MODELING URBAN METEOROLOGY OVER IDEALISED CITIES. COMPARISON BETWEEN RESULTS OF URBAN PARAMETERIZATION IMPLEMENTED IN MESOSCALE MODEL AND HORIZONTAL SPATIAL AVERAGE PROPERTIES OBTAINED USING CFD SIMULATIONS

Air quality inside the urban canopy layer (UCL) is important because here is where people live and a significant part of the emissions are located. In this way, the modeling of UCL is also important. Different factors such as the increase of urban population and the improvement of computational power, has produced an increasing interest on urban mesoscale modeling since mid 1990s. However, the modeling of urban boundary layer is difficult because it is influenced by the complex morphology of a city (buildings, cars, gardens) with different mechanical and thermal/radiative properties. In addition, the domain of mesoscale models has a horizontal extension of several tens of kilometers (the whole city and its surrounding area) and, for computational reasons, it is not possible to solve explicitly the flow around buildings. Therefore, urban parameterizations are necessary for high resolution mesoscale simulations. On the other hand, Computational Fluid Dynamics (CFD) models can solve explicitly the flow around buildings but their simulation domains cannot cover the whole city. In this work, focused on mechanical effects produced by buildings, CFD simulations and the horizontal spatial average of the different flow properties are used to assess the performance of an urban parameterization implemented on a mesoscale model and find its strengths and weaknesses. Horizontal spatial average of the CFD results around the buildings are made in order to compare with similar mesoscale variables corresponding to a column of computational cells over a urban zone with the same characteristics as the CFD configuration. In this case, the city is represented by an array of cubes