Spatial modeling of residential woodsmoke across a non-urban upstate New York region

Residential wood combustion is an important contributor to heating-season ambient fine particle concentrations in many regions of North America. We applied a measurement and modeling approach to evaluate the effect of wood combustion on local and regional air quality in a non-urban setting with complex topography-the Adirondacks region of New York State. Spatially resolved topographic, census, property assessment, and emissions survey data were used to model spatial variability of woodsmoke fine particulate matter (PM2.5) emissions across the region. This spatial emissions model was then used to locate fixed-site monitors and to design a mobile-monitoring campaign. Measurements using a dual-wavelength Aethalometer(TM) enabled the discrimination of woodsmoke from other combustion sources. By comparing these Aethalometer measurements with nephelometer measurements of PM2.5 we determined that the majority of the PM2.5 measured in the Adirondacks during calm, cold winter nights came from wood combustion. We used the measurements of spatial variability in woodsmoke PM2.5 concentrations, the spatial emissions estimates, and additional geographic covariates to develop a spatial model of woodsmoke PM2.5. Considering those living in areas in the upper tertile of modeled woodsmoke concentrations as "exposed," the model estimated that about 20% (similar to 130,000 people) of the population in the study area were exposed to elevated woodsmoke. Further, the model demonstrates that U.S. census information can be combined with additional survey and property assessment data to provide a broadly applicable estimate of woodsmoke spatial patterns and population exposure. This approach is a promising method for screening potential woodsmoke problems, including those areas with complex terrains.