High-resolution spatiotemporal estimates of total and wildfire PM 2.5 in the contiguous United States, 2018-2024: Implications for population exposure and mortality burden

Wildfire smoke has become an increasing public health concern in the contiguous United States (CONUS). Previous studies have provided high-resolution estimates of either total PM2.5 or wildfire-related PM2.5 during the past two decades, but few have focused on the most recent wildfire-intensive years, and jointly estimated both to quantify wildfire contributions to total PM2.5 exposure. This study developed a high-resolution machine learning framework to estimate daily 1-km concentrations of total and wildfire-related PM2.5 across the CONUS from 2018 to 2024. This framework incorporates chemical precursors and wildfire smoke density as predictors and relies on publicly accessible data sources enabling future updates. The model achieved high performance with cross-validated R2 of 0.72 and 0.89 for daily and annual estimates, respectively. Elevated total PM2.5 concentrations were observed along the west coast and in several metropolitan areas across the CONUS. Wildfire smoke contributed substantially to total PM2.5 concentrations during major wildfire years, accounting for approximately 35% of PM2.5 along the west coast in 2020, and over 25% in the eastern U.S. in 2023. Wildfire-related PM2.5 mortality accounted for 20% of PM2.5-attributable deaths along the west coast in 2020 (30 times that in 2019), and 24% in the northeastern U.S. in 2023 (14 times that in 2019). These findings highlight the growing role of wildfire smoke as a major driver of total PM2.5 exposure and associated health risks in both western and eastern U.S., underscoring the need for sustained nationwide public health interventions to reduce wildfire smoke exposure.