Space-time characterization of fire-related air pollutants over Portugal.

Air pollution has significant and severe impacts on human health, the environment, materials, and the economy, emerging as a key issue for microclimate and air quality regulation. Hence, the spatial and temporal characterization of air pollutants and their relationship with meteorological constraining factors is of utmost importance, particularly under a climate change perspective. Air pollutants’ spatial and temporal characterization over the Iberian Peninsula is performed, focusing particularly on the emissions of Particulate Matter (PM) and Carbon Monoxide (CM) during wildfire events in 2012-2023. This will be performed based on the Copernicus Atmosphere Monitoring (CAMS) data, to profit from the added value of having reliable and gridded information on the atmosphere composition and its related processes, anywhere in the world. After a preliminary analysis, air quality (AQ) forecasts are produced to model AQ environmental emergencies. The rationale is to develop a methodology to forecast air pollutants' exceedances, without being limited to areas closer to monitoring AQ stations. To achieve this goal, Machine Learning (ML) methods are applied to find the most efficient model architecture predicting pollutants’ concentration a few days ahead.

Space-time patterns reveal a good agreement between CAMS data and extreme fire events, with this agreement being clearer for maximum concentrations measured by CAMS pollutants such as CO, PM10, and PM2.5, as well as the exceedances of pollutant thresholds during fire activity periods, over affected regions. ML models reveal coefficient of determination (R) values ranging from 0.76 to 0.85 for their forecasts, revealing high accuracy in predicting PM10 exceedances, with precision levels up to 0.91.

Model results reveal the potential to develop an air quality tool over regions less covered by the national air quality monitoring network.

Acknowledgements: This study is partially supported by the Portuguese Fundação para a Ciência e a Tecnologia (FCT) I.P./MCTES on behalf of DHEFEUS -2022.09185.PTDC and the project FAIR- 2022.01660.PTDC).