Search for a command to run...
At neighbourhood scale, the urban climate is governed by multi-scale exchanges of momentum, heat, and water vapour occurring at the canopy top and between districts through turbulent diffusion and advection. At the scale of several neighbourhoods, individual buildings cannot be explicitly resolved using for example an immersed boundary method in atmospheric models, and their parameterisation through a simple roughness length is insufficient to accurately represent turbulent exchanges within the roughness sublayer. An alternative consists in modelling the urban canopy as a porous medium, following the drag–porosity approach commonly used for vegetation canopies, in order to improve the representation of wind dynamics inside the roughness sublayer. However, unlike vegetation canopies, urban canopies are characterised by solid volumes, sharp edges and strong spatial heterogeneity, which strongly modulate the dominant turbulent motions responsible for turbulent exchanges. Instead of assuming a horizontally homogeneous porosity field as in vegetation canopies, we investigate an adaptation of the drag–porosity approach for urban canopies by concentrating the porosity at building locations, explicitly accounting for the three-dimensional urban morphology at the metre scale. This so-called object-based porosity approach is evaluated using large-eddy simulations of the flow over a staggered array of buildings in neutral thermal stratification. We analyse the differences in wind dynamics obtained by representing the urban canopy through a drag–porosity approach and the object-based porosity method, in comparison with explicit building-resolving configurations from literature. Flow statistics and conditional averaging show that, compared to the homogeneous drag–porosity approach, the object-based formulation yields a more realistic representation of turbulent momentum exchanges at canopy top and better captures the dominant coherent structures within the roughness sublayer.