Sensitivity of an urban sub grid scheme to the canyons orientations over a wide metropolitan area

Abstract: In mesoscale atmospheric modeling, adequate reconstruction of the meteorological field over cities, is essential for an accurate determination of the parameters governing transport and dispersion of pollutants in the urban boundary layer. In recent years, many efforts have been made by several authors to improve and integrate the description of the city effects within the numerical weather prediction (NWP). Among the various schemes proposed in the past, we recall the single-layer model of Masson (2000) and the multi-layer model of Martilli (2002). Between the two, the more easily implemented in a numerical model seems to be the Town Energy Budget (TEB) of Masson (2000). The TEB has been used several times in mesoscale numerical models. For example, Lemonsu and Masson (2002) described the interaction of the summer breeze with the city of Paris (France), while Freitas et al. (2007) investigated the breeze circulation during winter over the metropolitan area of São Paulo (Brazil).

The TEB scheme is based on the canyon concept (e.g., Oke, 1987). The canyon is the space contained between two facing buildings, and it is delimited by four surfaces, respectively, roof, two walls and road. In accordance with the definitions adopted by Porson et al. (2009) the TEB scheme can be defined as a three facet model (hereinafter TFM). Unlike the TFM, the four facet model (hereinafter FFM), in the radiative and thermal budget, takes into account both the walls of the canyon and their mutual orientation as well. Porson et al. (2009) showed that the TFM exhibits a good agreement with the FFM, except for an error that comes from considering the average orientation of the urban canyons instead of the actual one. It is not still clear, however, how large the amplitude of the error associated with a TFM description is on a wide area (i.e., the domain of influence of a surface node in the computational grid) where several canyons with different geometric and thermal property, characterized by different orientations are present. In order to describe a complex urban area, like a typical European city, we want to investigate how much the averaging over the street canyon orientation affects the accuracy of the energy budget.

To do this, an urban sub-grid scheme based on TEB model was coupled with the Regional Atmospheric Modelling System (RAMS, version 6.0, Pielke et al. 1992, Cotton et al. 2003) to perform an investigation over the urban area of Rome (Italy). The TEB original scheme was modified in order to take into account the heat budget of both the canyon walls and their own orientation. An urban digital elevation model for the city of Rome (Italy) with an average resolution of 1.5 meters, provided all the necessary geometrical information, while, for the sake of simplicity, for the emissivity and other thermal characteristics a single set of parameters was used. From this digital model, by a shape finding algorithm, the canyon surfaces approximating the urban buildings were extracted and displaced according the FFM geometry. This database has been included in the Rams 6.0, to take into account each canyon. In order to verify the capability of FFM and TFM to reproduce the urban energy balance, two model runs are then performed to simulate typical summertime atmospheric conditions as most favorable to observe the urban heat island (UHI) effects. In the first one, the modified TEB was used, while in the second, the unmodified TEB scheme (that, like a TFM, averaged over the street-canyon orientation) was adopted.

Time series of air temperature taken at a meteorological site located in the city centre of Rome, were compared with the numerical results of both simulations. It was verified that the use of the TFM applied to areas where there is an uniform canyons orientation distribution, does not produce a significant difference in comparison with FFM. In contrast, in areas characterized by an non-uniform canyons orientation distribution, the canyon orientation produces a not negligible error. This analysis has allowed us to ascertain that, for the purpose of atmospheric modeling, the choice of proper urban sub-grid scheme must be supported by the knowledge of the urban surface texture.