Concentrations of VOCs
The target compounds for measurement were 101 components of VOCs (31 alkanes, 12 alkenes, 18 aromatic hydrocarbons, 28 organic halogen compounds, 11 oxygenated compounds, and one nitrogenous compound). Table 3 shows the concentrations of VOCs at sampling sites. The data lower than the method detection limit (MDL) was evaluated at half the concentration of the MDL value. Methylcyclopentane and bromomethane were detected in less than half of the samples, therefore a median value was reported as a half of the MDL. cis-1,2-Dichloroethylene, 1,1-dichloroethane, 1,1,2-trichloroethane, 1,1,2,2-tetrachloroethane, 1,2-dibromoethane, trans-1,3-dichloropropene, cis-1,3-dichloropropene, hexachloro-1,3-butadiene, m-dichlorobenzene, and benzyl chloride were not detected, therefore mean and median values were reported as a half of the MDL. The difference between the highest and the second highest median values in the sampling sites was examined statistically using a Wilcoxon test, one of non-parametric statistical tests. As a result, the substances that the difference was significant at a 1% level were n-butane, isobutane, cyclopentane, n-hexane, n-nonane, n-undecane, n-dodecane, n-tridecane, n-tetradecane, n-hexadecane, 2-methyl-1,3-butadiene, styrene, 1,2,4-trimethylbenzene, dichloromethane, vinyl chloride monomer, 1,1-dichloroethylene, 1,2-dichloroethane, 1,1,1-trichloroethane, tetrachloroethylene, formaldehyde, acetaldehyde, and 1-butanol. In industrial area (Takasago), n-hexadecane, styrene, vinyl chloride monomer, and 1,2-dichloroethane showed a high concentration. In roadside area (Ashiya), n-hexane, n-undecane, n-dodecane, tetrachloroethylene, formaldehyde, acetaldehyde, and 1-butanol showed a high concentration. Takasago is typically an industrial area where many manufacturing factories are located around the sampling site. One of them is a polyvinylchloride factory. Vinyl chloride monomer is used in the production of polyvinylchloride (Na et al. 2001). 1,2-Dichloroethane is intermediate for synthesis of vinyl chloride monomer (Incavo 1996; Sotowa et al. 1999). These results suggested that the exhaust gas from this factory influenced the high concentration of these chemicals only in Takasago. n-Hexane and n-undecane are components of gasoline and n-undecane, n-dodecane, and n-hexadecane are ones of light diesel oil. Additionally, styrene, formaldehyde, and acetaldehyde are the substances that the emission from automobile is large in Japan (Ministry of the Environment, Japan 2010a, b, c). These results suggested that the emission from automobile influenced the high concentration of n-hexadecane and styrene only in Takasago and n-hexane, n-undecane, n-dodecane, formaldehyde, and acetaldehyde only in Ashiya. The reason for the higher concentrations of tetrachloroethylene and 1-butanol only in Ashiya needs more investigation. In other substances, n-butane showed a high concentration only in Sumoto urban area. The sampling site in Sumoto is located in the center of the city, and there is a lot of traffic around the site. Olson et al. (2009) have reported that concentrations of 55 VOCs were measured near the roadway and n-butane was one of the highest concentration chemicals of individual VOCs. These results suggested that the emission from automobile influenced the high concentration of this chemical only in Sumoto.
Evaluation on the hazardousness and the photochemical reactivity of VOCs
The excess cancer incidences were calculated using VOCs measured data and the unit risk values to evaluate human health risk for carcinogenic effect. Table 1 shows the unit risk values for the measured substances. The unit risk value for benzene is set to be in the 2.2 × 10−6 per μg m−3 to 7.8 × 10−6 per μg m−3 range. In this study, 5.0 × 10−6 per μg m−3 of a median was used to calculate the excess cancer incidence. Figure 2 shows the health risk estimation for VOCs by excess cancer incidence. For formaldehyde, acrylonitrile, and acetaldehyde, the excess cancer incidences were more than 10−5 of the level of concern for carcinogenic effect at one or more sites. Particularly, for formaldehyde, they were in the 10−4 to 10−5 range at all sites. This result indicated the need to reduce formaldehyde emissions from the standpoint of their carcinogenic risk. The hazard quotients were calculated using VOCs measured data and the reference concentration to evaluate human health risk for noncarcinogenic effect. The reference concentration values for the measured substances are shown in Table 1. Figure 3 shows the health risk estimation for VOCs by the hazard quotient. Although the hazard quotients of acrylonitrile, 1,3-butadiene, acetaldehyde, and toluene were higher than those of other substances, they were less than 1 of the level of concern for noncarcinogenic effect at all sites. This suggested that the need for reduction measures of the noncarcinogenic effect is small at this time. The ozone productions were calculated using VOCs measured data and the MIR values to evaluate the photochemical reactivity. The MIR values for the measured substances are shown in Table 2. Figure 4 shows the calculation results of the ozone productions for VOCs. Toluene had a high percentage of the ozone production at all sites. This result indicated the need to reduce toluene emissions from the standpoint of the photochemical oxidant formation.
Concentration trend of the selected substances
As a result of the above evaluation, toluene and formaldehyde were selected as the substances requiring emissions reduction. The state and characteristics of environmental pollution were considered from the concentration trends of these substances.
The trends of the annual mean concentrations for toluene at sampling sites are shown in Fig. 5. The interannual change tendencies of toluene concentrations were different at a regional level. It was determined that the annual mean concentrations generally continued to be flat in Nishiwaki, Toyooka, Sumoto, and Ashiya, and exhibited an increasing tendency in Sanda. In Takasago, the annual mean concentration in 2007 was relatively higher than the others. It could be that the concentration tendency in Sanda is due to the influence of vehicle emission owing to the increase in traffic around the sampling site because of rapid urbanization over the last few years, that the high concentration in 2007 in Takasago is due to the influence of industrial fume because of the presence of factories that emitted toluene around the sampling site, and that many of the predominant wind directions at the time of measurement were on the side of the factories in 2007. Toluene concentration in roadside area (Ashiya) was not high compared with the other areas. This result was in contrast with the previous study that the highest concentration of toluene was observed at the site exposed directly to road traffic emissions (Kerbachi et al. 2006). The trends of the seasonal concentrations for toluene at sampling sites are shown in Fig. 6. The seasonal concentration trend differed according to the sampling site. The differences in toluene concentrations between summer and winter are uncertain at all sites, although Parra et al. (2009) have reported that toluene concentration was higher in winter than in summer due to higher atmospheric stability. Toluene is emitted from many sources such as vehicle exhaust, solvent use, gasoline evaporation, fossil fuel combustion, landfill, tobacco smoke, or adhesive (Baek and Jenkins 2004; Filella and Penuelas 2006; Kim and Kim 2002; Na et al. 2004; Sack et al. 1992). Therefore, it was presumed that the interannual and seasonal trends of toluene concentrations were different at a regional level because the conditions of the above emission sources, wind direction, and wind speed had a significant influence on the toluene concentration.
The trends of the annual mean concentrations for formaldehyde at the sampling sites are shown in Fig. 7. Formaldehyde concentrations have been reduced at all sites in 2009 compared in the previous year. Especially, in Ashiya, it exhibited relatively more reduction rate than others. This could be because of the decrease in traffic. The comprehension of the future concentration trend is required to identify the exact reason. The trends of the seasonal concentrations for formaldehyde at sampling sites are shown in Fig. 8. The seasonal concentrations increased from spring to summer and decreased from summer to winter at all the urban sites. The seasonal concentrations increased from spring to summer and decreased from summer to winter at all the urban sites. There were significant differences between formaldehyde concentrations in summer and winter at above sites except Ashiya as roadside area (1% significance, Wilcoxon test). The exact reason for this needs more investigation. The above seasonal patterns of formaldehyde in urban areas were consistent with those from other studies (Ceron et al. 2007; Ho et al. 2002; Pang and Mu 2006). For example, Ceron et al. measured the concentrations of 5 carbonyl compounds including formaldehyde in a semi-urban site and found that the levels of carbonyl compounds showed a strong seasonal trend with the following relative abundance: summer > autumn > winter because photochemical activity and temperature played an important role in the creation of carbonyl compounds in the sampling site during the summer. Ho et al. determined ambient levels of five carbonyl compounds including formaldehyde in urban area and found that the mean concentrations of formaldehyde and acetaldehyde were significantly higher in summer because they were produced photochemically at a higher level in summer. Therefore, it was quite possible that the formaldehyde concentrations in summer were significantly higher than those in winter at the urban sites because the formation of photochemical oxidant had a significant impact on the formaldehyde concentration in summer.