Pollutant dispersion in tree-lined canyons: aligning wind-tunnel experiments with large-eddy simulation

Vegetation is increasingly integrated into urban environments to improve microclimate and air quality. However, the role of trees in pollutant dispersion within street canyons remains complex and debated. This study presents a detailed comparison between high-resolution wind tunnel experiments and large-eddy simulations (LES) using uDALES. Such a comparison is essential to validate the physical realism of numerical models and assess their ability to capture the complex, vegetation-induced interactions that govern urban ventilation and pollutant transport. The analysis includes velocity and concentration measurements in an idealised street canyon with varying tree densities. Special attention is given to the definition of reference scales for the flow and scalar fields, ensuring consistency across methods through the use of friction velocity and canyon geometry as scaling quantities. The tree drag length also proves crucial for aligning the aerodynamic resistance of physical and simulated vegetation. The simulations reproduce key features of the mean flow, both with and without trees—notably the structure and extent of large-scale recirculation cells—but tend to underestimate turbulent kinetic energy and scalar fluxes. This suggests that tree models acting solely as momentum sinks do not fully capture vegetation-induced turbulence. Despite these limitations, the LES captures the dominant ventilation mechanisms at the canyon roof. The estimated bulk exchange velocity shows no systematic dependence on tree number or drag intensity, in contrast with previous studies, highlighting the need for context-specific assessments of urban vegetation effects.