Numerical Simulations of the Microscale Heterogeneities of Turbulence Observed on a Complex Site

We simulate the microscale heterogeneities of turbulent variables observed at a complex site for different wind directions. The atmospheric computational fluid dynamics (CFD) results are compared with an ensemble of 36 months of data collected at the experimental site SIRTA “Site Instrumental de Recherche par Télédétection Atmosphérique”, located near Paris (France) in a semi-urban environment. The experimental data show that the normalized turbulent kinetic energy (TKE) k/U ² (where k is TKE and U is the wind speed) at 10-m height, for two different locations, is highly dependent on wind direction and strongly influenced by trees. These measurements show a strong increase of the normalized TKE downstream of the forest canopies with a large variability within the 36-month period in part due to the variation of the tree foliage. The numerical simulations are carried out using the CFD code Code_Saturne with the standard k?ε closure, in neutral stratification. The buildings are taken into account explicitly in the mesh and the forested areas are modelled with two approaches: the classical roughness wall law and a drag porosity. A comparison has been performed between the calculated values and the median of measured values of the normalized TKE and the normalized friction velocity, for each wind sector of 10°. A very good agreement is obtained with the drag porosity model, whereas the classical roughness law leads to a strong underestimation downstream of the forested areas. However, this large improvement of the results using the drag porosity model can only be obtained with a refinement of the grid, especially in forested areas, and an accurate land-use map.