Modeling Daily Plume Specific Smoke Concentrations for Health Effects Studies with Estimates of Fire Size, Plume Age, and Fuel Type

SAM D FAULSTICH , Matthew J. Strickland, Yan Liu, Marcela Loria-Salazar, Xia Sun, Ash B. Cale, Heather A. Holmes

Inhaling smoke PM2.5 can cause adverse health effects ranging from acute (e.g., lung irritation) to chronic (e.g., lung cancer). Acute health effects have immediate implications for public health, requiring rapid response to minimize harm during an exposure window. Estimating acute health effects requires short-term (e.g., daily) estimates of fire-specific smoke PM2.5 concentrations at ground level. Any temporal discrepancy (e.g., missing fire emissions information) may result in underestimated smoke exposure in an epidemiology study. This paper introduces a method to estimate daily fire-specific PM2.5 smoke concentration at ground level in the western US from 2007-2019 to provide smoke characterizations (i.e., exposure estimates) for time-series studies investigating acute health effects. The smoke exposure model incorporates data on fire characteristics, such as fuel type, fire size, and fire distance, enabling a more detailed analysis of health impacts. This method utilizes updated fire emissions information as inputs to an atmospheric dispersion model, which determines the concentration and location of wildfire smoke after transport. These results are combined with a Bayesian time-series model to determine the smoke-specific portion of PM2.5 measured from nine ground-based EPA monitors in the western US. The Bayesian model includes meteorology and season to estimate the background PM2.5 concentrations. Using this data set with retained fire characteristics provides valuable insight into the differences between PM2.5 concentrations at different locations. We found that fires with the largest burned area during the study period (1,753-1,850 km^2) affected six of our nine stations, showing how widespread smoke impacts from large fires can be. The Lindon, UT station was impacted by the greatest number of fires over the period (398), but the average smoke PM2.5 concentration per fire was ~2 ug m^(-3)and the highest smoke PM2.5 concentration was 35 ug m^(-3). In comparison, the Carson City, NV station was impacted by less than half the number of fires over the study period (177), but the average smoke PM2.5 concentration per fire was three times higher (~6 ugm^(-3)), and the highest smoke PM2.5 concentration was 159 ug m^(-3). These examples highlight two significantly different smoke exposure conditions that could plausibly lead to different health outcomes. Being able to investigate the health effects of the fire-specific smoke characteristics improves our understanding of the impacts of smoke exposure and ensures that management strategies are mitigating all possible outcomes of wildfires, including transported smoke.