Experiments on the flow over a hill covered by a canopy in stably stratified conditions

Abstract

It has long been suspected that thermo-topographic flows, especially gravity currents, within vegetation canopies on complex terrain are one of the main reasons behind the failure to reconcile micrometeorological and biometric estimates of canopy-atmosphere exchange at many sites. However, the physical mechanisms governing the initiation and the scaling of these flows remain poorly understood. Here we present the results of a novel wind tunnel study that looks in detail at the flow within and above an open canopy in stably stratified conditions and investigates the physical mechanisms responsible for gravity currents within canopies. The wind tunnel simulations demonstrate that gravity currents are established through a complex balance of competing forces on the flow within the canopy. Three forcing terms act on the flow in the canopy as it passes over the hill. First is the hydrodynamic pressure gradient associated with the boundary layer flow aloft; second, a hydrostatic pressure gradient associated with the displacement of temperature and density surfaces by the hill, and finally a thermal wind term, where a streamwise pressure gradient is caused by changes in the depth of the temperature perturbations to the flow. The net balance of these forces is opposed by the canopy drag. Gravity currents, however, do not appear unless the turbulence, which supports the transport of momentum into the canopy, is also reduced. This suppression occurs preferentially deep within the canopy due to a Richardson number cut-off effect, which is directly linked to the different transport mechanisms of heat and momentum across the boundary layers on the canopy elements. The gravity current first appears at the ground surface, despite cooling profiles that are concentrated in the upper canopy. Once initiated, a gravity current can propagate substantial distances away from the triggering topography, driven by the thermal wind term. If shown to be robust these results have widespread implications for the micrometeorology, atmospheric boundary layer and numerical weather prediction communities.