Comparing Monthly Ozone Concentrations in the Sao Paulo Metropolitan Area: Seasonal Variability or Land Use Characteristics?

Health and air quality management in large metropolitan areas have always presented a challenge for local authorities and environmental agencies throughout the world, especially when it comes to developing countries. In Brazil, this can be verified in the São Paulo Metropolitan Area (SPMA), with about 20 million residents and approximately 9 million light and heavy-duty vehicles. It is located at about 750m above sea level, surrounded by high hills of up to 1200m, which act as natural barriers for air flow, potentially trapping atmospheric pollutants which contribute to degrade air quality. Tropospheric ozone (O3) remains as one of the most hazardous air pollutants in the city, exceeding the Air Quality Standards and with no significant tendency of decrease in the last years, as shown by State Environmental Agency (CETESB) data. Most pollutants in this area tend to have a seasonal cycle with a minimum during Summer, when the atmosphere is more turbulent, and a maximum in the more stagnant wintertime conditions, under the influence of the South Atlantic Subtropical Anticyclone. However, since O3 is produced photochemically in the atmosphere when nitrogen oxides and volatile organic compounds emitted by vehicle exhaust react with solar radiation, it shows an almost opposite seasonal cycle than most pollutants, with a minimum in Autumn and early and mid-Winter, and a maximum in Spring, with a secondary maximum in late Summer. Previous works have related this cycle to a decrease in incoming solar radiation and temperature during Winter, and an increase in Spring, while in mid-Summer cloud cover increase due to local convection, somewhat counter-acting the increase in radiation. Because O3 is a secondary pollutant, many factors aside from the atmospheric conditions influence its formation and concentration. Precursor emission pattern characteristics are of vital importance on its concentrations, since O3 can be formed and consumed by NOx and VOCs (especially in Brazil, where there is an abundance of VOCs in the atmosphere due to the widespread use of ethanol as vehicle fuel). In this context, the distance of the monitoring site from the main roads is of key importance to understand the spatial variability of ozone, which typically presents its highest concentrations in places away from the main city activities, such as urban parks and residential areas. Different kinds of land use and land cover, which affects road density and amount of vehicles in a given neighbourhood, must be monitored in order to understand more completely how population can be exposed to this pollutant. We analysed hourly data ranging from 1996-2011 from four different ozone monitoring stations from CETESB's automatic network, representing four different categories: vehicular, commercial, residential and background, by calculating monthly means. Typical concentrations for Summer (January), Autumn (April), Spring (October) and Winter (July) months were compared in all stations, and then compared to the seasonal cycle, in order to assess if the differences in concentrations due to land use, observed when comparing different stations, were greater than the control exerted by the atmosphere, observed in the seasonal cycle. Results pointed that ozone concentrations were lowest in the vehicular station, followed by the commercial station (which presented the greatest variability), then the background station (located in a city park), and highest in residential areas. Concentrations were 41% higher in the residential area compared to the vehicular station in Winter, with monthly averages ranging from 29.43µg/m3 (residential station) to 17.48µg/m3 (vehicular station), with a minimum difference between the residential and vehicular station observed in Autumn (22%). When looking at the seasonal cycle in each station, the Spring maximum is 46% higher than the Winter minimum for the commercial station, with a range of 20.82µg/m3 to 38.52µg/m3, respectively, and 37% higher in the background station. The highest monthly averages in the vehicular station (28.74 µg/m3, in Spring) were of absolute values similar to the lowest monthly averages in background and residential stations (28.81µg/m3 and 29.43µg/m3, respectively, in Winter), showing that, in the seasonal scale, ozone concentrations in the city can be affected considerably by land use, through the emission of NOx and VOCs, almost as much as by seasonal variation, through the availability of solar radiation and high temperatures associated to ozone production. It was possible to conclude that, in the seasonal scale, temporal variability of ozone can be associated more strongly to atmospheric variability (if people are exposed to chronic, long-term high concentrations throughout the year), whereas the absolute concentration values (if people are exposed to acute, short-term high concentrations during the month) are more closely related to the land use in the vicinity of the monitoring point. Future studies will be carried out in this project, using atmospheric modelling (with WRF/Chem) in the area of study. By changing land use factors in the model, replacing a built-up area inside the city by forest or grassland, we will be able to assess concentrations differences. We expect that concentrations should increase in areas where city parks are to be built if they are bordered by main roads with precursor emission, exposing the visiting population to higher ozone concentrations than before.