MODELLING PHOTOCHEMICAL AIR POLLUTION IN SÃO PAULO, BRAZIL Ana G. Ulke 1 and M. Fátima Andrade 2 1 Dept. of Atmospheric Sciences, Faculty of Sciences, University of Buenos Aires, Buenos Aires, Argentina 2 Department of Atmospheric Sciences, Institute of Astronomy and Geophysics, University of São Paulo, São Paulo, Brazil A photochemical airshed model (CIT model, Harley et al., 1993) has been used to study the transformation, transport and removal of the pollutants that cause the air quality deterioration in São Paulo, Brazil. The region is located near the ocean in a complicated topography. The urban core is one of the megacities of the world and the Metropolitan Area of São Paulo is the largest industrialized region in Latin America. Approximately 90% of the ozone precursors are emitted to the atmosphere by the vehicular fleet. The federal imposed limits on ozone and nitrogen oxides have been violated several times every year (Massambani and Andrade, 1994). Meteorological and air quality field measurements were available to perform model simulations over the period 16-19 February 1989. During this four-day interval, no track of fronts was observed over the region, which was under the influence of a high-pressure system. The flow pattern showed the typical development of local circulations (Silva Días and Machado, 1997). A multiday simulation was carried out to avoid the influence of inaccuracies on initial and boundary conditions. The modeling domain, centered in 23.6° S, 46.5° W, covered a area, with 5- km resolution. The vertical extent was 1100m, divided in five cells with varying depth. Meteorological and air quality fields were generated based on measurements, applying objective analysis procedures. The wind field on afternoon hours of the final day of the simulation along with the topography is depicted in Figure 1. Different emission scenarios were considered in order to analyze uncertainties in the emission pattern. There is a lack of accurate information related to the emissions in São Paulo. The 1989 actual fleet scenario (AF) considered the official emission inventory (CETESB, 1990). Alternative scenarios duplicated the nitrogen oxides contributions (NOx2) and the hydrocarbon emissions (HC2). Figures 2 to 4 present the predicted ground level ozone concentrations for Feb. 19, 15 LT. The urban core of São Paulo exhibited inhibition of ozone formation due to high nitrogen oxides levels, while the suburban area showed larger ozone concentrations, the product of photochemistry and transport. The ozone levels predicted by the NOx2 scenario were the smallest. The HC2 scenario led to the greatest downwind ozone concentrations. This alternative resulted also in a satisfactory agreement with observed values. The application of the model demonstrated the need of an accurate emission inventory for improved predictions of the pollutant concentrations. An alternative approach for the turbulence parameterization was introduced in the model. The scheme gives a continuous transition between the different regimes in the atmospheric boundary layer (Ulke, 2000). Figures 5 and 6 show the evolution of the vertical distributions of ozone during the two final days of the period, obtained with the original and the alternative schemes. The selected grid point (X=27.5, Y=15) represents a typical urban location. The alternative parameterization led to greater concentration levels and less uniform distributions. Air Pollution Modeling and Its Application XIV, Edited by Gryning and Schiermeier, Kluwer Academic/Plenum Publishers, New York, 2001 693