A wind tunnel study of stably stratified flow on a ridge covered with a tall plant canopy

A large fraction of the 400+ Flux Tower sites that comprise the ‘Fluxnet' suffer from advective flows at night, which are now known to be caused by gravity currents. These flows are poorly understood, not least because there are no field experiments that have followed the development of the turbulent wind and temperature fields over a hill of simple geometry under stable conditions. As a first attempt to do this without the enormous expense of a multi-tower field study we have performed a novel wind tunnel experiment where a heated model canopy on a low two-dimensional ridge was mounted upside down on the roof of the tunnel to generate a stable flow.

The canopy is made of bluff metal strips that can be heated independently of the ‘soil' surface. Their geometry is identical to a model whose turbulence statistics are well understood. Low wind velocities are needed to obtain realistic Froude Numbers but a comparison of mean and turbulent statistics at high and low windspeeds indicated that the flow maintains a sufficiently high effective Reynolds Number at the lowest windspeeds studied.

The onset of a gravity current that is downslope on both sides of the hill is controlled by a Froude No. (Fr) that can be expressed in flux units. Scale analysis shows that this Fr depends on the length of the hill and the temperature anomaly of the gravity current but not on the magnitude of the hill slope. On the upwind hillside, the flow within the canopy is in the opposite direction to the main flow above. On the downwind side, the gravity current removes the separated flow region that is present in neutral stratification. As Fr is reduced from near neutral, the downslope flow appears first as a thin layer at the surface but grows to fill the canopy as Fr falls further.

The stably stratified flow has several unexpected features. First, the flow within the canopy is clearly turbulent but supports little or no momentum flux, indicating that the turbulent fluctuations within the canopy are mainly irrotational. Second, the gravity current within the canopy penetrates upwind on the flat surface a distance of -10L from the hill crest, where the hill width is 4L. Third, there is a significant region of enhanced heat flux at the upwind terminus of the gravity current.

These results are of potential importance to the design and interpretation of flux tower experiments. In particular the decoupling of the within- and above-canopy flow and its dependence on Fr suggests physically based criteria for the onset of nighttime advective flows at flux tower sites. Furthermore, the significant extent of upwind penetration of the gravity current, which appears to be driven by the thermal wind term, is information that could be critical in tower siting.