Modelling urban turbulent heat exchanges between the urban canopy and the atmosphere: differences between CFD and surface energy balance models

Urban microclimate models were initially designed to improve weather forecasting and climate modelling. They have been used to assess the large scale deployment mitigation techniques for many years. However, urban climate models necessarily introduce many simplifications and approximations, despite the increase in available computational resources. Here, we present a comparison between the wind velocity computed with a 2-D urban canopy parameterization, and with a computational fluid dynamic model (CFD) by means of FLUENT-ANSYS in 2-D and 3-D. The urban canopy parameterization is the urban tile of the Community Land Model (Oleson et al. 2010). We considered an urban canyon in Milan in peak summer conditions, having equal height and width, and subject to wind having two possible velocities (2, or 4 m s-1). In both cases, we modelled the canyon considering or neglecting the natural convection. In the considered conditions, although urban canopy parameterizations are not capable of distinguishing the wind direction within the urban fabric, they get the magnitude of wind velocity, achieving intermediate values between those computed with 2-D and 3-D CFD modelling.