Source analysis of volatile organic compounds by positive matrix factorization in urban and rural environments in Beijing

This paper applies advance receptor model positive matrix factorization (PMF) source analysis to 1-h resolution VOC data collected at Yufa (rural site) and Peking University, or PKU (urban site), in Beijing. A range of major VOC sources was identified, including vehicle emissions, liquefied petroleum gas, coal combustion, and biogenic emissions. Vehicle activities contributed approximately 62% of VOC loading and 55% of ozone forming potential at the PKU site, compared to 38% of VOC loading and 42% of ozone forming potential at Yufa. These results indicate that the control of vehicle emissions is essential to alleviating VOC pollution, particularly in urban Beijing. We found that VOC emission strengths are relatively consistent throughout the day in the rural area; variation of mixing height therefore is a controlling factor for ground-level VOC concentration. In the urban area, both vehicle activity and variation of mixing height strongly impact VOC levels. Local sources within Beijing appeared to contribute most of the VOCs recorded at both urban and rural areas. However, as some of the VOC species are quite reactive, VOC emitted from distant sources would have been depleted during transportation, concurrent with the formation of secondary pollutants such as ozone and sulfate. Such depletion would mean that this source apportionment analysis would tend to overestimate the ozone forming capacity of local VOC sources compared to emissions from distant sources. Nevertheless, this study illustrates that high-resolution VOC measurements, especially those with a sampling frequency on the scale of less than 1 day, combined with PMF, can make a strong contribution to our understanding of pollutant emissions and transport characteristics and is a useful tool with which to formulate effective pollution control strategies.