Fire-Pollutant-Atmosphere Components and Its Impact on Mortality in Portugal During Wildfire Seasons

Ediclê de Souza Fernandes Duarte; Vanda Salgueiro; Maria João Costa; Paulo Sérgio Lucio; Miguel Potes; Daniele Bortoli; Rui Salgado

doi:10.1029/2023GH000802

Fire-Pollutant-Atmosphere Components and Its Impact on Mortality in Portugal During Wildfire Seasons

Geohealth. 2023 Oct 6;7(10):e2023GH000802. doi: 10.1029/2023GH000802. eCollection 2023 Oct.

Authors

Ediclê de Souza Fernandes Duarte^{1

2

3}, Vanda Salgueiro^{1

2

3}, Maria João Costa^{1

2

3}, Paulo Sérgio Lucio⁴, Miguel Potes^{1

2

3}, Daniele Bortoli^{1

2

3}, Rui Salgado^{1

2

3}

Affiliations

¹ Instituto de Ciências da Terra-ICT (Pólo de Évora) Instituto de Investigação e Formação Avançada (IIFA) Universidade de Évora Évora Portugal.
² Earth Remote Sensing Laboratory (EaRSLab) Instituto de Investigação e Formação Avançada (IIFA) Universidade de Évora Évora Portugal.
³ Departamento de Física Escola de Ciências e Tecnologia (ECT) Universidade de Évora Évora Portugal.
⁴ Departamento de Ciências Atmosféricas e Climáticas Universidade Federal do Rio Grande do Norte Natal Brazil.

Abstract

This study analyzed fire-pollutant-meteorological variables and their impact on cardio-respiratory mortality in Portugal during wildfire season. Data of burned area, particulate matter with a diameter of 10 or 2.5 μm (μm) or less (PM₁₀, PM_2.5), carbon monoxide (CO), nitrogen dioxide (NO₂), ozone (O₃), temperature, relative humidity, wind speed, aerosol optical depth and mortality rates of Circulatory System Disease (CSD), Respiratory System Disease (RSD), Pneumonia (PNEU), Chronic Obstructive Pulmonary Disease, and Asthma (ASMA), were used. Only the months of 2011-2020 wildfire season (June-July-August-September-October) with a burned area greater than 1,000 ha were considered. Principal component analysis was used on fire-pollutant-meteorological variables to create two indices called Pollutant-Burning Interaction (PBI) and Atmospheric-Pollutant Interaction (API). PBI was strongly correlated with the air pollutants and burned area while API was strongly correlated with temperature and relative humidity, and O₃. Cluster analysis applied to PBI-API divided the data into two Clusters. Cluster 1 included colder and wetter months and higher NO₂ concentration. Cluster 2 included warmer and dried months, and higher PM₁₀, PM_2.5, CO, and O₃ concentrations. The clusters were subjected to Principal Component Linear Regression to better understand the relationship between mortality and PBI-API indices. Cluster 1 showed statistically significant (p-value < 0.05) correlation (r) between RSDxPBI (r _RSD = 0.58) and PNEUxPBI (r _PNEU = 0.67). Cluster 2 showed statistically significant correlations between RSDxPBI (r _RSD = 0.48), PNEUxPBI (r _PNEU = 0.47), COPDxPBI (r _COPD = 0.45), CSDxAPI (r _CSD = 0.70), RSDxAPI (r _CSD = 0.71), PNEUxAPI (r _PNEU = 0.49), and COPDxAPI (r _PNEU = 0.62). Cluster 2 analysis indicates that the warmest, driest, and most polluted months of the wildfire season were associated with cardio-respiratory mortality.

Keywords: atmospheric pollutants; cardio‐respiratory mortality; environmental health; environmental risks; multivariate statistical techniques.