LAUSR

Abstract The identification of air pollution sources is a challenging task due to the confounding effects of local and transboundary emitters. While the Potential Source Contribution Function (PSCF) is a useful tool for trajectory analysis and source identification, it can overestimate the impact of less polluted areas in which a trajectory event passed through. In this work, we propose incorporating air pollutant concentration fields from reanalysis data to improve the PSCF method's performance that avoids the overcontribution of less polluted areas. We develop the 3D-PSCF-CONC based on an updated three-dimensional version of the PSCF method (3D-PSCF). We used as a test case the Metropolitan Area of Sao Paulo (MASP) in Brazil, one of the most polluted regions in the world, for comparing the results of the conventional and the new PSCF method. A monitoring station in Vitoria - Espirito Santo, Brazil which has smaller local emissions than MASP was used as an auxiliary test case. Hourly carbon monoxide (CO) and sulfur dioxide (SO2) concentrations from 2015 to 2019 of monitoring stations were analyzed. The daily average concentrations of CO and SO2 were used as a threshold on the probability calculations. An overall of 1825 backward trajectories from HYSPLIT version 4 was processed using the three models. CO and SO2 surface concentrations from MERRA-2, with a resolution of 0.5° x 0.625° was incorporated in the 3D-PSCF-CONC to correct the potential contribution calculation based on the air pollutant concentration field. Our results revealed that the 3D-PSCF-CONC reached a more consistent identification of local and regional sources, rather than emphasizing mostly the regional sources as the other PSCF methods. The 3D-PSCF-CONC reduced the potential contribution of less polluted and high precipitation areas. The analysis suggests a contribution of vehicular sources close to the receptor site in MASP and long-range transport of industrial emissions and biomass burning from the land clearing of sugarcane production. This work presents an important tool for understanding the air pollution process and source identification.