Quantitative Analysis of Variability and Uncertainty in On-Road and Non-Road Mobile Source Emission Factors

Abstract

The goal of this research is to demonstrate a methodology for quantifying variability and uncertainty in mobile source emissions. Emission factors and emission estimates are subject to both variability and uncertainty. Variability in emissions deals with real differences in emissions among multiple emission sources at any given time or over time for any individual emission source. Variability is the heterogeneity of values of a quantity with respect time space or across a population. Uncertainty in emissions on the other hand implies the lack of knowledge regarding the true value of emissions. In this research variability and uncertainty are treated separately since their sources are different and as such they affect the decision making process in a different way. For example, sources of variability in mobile source emissions include: vehicle make; ambient temperature; vehicle model; fuel used; vehicle age; and/or driving behavior. Sources of uncertainty may include: small sample sizes; lack of precision and/or accuracy in measurements; non-representativeness; or lack of data. In this work a methodology for simultaneous characterization of variability and uncertainty in mobile source emission factors is described. Variability is characterized by fitting probability distributions to the data sets. Uncertainty due to random sampling error is characterized by using the parametric Bootstrap technique.This methodology is demonstrated for emission factors for three categories: (1) Onroad mobile source exhaust air toxic emissions (2) nonroad lawn and garden equipment emissions and (3) nonroad construction farm and industrial equipment emissions. For the first category a database of vehicular exhaust emissions developed by the California Air Resources Board (CARB) was used. For the second and third categories emission factor databases were developed by reviewing reports and/or technical papers from U.S. Environmental Protection Agency (EPA), Southwest Research Institute (SwRI), CARB and Society of Automotive Engineers (SAE).The main results regarding the demonstrated methodology and related statistical analysis in this research include: (1) emission factor groupings were determined statistically (e.g. by engine size for lawn and garden engines); (2) substantial inter-engine variability in emissions was discovered; and (3) large ranges of uncertainty in mean emissions were found in many cases. In particular, uncertainty ranges for the mean emission factors were in excess of ±20 percent in most cases and as high as minus 49 percent to plus 89 percent.