Aerosol emissions from wildfires and prescribed burns: implications for firefighter exposure and air quality

[eng] Wildland fires, including both wildfires and prescribed burns, are becoming increasingly frequent and intense across Southern Europe due to climate and land-use changes, posing growing threats to air quality and human health. However, estimating health impacts of aerosols emitted by wildfires remains a challenge, partly due to uncertainties in wildfire emissions and exposure assessment. Understanding pollutant concentrations in smoke-filled environments is key to developing mitigation strategies that protect exposed populations, public health and the environment. This doctoral thesis aims to assess the impacts of smoke emissions on ambient air and firefighters’ personal exposure to fine particulate matter (PM2.5), black carbon (BC) and polycyclic aromatic hydrocarbons (PAHs). Between 2022 and 2024, experimental field campaigns were conducted in two of Spain’s most fire-prone areas: Galicia (NW Spain) and Catalonia (NE Spain), using a combination of fixed monitors and portable sampling devices. - Air quality monitoring focused on ambient PM2.5 concentrations in Galicia using a network of PurpleAir monitors strategically deployed at firefighter stations during summer seasons over three consecutive years. Data from the 2022 campaign, marked by the highest number of large wildfires (>500 ha), was analyzed applying source apportionment tools to quantify wildfire impacts. - Personal exposure monitoring took place in Catalonia during prescribed burns and wildfires, using portable real-time PM2.5 and BC monitors (AirBeam2 and Micro-aethalometers AE51 and MA200) worn by firefighters. PAHs were analyzed in PM2.5 filters collected from BC monitors. Additionally, PAHs were determined in silicone wristbands (SWBs) worn by firefighters as passive air samplers, and air concentrations were calculated using a compound-specific kinetic uptake model. Different methods of source apportionment were applied based on the types of datasets generated: - For spatially resolved ambient PM2.5 concentrations, quasi-empirical orthogonal functions (QEOFs) obtained with Positive Matrix Factorization (PMF) were used in combination with HYSPLIT back-trajectory and Concentration Weighted Trajectory (CWT) analyses. Results from the QEOF-PMF analysis identified 18 wildfire events, with the Folgoso do Courel fire as a dominant source of PM2.5 spikes (>300 μg/m3; 1-hour mean), and also revealed transboundary smoke transport from Portugal. Personal exposure data showed significant and comparable PM2.5 and BC concentrations during wildfires and prescribed burns (wildfire mean PM2.5 = 152 μg/m3 vs. 110–145 μg/m3 for prescribed burns). BC/PM2.5 ratios indicated varying contributions of mineral aerosols to the emissions mix originating from firefighting tasks. Among these, mop-up tasks were identified as a significant contributor to PM2.5 exposures, with PM2.5 peak concentrations reaching up to 1190 μg/m3 (1-minute average). - For personal BC exposure concentrations, multi-wavelength BC data and individual PAH profiles were used to determine task-specific and source-specific contributions. Source apportionment of BC data distinguished task-specific exposure profiles: line operators (managing fire perimeters) were primarily exposed to biomass-burning BC (𝐵𝐶𝑏𝑏 = 61%), while torchers (initiating fires using fossil-fuel drip torches) were predominantly exposed to fossil-fuel BC (𝐵𝐶𝑓𝑓 = 77%). In addition, PAH analysis from filters and SWBs confirmed these task-specific exposure patterns. Torchers experienced the highest toxic burdens (mean ΣPAHs = 394 ng/m³, benzo[a]pyrene (BaP) = 56 ng/m³ on filters; ΣPAHs = 5004 ng/m³, BaP = 29 ng/m³ in SWBs), exceeding excess lifetime cancer risk (ELCR) thresholds. In contrast, liners showed lower exposures (mean ΣPAHs = 115 ng/m³, BaP = 12 ng/m³ on filters; ΣPAHs = 1642 ng/m³, BaP = 5 ng/m³ in SWBs. In addition, diagnostic isomer ratios (e.g., benzo[a]pyrene/benzo[e]pyrene indicated the relative distance from fresh emissions, while the presence of retene pointed to pine wood combustion. Overall, the integration of fixed-site ambient and portable personal real-time monitoring with passive sampling offered complementary insights into wildfire smoke exposure. The combination of these tools proved effective in identifying spatial and task-specific pollutant patterns, and was considered a valuable tool for exposure assessment. Furthermore, these methods may also contribute to the validation of smoke dispersion model and the development of mitigation strategies addressing environmental and health concerns related to wildland fires.