Existing theories on the turbulence structure in the atmospheric surface layer are based on the hypothesis of the terrain homogeneity. However, the surface of the Earth is covered with roughness elements which disturb the turbulent flow and influences the processes that govern the exchange of momentum, heat, and mass between the surface and the planetary boundary layer (PBL). The structures contribute to about 40 to 70% of energy and matter transport between the surface and the atmosphere (Barthlott et al., 2007). In this study, we investigate experimentally and numerically the impact of terrain complexity on the near-surface coherent structures properties using large-eddy simulation (LES) performed with and without forest canopy and turbulence data collected by the SIRTA observatory.
2. EXPERIMENTAL STUDY
Since April 2005, turbulence data have been collected at the French ground-based remote sensing atmospheric observatory, SIRTA, located in Palaiseau, which operates a 30 m instrumented mast with two anemometers at and 30 m heights.
We classify our dataset following the wind direction:
· Wind direction 320°-40°: Close forest
· Wind direction 100°-170°: Distant forest
· Wind direction 170°-260°: Buildings
· Wind direction 260°-320°: Open field
Note that there is an open "green" area with at least 3-4h fetch next to the mast in all directions, with h the roughness element height.
2.1 Turbulent fluxes and turbulent kinetic energy
The friction velocity, u*, strongly depends on the wind direction. We can notice a significant difference between the measurements at 10 and 30 m particularly for wind sectors corresponding to upstream distant forest and buildings. For most homogeneous sectors, the values of u* measured at 10 and 30 m are very similar. For the north sector, the two anemometers are located within the wake zone where strong shear-induced mixing occurs, homogenizing the vertical profile of u* at the two levels. The friction velocity u* is also strongly dependent on stratification. Indeed, for very unstable conditions, the differences are minimum between 10 and 30 m heights and wind sector which is consistent with the fact that the source of turbulence is essentially of thermal origin. This difference between 10 and 30 m becomes significant for unstable, stable and very stable conditions. Under such conditions, the main source of turbulence is mechanically induced by shear and advection is dominant favouring turbulence transport to the measurement tower.
2.2 Coherent structures
Coherent structures in the atmospheric surface layer contribute to about 40 to 70% of energy and matter transport between the surface and the atmosphere. The identification of coherent structures consists in detecting ramp-like pattern in the time series of the temperature fluctuations with a wavelet analysis. We can notice that the terrain heterogeneity seems to have no impact on coherent structures occurrence, only the stability seems to have an impact on coherent structures occurrence. Similar results have been found for their contribution to the total fluxes. However, our dataset does not allow the extension of this concept to normalized fetch distance x/h smaller than 3-4.
3. NUMERICAL STUDY
3.1 LES simulation with the model ARPS
The LES simulations were performed using the numerical model ARPS (Dupont and Brunet, submitted). The grid is orthogonal on the horizontal direction and stretched on the vertical one. The model solves the conservation equation for the three wind velocity components, the pressure, the potential temperature and water. 3D simulations were performed within this m3 domain, corresponding to grid points in the x, y and z directions, respectively. In order to reproduce the experimental conditions, we extract the data simulated at 1Hz at 10 and 30 m. The forest height is set at 20m with a frontal area density profile, LAI = 2.
3.2 Turbulent fluxes and coherent structures
We first studied the averaged turbulence variables such as the turbulent kinetic energy or the friction velocity. In the heterogeneous case, the flow is more turbulent then for the homogeneous case, so, the forest barrier has a strong impact on the turbulence variables.
For the coherent structures analysis, we used the same detection technique. The uncertainties associated with the mean values are large (about 20-25%). Note that the properties of the coherent structures are slightly different between homogeneous and heterogeneous cases. But, the uncertainty is large (~25%), so, the effect of the forest on the coherent structures is not evident. Despite the limited number of data simulated in this numerical study, the comparison with the experimental investigation (Fesquet et al.) is interesting. The FO is rather well predicted by the numerical model with similar values. We notice a contribution to the momentum flux larger in the numerical study. These differences between LES and experimental study are probably due to the fact that the simulations correspond to "idealised" conditions with fixed parameters (for example the LAI), neutrally stratified and hence, are not exactly representative of the "real" conditions of the SIRTA laboratory. Anyway, we find a good agreement between observations and simulations.
This study investigates experimentally and numerically the impact of terrain complexity on the coherent structures properties using large-eddy simulation (LES) performed with and without forest canopy and turbulence data collected by two sonic anemometers at 10 and 30m height on the 30-m tower. Using an objective detection technique based on wavelet transforms of fluctuations time series of atmospheric variables (vertical wind component or temperature) measured by the SIRTA and simulated by LES, the study shows that whatever the upstream complexity of the terrain, the coherent structures display turbulence properties which are independent of the complex nature of the terrain. These results are fundamental since their "universal" properties may probably facilitate their parameterization in the numerical models even in presence of surface heterogeneities.
Barthlott C., Drobinski P., Fesquet C., Dubos T., Pietras C., 2007. Long-term study of coherent structures in the atmospheric surface layer. Boundary-Layer Meteorol. 125, 1-24.
Dupont S., Brunet Y. Influence of foliar density profile on canopy flow: a large-eddy simulation study. (submitted).
Fesquet C., Drobinski P., Barthlott C., Dubos T. Impact of terrain heterogeneity on near-surface turbulence structure. (submitted).