Source apportionment of hourly-resolved ambient volatile organic compounds: Influence of temporal resolution

Abstract

High temporal-resolution VOC concentration data can provide detailed and important temporal variations of VOC species and emission sources, which is not possible when using coarse temporal-resolution data. In this study, we utilized the positive matrix factorization (PMF) model to conduct source apportionment of hourly concentrations of nineteen VOC species and CO measured at the Mong Kok air quality monitoring station, operated by the Hong Kong Environmental Protection Department, from January 2013 to December 2014. The PMF analysis of the hourly dataset (PMF_Hourly) identified five sources, including liquefied petroleum gas (LPG) (contribution of 45%), gasoline exhaust (21%), combustion (20%), biogenic emission (9%), and paint solvents (6%). The diurnal patterns of VOC emissions from identified sources are likely to be affected by the strength of emissions, variation of the planetary boundary layer height, and photochemical reactions. In addition, the PMF analyses of hourly and 24-hour averaged data of the hourly-resolved data (PMF_Hourly and PMF_Daily) were generally comparable, but the time series of VOC emissions from PMF_Hourly could not be well captured by PMF_Daily for two local VOC sources of gasoline exhaust and LPG. This study highlights the benefit of high temporal-resolution measurement data in apportioning VOC sources, hence providing critical information on VOC emission sources (e.g., diurnal variations) for controlling VOC emissions effectively.