Source apportionment of VOCs based on photochemical loss in summer at a suburban site in Beijing

Abstract

Volatile organic compounds (VOCs) are critical precursors of photochemical smog. Quantitatively evaluating the VOC sources and their contributions to ozone formation provides valuable information for photochemical pollution abatement. However, due to the fast oxidization of VOCs during transport in the atmosphere, source apportionment via receptor models using measured VOC data inevitably has a systematic deviation. In this study, we conducted a field measurement of VOCs at a suburban site in Beijing and evaluated the influences of photochemical oxidization on VOC source contributions. The initial VOC concentration, defined as the sum of measured VOC concentration and their photochemical losses. The real-time photochemical age-based parameterization method and a sequential reaction model were applied to characterize the photochemical loss of VOCs. Also, the photochemical loss of VOCs was considered in the he positive matrix factorization (PMF) model. Photochemical losses, on average, accounted for 5.6% of total VOCs (TVOCs) and 59.2% of alkenes, indicating that the impact of photochemical reaction cannot be ignored. Alkanes were the most significant contributors to both measured and initial TVOCs, while the alkenes and oxygenated volatile organic compounds (OVOCs), especially ethylene and isoprene, contributed the most to ozone formation potential. By adopting the positive matrix factorization model, seven VOC sources were identified for both measured and initial TVOCs. Vehicular emissions, industrial sources, and biomass burning were the top three contributors to ambient TVOC. Photochemistry significantly influenced the results of source apportionment, e.g., the contribution of biogenic sources (+4.4%), industrial emissions (+2.5%) and solvent usage (+2.5%) based on initial VOCs were larger than that of measured VOCs, while the contribution of gasoline vehicular emissions (−9.9%) was lower than that of measured VOCs. Our results highlight the necessity of adequately considering the photochemical loss of VOCs for accurately apportioning the emission sources.