Source Apportionment

Abstract

Source apportionment models use variations in air pollutant composition, time, and space to infer source types and quantify their contributions. Marker species are sought in source emissions that minimize collinearity among the pollutant patterns found in many source types. Engine exhaust, biomass burning, geological material, sea salt, and secondary sulfates and nitrates can be separated and quantified with commonly measured elements, ions, and carbon fractions. More detailed organic compounds, isotopic abundances, and particle morphology are usually needed to distinguish source subtypes, such as diesel and gasoline engine exhaust. The data sets are noisy, and uncertainties in source contribution estimates must be assessed. At a minimum, variability in the chemical patterns in emissions and ambient concentrations should be quantified and propagated. Sample sizes should encompass a range of source contributions, which can be accomplished by locating sampling sites and selecting measurement periods where large contributions are expected. A “weight of evidence” approach assesses uncertainties by the application of several source apportionment techniques with reconciliation of discrepancies for the source contribution estimates.