In the austral spring, biomass fires affect a vast area of South America each year. We combined in situ ozone (O3) data, measured in the states of São Paulo and Paraná, Brazil, in the period 2014-2017, with aerosol optical depth, co-pollutants (NOx, PM2.5 and PM10) and air backtrajectories to identify sources, transport and geographical patterns in the air pollution data. We applied cluster analysis to hourly O3 data and split the investigation area of approximately 290,000 km2 into five groups with similar features in terms of diurnal, weekly, monthly and seasonal O3 concentrations. All groups presented a peak in September and October, associated with the fire activities and enhanced photochemistry. The highest mean O3 concentrations were measured inland whilst, besides having lower concentrations, the coastal group was also associated with the smallest diurnal and seasonal variations. The latter was attributed to lower photochemical activity due to frequently occurring overcast weather situation. The mean annual regional contribution of O3 over the area was 61 μg/m3, with large seasonal and intersite variabilities (from 35 to 84 μg/m3). The long-range transport of smoke contributed with between 23 and 41% of the total O3 during the pollution events. A pollution outbreak in September 2015 caused many-fold increases in O3, PM2.5 and PM10 across the investigation area, which exceeded the World Health Organisation recommendations. We show that the regional transport of particulates and gas due to biomass burning overlays the local emissions in already highly polluted cities. Such an effect can outweigh local measures to curb anthropogenic air pollution in cities.
Keywords: Agricultural fire; Air quality; Cluster analysis; Short-lived climate forcer; Transboundary pollution.
Copyright © 2019 Elsevier Ltd. All rights reserved.