Análise de tendências de poluentes atmosféricos no estado do Paraná com dados de reanálise do Copernicus Atmosphere Monitoring Service

Abstract

Reanalysis products are an accessible alternative for obtaining air pollutant concentrations in the absence of in situ data. This study estimated the long-term trends (2003-2019) of CO2, CH4, O3, CO, PM10, PM2.5 and black carbon (BC) concentrations using data from the Copernicus Atmosphere Monitoring Service (CAMS) reanalysis in Cascavel, Guarapuava, Ponta Grossa, Grande Curitiba and Londrina (PR). The data was meteorologically normalized and trends were obtained using the Theil-Sen estimator. O3, CO and PM10 concentrations from CAMS in Grande Curitiba and Londrina were compared with in situ measurements. CAMS overestimated monthly O3 concentrations by 171.8% (Grande Curitiba) and 44.0% (Londrina). On the other hand, the monthly averages of CO and PM10 were underestimated (86.1% Grande Curitiba [CO] and 17.3% Londrina [PM10]). For O3 and PM10, the correlation coefficient (r) between CAMS and in situ data on a monthly basis exceeded that on a daily basis, with the greatest variations for O3 (0.37 vs. 0.55) in Grande Curitiba and PM10 (0.54 vs. 0.81) in Londrina, suggesting the use of CAMS data on a monthly basis. In Grande Curitiba, the correlation for CO was weak (r < 0.2). The positive trends in CO2 and CH4 followed global trends and received local contributions, such as the increase in the vehicle fleet and in final waste disposal and effluent treatment. The increased use of gasohol and hydrated ethanol in Paraná may have contributed to the positive O3 trends. With regard to CO, the negative trends (up to 2013) were related to industrial emissions, while the positive trends were probably due to CAMS using outdated emissions from the vehicle sector. Similarly, the negative (up to 2011) and positive trends for PM10 and PM2.5 were associated with emissions in both sectors, as were the negative trends for BC. Despite the potential of using CAMS data in the study region, some caveats should be considered, particularly in epidemiological applications where high accuracy is required to assess health impacts. This highlights the importance of preparing reliable regional inventories of anthropogenic emissions of atmospheric pollutants, so that they can be used in reanalysis products. Finally, before using CAMS data, it is advisable to validate it with in situ data.