Statistical Framework Using GIS to Estimate Unpaved Road VMT for ${\mathrm{PM}}_{10}$ Emission Inventories

The latest Clean Air Act Amendments mandate that every area in the nation meet standards for six criteria pollutants—ozone, sulfur dioxide, carbon monoxide, nitrogen dioxide, lead, and particulate matter—both with an aerodynamic diameter less than 10 μm $\left({\mathrm{PM}}_{10}\right)$ and with an aerodynamic diameter less than 2.5 μm $\left({\mathrm{PM}}_{2.5}\right)\text{.}$ Regions in nonattainment of these standards must produce an emission inventory for the violating pollutant. By both California and federal air quality standards, all counties in the San Joaquin Valley (SJV) air basin were ${\mathrm{PM}}_{10}$ nonattainment areas in 2001. Reentrained dust created by vehicles is considered a large contributor to ${\mathrm{PM}}_{10}$ emissions in the SJV. However, current methods used to estimate ${\mathrm{PM}}_{10}$ emissions from vehicle activity on unpaved roads do not account for spatial or seasonal variability. Using statistical models and GIS, we introduce a new framework for estimating vehicle miles traveled on unpaved roads. Because a large portion of the SJV contains agricultural land, our models distinguish between vehicular traffic on unpaved roads associated with harvest and nonharvest activities. Using the method on counties in the SJV, we find that the new methodology results in approximately 29% less vehicle miles traveled on unpaved roads in San Joaquin County than calculated by the current California Air Resources Board (CARB) method, and approximately 40% less vehicle miles traveled in Fresno County. Furthermore, the results indicate that integrating GIS methods in ${\mathrm{PM}}_{10}$ inventory procedures greatly enhanced the overall process of model development and estimation.