Hindcasting fine particulate matter in Alaska, US, during wildfire seasons
Year: 2025
Type: article
Abstract: Background Alaska’s (US) frequent wildfire activity affects air quality across large regions, highlighting the need for reliable surface-level fine particulate matter (PM2.5) concentration estimation methods in areas with limited monitoring coverage. Aims We aim to develop a computationally efficient machine learning approach for estimating PM2.5 concentrations, ensuring model accuracy remains above a coefficient of determination (R2) threshold of 0.80. Methods We constructed random forest models using data from May–September 2019–2022. Inputs included land cover, land use, meteorological variables, fire activity metrics, aerosol optical depth (AOD) observations and measurements from low-cost and regulatory-grade monitors. Key results All models demonstrated strong explanatory power. The top-performing model, which incorporated AOD, achieved a five-fold cross-validated R2 of 0.865. Models relying solely on active fire data (excluding AOD) also performed well, with an R2 of 0.833. Conclusions Models based on fire activity metrics, such as total fire count, can reliably estimate regional PM2.5 levels. Incorporating additional variables like wind patterns yielded only marginal improvements, underscoring the value of streamlined modeling approaches. Implications These approaches can enhance air quality monitoring in rural and fire-prone regions by generating consistent PM2.5 estimates. With expanded observations and strategic deployment of low-cost sensors, these models could support near-real-time monitoring and long-term wildfire smoke exposure research.
DOI: 10.1071/wf25087
Authors: Allison Bredder, Tatiana V. Loboda, Dong Chen
Institution: University of Maryland, College Park
Language: English
Subfield: Environmental Engineering
Field: Environmental Science
Domain: Physical Sciences
SDGs: Life on land, Climate action
Open Access status: hybrid
OpenAlex ID: W7117388654