Impact of regional biomass burning and meteorological conditions on ambient PM2.5 and oxidative potential in São Paulo, Brazil

The regional and long-range transport of biomass burning plumes can significantly influence air pollutant concentrations in urban environments. In Brazil, recurring fires in the central and northern regions, as well as within the state of São Paulo, contribute to elevated PM_2.5 levels across the southern and southeastern areas. During the sampling period of this study (May 2021–August 2022), a large forest fire in the Metropolitan Area of São Paulo (MASP) significantly impacted PM_2.5 concentrations at the sampling site near downtown São Paulo, a region already heavily affected by vehicular emissions. PM_2.5 was collected using a Partisol 2025i Sequential Air Sampler. Elemental composition was determined via energy-dispersive X-ray fluorescence (EDXRF) and the Lab Organic Carbon-Elemental Carbon (OC-EC) Aerosol Analyzer. Emission sources were identified using Positive Matrix Factorization (PMF), and oxidative potential was assessed through electron spin resonance (OP^ESR) and dithiothreitol (OP^DTT) assays. On August 22, 2021, a major fire occurred at Juquery State Park in Franco da Rocha, burning approximately 1,175 hectares (more than half the park's vegetation) and causing severe damage to local flora and fauna. Soot from the fire was deposited across the city of São Paulo. The highest 24-hour PM_2.5 concentrations of the entire campaign were recorded during this episode. Between August 22 and 25, average peak concentrations of PM_2.5, OC, and biomass burning-related black carbon (BC_BB) reached 43.3, 23.5, and 2.2 µg m^-3, respectively. Trajectory analyses using NOAA’s HYSPLIT model indicated that winds from the northwest, north, and northeast transported the smoke plume from the park to the sampling site. The highest PM_2.5 and BC concentrations were traced back to the park region, confirming the fire’s impact. After it was controlled, pollutant levels declined, helped by 27.3 mm of rainfall on August 28, following 10 days without precipitation. Atmospheric stability likely also contributed to the four-day duration of the fire. During the event, volume-based OP^ESR (OP^ESRv) values decreased while OP^DTTv remained elevated, following the trends of PM_2.5, OC, and BC_BB. Multiple linear regression (MLR) analysis showed significant associations between OP^ESRv and BC_BB, fossil fuel-related BC (BC_FF), EC, and several elements, including S, K, Mg, Cu, Br, Ni, Si, Al, Pb, and Cr. OP^DTTv showed similar associations, although EC, S, Cu, Na, and Zn were not statistically significant. OC was not significantly associated with OP^DTTv, possibly due to a lower redox activity or a stronger influence of metal components. Source-based MLR indicated vehicle emissions and secondary aerosols as the major contributors to OP^ESRv, while OP^DTTv was primarily driven by biomass burning, vehicular emissions, and marine aerosols. Therefore, these results highlight the combined influence of local, regional, and long-range sources, as well as meteorological conditions, on urban air quality.