|
Abstract:
|
The negative health impacts of exposure to fine particulate matter, or PM2.5, are well-documented as is the relationship between wildfire smoke and increased pollution levels. However, less well-researched is the causal effect of fire smoke on PM2.5 concentrations as determined by a potential outcomes framework from the Rubin Causal Model. We present a study on the causal effect of wildfire smoke on PM2.5 concentrations in California in 2008. The stringency of the potential outcomes framework does not readily lend itself to problems that are environmental and spatial in nature. In this project, we address those difficulties and present a potential outcomes model for PM2.5 that incorporates computer model data via a spatial hierarchical spatial downscaling model. Our novel approach makes use of monitoring data from the US Environmental Protection Agency's Federal Reference Method sites, satellite-visible smoke plume images from the NOAA Hazard Mapping System and numerical model output from CMAQ. Our results provide insight into using causal inference methods with spatial data and a method that can be extended to investigating the causal effects of wildfire smoke on mortality.
|
