A comparison of two dry deposition parameterizations for O3 within an atmospheric mesoscale eulerian air quality model:: Madrid case study

The importance of including an adequate parameterization of the deposition processes in the simulation of three dimensional pollution fields in mesoscale models is well known. This is essential for improving the determination of the flux removal and for allowing longer simulation periods of the atmospheric processes. In addition, appropiate deposition patterns allow more realistic diagnostics of the impact of different pollutants on different types of terrain, such as the urban surface type. In the present study, we have implemented a resistance deposition model into an Air Quality System (ANA) (San Jose et al., 1996) applied over a large city such as Madrid (Spain). The model domain is 80x100 km which is much larger than the actual urban domain. The ANA model is composed of five different modules: a three dimensional meteorological module which predicts the wind, temperature and humidity fields every time step; a module which provides automatic landuse classification; an emission module, which produces the emissions every hour and with a high spatial resolution (250 x 250 m) using landuse information (for biogenic emissions) from the Landsat-5 satellite image; a photochemical module, which is based on the CBM-IV (Gery et al, 1989) mechanism and solved numerically by following the SMVGEAR (Jacobson and Turco, 1993) method and finally, a deposition module which is based on the resistance approach. Two different deposition approaches have been used in this study; the Wesely (1989) and the Erisman et al. (1994) parameterizations of surface resistances. Comparison between model results using these parameterizations and observations at different stations shows that both approaches improve the simulations of the air concentrations but the Erisman et al. (1994) approach obtains better results than Wesely (1989) under stable turbulent stratification conditions.