Flow divergence and density flows above and below a deciduous forest. Part I. Non-zero mean vertical wind above canopy

Ecosystem-atmosphere exchange flux measurements above tall vegetation in hilly terrain are well known to suffer from systematic underestimates of nighttime fluxes of CO2 and other scalars with significant sources in or below the canopy. This bias is commonly attributed to advection driven by thermotopographic density flows and the resulting horizontal flow divergence. Flux correction methods have been proposed based on this notion. To examine the structure and dynamics of horizontal flow divergence and vertical convergence mean vertical velocities are analyzed. These are derived from sonic anemometers at 1.8 and 1.3 times canopy height for 3 years above a deciduous forest in hilly terrain at the Morgan-Monroe State Forest (Indiana, USA) Fluxnet site. Measured vertical velocities are linked to forcing parameters represented in the equations of motion and heat for sloping terrain. In the leaf-off season, the data suggest that the dynamics and daily patterns of horizontal flow divergence (implied from vertical convergence through continuity) are entirely consistent with the hypothesis that the divergence is driven by thermotopographic density flows. However, in the vegetative season with a full canopy, a more complex picture emerges, suggesting strong dynamic and thermal decoupling of the horizontal divergence below canopy from flow conditions above. Thus we conclude that flux correction methods based on above-canopy conditions alone may significantly misrepresent scalar transport below canopy during the vegetative season and should be avoided. (c) 2005 Elsevier B.V. All rights reserved.