Pediatric Risk Mapping From Co-Exposure to Extreme Temperatures and Air Pollutants

Exposure to extreme temperatures and fine particulate matter is hazardous to human health, especially among children. With growing evidence on the impact of these environmental hazards on child health-including increased risks of respiratory illnesses, heat-related illnesses, and developmental challenges-there is a substantial need to identify high-risk areas for potential adverse pediatric health outcomes. In this study, we developed pediatric vulnerability indices at the census tract scale across the contiguous United States (CONUS) using five dimensionality reduction methods, including unsupervised machine learning and deep learning approaches. We integrated these indices with data on extreme temperature (2012-2024), PM_2.5, and black carbon exposures from 2010 to 2020, enabling spatial analysis of environmental hazards and pediatric vulnerability. Among the five methods tested, principal component analysis (PCA) was selected for its balanced representation of 12 vulnerability factors, which explained 23% of the variance in the data. We observed that co-exposure to extreme temperature and air pollutant exposures were highest in the West, Southwest, and certain parts of the South and Northeast. Approximately 36% of the CONUS showed statistically significant hotspots of co-exposures to higher temperatures and air pollutants, concentrated in the West, Northwest, and Southwest. High-risk areas for pediatric vulnerability and co-exposure to environmental hazards were identified in both urban and rural communities, including indigenous lands and agricultural regions. These findings can aid policymakers and public health officials as a preliminary resource in developing heat action plans and allocating cooling centers to protect children living in the most affected communities.