[1] Beijing has long suffered from serious ground-level ozone pollution, and volatile organic compounds (VOCs) play a key role in ozone formation. To understand the chemical speciation of VOCs in Beijing, nonmethane hydrocarbons (NMHCs) and oxygenated VOCs (OVOCs) were measured in summer in Beijing and nearby provinces (VOCs in this work means NMHCs+OVOCs). A variation of VOC mixing ratios and chemical speciation from 2004 to 2006 was observed at an urban site in Beijing. The typical VOC species, e.g., propane, propene, and toluene, had comparable or lower mixing ratios than levels found in other cities that previously hosted the Olympic Games, while the mixing ratios for isoprene were higher. The chemical compositions of VOCs within Beijing were heavily influenced by vehicular emissions and differed from those obtained in Tianjin and Hebei Province. OVOCs were an important component, accounting for 54% and 37% in the VOC mixing ratio in 2005 and 2006, respectively, and about 40% of the OH loss rates. The main reactive VOC compounds were aldehydes and alkenes. By using isoprene chemistry and the ratio of ethylbenzene to mp-xylene, the initial mixing ratios of VOCs were estimated. The VOCs had similar variation patterns to ambient ozone and peroxyacetyl nitrate (PAN) concentrations. The correlation between daily maximum ozone concentrations and initial VOCs revealed that ozone formation was sensitive to VOCs for both urban (Peking University, PKU) and rural (Yufa) sites. A reduction in NO x would lead to a decrease in ozone at Yufa, but would cause increased ozone at the PKU site.

Abstract. China implemented systematic air pollution control measures during the 2008 Beijing Summer Olympics and Paralympics to improve air quality. This study used a versatile mobile laboratory to conduct in situ monitoring of on-road air pollutants along Beijing's Fourth Ring Road on 31 selected days before, during, and after the Olympics air pollution control period. A suite of instruments with response times of less than 30 s was used to measure temporal and spatial variations in traffic-related air pollutants, including NO x , CO, PM 1.0 surface area (S(PM 1 )), black carbon (BC), and benzene, toluene, the sum of ethylbenzene, and m-, p-, and o-xylene (BTEX). During the Olympics (8-23 August, 2008), on-road air pollutant concentrations decreased significantly, by up to 54% for CO, 41% for NO x , 70% for SO 2 , 66% for BTEX, 12% for BC, and 18% for S PM 1 , compared with the pre-control period (before 20 July). Concentrations increased again after the control period ended (after 20 September), with average increases of 33% for CO, 42% for NO x , 60% for SO 2 , 40% for BTEX, 26% for BC, and 37% for S(PM 1 ), relative to the control period. Variations in pollutants concentrations were correlated with changes in traffic speed and the number and types of vehicles on the road. Throughout the measurement periods, the concentrations of NO x , CO, and BTEX varied markedly with the numbers of light-and medium-duty vehicles (LDVs and MDVs, respectively) on the road. Only after 8 August was a noticeable relationship found between BC and S(PM 1 ) and the number of heavy-duty vehicles (HDVs). Additionally, Correspondence to: T. Zhu (tzhu@pku.edu.cn) BC and S(PM 1 ) showed a strong correlation with SO 2 before the Olympics, indicating possible industrial sources from local emissions as well as regional transport activities in the Beijing area. Such factors were identified in measurements conducted on 6 August in an area southwest of Beijing. The ratio of benzene to toluene, a good indicator of traffic emissions, shifted suddenly from about 0.26 before the Olympics to approximately 0.48 after the Olympics began. This finding suggests that regulations on traffic volume and restrictions on the use of painting solvents were effective after the Olympics began. This study demonstrated the effectiveness of air pollution control measures and identified local and regional pollution sources within and surrounding the city of Beijing. The findings will be invaluable for emission inventory evaluations and model verifications.

No abstract