Apportionment of Vehicle Fleet Emissions by Linear Regression, Positive Matrix Factorization, and Emission Modeling

Real-world emission factors for different vehicle types and their contributions to roadside air pollution are needed air-quality management. Tunnel measurements have been used estimate several using linear regression or receptor-based source apportionment. However, the accuracy uncertainties of these methods not sufficiently discussed. This study applies four derive from tunnel in Hong Kong, China: (1) simple regressions (SLR); (2) multiple (MLR); (3) positive matrix factorization (PMF); (4) EMission FACtors Kong (EMFAC-HK). Separable include those fueled by liquefied petroleum gas (LPG), gasoline, diesel. PMF was most useful, as it simultaneously seeks profiles contributions. Diesel-, gasoline-, LPG-fueled emissions accounted 52%, 10%, 5% PM2.5 mass, respectively, while ammonium sulfate (~20%), nitrate (6%), road dust (7%) were also large contributors. MLR exhibited highest relative uncertainties, typically over twice determined SLR. EMFAC-HK has lowest due its assumption a single average factor each pollutant category under specific conditions. The SLR comparable.