MODELING THE TURBULENCE STRUCTURE IN THE CANOPY LAYER

Citation
Sh. Shen et My. Leclerc, MODELING THE TURBULENCE STRUCTURE IN THE CANOPY LAYER, Agricultural and forest meteorology, 87(1), 1997, pp. 3-25
Citations number
34
Categorie Soggetti
Metereology & Atmospheric Sciences",Agriculture,Forestry
ISSN journal
01681923
Volume
87
Issue
1
Year of publication
1997
Pages
3 - 25
Database
ISI
SICI code
0168-1923(1997)87:1<3:MTTSIT>2.0.ZU;2-Z
Abstract
The large-eddy simulation (LES) is used to investigate the canopy flow structure and the transfer of TKE, momentum and heat within and above a tall canopy in full leaf, This paper reports on a comparison of thr ee simulations performed, one in neutral conditions and two with diffe rent thermal loads on the canopy layer, The instantaneous flow structu res observed in the upper crown region, the horizontally averaged stat istics of the flow, the second-order equations of the Reynolds stress, TKE and heat fluxes, and the quadrant analyses of the Reynolds stress and heat flux are examined for three atmospheric conditions. The LES results of vertical profiles, temperature, momentum and heat fluxes, v elocity variances and skewnesses generally agree well with observation s. The effects of buoyancy forces on the velocity variances and skewne sses are discussed. LES results of instantaneous turbulent flow fields reveal basic characteristics of turbulence structures particularly ne ar the top of the canopy layer. The LES simulation results show that n ear the treetop, shear production constitutes the main source of TKE. The pressure transport, together with the canopy drag are important en ergy source and sink respectively throughout most of the canopy. In th e bottom two-third of the canopy layer, the pressure transport term ex ceeds other production terms. The turbulent transport term shows an ex port of TKE from the upper canopy to the region above the canopy and t o the deeper regions of the forest. Thermal effects inside the canopy layer are generally small in that budget in unstable conditions, excep t for the buoyancy production term which peaks at about two-thirds of the layer. This is in contrast with the corresponding term in the Reyn olds stress budget where buoyancy production is of almost the same mag nitude as the turbulent transport term in unstable conditions, corresp onding to about 15% of the shear production term near the canopy top. In the Reynolds stress budget of the upper canopy region, pressure des truction is balanced by shear production, Turbulent transport and subg rid-scale effects are largest near the canopy top, Thermal effects imp act mostly on buoyancy production and on the turbulent transport term. These effects are most pronounced near the top of the canopy layer, I n the vertical heat flux budget, gradient production and pressure cova riance are the main production and destruction of heat flux terms resp ectively and the gradient production is the largest source term produc ing an upward heat flux in the crown layer and a downward heat flux in the lower canopy, Unlike its corresponding terms in the TKE and Reyno lds stress budgets, this term becomes large and negative in the lower canopy. The importance of the turbulent transport term increases subst antially with increasing. thermal instability. The magnitude of the pr essure covariance, gradient production, turbulent transport and buoyan cy production increases with increased thermal load on the canopy laye r. Quadrant analysis obtained from our LES results both for momentum a nd heat fluxes show that above and below the treetop, the contribution to momentum and to a lesser extent to heat transfer, is chiefly contr olled by the sweep/ejection mechanisms in the cases studied. In these regions, the ejections contribution increases with unstable conditions while the sweeps contribution to the Reynolds stress decreases. Resul ts are similar although more subdued in the case of the heat flux. For both momentum and heat flux. in the lower canopy, thermal stability s ignificantly alters the structure of the flow mostly by shifting the p redominance of the inward interaction quadrant in neutral conditions t o the outward interaction quadrant in unstable conditions. (C) 1997 El sevier Science B.V.