Assessing the effects of viaduct configurations on traffic pollutant transport: A large-eddy simulation study comparing quasi-two-dimensional and three-dimensional urban canyons

Viaducts are widely used across high-density urban environments. Although their influence on pollutant dispersion has been studied, the majority of studies use quasi-two-dimensional (quasi-2D) street canyon models, with few directly comparing them to three-dimensional (3D) ones. This study explores these differences by conducting large-eddy simulations using both quasi-2D and 3D models to assess traffic pollutant dispersion in street canyons with urban viaduct configurations under perpendicular approaching wind. Four configurations are examined, including ground-level roads only, viaducts without noise barriers, and viaducts with bilateral noise barriers of 2 or 4m in height. Five important parameters, including mean wind velocity, pollutant concentration, indoor personal intake fraction (P_IF), vertical mass fluxes, and pollutant exchange velocity, are employed to evaluate the differences between quasi-2D and 3D models. The findings show that quasi-2D models with various viaduct configurations yield mean indoor P_IF for buildings away from the road and overall street canyon pollutant concentrations comparable to those from 3D models. However, quasi-2D models with viaduct configurations tend to overestimate spatially averaged pollutant concentrations at pedestrian-level zones by up to 1.8 times and overestimate mean indoor P_IF in near-road buildings by as much as 54.8% compared to the 3D cases with viaduct configurations being investigated. Moreover, 3D models with viaduct configurations exhibit better pollutant removal efficiency, with the dimensionless pollutant exchange velocity through the roof level up to 33.6% higher than in quasi-2D models.