Experimental and Numerical Study of Wind and Turbulence in a Near-Field Dispersion Campaign at an Inhomogeneous Site

Abstract

We present a detailed experimental and numerical study of the local flow field for a pollutant dispersion experimental program conducted at SIRTA (Site Instrumental de Recherche par Teledetection Atmospherique), a complex and intensively instrumented site in a southern suburb of Paris. Global analysis of continuous measurements over 2 years highlights the impact of terrain heterogeneity on wind and turbulence. It shows that the forest to the north of the experimental field induces strong directional shear and wind deceleration below the forest canopy height. This directional shear is stronger with decreasing height and decreasing distance from the forest edge. Numerical simulations are carried out using Code_Saturne, a computational fluid dynamics code, in Reynolds-averaged Navier–Stokes mode with a standard $$k{-}\varepsilon $$ closure and a canopy model, in neutral and stable stratifications. These simulations are shown to reproduce globally well the characteristics of the mean flow, especially the directional wind shear in northeasterly and northwesterly cases and the turbulent kinetic energy increase induced by the forest. However, they slightly underestimate wind speed and the directional shear of the flow below the forest canopy height. Sensitivity studies are performed to investigate the influence of leaf area density, inlet stability condition, and roughness length. These studies show that the typical features of the canopy flow become more pronounced as canopy density increases. Performance statistics indicate that the impact of the forest and adequate inlet profiles are the most important factors in the accurate reproduction of flow at the site, especially under stable stratification.