【摘要】 Application of the linear regression between vehicle types/numbers and PAH contents (Particle + Vapor) within a radius of 60 and 300 miles can be employed to assess the impact of long range atmospheric transportation (LRAT) to the level of PAHs in the rural atmosphere of Whitbourne， UK. In general， higher Rvalues coupled with lower Pvalues were found in those samples within a radius of 300 miles， whereas lower Rvalues were detected in those samples within a radius of 60 miles. Rvalues of PAH concentrations vs. all vehicle numbers within radius of 300 miles ranged from 0.51 to 0.92 with the arithmetic mean of 0.75 ± 0.12， whilst those Rvalues within a radius of 60 miles ranged from 0.05 to 0.61 with the arithmetic mean of 0.11 ± 0.28. This supports the hypothesis that PAHs in Whitbourne undergo LRAT.
【关键词】 Polycyclic aromatic hydrocarbons; Linear regression; Vehicle types and numbers; Whitbourne
INTRODUCTION
Polycyclic aromatic hydrocarbons (PAHs) has been widely acknowledged as ubiquitous carcinogenic in the atmospheric environment， generally containing two to eight benzene rings and can be produced in natural (i.e. volcanic eruptions， forest fires) and anthropogenic (i.e. traffic emissions， industrial activities， aluminium production， domestic heating and tobacco smoking) processes (Harvey 1991; Menzie et al 1992; Mastral and Callén 2000; Sanderson et al 2000; Korenaga et al 2001; Ohura et al 2004; Repace et al 2004). PAHs were detected in a wide range of environmental compartments in remote areas such as the Artic as consequences of agricultural practice， accidental spillage and uncontrolled release of industrial effluents (Halsall et al 1994 and 2000; Drr et al 1996) and exist mainly in the vapor phase. Therefore， they have atmospheric lifetimes that are relatively short， on the order of hours or days (Brubaker et al 1998; Atkinson et al 1994). If the major source of PAHs in urban air is anthropogenic emission， one might assume that the atmospheric PAH levels would positively correlate with local human population (Hafner et al 2005).
Various research efforts have noted such a relationship between human populations and atmospheric concentrations of PAHs (Dann et al 1998; Hafner et al 2005; Prevedouros et al 2004). For instance， a US group found that there was a strong correlation (r2 = 0.84 within 95% of prediction limits) between atmospheric PAH contents and human populations for all continental sites (Hafner et al 2005). In France， fiftyeight weekly samples of atmospheric bulk deposition (dry and wet) were collected at six specific sites including urban， semirural， rural and forest sites over a year (Garban et al 2002). The results reveal that mean PAH concentrations decreased by twothirds at sampling site located about 50 km far from Paris. Another US study discovered higher concentrations of phenanthrene， anthracene， floranthene， pyrene， benzo(a)anthracene and chrysene in spruce needles taken from monitoring sites adjacent to the city of Fairbanks than samples taken from more rural areas (Howe et al 2004). There is a general assumption that the level of PAHs extracted from the vegetation reflects the timeintegrated level of PAHs in the atmosphere. Taken together， these data indicate that local activities may produce PAHs from either combustion or petrogenic sources adjacent to the city of Fairbanks.
Influence of vehicle population on atmospheric levels of carcinogenic polycyclic aromatic hydrocarbons For decades， Long range atmospheric transportation(LRAT) has been the controversial subject of atmospheric transport and fate of PAHs in pristine areas (Carrera et al 2001; Fernandez et al 2002; Gustafson and Dickhut 1997). Owing to the discrepancy between the short atmospheric lifetimes of PAHs derived from field observations and its LRAT impacts found in remote regions， it would be very useful to have study focuses an attention on the evaluation of LRAT in comparison with previously reported field and experimental data. The main objective of this study is to provide data analytical strategies that allow this contradict to be comprehensively evaluated. Firstly， oneday back trajectories have been used to assign air masses of PAHs to the compass point in Whitbourne sampling site. Back trajectories are based on the prediction of using BADC (British Atmospheric Data Centre; http://badc.nerc.ac.uk/home/index.html) model as a part of NERC (Natural Environment Research Council) project.
Secondly， PAH data sets have been further separated into two categories， namely SRAT Short range atmospheric transportation(SRAT) and LRAT; within 60 miles and 300 miles of radius from Whitbourne respectively (Figure 1). On the assumption that PAHs introduced into atmosphere by petroleum combustion， then the correlation coefficients (Rvalue) and Pvalue of linear regression between the concentrations of PAHs and vehicle population within radius of 60 and 300 miles may be used as indicators for evaluating the impact of LRAT in Whitbourne. In addition， the correlation coefficients between the concentration of PAHs and the vehicle numbers (National Atmospheric Emissions Inventory; http://www.naei.org.uk/) categorized by its type and the distance from count point location to Whitbourne have been illustrated and discussed. Since Rvalue and Pvalue of the linear regressions reflect simultaneously the effect of petrol and/or mobile engine type coupling with its operating condition ［i.e. cold/hot ignition temperature， exhausted gas recirculation system(EGR)］， data interpretation has to be performed with great caution.Figure 1 Wind direction sector within radius of 60 and 300 miles of Whitbourne， United Kingdom.
2 MATERIALS AND METHODS
2.1 Sampling site
The field study was conducted at Whitbourne， which is located 8 km north west of Worcester city centre (West Midlands of UK). There are about 542000 people living in a 174000 hectare area of Worcestershire， with the population density of 3.12 people km2. The sampling location was on farmland， about 2 km away from the nearest main road. There are no major anthropogenic emission source， like industrial emission and vehicular emission in the immediate vicinity of the site. But vegetative burning， biogenic sources， soil dust， and regional source would be the major PM pollutants in this area.
2.2 Vehicle types and numbers
To achieve these goals， vehicle types and numbers of each city and town were compiled from the UK National Atmospheric Emissions Inventory (NAEI) sponsored by Department of Environmental Food and Rural Affairs (DEFRA) (http://www.naei.org.uk/data_warehouse.php). Vehicle numbers were estimated by using count point locations and annual average daily vehicle flows (AADF) in 2000. Vehicle types were categorized into seven types， namely， All vehicles (total number of vehicles)， CAR (Cars)， BUS (Buses)， light goods vehicles(LGV)， rigid heavy goods vehicles(HGVr)， articulated heavy goods vehicles(HGVa) and Moto (Motorcycles). Estimation of traffic levels uses information from both manual and automatic counts and performed at major and minor roads. The major roads are split into five road classes: motorways， trunk roads and principal roads with the latter two pided into urban and rural roads.
Urban roads are defined as those within the boundaries of the Urban Area polygons for settlements of 10000 populations or more， based on the 2001 Population Census. On the outskirts of urban areas， bypasses are normally treated as rural even if part of the road may lie within the urban area polygon. Conversely， roads between urban areas with short lengths outside the polygons are normally treated as urban. Minor roads are pided into six classes: B class， C class and U (unclassified) roads， each subpided into urban and rural. The traffic data sets were further categorized into two types (i.e. SRAT and LRAT) according to sampling location. Short range atmospheric transportation (SRAT) of PAHs was designated as target compounds being emitted from any sources to the receptor site within a radius of 60 miles from the monitoring site in Whitbourne. Whereas for long range atmospheric transportation (LRAT) of PAHs was a result of numerous pollutants being released within a 300mileradius circle centered on the sampling location (Figure 1).
2.3 PAH data sets
In addition， PAH data sets (Ac: acenaphthylene， Ace: acenaphthene， Fl: fluorine， Ph: phenanthrene， MePh3: 3methylphenanthrene， MePh9: 9methylphenanthrene， MePh1: 1methylphenanthrene， MePh3: 2methylphenanthrene， Fluo: floranthene， Pyr: pyrene， Ret: retene， B(a)A: benzo(a)anthracene， Chry: chrysene， B(b， j， k)F: benzo(b， j， k)fluoranthene， B(e)P: benzo(e)pyrene， B(a)P: benzo(a)pyrene， Per: perylene， Ind: indeno(1，2，3cd)pyrene， B(g， h， i)P: benzo(g， h， i)perylene， An: anthanthrene， D(a， h)A: dibenzo(a， h)anthracene， Cor: coronene) were obtained from a PhD thesis entitled “Atmospheric Chemistry of SemiVolatile Organic Compounds in Urban and Rural Air” (Pongpiachan， 2006). To make a valid comparison between vehicle numbers/types and PAH concentrations， wind direction was separated into six categories， namely northeast (NE)， east (E)， southeast (SE)， southwest (SW)， northwest (NW) and north (N) wind sectors based on the analysis of oneday back trajectories (Table 1). Note that PAH concentrations were calculated as a bulk of vapor and particulate phase in order to avoid any impacts of sampling bias on gas/particle partitioning. Table 1 Location of UK cities in each wind section.（略）
3 RESULTS
Table 2 and Table 3 display the vehicle distribution in each corresponding wind sector within a radius of 60 miles and 300 miles respectively. The arithmetic mean concentrations of all vehicles within radius of 60 miles detected in each wind sector decreased in the following order: NE (26154±24915) &> E (19893±23308) &> N (17818±12606) &> SE (17081±20404) &> SW (8904±9149) &> NW (6396 ± 3695). On the other hand， arithmetic mean concentrations of all vehicles within radius of 300 miles detected in each wind sector decreased in the following order: SE (26334 ± 33435) &> NE (20838 ± 26068) &> E (20695 ± 28693) &> NW (17479 ± 23507) &> SW (16283 ± 14785) &> N (11420 ± 14703). Table 2 Distribution list of vehicles from different sections within a radius of 60 miles of Whitbourne (Worcester)（略）Table 3 Distribution list of vehicles from different sections within a radius of 300 miles of Whitbourne (略).
The fact that the highest level of vehicles within radius of 60 miles were shown in NE wind sector， whereas those of the highest level of vehicles within radius of 300 miles were displayed in SE wind sector， is attributable to the presence of more human population in south eastern city such as London. Nevertheless， the maximum arithmetic mean of vehicle numbers obtained within radius of 60 miles were broadly in line with those counted within radius of 300 miles， for instance， 26154 ± 24915 in NE wind sector within radius of 60 miles; 26334 ± 33435 in SE wind sector within radius of 300 miles. Hence， the vehicle levels of the areas surveyed within radius of 60 and 300 miles of the sampling site in Whitbourne were almost homogeneous. To assess the impact of LRAT to PAH level in Whitbourne， the linear regression between vehicle numbers and PAH contents within radius of 60 and 300 miles was applied. The regression parameters are displayed in Table 4. In general， higher Rvalues are found in those samples within radius of 300 miles， while lower Rvalues are detected in those samples within radius of 60 miles. For example， Rvalues of PAH concentrations vs all vehicle numbers within radius of 300 miles ranged from 0.51 to 0.92 with the arithmetic mean of 0.75 ± 0.12， whereas those Rvalues within radius of 60 miles ranged from 0.05 to 0.61 with the arithmetic mean of 0.11 ± 0.28 (Figure 25). The reason that Rvalues of PAH concentrations vs all vehicle numbers within radius of 300 miles were generally higher than those of Rvalues within radius of 60 miles， is due to the fact that local traffic emissions are of minor importance， and thus it is becoming evident that PAHs in Whitbourne undergo LRAT.
Despite the relatively high Rvalues detected within radius of 300 miles， none of those within radius of 60 miles were higher than 0.7. The remarkable differences of R and pvalues between both radiuses have been further investigated. For instance， the highest Rvalues of linear regression between PAH contents and CAR number within radius of 300 miles in decreasing order are: Fl， Fluo， D(a， h)A， Cor， Chry， 3MePh and Pyr (Raverage = 0.76 ± 0.11， Paverage = 0.15 ± 0.10). These compounds are known to strongly associate with petroleum combustion sources and could again indicate the importance of LRAT to the PAH concentrations in Whitbourne atmosphere.Unlike the regression for CAR， BUS and LGV， the regression parameters for HGVa and HGVr are insignificant (P &> 0.50)， and overall， there seems to be little difference of Rvalues in accordance with two criteria: namely “within radius of 300 miles” and “within radius of 60 miles” . A likely reason for this dramatic drop in correlation is mainly connected with the reduction of accumulation mode particles， which are subject to LRAT (Allen et al 2001; Friedlander 2000). It is well known that the diesel emissions have been reduced progressively by the adequate application of engine modifications over the last twenty years in response to the restrictive emission legislation (Desantes et al 2005). Despite the fact that particulate emissions from diesel engines have been reduced successfully， there is increasing aware regarding whether current a stateoftheart Heavy Duty Diesel engine is increasing the number and/or reducing the size of the emitted particles (Bagley et al 1996; Baumgard et al 1996; Bertola et al 2001). For instance， increasing injection pressure was generally found to reduce the accumulation mode particle number and favor the formation of nuclei mode particles. Other engine parameters such as start of injection， engine speed and load conditions were also found to significantly decrease particle size distributions (Desantes et al 2005).
On the other hand， motorcycle (Moto) displays the opposite trend. For example， the highest arithmetic mean of Rvalue (0.91 ± 0.14) coupling with the lowest arithmetic mean of pvalue (0.09 ± 0.14) were observed in the linear regressions between PAHs and Moto numbers within radius of 300 miles (Table 4). By contrast， none of PAHs significantly correlated with Moto numbers within radius of 60 miles. These results suggest that i) Moto emissions possibly contain high levels of accumulation mode particles and ii) Whitbourne atmosphere might be significantly subject to these Moto accumulation mode particles. In general， 2stroke motorcycles are often run by a mixture of gasoline and a lubricant， causing emission of accumulation mode particle. The most recent findings from Yang et al show that size distribution of PAHs in emissions of a 2Stk/Cb have two significant modes that peak at &< 0.1 and 0.18 – 0.32 . The emission of a 2stroke motorcycle is greater than that of a 4stroke motorcycle for most regulated air pollutants and volatile organic compounds (Chan et al 1995; Tsai et al 2003; Leong et al 2002). Besides， the 2Stk/Cb motorcycle had the largest total B(a)P equivalent emission factor of 10.8 μg km-1， indicating that the emission exhaust from the 2Stk/Cb motorcycle was most carcinogenic (Yang et al 2005). According to compendium of motorcycling statistics reported by DEFRA， 2.7% of households in Great Britain owned a motorcycle in 2003， with ownership being more common among households that also owned one or more cars (For more detail see http://www.dft.gov.uk/stellent/groups/dft_transstats/do cuments/downloadable/dft_transstats_032250.pdf). There are about 1.52 million motorcycles in Great Britain， including motorcycles in other tax classes， those only licensed in summer， and those evading tax. Furthermore， the top 2 models of new motorcycle registration in UK 2003 are dominated by 2stroke carburetor engine (50 – 150 cc)， namely Piaggio NRG Scooter and Piaggio Zip50 Scooter (Source: Motorcycle Industry Association Report 2003: http://www.mcia.co.uk/Content/). Table 4 Arithmetic mean of R and p values of linear regression between vehicle number.and PAH contents within radius of 300 miles and 60 miles（略）
5 CONCLUSIONS
Overall， it is obvious that atmospheric PAHs are associated with traffic emissions via LRAT， as represented by high correlation coefficients of linear regressions between vehicle populations and atmospheric PAH concentrations within a 300miles radius of the sampling site. Additional anthropogenic emission from local residential wood combustions， incinerators and aluminum smelters are also important drivers for PAH pollution. Physical characterization (i.e. particle number concentrations and size distributions) as well as chemical characterization is crucial for a better estimation of possible adverse effects on human health as well as for a better understanding of LRAT mechanisms.
ACKNOWLEDGEMENTS
The authors would like to thank National Center of Excellence for Environmental and Hazardous Waste ManagementPSU Satellite Center， Prince of Songkla University for the continued financial support of this work.
参考文献
1 Allen G， Nemitz E， Shi JP， Harrison MR， Greenwood JC. Size distributions of trace metals in atmospheric aerosols in the United Kingdom. Atmos Environ 2001; 35(27): 45814591.
2 Atkinson R， Arey J. Atmospheric chemistry of gasphase polycyclic aromatic hydrocarbons: formation of atmospheric mutagens. Environ Health Perspect 1994; 102: 117126.
3 Bagley ST， Baumgard KJ， Gratz LD， Johnson JH， Leddy DG. Characterization of fuel and aftertreatment device effects on diesel emissions. Research Report 76 Health Effects Institute， USA 1996.
4 Baumgard KJ， Johnson JH. The effect of fuel and engine design on diesel exhaust particle size distributions. SAE Technical paper series， 960131.1996.
5 Bertola A， Schubiger R， Kasper A， Matter U， Forss A M， Mohr M， et al. Characterization of diesel particulate emissions in heavy duty DI diesel engines with common rail fuel injection. Influence of injection parameters and fuel composition. SAE Technical paper series， 2001013573.2001
6 Brubaker WW， Hites RA. OH reaction kinetics of polycyclic aromatic hydrocarbons and polychlorinated dibenzopdioxins and dibenzofuran. J Phys Chem 1998; 102: 915921.
7 Carrera G， Fernandez P， Vilanova RM， Grimalt JO. Persistent organic pollutants in snow from European high mountain areas. Atmos Environ 2001; 35: 245–254.
8 Chan CC， Nien CK， Tsai CY， Her GR. Comparison of tailpipe emissions from motorcycles and passenger cars， J Air Waste MA 1995; 45: 116–124.
9 Dann， T. Ambient air measurements of polycyclic aromatic hydrocarbons (PAH)， polychlorinated dibenzopdioxins (PCDD) and polychlorinated dibenzofurans in Canada (19871997); Analysis and Air Quality pision， Environment Canada:Ottawa， ON， 1998.
10 Desantes JM， Bermúdez V， García JM， Fuentes E. Effects of current engine strategies on the exhaust aerosol particle size distribution from a HeavyDuty Diesel Engine. J Aerosol Sci 2005; 36(10): 12511276.
11 Drr G， Hippelein M， Kaupp H， Hutzinger O. Baseline contamination assessment for a new resource facility in Germany: Part VI: Levels and profiles of PAH in ambient air. Chemosphere 1996; 33: 1569–1578.
12 Fernandez P， Grimalt JO， Vilanova RM. Atmospheric gasparticle partitioning of polycyclic aromatic hydrocarbons in high mountain regions of Europe. Environ Sci Technol 2002; 36: 1162–1168.
13 Friedlander SK. Smoke， dust， and haze. Fundamentals of aerosol dynamics.2nd ed. Oxford University Press， New York;2000，p.360.
14 Garban B， Blanchoud H， MotelayMassei A， Chevreuil M， Ollivon D. Atmospheric bulk deposition of PAHs onto France: trends from urban to remote sites. Atmos Environ 2002; 36(34): 53935403.
15 Gustafson KE， Dickhut RM. Gaseous exchange of polycyclic aromatic hydrocarbons acroos the airwater interface of southern Chesapeake Bay. Environ Sci Technol 1997; 31: 1623–1629.
16 Hafner WD， Carlson DL， Hites RA. Influence of local human population on atmospheric polycyclic aromatic hydrocarbon concentrations. Environ Sci Technol 2005; 39(19): 73747379.
17 Halsall CJ， Coleman PJ， Davis BJ， Burnett V， Waterhouse KS， HardingJones P， Jones KC. PAHs in UK urban air. Environ Sci Technol 1994; 28: 2380–2386.
18 Halsall CJ， Sweetman AJ， Barrie LA， Jones KC. Modelling the behaviour of PAHs during atmospheric transport from the UK to the Arctic. Atmos Environ 2000; 35 (2): 255267.
19 Korenaga T， Liu XX， Huang ZY. The influence of moisture content on polycyclic aromatic hydrocarbons emission during rice straw burning. Chemosphere – Global Change Science 2001; 3: 117–122.
20 Leong ST， Muttamara S， Laortanakul P. Influence of benzene emission from motorcycles on Bangkok air quality. Atmos Environ 2002; 36: 651661.
21 Mastral AM， Callén MS. A review on polycyclic aromatic hydrocarbon (PAH) emissions from energy generation. Environ Sci Technol 2000; 34: 3051–3057.
22 Menzie CA， Potocki BB， Santodonato J. Exposure to carcinogenic PAHs in the environment. Environ Sci Technol 1992; 26(7): 1278–1284.
23 Ohura T， Amagai T， Fusaya M， Matsushita H. Spatial distributions and profiles of atmospheric polycyclic aromatic hydrocarbons in two industrial cities in Japan， Environ Sci Technol 2004; 38: 49–55.
24 Prevedouros K， BrorstromLunden E， Halsall CJ， Jones KC， Lee RGM， Sweetman AJ. Seasonal and longterm trends in atmospheric PAH concentrations: evidence and implications. Environ Pollut 2004; 128: 1727.
25 Repace J. Respirable particles and carcinogens in the air of Delaware hospitality venues before and after a smoking ban. J Occup Environ Med 2004; 46: 887–905.
26 Sanderson EG， Farant JP. Use of benzo(a)pyrene relative abundance ratios to assess exposure to polycyclic aromatic hydrocarbons in the ambient atmosphere in the vicinity of a Sderberg aluminum smelter. J Air Waste MA 2000; 50: 2085–2092.
27 Tsai JH， Chiang HL， Hsu YC， Weng HC， Yang CY. The speciation of volatile organic compounds (VOCs) from motorcycle engine exhaust at different driving modes. Atmos Environ 2003; 37: 2485–2496.
28 Yang HH， Hsieh LT， Lui HC， Mi HH. Polycyclic aromatic hydrocarbon emissions from motorcycles. Atmos Environ 2005; 39(1): 1725.