Testing a Multilayer Urban Parameterization Against Spatially Averaged Profiles of Wind and Turbulence from a Microscale CFD Model

A thorough validation of multilayer urban canopy parameterizations (UCP) in a real urban environment has always been a challenging task due to the strong heterogeneity of the flow and turbulent fields in cities. By definition, UCPs embedded in mesoscale models provide vertical profiles of meteorological variables that represent horizontal spatial averages over areas comparable with the grid cell size (typically of the order of 10⁶-10⁵ m², e. g. much larger than the typical building dimensions). A direct measure of such horizontal averages is impossible to obtain from point measurements in the urban canopy layer, in particular for variables like wind speed and direction or turbulent kinetic energy (TKE) that are strongly affected by channeling and blocking effects in the street canyons. Nowadays, the accuracy of microscale models in urban areas is becoming enough to provide realistic flow fields that can be used to derive such spatial averages. In this contribution, a CFD microscale model is used to simulate the flow field over a domain comparable with the typical resolution of mesoscale models, around a square in the southern part of the city of Madrid (Spain).  Highly detailed urban morphology data are used to represent the buildings. The model is run in neutral conditions, for 16 different wind directions, with 22.5 Degrees intervals. To assess the validity of the CFD, results are compared to wind and turbulent measurements recorded in two points around the square. Then, the horizontal spatial averages of the two wind components, TKE and turbulent momentum fluxes are used to validate the vertical profiles obtained from a series of mesoscale simulations carried on with the model WRF and the multilayer UCP BEP-BEM, run at high vertical resolution. Different urban morphological inputs are used for the UCP, from standard values derived for the corresponding urban climate zone to the detailed morphological characteristics of the area, to assess the sensitivity to this parameter.  In addition to this, a passive tracer, emitted at street level is added both in the CFD and in the mesoscale simulations, so to evaluate the ability of the model to reproduce pollutant dispersion. Finally the analysis of the CFD results is pushed further to investigate the importance of the dispersive fluxes and the spatial variability of flow and pollutant in the domain, so to get indications about their possible modelling in the contest of mesoscale models.