were taken both with a rake of pressure impact probes, from which the horizontal wind speed was deduced and so-called Cobra probes, which measure all three wind components. With these rakes, the vertical resolution was 3-5 times finer than the resolution in the full-scale experiment. Whereas the full-scaled experiment was focused on the interval between -1.5 and 1.5 hc, the scaled experiment was extended to x = 3.5 hc, with a horizontal resolution of 0.375 hc, where x is the horizontal distance downwind of the forest edge. The high vertical resolution in the impact probe rake indicated that near and above the top of the trees at x = 1.5 hc, strong gradients in third-order moment (from positive to negative) indicates a presence of a mixing layer between the low wind speed inside the canopy and the relatively much faster wind speed above the forest. The fine horizontal resolution allowed for a detailed description of the flow evolution in a range where the flow is changing suddenly, which is of great benefit for the understanding of the different physical processes involved.

A difficulty in scaled wind tunnel studies concern the limitation on accurately reproducing the turbulence in the inflow, as well as flow evolution in the absence of a model forest. Even if WindEEE is a large facility and the flow is highly turbulent, the horizontal turbulent length scale in the inflow was an order of magnitude lower than the length scale in the inflow to the full-scale experiment. Implications of such differences in the background flow are also addressed in the presentation.

The presentation ends with a suggestion on how to make the results of scaled experiments more readily comparable to each other by more consistently reporting either the surface pressure gradient or the observed drag force, for the benefit of faster progress in boundary-layer vegetation interaction studies.

References:

Belcher S.E., Harman, I.N., Finnigan, J.J. (2012), The Wind in the Willows: Flows in Forest Canopies in Complex Terrain, Annu. Rev. Fluid Mech. 2012. 44:479–504, 10.1146/annurev-fluid-120710-101036

Dellwik, E., Bingol, F., Mann J (2014) Flow distortion at a dense forest edge, Quarterly Journal of the Royal Meteorological Society, 140, 676-686, doi 10.1002/qj.2155