- Email: [email protected]
Polycyclic aromatic hydrocarbons in the surface soil of Shanghai, China: Concentrations, distribution and sources
Organic Geochemistry 41 (2010) 355–362
Contents lists available at ScienceDirect
Organic Geochemistry journal homepage: www.elsevier.com/locate/orggeochem
Polycyclic aromatic hydrocarbons in the surface soil of Shanghai, China: Concentrations, distribution and sources Ying Liu a, Ling Chen a,*, Jianfu Zhao a,b, Yanping Wei a, Zhaoyu Pan a, Xiang-Zhou Meng a, Qinghui Huang b, Weiying Li a a b
State Key Laboratory of Pollution Control and Resources Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China Key Laboratory of Yangtze River Water Environment, Ministry of Education, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China
a r t i c l e
i n f o
Article history: Received 3 June 2009 Received in revised form 7 November 2009 Accepted 23 December 2009 Available online 4 January 2010
a b s t r a c t We quantified 18 polycyclic aromatic hydrocarbons (PAHs) in 54 surface soil samples covering an area of 6400 km2 in Shanghai. An isopleth map of total concentrations of the 18 PAHs, which was constructed using an ordinary Kriging approach with log transformed data, clarified the regional variability and identified regional hot spots in the urban and industrial areas of Shanghai. These hot spots all suffer from high PAH pollution, suggesting that local human activities (e.g., vehicular traffic, petrochemical industry and coal combustion) may be the main contributors. Coal or oil fired power plants and their locations seem to be a significant factor controlling the PAH concentrations in surface soil. The higher molecular weight PAHs are often accumulated near pollution sources and are more heterogeneous in Shanghai soil, because they are less easily transported and biodegraded than 2 ring PAHs. The total concentrations are not correlated with soil total organic carbon. We successfully applied hierarchical cluster analysis (HCA) and principal components analysis (PCA) based on a centered log ratio procedure to explore the characteristics and possible sources of soil PAHs. The high PAH contamination in the Shanghai surface soil is mainly attributed to the contribution of pyrogenic sources (vehicular traffic pollution and combustion of coal and biomass). Furthermore, we applied PAH percentages by ring number, isopleth maps of total concentrations of 18 PAHs and source diagnostic ratios of PAHs to help assign the pyrogenic sources in Shanghai soils. Such map based approaches have only rarely been applied in investigations published in Organic Geochemistry. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Polycyclic aromatic hydrocarbons (PAHs) are common contaminants that are difficult to biodegrade in the environment. They are derived from natural processes (e.g., forest fires) and/or anthropogenic activities (e.g., combustion of coal, waste incineration, vehicular traffic, spillage of petroleum). Some PAHs have been recommended as priority pollutants by the United States Environmental Protection Agency (US EPA) and as persistent toxic substances (PTS) by the United Nations Environment Programme (UNEP) due to their persistent, toxic, mutagenic and carcinogenic characteristics (Zedeck, 1980; NRC, 1983; UNEP, 2002). In recent years, there has been an increasing focus on regional soil contamination by PAHs, especially in China (Cai et al., 2008). PAHs are easily accumulated in soils over many years due to their biochemical persistence and hydrophobicity (Johnsen et al., 2005). As a result, soils become an important reservoir of PAHs. In addition, PAHs in
* Corresponding author. Tel./fax: +86 21 65984261. E-mail addresses: [email protected] (Y. Liu), [email protected] (L. Chen). 0146-6380/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.orggeochem.2009.12.009
soils may further accumulate in vegetables and other biota via food chains (Kipopoulou et al., 1999; Li et al., 2008), leading to direct or indirect exposure in humans. Therefore, studies on soil contamination by PAHs are necessary to minimize the risk of human exposure and environmental pollution. Shanghai is one of the most comprehensively industrial and commercial cities in China, ranking first in population and population density. Recently, many studies have been conducted in Shanghai on the levels and sources of PAHs in the atmosphere (Feng et al., 2006; Cheng et al., 2007), sediments (rivers and estuary) (Bouloubassi et al., 2001; Liu et al., 2008), road dust (Ren et al., 2006; Liu et al., 2007a) and urban soils (Jiang et al., 2009). However, few data have been reported on the spatial distribution and sources of PAHs in Shanghai soils, especially on the large scale. Regional soil PAH contamination mainly originates from dry and wet atmospheric deposition. PAH concentrations in road dust vary from 3100–32,000 ng/g in the urban area of Shanghai, which have been attributed to emissions from vehicles (Ren et al., 2006) and a mixture of traffic and coal combustion (Liu et al., 2007a). Concentrations of atmospheric PAHs range from 0.07–270 ng/m3, with major sources being fossil fuel combustion, coal burning, industrial
356
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
Fig. 1. Sampling locations (a) and isopleth map of total concentrations of 18 PAHs (b) in surface soils of the Shanghai area.
furnaces and engine exhaust emissions (Feng et al., 2006; Cheng et al., 2007). PAH concentrations in urban soils have been found to range from 442–19,700 ng/g, with PAHs attributed primarily to combustion (Jiang et al., 2009). Further studies are needed to understand and explore combustion sources of these PAHs. The soil is a sink and source for atmospheric PAHs and a source for sedimentary PAHs. Our previous work has shown that a river segment near Site 606 (Fig. 1a) is heavily polluted by PAHs and coal burning is one of the contributors of pyrogenic PAHs (Liu et al., 2008). Therefore, studies on the concentrations and spatial distribution of soil PAHs are useful to further understand the distribution of sedimentary PAHs in Shanghai. The objectives of the present study are to determine the concentrations and spatial distributions of PAHs in Shanghai soils, and to explore their characteristics and possible combustion sources. Such map based approaches have only rarely been applied in investigations published in Organic Geochemistry. The results of this study will input new data to the global PAH database and provide valuable information for regulatory actions to improve the environmental quality of the Yangtze River Delta.
2.2. Sample extraction and cleanup
2. Materials and methods
Sixteen PAHs characterized by the US EPA as priority pollutants were analyzed, including naphthalene (Nap), acenaphthylene (AcNy), fluorene (Fl), acenaphthene (AcNe), phenanthrene (PhA), anthracene (An), fluoranthene (FlA), pyrene (Py), benz[a]anthracene (BaA), chrysene (Chy), benzo[b]fluoranthene (BbF), benzo[k]fluoranthene (BkF), benzo[a]pyrene (BaP), indeno[1,2,3cd]pyrene (IP), benzo[ghi]perylene (BghiP) and dibenz[a,h]anthracene (DBahA). In addition, benzo[e]pyrene (BeP), perylene (Pery), 1-methylnaphthalene and 2-methylnapthalene were measured qualitatively. Concentrations of the two methylnaphthalene isomers were combined and reported as total methylnaphthalenes (MNap). PAH analysis was carried out by high performance liquid chromatography (HPLC) with a photodiode array detector. Identification of PAH compounds was based on chromatographic retention time and the ultraviolet spectra of PAH standards. The quantification was performed by the external standard method. The ultraviolet measurement wavelengths included 218 nm (NaP), 223 nm (MNaP), 226 nm (AcNe and AcNy), 249 nm (IP), 254 nm (Fl, PhA and An), 266 nm (Chy), 286 nm (FlA and BaA), 300 nm (BbF, BkF, BaP, DBahA and BghiP), 330 nm (BeP), 334 nm (Py) and 433 nm
2.1. Study area and sample collection Shanghai is located in the east of the Yangtze River Delta, as shown in Fig. 1a. Bordered by Jiangsu and Zhejiang provinces in the west, Shanghai is bounded by the East China Sea to the east and Hangzhou Bay to the south. North of the city, the Yangtze River flows into the East China Sea. Shanghai has a total area of 6340 km2, including a land area of 6218 km2 and a water area of 122 km2. As is typical of subtropical areas, Shanghai has a marine monsoon climate with a clear division of four seasons, enough sunshine and ample rainfall. The average annual temperature and rainfall are about 15.8 °C and 1149 mm, respectively. We collected 54 surface soil samples from Shanghai in April 2007 using the systematic grid method. Each 5–10 cm deep sample was obtained using a Luoyang shovel and consisted of five* subsamples from the surrounding area of each sampling site (within 10 m2). All samples were air dried in the dark and sieved to <0.076 mm (200 mesh) after removing stones and residual roots, then stored at 4 °C until the PAHs were analyzed.
The five sub-samples were mixed together to obtain a representative sample for each site and each mixed sample was analyzed in duplicate. Sample extraction and cleanup followed Method 3541 (automated Soxhlet extraction), Method 3630C (silica gel cleanup) and Method 3660B (sulfur cleanup) recommended by the US EPA (1996). Five grams of soil sample were blended with anhydrous sodium sulfate (1:1, w:w) and extracted by a FOSS Soxtec Avanti 2050 automatic system using 70 ml of a mixture of acetone and hexane (1:1, v:v), a hot plate temperature of 160 °C and an extraction time of 60 min for boiling and 60 min for rinsing. The extract was concentrated by a rotary evaporator, and the extract solvent was exchanged to cyclohexane. The detailed cleanup procedure is reported elsewhere (Liu et al., 2007b, 2008). Briefly, the extracts were further cleaned using a chromatographic column packed with 5 g activated silica gel and the eluate was finally concentrated to 1 ml. 2.3. Sample analysis and quality control
357
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
(Pery). Detailed measurement procedures are described elsewhere (Liu et al., 2007b, 2008). All analyses were calculated on a dry weight (dw) basis. Method detection limits (MDLs) for all targets ranged from 1–19 ng/g dw. The recoveries for spiked PAHs varied from 87%–113%. Each mixed sample was analyzed in duplicate and the relative standard deviation was <20%. Total organic carbon (TOC) analysis was performed with the Shimadzu TOC-Vcpn analyzer with the solid sample module (SSM-5000A). The overall precision of measurements was less than 3% (n = 3).
ker and Macdonald, 2003). Finally, the ordinary Kriging approach with log transformation was applied to data interpolation for the total concentrations of 18 PAHs, and inverse distance weighted interpolation was used for source diagnostic ratios of PAHs (after omitting the data with undetectable values of PAH), using the Geostatistical Analyst tool of ESRI ArcMap 9.2 software to draw isopleth maps of Shanghai.
2.4. Data analysis
The concentrations of individual PAHs (P3 rings) show large variations (Table 1). The highest concentration of a single PAH is of FlA, which ranges from undetected to 5200 ng/g dw, and the maximum concentrations for the P4 rings PAHs and PhA were more than 1000 ng/g dw. The minimum concentrations for all PAHs analyzed were less than the method detection limits, except for PhA. However, the 2 ring PAHs Nap and MNap have a smaller range between the highest and lowest concentrations and similar trends were found between the average and median concentrations as compared with other PAHs. Hence, the residual levels of 2 ring PAHs are more homogeneous in the surface soils than those of other PAHs. Zhang and Selinus (1998) suggested that the probability distributions of environmental geochemical variables for major elements (i.e., O and Si) generally follow normal distributions, but trace elements tend to follow log normal distributions. In this study, the concentrations of trace PAHs in soils follow similar probability distributions. The K–S test indicated that the concentrations of Nap and MNap were normally distributed (p < 0.05), as shown in Table 1, while the other PAHs followed log normal or positively skewed distributions. By comparison, Nap and MNap have low boiling points (relatively easy to volatilize) and good water solubility (tend to transport with water). As a result, they can be biodegraded and transported easily. Zhang et al. (2008) also believed that vertical transport of PAHs in soils were controlled by the nature of PAHs (i.e. log Koc, molecular weight). Other PAHs, despite
Before statistical analysis, analytical replicates were averaged and undetectable values were assigned the value of half of the MDLs. AcNy was eliminated from the dataset as it was undetectable in all samples. Statistical analyses, including the Kolmogorov–Smirnov (K–S) test, descriptive statistics, hierarchical cluster analysis (HCA) and principal components analysis (PCA), were performed using SPSS 13.0 for Windows. Before HCA and PCA, PAH data were normalized using a centered log ratio procedure (Bonn, 1998; Yunker and Macdonald, 2003; Pollard et al., 2006). Given a sample S ¼ ðx1 ; x2 ; . . . ; xN Þ, where the xi are the percentages of N species, the transformed sample is expressed as Str ¼ ðln½x1 =g; ln½x2 =g; . . . ; ln½xN =gÞ, where g is the geometric mean of sample S, which is calculated using the pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi equation g s ¼ N x1 x2 xN . Autoscaling (scale to variable mean and standard deviation) was then applied. To explore the structure between variables (PAH), the normalized dataset was hierarchically clustered using the weighted average linkage between the groups and the Pearson correlation for the cluster intervals (Kavouras et al., 2001; Zhang et al., 2006). In addition, to further understand the relationship between variables (PAH) and identify major sources of PAHs, we applied PCA to extract the principal components (PCs) with a varimax rotation. The rotation maximizes or minimizes the loading of each variable on each PC while preserving its trend (Kavouras et al., 2001; Yun-
3. Results and discussion 3.1. Concentration and probability distribution of PAHs
Table 1 Descriptive statistics of PAHs in surface soils of Shanghai. No.
PAHs
Abbrev.
No. of rings
Distribution
Nap MNap AcNe Fl PhA An FlA Py Chy BaA BbF BkF BaP BeP Pery DBahA IP BghiP
2 2 3 3 3 3 4 4 4 4 5 5 5 5 5 5 6 6
Normal Normal Skewed Skewed Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Skewed Skewed Skewed Lognormal Lognormal Lognormal Lognormal
MDLa
Min.
Max.
Meanb
Median
Geo.c
DF (%)d
NDe ND ND ND 9.5 ND ND ND ND ND ND ND ND ND ND ND ND ND 94 62.4
110 90.3 188 243 2300 807 5190 4290 2230 2440 3590 1380 3130 2790 2620 1190 2260 2680 37,400 31,900
28.6 30.5 10.0 19.9 130 28.3 259 220 127 124 177 72.8 181 129 139 59.5 118 144 2000 1700
24.7 25.4 ND ND 34.3 2.8 40.5 30.7 22.5 16.9 29.5 12.0 30.6 ND 26.9 6.5 15.0 20.9 390 314
21.6 23.7 2.40 12.8 43.8 3.9 44.4 41.2 25.1 20.4 25.8 12.7 33.6 24.4 32.7 10.4 19.8 25.0 526 405
94 94 35 24 100 65 91 91 93 80 81 80 80 50 61 52 72 80 100 100
(ng/g dw) 1 Naphthalene 2 Methylnaphthalene 3 Acenaphthene 4 Fluorene 5 Phenanthrene 6 Anthracene 7 Fluoranthene 8 Pyrene 9 Chrysene 10 Benz[a]anthracene 11 Benzo[b]fluoranthene 12 Benzo[k]fluoranthene 13 Benzo[a]pyrene 14 Benzo[e]pyrene 15 Perylene 16 Dibenz[a,h]anthracene 17 Indeno[1,2,3-cd]pyrene 18 Benzo[ghi]perylene P 18 PAHs P 16 PAHs (US EPA) a b c d e
MDL, method detection limit. Mean, arithmetic mean. Geo., geometric mean. DF, detectable frequency. ND, not detected.
4.0 3.6 1.9 18.3 1.1 1.8 5.1 7.1 2.2 4.9 1.9 3.1 6.5 19.1 17.1 6.4 6.4 6.1 – –
358
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
low concentrations in the environment, are usually enriched in heavy traffic or industrial areas, which may lead to a positively skewed distribution (Table 1). However, the concentrations of AcNe, Fl, BeP, Pery and DBahA do not follow log normal distributions, because the detectable frequencies of those compounds are low (<55%) and the undetectable values were assigned values of half of the MDLs (Table 1). In general, the arithmetic mean is used for normally distributed parameters, such as Nap and MNap, while the geometric mean is used for parameters that follow a log normal or positively skewed distribution, such as Fl, PhA and BaP. P The total concentrations of 18 PAHs ( 18 PAHs) range from 94–37,300 ng/g dw with a geometric mean of 526 ng/g dw. In P 80% of samples, 18 PAHs is <1000 ng/g dw (Fig. S1). The K–S P test also indicated that 18 PAHs followed a log normal distribution. 3.2. The relationship between total PAHs and total organic carbon (TOC) in soil TOC is an important controlling factor of sorption of PAHs in soil and may pre-determine the level of soil contamination with PAHs (Tang et al., 2005). In this study, soil TOC varies greatly from 4.25– 46.1 mg/g dw. Pearson correlation analysis shows that the total concentrations of 18 PAHs in soils are not significantly correlated with soil TOC (r = 0.062 and p = 0.65, see Fig. S2). Jiang et al. (2009) also reported a similar result when they researched the relationship between PAHs and TOC in urban soil of Shanghai and believed the result to be due to non-equilibrium adsorption between TOC and PAHs in soils. In our work, the inputs of PAHs to each site vary and not all adsorption of PAHs on TOC reaches equilibrium. Therefore, the effect of soil TOC on the amount of PAH accumulated in soil is ignored in the following spatial distribution of PAHs. 3.3. Spatial distribution of total PAHs Many interpolation methods can be used to analyze the spatial distribution of PAH levels in the environment. A surface tension spline function was used for dispersive interpolation and spatial distribution contouring of PAHs (Ye et al., 2006). Geostatistics can provide an optimal interpolation approach known as Kriging. This theory has been developed and applied widely by geologists to estimate mineral reserves. Geochemists have also tried to use the Kriging approach for the spatial distribution analysis of heavy metals (Zhang and Selinus, 1997) and PAHs (Wang et al., 2003). Kriging methods work best when the data are approximately normally distributed. In addition, the log transformation is often used where the data has a positively skewed distribution and there are some very large values in order to avoid a proportional effect (Zhang and Selinus, 1997; Johnston et al., 2004). In the present study, the total PAH concentration follows a lognormal distribution. We therefore applied the ordinary Kriging method with log transformation to predict the PAH values in the non-sampled area in Shanghai. Fig. 1b shows the spatial distribution contours of total PAH concentrations in the surface soils, residential and industrial areas and main power plants sites. Isopleth mapping has proved to be a valuable approach for identification of regional hot spots for PAH pollution (Lehndorff and Schwark, 2009). In this study, three regions with high PAH concentration were delimited based on the region function, here designated as regions A, B and C. Regions A and B are located in urban areas, and Region C is situated in a petrochemical industrial area. These regions are severely influenced by human activity. Region A is located in the central part of Shanghai with heavy traffic. As a result, the surrounding soil is contaminated by road dust, which is an agglomeration of various traffic related PAH sources,
e.g., automobile exhaust, lubricating oil, gasoline, diesel fuel, weathered materials of street surfaces and tire particles (Murakami et al., 2005). Liu et al. (2007a) reported that 16 PAH concentrations in road dust from the central part of Shanghai (Region A) were very high, with a range from 6875–32,573 ng/g dw, which suggests heavy contamination in the surface soils of Region A. The high PAH concentrations in Region B may be associated with regional industrial pollution. Region B is situated in an old industrial district with many manufacturing facilities, e.g., chemical and coke oven plants. Nowadays, some plants with heavy pollution have been encouraged to move elsewhere, and a new residential and financial district is gradually forming. However, traffic pollution is replacing industrial pollution and continues to contaminate the surrounding soils. We note that coal fired power plants in Region B contribute to PAH contamination in soil. Region C adjoins the sea and Shanghai’s principle petrochemical station and its oil refinery are located in this region. The high PAH contents in the soil may be due to contamination from the petrochemical industry. To understand the effect of power plants on the spatial distribution of PAHs, we show the main large scale coal or oil fired power plants in Shanghai in Fig. 1b. Clearly, the power plants are situated in the higher PAH pollution areas. This indicates that power plants seem to be an important factor controlling the PAH contents of surface soils in Shanghai. 3.4. PAH composition comparison using multivariate statistic analysis The relationship of PAH compositions and sources in surface soils of Shanghai were examined using two multivariate statistical analysis methods. AcNy, AcNe and Fl were excluded in the analysis due to their low detectable frequencies (Table 1). The normality of the data is an important factor when interpreting analytical results, since many inferences are based on the assumption of normality. Multivariate statistical applications are still almost exclusively based on correlation or covariance matrices determined directly from raw data. However, normalizing data can improve the explanatory power of a multivariate analysis (Kucera and Malmgren, 1998). In a concentration based multivariate model, the samples with the highest concentrations usually project separately as single samples or small groups of samples (generally with positive projections), while the remaining (lower concentrations) samples project in a group near the origin and make little or no contribution to the model. If normalization to the total PAH concentration has been used, PAH concentration data are transformed into compositional data with the constant sum constraint (CSC). The CSC introduces spurious negative correlations between major variables and positive correlations between minor ones (Johansson et al., 1984). The log ratio transformation can effectively reduce the CSC from compositional data (Kucera and Malmgren, 1998), and the centered log ratio transformed values have been used as inputs into PCA (Bonn, 1998; Yunker and Macdonald, 2003; Pollard et al., 2006). In this paper, we applied a pattern based approach to multivariate statistical analysis using the centered log ratio procedure to produce a normalized dataset as input into HCA and PCA. 3.4.1. Hierarchical cluster analysis Hierarchical cluster analysis (HCA) was used to explore structure of the normalized dataset. No significant difference was found when different linkage methods were used based on the results of the HCA. Fig. 2 shows the results presented in the form of a dendrogram. Sixteen variables (PAHs) were classified into three major groups. The first group is composed of low molecular weight PAHs with 2–3 rings and alkyl substituted PAHs, including Nap, PhA and MNap. The second group includes DBahA, BeP, An and Pery. Detectable frequencies of the four compositions were less than 70% (Table 1). In this group, PAH compositions with similar detectable
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
Fig. 2. Hierarchical dendrogram for PAHs using average linkage between groups and Pearson correlation as a measure interval.
frequencies were clustered together, e.g., DBahA and BeP. The third group, including most 4–6 rings PAHs, was subdivided into two subgroups, namely 4 ring PAHs (FlA, Py, Chy and BaA) and 5–6 ring PAHs (BbF, BkF, BaP, BghiP and IP). These PAHs are usually detected in pyrogenic sources, e.g., combustion of coal, wood, vehicle fuel and waste tire (Levendis et al., 1998; Wang et al., 2007). 3.4.2. Principal components analysis Principal components analysis (PCA) was performed to further understand the relationship between PAH compositions and possible chemical sources for each factor. Factor loadings of the normalized dataset are listed in Table S1 and plotted in Fig. 3a. Most of the variance (76.2%) of the normalized dataset was explained by the first four factors. Factor 1 explains 34.0% of the total variance. High negative loadings of MNap, Nap and PhA and high positive loadings of BbF, BkF, BaP, IP and BghiP were observed in this factor (Table S1). These PAH compositions are clustered on the left and right sides of the x-axis in Fig. 3a, respectively. MNap, Nap and PhA belongs to the low molecular weight PAHs with 2–3 rings or alkyl substituted PAHs, which are abundant in petrogenic source mainly caused by petroleum spills (Marr et al., 1999; Dobbins
359
et al., 2006). Samples with high negative score have a dominance of 2–3 rings PAHs, which originated mainly from petrogenic source. BghiP is identified as a tracer of auto emissions because it was found to be enriched in a traffic tunnel along with BaP (Harrison et al., 1996; Larsen and Baker, 2003; Boonyatumanond et al., 2007). IP is found in petroleum as a tracer of diesel combustion (Kavouras et al., 2001). The high level of BkF relative to other PAHs is suggested to indicate diesel vehicles (Larsen and Baker, 2003). BkF and its isomers (e.g., BbF and BaP) are dominant compounds in particulate samples of roadside air (Boonyatumanond et al., 2007). Samples with high positive score have a dominance of 5–6 rings PAHs, and vehicular traffic pollution is a major contributor of PAHs contamination in the samples. To sum up, the loadings for Factor 1 are primarily due to the separation between petrogenic source and vehicular traffic pollution. Factor 2, accounting for 16.5% of total variance, is negatively dominated by MNap and Nap and positively by FlA, Py, Chy and BaA (Table S1). In Fig. 3a, MNap and Nap are projected in the lower left quadrant, and they are related to petrogenic source. FlA, Py, Chy and BaA are projected in the upper right quadrant. The four compositions are high molecular weight PAHs with 4 rings as markers of coal combustion (Duval and Friedlander, 1981). The furnace effluents from combustion of pulverized coals contained FlA and Py (Levendis et al., 1998). Likewise, the semi-volatile PAHs of FlA and Py are attributed to unburned fossil fuels (Kavouras et al., 2001). Contents of the four PAHs in soils mainly reflect the contribution of combustion of coal and biomass. Therefore, the loadings for Factor 2 are primarily due to the separation between petrogenic source and combustion of coal and biomass. Factors 3 and 4 are responsible for 14.5% and 11.2% of the variance, respectively (Table S1). Factor 3 is heavily weighted by BeP and DBahA, while Factor 4 is dominated by An and Pery. However, no chemical sources with such a PAH profile have been reported until now. They might be likely to represent ‘‘noise” due to their low detectable ratios (50–65%). It is interesting that the PCA result in this paper is compared with the PCA result of sedimentary PAHs in our previous work, and PAH compounds with high positive loadings in extracted factors are parallel between in soil and in sediment (Liu et al., 2009). This indicates that soil PAHs and sedimentary PAHs in Shanghai have common sources. Fig. 3b shows a plot of the scores of samples for the first two principal components and stack column graphs of the percentages that 2–6 ring PAHs account for in the samples. In Fig. 3b, the urban influenced samples (I) project in the upper right quadrant, indicat-
Fig. 3. Loading plot (a) of the first two principal components and score plot (b) of each sample. The variance accounted for by each PC is shown after the axis label. Stack column graphs show percentages accounted for by 2–6 ring PAHs in the samples.
360
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
ing that these samples suffered greatly from contamination by vehicular traffic and combustion of coal and biomass. Likewise, PAHs from vehicular traffic and combustion of coal and biomass belong to pyrogenic PAHs and should have higher proportions of 4–6 rings PAHs, which agrees with their stack column graphs (I). Furthermore, these samples (I) were collected from heavy PAH contaminated regions in Fig. 1b. The regions are close to power plants and located in urban or industrial areas where vehicular traffic pollution is obvious. We note that Site 901 is situated in a petrochemical industrial region, where vehicular traffic pollution is not prominent in theory, but there is a similar stack column as other sites, e.g., Sites 604 and 405. According to photos pictured near Site 901 (Fig. S3), exhaust from petrochemical processing usually includes combustion products from hydrocarbons gases. The treatment of the exhaust by direct combustion produces high molecular weight PAHs and contaminates the surrounding soils, as does combustion of gasoline or diesel fuel in vehicles. Therefore, it is reasonable that the high PAH contamination in the Shanghai soil is due to the contribution of vehicular traffic and combustion of coal and petroleum.
In Fig. 3b, samples from Chongmingdao Island (II) are projected in the upper left quadrant, illustrating a weak contribution of vehicular traffic and a strong contribution of combustion of coal and biomass. Chongmingdao Island is an isolated region currently without linkage of a bridge or tunnel to the mainland of Shanghai. The isolation limits development of vehicular traffic, leading to pollution from vehicular traffic on the island being less than that on the mainland. Nowadays, Chongmingdao Island supports large scale agriculture, especially green agriculture and industrial development polluting the environment is limited. Here, the strong contribution of combustion of coal and biomass is attributed to straw combustion. In local agricultural activity, rice straw is burned in situ to release nutrients for the next growing season. The concentrations of atmospheric PAHs with lighter molecular weights, e.g., Py, FlA, PhA, Fl and Nap (2–4 ring PAHs), are obviously higher on rice straw burning days than on non-burning days, but no concentration difference is observed for higher molecular weight PAHs, e.g., 5–6 ring PAHs (Yang et al., 2006). The stack column graphs of Site A02 and Site A04 (Fig. 3bII) also reflect the dominance of 2–4 ring PAHs. In the samples from III (Sites 201, 301
Fig. 4. Isopleth maps of source diagnostic ratios of PAHs in surface soils of the Shanghai area.
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
361
and 302), the PAH concentration is low, even less than the MDLs and the total concentrations of PAHs are <150 ng/g dw. Overall, multivariate statistical analysis based on the centered log ratio procedure unveils the relationships of PAH compositions and identifies vehicular traffic and combustion of coal and biomass as dominant contributors of PAH contamination in soils. The spatial distribution contour of total PAH concentrations and stack column graphs of 2–6 ring PAH percentages further confirm the results of the multivariate statistic analysis.
Acknowledgements
3.5. Source diagnostic ratios of PAHs
Appendix A. Supplementary material
The relative abundances or diagnostic ratios are valid indicators of PAH sources because isomer pairs are diluted to a similar extent upon mixing with natural particulate matter and are distributed similarly to other phases as they have comparable thermodynamic partitioning and kinetic mass transfer coefficients (Dickhut et al., 2000). Diagnostic ratios of PAHs, such as ratios of An/(PhA + An), FlA/(FlA + Py), BaA/(BaA + Py) and IP/(IP + BghiP), can be applied to identify the possible emission sources, as noted in Fig. 4 (Yunker et al., 2002). Isopleth maps of source diagnostic ratios of PAHs are shown Fig. 4. The ratio of An/(An + PhA) ranges from 0.01–0.26, FlA/(FlA + Py) from 0.11–0.77, BaA/(BaA + Chy) from 0.30–0.60 and IP/(IP + BghiP) from 0.30–0.56. The results of these ratios indicate that the combustion of biomass, coal and petroleum are the major sources of PAHs in Shanghai. The region with higher PAH contamination is companied with a higher ratio of An/(An + PhA) (Fig. 4a). The ratios of FlA/(FlA + Py) and BaA/(BaA + Chy) show that combustion of coal and biomass are the predominant sources (Fig. 4b and c), but that of IP/(IP + BghiP) shows petroleum combustion is the major source (Fig. 4d). This may be because FlA, Py, BaA and Chy originate mainly from the contribution of combustion of coal and biomass, while IP and BghiP source chiefly from vehicular traffic pollution (e.g. combustion of fuel oils), according to PCA results. Therefore, the results from source diagnostic ratios of PAHs can reflect basically the contribution of combustion sources of PAHs.
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.orggeochem.2009.12.009.
4. Conclusions Although the total PAH concentrations ranged from 94– 37,300 ng/g dw, most surface soils in Shanghai have low PAH contamination. The concentrations of total PAH and most individual PAHs follow log normal distributions except for the Nap and MNap components. Isopleth maps are a valid approach to identify local hot spots in PAHs contamination and three regions with high PAH contamination were delimited using the ordinary Kriging method with log transformation. The surface soils with high PAHs concentrations are found in the urban area and petrochemical industrial area and anthropological influence is the major contributor of PAHs contamination in Shanghai surface soil. The ecological risk from PAH contamination should be considered for the persons living in the three regions. A centered log ratio transformation is an effective normalization procedure for PAH concentration data. Two multivariate statistic analysis methods based on this data normalization procedure have been successfully applied to explore the characteristics and to identify PAH sources. Vehicular traffic and combustion of coal and biomass are the main sources of PAHs in Shanghai surface soil. The results from multivariate statistical analysis are combined with PAH percentages by ring number and isopleth maps of total concentrations of 18 PAHs and source diagnostic ratios of PAHs, which further confirm the sources of PAHs. Such map based approaches are one of valid path in identification of PAH sources.
This work was supported by the National Natural Science Foundation of China (No. 20907034), Program for Young Excellent Talents in Tongji University (No. 2008KJ022), the Foundation of the State Key Laboratory of Pollution Control and Resource Reuse (No. PCRRK09001), and the National Key Technology R&D Program (No. 2008BAC46B02). We also thank Mark Bernard Yunker, Lorenz Schwark, Xue-Tong Wang and an anonymous reviewer for helpful comments and suggestions.
Associate Editor—Mark Yunker References Bonn, B.A., 1998. Polychlorinated dibenzo-p-dioxin and dibenzofuran concentration profiles in sediment and fish tissue of the Willamette Basin, Oregon. Environmental Science and Technology 32 (6), 729–735. Boonyatumanond, R., Murakami, M., Wattayakorn, G., Togo, A., Takada, H., 2007. Sources of polycyclic aromatic hydrocarbons (PAHs) in street dust in a tropical Asian mega-city, Bangkok, Thailand. Science of the Total Environment 384 (1– 3), 420–432. Bouloubassi, I., Fillaux, J., Saliot, A., 2001. Hydrocarbons in surface sediments from the Changjiang (Yangtze River) estuary, East China Sea. Marine Pollution Bulletin 42 (12), 1335–1346. Cai, Q.-Y., Mo, C.-H., Wu, Q.-T., Katsoyiannis, A., Zeng, Q.-Y., 2008. The status of soil contamination by semivolatile organic chemicals (SVOCs) in China: a review. Science of the Total Environment 389 (2–3), 209–224. Cheng, J., Yuan, T., Wu, Q., Zhao, W., Xie, H., Ma, Y., Ma, J., Wang, W., 2007. PM10bound polycyclic aromatic hydrocarbons (PAHs) and cancer risk estimation in the atmosphere surrounding an industrial area of shanghai, China. Water, Air, and Soil Pollution 183 (1–4), 437–446. Dickhut, R.M., Canuel, E.A., Gustafson, K.E., Liu, K., Arzayus, K.M., Walker, S.E., Edgecombe, G., Gaylor, M.O., MacDonald, E.H., 2000. Automotive sources of carcinogenic polycyclic aromatic hydrocarbons associated with particulate matter in the Chesapeake Bay region. Environmental Science and Technology 34 (21), 4635–4640. Dobbins, R.A., Fletcher, R.A., Benner Jr., B.A., Hoeft, S., 2006. Polycyclic aromatic hydrocarbons in flames, in diesel fuels, and in diesel emissions. Combustion and Flame 144 (4), 773–781. Duval, M.M., Friedlander, S.K., 1981. Source resolution of polycyclic aromatic hydrocarbons in the Los Angeles atmosphere: application of a CMB with first order decay. US EPA Report EPA-600/2-81-161, US Government Printing Office, Washington, DC. Feng, J., Chan, C.K., Fang, M., Hu, M., He, L., Tang, X., 2006. Characteristics of organic matter in PM2.5 in Shanghai. Chemosphere 64 (8), 1393–1400. Harrison, R.M., Smith, D.J.T., Luhana, L., 1996. Source apportionment of atmospheric polycyclic aromatic hydrocarbons collected from an urban location in Birmingham, UK. Environmental Science and Technology 30 (3), 825–832. Jiang, Y.F., Wang, X.T., Wang, F., Jia, Y., Wu, M.H., Sheng, G.Y., Fu, J.M., 2009. Levels, composition profiles and sources of polycyclic aromatic hydrocarbons in urban soil of Shanghai, China. Chemosphere 75 (8), 1112–1118. Johansson, E., Wold, S., Sjoedin, K., 1984. Minimizing effects of closure on analytical data. Analytical Chemistry 56 (9), 1685–1688. Johnsen, A.R., Wick, L.Y., Harms, H., 2005. Principles of microbial PAH-degradation in soil. Environmental Pollution 133 (1), 71–84. Johnston, K., Hoef, J.M.V., Krivoruchko, K., Lucas, N., 2004. Using ArcGIS geostatistical analyst (GIS by ESRI). In: ArcGIS, ESRI, p. 300. Kavouras, I.G., Koutrakis, P., Tsapakis, M., Lagoudaki, E., Stephanou, E.G., Von Baer, D., Oyola, P., 2001. Source apportionment of urban particulate aliphatic and polynuclear aromatic hydrocarbons (PAHs) using multivariate methods. Environmental Science and Technology 35 (11), 2288–2294. Kipopoulou, A.M., Manoli, E., Samara, C., 1999. Bioconcentration of polycyclic aromatic hydrocarbons in vegetables grown in an industrial area. Environmental Pollution 106 (3), 369–380. Kucera, M., Malmgren, B.A., 1998. Log ratio transformation of compositional data – a resolution of the constant sum constraint. Marine Micropaleontology 34, 117– 120. Larsen III, R.K., Baker, J.E., 2003. Source apportionment of polycyclic aromatic hydrocarbons in the urban atmosphere: a comparison of three methods. Environmental Science and Technology 37 (9), 1873–1881. Lehndorff, E., Schwark, L., 2009. Biomonitoring airborne parent and alkylated threering PAHs in the Greater Cologne Conurbation II: regional distribution patterns. Environmental Pollution 157 (5), 1706–1713.
362
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
Levendis, Y.A., Atal, A., Carlson, J.B., 1998. On the correlation of CO and PAH emissions from the combustion of pulverized coal and waste tires. Environmental Science and Technology 32 (23), 3767–3777. Li, Y.T., Li, F.B., Chen, J.J., Yang, G.Y., Wan, H.F., Zhang, T.B., Zeng, X.D., Liu, J.M., 2008. The concentrations, distribution and sources of PAHs in agricultural soils and vegetables from Shunde, Guangdong, China. Environmental Monitoring and Assessment 139 (1–3), 61–76. Liu, M., Cheng, S.B., Ou, D.N., Hou, L.J., Gao, L., Wang, L.L., Xie, Y.S., Yang, Y., Xu, S.Y., 2007a. Characterization, identification of road dust PAHs in central Shanghai areas, China. Atmospheric Environment 41 (38), 8785–8795. Liu, Y., Chen, L., Tang, Y., Huang, Q., Zhao, J., 2007b. Determination of trace polycyclic aromatic hydrocarbons in surface sediments of Huangpu River by high performance liquid chromatography. Chinese Journal of Chromatography 25 (3), 356–361. Liu, Y., Chen, L., Zhao, J.F., Huang, Q.H., Zhu, Z.L., Gao, H.W., 2008. Distribution and source of polycyclic aromatic hydrocarbons in surface sediments of rivers and an estuary in Shanghai, China. Environmental Pollution 154 (2), 298–305. Liu, Y., Chen, L., Huang, Q.H., Li, W.Y., Tang, Y.J., Zhao, J.F., 2009. Source apportionment of polycyclic aromatic hydrocarbons (PAHs) in surface sediments of the Huangpu River, Shanghai, China. Science of the Total Environment 407 (8), 2931–2938. Marr, L.C., Kirchstetter, T.W., Harley, R.A., Miguel, A.H., Hering, S.V., Hammond, S.K., 1999. Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions. Environmental Science and Technology 33 (18), 3091–3099. Murakami, M., Nakajima, F., Furumai, H., 2005. Size- and density-distributions and sources of polycyclic aromatic hydrocarbons in urban road dust. Chemosphere 61 (6), 783–791. Pollard, A.M., Blockley, S.P.E., Lane, C.S., 2006. Some numerical considerations in the geochemical analysis of distal microtephra. Applied Geochemistry 21 (10), 1692–1714. NRC, 1983. Polycyclic Aromatic Hydrocarbons: Evaluation of Sources and Effects. National Academy Press, Washington, DC. Ren, Y., Zhang, Q., Chen, J., 2006. Distribution and source of polycyclic aromatic hydrocarbons (PAHs) on dust collected in Shanghai, People’s Republic of China. Bulletin of Environmental Contamination and Toxicology 76 (3), 442–449. Tang, L., Tang, X.Y., Zhu, Y.G., Zheng, M.H., Miao, Q.L., 2005. Contamination of polycyclic aromatic hydrocarbons (PAHs) in urban soils in Beijing, China. Environment International 31 (6), 822–828.
UNEP, 2002. Europe regional report. Global Environment Facility, Switzerland. USEPA, 1996. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods SW-846. Office of Solid Waste and Emergency Response, Washington, DC, USA. Wang, X.J., Zheng, Y., Liu, R.M., Li, B.G., Cao, J., Tao, S., 2003. Kriging and PAH pollution assessment on the topsoil of Tianjin area. Bulletin of Environmental Contamination and Toxicology 71 (1), 189–195. Wang, Z., Li, K., Lambert, P., Yang, C., 2007. Identification, characterization and quantitation of pyrogenic polycyclic aromatic hydrocarbons and other organic compounds in tire fire products. Journal of Chromatography A 1139 (1), 14–26. Yang, H.-H., Tsai, C.-H., Chao, M.-R., Su, Y.-L., Chien, S.-M., 2006. Source identification and size distribution of atmospheric polycyclic aromatic hydrocarbons during rice straw burning period. Atmospheric Environment 40 (7), 1266–1274. Ye, B., Zhang, Z., Mao, T., 2006. Pollution sources identification of polycyclic aromatic hydrocarbons of soils in Tianjin area, China. Chemosphere 64 (4), 525– 534. Yunker, M.B., Macdonald, R.W., 2003. Alkane and PAH depositional history, sources and fluxes in sediments from the Fraser River Basin and Strait of Georgia, Canada. Organic Geochemistry 34 (10), 1429–1454. Yunker, M.B., Macdonald, R.W., Vingarzan, R., Mitchell, R.H., Goyette, D., Sylvestre, S., 2002. PAHs in the Fraser River basin: a critical appraisal of PAH ratios as indicators of PAH source and composition. Organic Geochemistry 33 (4), 489– 515. Zedeck, M.S., 1980. Polycyclic aromatic hydrocarbons: a review. Journal of Environmental Pathology and Toxicology 3 (5–6), 537–567. Zhang, C., Selinus, O., 1997. Spatial analyses for copper, lead and zinc contents in sediments of the Yangtze River basin. Science of the Total Environment 204 (3), 251–262. Zhang, C., Selinus, O., 1998. Statistics and GIS in environmental geochemistry – some problems and solutions. Journal of Geochemical Exploration 64 (1–3), 339–354. Zhang, H.B., Luo, Y.M., Wong, M.H., Zhao, Q.G., Zhang, G.L., 2006. Distributions and concentrations of PAHs in Hong Kong soils. Environmental Pollution 141 (1), 107–114. Zhang, Y.J., Zhu, S.Q., Xiao, R., Wang, J., Li, F.S., 2008. Vertical transport of polycyclic aromatic hydrocarbons in different particle-size fractions of sandy soils. Environmental Geology 53 (6), 1165–1172.
Contents lists available at ScienceDirect
Organic Geochemistry journal homepage: www.elsevier.com/locate/orggeochem
Polycyclic aromatic hydrocarbons in the surface soil of Shanghai, China: Concentrations, distribution and sources Ying Liu a, Ling Chen a,*, Jianfu Zhao a,b, Yanping Wei a, Zhaoyu Pan a, Xiang-Zhou Meng a, Qinghui Huang b, Weiying Li a a b
State Key Laboratory of Pollution Control and Resources Reuse, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China Key Laboratory of Yangtze River Water Environment, Ministry of Education, College of Environmental Science and Engineering, Tongji University, Shanghai 200092, China
a r t i c l e
i n f o
Article history: Received 3 June 2009 Received in revised form 7 November 2009 Accepted 23 December 2009 Available online 4 January 2010
a b s t r a c t We quantified 18 polycyclic aromatic hydrocarbons (PAHs) in 54 surface soil samples covering an area of 6400 km2 in Shanghai. An isopleth map of total concentrations of the 18 PAHs, which was constructed using an ordinary Kriging approach with log transformed data, clarified the regional variability and identified regional hot spots in the urban and industrial areas of Shanghai. These hot spots all suffer from high PAH pollution, suggesting that local human activities (e.g., vehicular traffic, petrochemical industry and coal combustion) may be the main contributors. Coal or oil fired power plants and their locations seem to be a significant factor controlling the PAH concentrations in surface soil. The higher molecular weight PAHs are often accumulated near pollution sources and are more heterogeneous in Shanghai soil, because they are less easily transported and biodegraded than 2 ring PAHs. The total concentrations are not correlated with soil total organic carbon. We successfully applied hierarchical cluster analysis (HCA) and principal components analysis (PCA) based on a centered log ratio procedure to explore the characteristics and possible sources of soil PAHs. The high PAH contamination in the Shanghai surface soil is mainly attributed to the contribution of pyrogenic sources (vehicular traffic pollution and combustion of coal and biomass). Furthermore, we applied PAH percentages by ring number, isopleth maps of total concentrations of 18 PAHs and source diagnostic ratios of PAHs to help assign the pyrogenic sources in Shanghai soils. Such map based approaches have only rarely been applied in investigations published in Organic Geochemistry. Ó 2009 Elsevier Ltd. All rights reserved.
1. Introduction Polycyclic aromatic hydrocarbons (PAHs) are common contaminants that are difficult to biodegrade in the environment. They are derived from natural processes (e.g., forest fires) and/or anthropogenic activities (e.g., combustion of coal, waste incineration, vehicular traffic, spillage of petroleum). Some PAHs have been recommended as priority pollutants by the United States Environmental Protection Agency (US EPA) and as persistent toxic substances (PTS) by the United Nations Environment Programme (UNEP) due to their persistent, toxic, mutagenic and carcinogenic characteristics (Zedeck, 1980; NRC, 1983; UNEP, 2002). In recent years, there has been an increasing focus on regional soil contamination by PAHs, especially in China (Cai et al., 2008). PAHs are easily accumulated in soils over many years due to their biochemical persistence and hydrophobicity (Johnsen et al., 2005). As a result, soils become an important reservoir of PAHs. In addition, PAHs in
* Corresponding author. Tel./fax: +86 21 65984261. E-mail addresses: [email protected] (Y. Liu), [email protected] (L. Chen). 0146-6380/$ - see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.orggeochem.2009.12.009
soils may further accumulate in vegetables and other biota via food chains (Kipopoulou et al., 1999; Li et al., 2008), leading to direct or indirect exposure in humans. Therefore, studies on soil contamination by PAHs are necessary to minimize the risk of human exposure and environmental pollution. Shanghai is one of the most comprehensively industrial and commercial cities in China, ranking first in population and population density. Recently, many studies have been conducted in Shanghai on the levels and sources of PAHs in the atmosphere (Feng et al., 2006; Cheng et al., 2007), sediments (rivers and estuary) (Bouloubassi et al., 2001; Liu et al., 2008), road dust (Ren et al., 2006; Liu et al., 2007a) and urban soils (Jiang et al., 2009). However, few data have been reported on the spatial distribution and sources of PAHs in Shanghai soils, especially on the large scale. Regional soil PAH contamination mainly originates from dry and wet atmospheric deposition. PAH concentrations in road dust vary from 3100–32,000 ng/g in the urban area of Shanghai, which have been attributed to emissions from vehicles (Ren et al., 2006) and a mixture of traffic and coal combustion (Liu et al., 2007a). Concentrations of atmospheric PAHs range from 0.07–270 ng/m3, with major sources being fossil fuel combustion, coal burning, industrial
356
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
Fig. 1. Sampling locations (a) and isopleth map of total concentrations of 18 PAHs (b) in surface soils of the Shanghai area.
furnaces and engine exhaust emissions (Feng et al., 2006; Cheng et al., 2007). PAH concentrations in urban soils have been found to range from 442–19,700 ng/g, with PAHs attributed primarily to combustion (Jiang et al., 2009). Further studies are needed to understand and explore combustion sources of these PAHs. The soil is a sink and source for atmospheric PAHs and a source for sedimentary PAHs. Our previous work has shown that a river segment near Site 606 (Fig. 1a) is heavily polluted by PAHs and coal burning is one of the contributors of pyrogenic PAHs (Liu et al., 2008). Therefore, studies on the concentrations and spatial distribution of soil PAHs are useful to further understand the distribution of sedimentary PAHs in Shanghai. The objectives of the present study are to determine the concentrations and spatial distributions of PAHs in Shanghai soils, and to explore their characteristics and possible combustion sources. Such map based approaches have only rarely been applied in investigations published in Organic Geochemistry. The results of this study will input new data to the global PAH database and provide valuable information for regulatory actions to improve the environmental quality of the Yangtze River Delta.
2.2. Sample extraction and cleanup
2. Materials and methods
Sixteen PAHs characterized by the US EPA as priority pollutants were analyzed, including naphthalene (Nap), acenaphthylene (AcNy), fluorene (Fl), acenaphthene (AcNe), phenanthrene (PhA), anthracene (An), fluoranthene (FlA), pyrene (Py), benz[a]anthracene (BaA), chrysene (Chy), benzo[b]fluoranthene (BbF), benzo[k]fluoranthene (BkF), benzo[a]pyrene (BaP), indeno[1,2,3cd]pyrene (IP), benzo[ghi]perylene (BghiP) and dibenz[a,h]anthracene (DBahA). In addition, benzo[e]pyrene (BeP), perylene (Pery), 1-methylnaphthalene and 2-methylnapthalene were measured qualitatively. Concentrations of the two methylnaphthalene isomers were combined and reported as total methylnaphthalenes (MNap). PAH analysis was carried out by high performance liquid chromatography (HPLC) with a photodiode array detector. Identification of PAH compounds was based on chromatographic retention time and the ultraviolet spectra of PAH standards. The quantification was performed by the external standard method. The ultraviolet measurement wavelengths included 218 nm (NaP), 223 nm (MNaP), 226 nm (AcNe and AcNy), 249 nm (IP), 254 nm (Fl, PhA and An), 266 nm (Chy), 286 nm (FlA and BaA), 300 nm (BbF, BkF, BaP, DBahA and BghiP), 330 nm (BeP), 334 nm (Py) and 433 nm
2.1. Study area and sample collection Shanghai is located in the east of the Yangtze River Delta, as shown in Fig. 1a. Bordered by Jiangsu and Zhejiang provinces in the west, Shanghai is bounded by the East China Sea to the east and Hangzhou Bay to the south. North of the city, the Yangtze River flows into the East China Sea. Shanghai has a total area of 6340 km2, including a land area of 6218 km2 and a water area of 122 km2. As is typical of subtropical areas, Shanghai has a marine monsoon climate with a clear division of four seasons, enough sunshine and ample rainfall. The average annual temperature and rainfall are about 15.8 °C and 1149 mm, respectively. We collected 54 surface soil samples from Shanghai in April 2007 using the systematic grid method. Each 5–10 cm deep sample was obtained using a Luoyang shovel and consisted of five* subsamples from the surrounding area of each sampling site (within 10 m2). All samples were air dried in the dark and sieved to <0.076 mm (200 mesh) after removing stones and residual roots, then stored at 4 °C until the PAHs were analyzed.
The five sub-samples were mixed together to obtain a representative sample for each site and each mixed sample was analyzed in duplicate. Sample extraction and cleanup followed Method 3541 (automated Soxhlet extraction), Method 3630C (silica gel cleanup) and Method 3660B (sulfur cleanup) recommended by the US EPA (1996). Five grams of soil sample were blended with anhydrous sodium sulfate (1:1, w:w) and extracted by a FOSS Soxtec Avanti 2050 automatic system using 70 ml of a mixture of acetone and hexane (1:1, v:v), a hot plate temperature of 160 °C and an extraction time of 60 min for boiling and 60 min for rinsing. The extract was concentrated by a rotary evaporator, and the extract solvent was exchanged to cyclohexane. The detailed cleanup procedure is reported elsewhere (Liu et al., 2007b, 2008). Briefly, the extracts were further cleaned using a chromatographic column packed with 5 g activated silica gel and the eluate was finally concentrated to 1 ml. 2.3. Sample analysis and quality control
357
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
(Pery). Detailed measurement procedures are described elsewhere (Liu et al., 2007b, 2008). All analyses were calculated on a dry weight (dw) basis. Method detection limits (MDLs) for all targets ranged from 1–19 ng/g dw. The recoveries for spiked PAHs varied from 87%–113%. Each mixed sample was analyzed in duplicate and the relative standard deviation was <20%. Total organic carbon (TOC) analysis was performed with the Shimadzu TOC-Vcpn analyzer with the solid sample module (SSM-5000A). The overall precision of measurements was less than 3% (n = 3).
ker and Macdonald, 2003). Finally, the ordinary Kriging approach with log transformation was applied to data interpolation for the total concentrations of 18 PAHs, and inverse distance weighted interpolation was used for source diagnostic ratios of PAHs (after omitting the data with undetectable values of PAH), using the Geostatistical Analyst tool of ESRI ArcMap 9.2 software to draw isopleth maps of Shanghai.
2.4. Data analysis
The concentrations of individual PAHs (P3 rings) show large variations (Table 1). The highest concentration of a single PAH is of FlA, which ranges from undetected to 5200 ng/g dw, and the maximum concentrations for the P4 rings PAHs and PhA were more than 1000 ng/g dw. The minimum concentrations for all PAHs analyzed were less than the method detection limits, except for PhA. However, the 2 ring PAHs Nap and MNap have a smaller range between the highest and lowest concentrations and similar trends were found between the average and median concentrations as compared with other PAHs. Hence, the residual levels of 2 ring PAHs are more homogeneous in the surface soils than those of other PAHs. Zhang and Selinus (1998) suggested that the probability distributions of environmental geochemical variables for major elements (i.e., O and Si) generally follow normal distributions, but trace elements tend to follow log normal distributions. In this study, the concentrations of trace PAHs in soils follow similar probability distributions. The K–S test indicated that the concentrations of Nap and MNap were normally distributed (p < 0.05), as shown in Table 1, while the other PAHs followed log normal or positively skewed distributions. By comparison, Nap and MNap have low boiling points (relatively easy to volatilize) and good water solubility (tend to transport with water). As a result, they can be biodegraded and transported easily. Zhang et al. (2008) also believed that vertical transport of PAHs in soils were controlled by the nature of PAHs (i.e. log Koc, molecular weight). Other PAHs, despite
Before statistical analysis, analytical replicates were averaged and undetectable values were assigned the value of half of the MDLs. AcNy was eliminated from the dataset as it was undetectable in all samples. Statistical analyses, including the Kolmogorov–Smirnov (K–S) test, descriptive statistics, hierarchical cluster analysis (HCA) and principal components analysis (PCA), were performed using SPSS 13.0 for Windows. Before HCA and PCA, PAH data were normalized using a centered log ratio procedure (Bonn, 1998; Yunker and Macdonald, 2003; Pollard et al., 2006). Given a sample S ¼ ðx1 ; x2 ; . . . ; xN Þ, where the xi are the percentages of N species, the transformed sample is expressed as Str ¼ ðln½x1 =g; ln½x2 =g; . . . ; ln½xN =gÞ, where g is the geometric mean of sample S, which is calculated using the pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi equation g s ¼ N x1 x2 xN . Autoscaling (scale to variable mean and standard deviation) was then applied. To explore the structure between variables (PAH), the normalized dataset was hierarchically clustered using the weighted average linkage between the groups and the Pearson correlation for the cluster intervals (Kavouras et al., 2001; Zhang et al., 2006). In addition, to further understand the relationship between variables (PAH) and identify major sources of PAHs, we applied PCA to extract the principal components (PCs) with a varimax rotation. The rotation maximizes or minimizes the loading of each variable on each PC while preserving its trend (Kavouras et al., 2001; Yun-
3. Results and discussion 3.1. Concentration and probability distribution of PAHs
Table 1 Descriptive statistics of PAHs in surface soils of Shanghai. No.
PAHs
Abbrev.
No. of rings
Distribution
Nap MNap AcNe Fl PhA An FlA Py Chy BaA BbF BkF BaP BeP Pery DBahA IP BghiP
2 2 3 3 3 3 4 4 4 4 5 5 5 5 5 5 6 6
Normal Normal Skewed Skewed Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Lognormal Skewed Skewed Skewed Lognormal Lognormal Lognormal Lognormal
MDLa
Min.
Max.
Meanb
Median
Geo.c
DF (%)d
NDe ND ND ND 9.5 ND ND ND ND ND ND ND ND ND ND ND ND ND 94 62.4
110 90.3 188 243 2300 807 5190 4290 2230 2440 3590 1380 3130 2790 2620 1190 2260 2680 37,400 31,900
28.6 30.5 10.0 19.9 130 28.3 259 220 127 124 177 72.8 181 129 139 59.5 118 144 2000 1700
24.7 25.4 ND ND 34.3 2.8 40.5 30.7 22.5 16.9 29.5 12.0 30.6 ND 26.9 6.5 15.0 20.9 390 314
21.6 23.7 2.40 12.8 43.8 3.9 44.4 41.2 25.1 20.4 25.8 12.7 33.6 24.4 32.7 10.4 19.8 25.0 526 405
94 94 35 24 100 65 91 91 93 80 81 80 80 50 61 52 72 80 100 100
(ng/g dw) 1 Naphthalene 2 Methylnaphthalene 3 Acenaphthene 4 Fluorene 5 Phenanthrene 6 Anthracene 7 Fluoranthene 8 Pyrene 9 Chrysene 10 Benz[a]anthracene 11 Benzo[b]fluoranthene 12 Benzo[k]fluoranthene 13 Benzo[a]pyrene 14 Benzo[e]pyrene 15 Perylene 16 Dibenz[a,h]anthracene 17 Indeno[1,2,3-cd]pyrene 18 Benzo[ghi]perylene P 18 PAHs P 16 PAHs (US EPA) a b c d e
MDL, method detection limit. Mean, arithmetic mean. Geo., geometric mean. DF, detectable frequency. ND, not detected.
4.0 3.6 1.9 18.3 1.1 1.8 5.1 7.1 2.2 4.9 1.9 3.1 6.5 19.1 17.1 6.4 6.4 6.1 – –
358
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
low concentrations in the environment, are usually enriched in heavy traffic or industrial areas, which may lead to a positively skewed distribution (Table 1). However, the concentrations of AcNe, Fl, BeP, Pery and DBahA do not follow log normal distributions, because the detectable frequencies of those compounds are low (<55%) and the undetectable values were assigned values of half of the MDLs (Table 1). In general, the arithmetic mean is used for normally distributed parameters, such as Nap and MNap, while the geometric mean is used for parameters that follow a log normal or positively skewed distribution, such as Fl, PhA and BaP. P The total concentrations of 18 PAHs ( 18 PAHs) range from 94–37,300 ng/g dw with a geometric mean of 526 ng/g dw. In P 80% of samples, 18 PAHs is <1000 ng/g dw (Fig. S1). The K–S P test also indicated that 18 PAHs followed a log normal distribution. 3.2. The relationship between total PAHs and total organic carbon (TOC) in soil TOC is an important controlling factor of sorption of PAHs in soil and may pre-determine the level of soil contamination with PAHs (Tang et al., 2005). In this study, soil TOC varies greatly from 4.25– 46.1 mg/g dw. Pearson correlation analysis shows that the total concentrations of 18 PAHs in soils are not significantly correlated with soil TOC (r = 0.062 and p = 0.65, see Fig. S2). Jiang et al. (2009) also reported a similar result when they researched the relationship between PAHs and TOC in urban soil of Shanghai and believed the result to be due to non-equilibrium adsorption between TOC and PAHs in soils. In our work, the inputs of PAHs to each site vary and not all adsorption of PAHs on TOC reaches equilibrium. Therefore, the effect of soil TOC on the amount of PAH accumulated in soil is ignored in the following spatial distribution of PAHs. 3.3. Spatial distribution of total PAHs Many interpolation methods can be used to analyze the spatial distribution of PAH levels in the environment. A surface tension spline function was used for dispersive interpolation and spatial distribution contouring of PAHs (Ye et al., 2006). Geostatistics can provide an optimal interpolation approach known as Kriging. This theory has been developed and applied widely by geologists to estimate mineral reserves. Geochemists have also tried to use the Kriging approach for the spatial distribution analysis of heavy metals (Zhang and Selinus, 1997) and PAHs (Wang et al., 2003). Kriging methods work best when the data are approximately normally distributed. In addition, the log transformation is often used where the data has a positively skewed distribution and there are some very large values in order to avoid a proportional effect (Zhang and Selinus, 1997; Johnston et al., 2004). In the present study, the total PAH concentration follows a lognormal distribution. We therefore applied the ordinary Kriging method with log transformation to predict the PAH values in the non-sampled area in Shanghai. Fig. 1b shows the spatial distribution contours of total PAH concentrations in the surface soils, residential and industrial areas and main power plants sites. Isopleth mapping has proved to be a valuable approach for identification of regional hot spots for PAH pollution (Lehndorff and Schwark, 2009). In this study, three regions with high PAH concentration were delimited based on the region function, here designated as regions A, B and C. Regions A and B are located in urban areas, and Region C is situated in a petrochemical industrial area. These regions are severely influenced by human activity. Region A is located in the central part of Shanghai with heavy traffic. As a result, the surrounding soil is contaminated by road dust, which is an agglomeration of various traffic related PAH sources,
e.g., automobile exhaust, lubricating oil, gasoline, diesel fuel, weathered materials of street surfaces and tire particles (Murakami et al., 2005). Liu et al. (2007a) reported that 16 PAH concentrations in road dust from the central part of Shanghai (Region A) were very high, with a range from 6875–32,573 ng/g dw, which suggests heavy contamination in the surface soils of Region A. The high PAH concentrations in Region B may be associated with regional industrial pollution. Region B is situated in an old industrial district with many manufacturing facilities, e.g., chemical and coke oven plants. Nowadays, some plants with heavy pollution have been encouraged to move elsewhere, and a new residential and financial district is gradually forming. However, traffic pollution is replacing industrial pollution and continues to contaminate the surrounding soils. We note that coal fired power plants in Region B contribute to PAH contamination in soil. Region C adjoins the sea and Shanghai’s principle petrochemical station and its oil refinery are located in this region. The high PAH contents in the soil may be due to contamination from the petrochemical industry. To understand the effect of power plants on the spatial distribution of PAHs, we show the main large scale coal or oil fired power plants in Shanghai in Fig. 1b. Clearly, the power plants are situated in the higher PAH pollution areas. This indicates that power plants seem to be an important factor controlling the PAH contents of surface soils in Shanghai. 3.4. PAH composition comparison using multivariate statistic analysis The relationship of PAH compositions and sources in surface soils of Shanghai were examined using two multivariate statistical analysis methods. AcNy, AcNe and Fl were excluded in the analysis due to their low detectable frequencies (Table 1). The normality of the data is an important factor when interpreting analytical results, since many inferences are based on the assumption of normality. Multivariate statistical applications are still almost exclusively based on correlation or covariance matrices determined directly from raw data. However, normalizing data can improve the explanatory power of a multivariate analysis (Kucera and Malmgren, 1998). In a concentration based multivariate model, the samples with the highest concentrations usually project separately as single samples or small groups of samples (generally with positive projections), while the remaining (lower concentrations) samples project in a group near the origin and make little or no contribution to the model. If normalization to the total PAH concentration has been used, PAH concentration data are transformed into compositional data with the constant sum constraint (CSC). The CSC introduces spurious negative correlations between major variables and positive correlations between minor ones (Johansson et al., 1984). The log ratio transformation can effectively reduce the CSC from compositional data (Kucera and Malmgren, 1998), and the centered log ratio transformed values have been used as inputs into PCA (Bonn, 1998; Yunker and Macdonald, 2003; Pollard et al., 2006). In this paper, we applied a pattern based approach to multivariate statistical analysis using the centered log ratio procedure to produce a normalized dataset as input into HCA and PCA. 3.4.1. Hierarchical cluster analysis Hierarchical cluster analysis (HCA) was used to explore structure of the normalized dataset. No significant difference was found when different linkage methods were used based on the results of the HCA. Fig. 2 shows the results presented in the form of a dendrogram. Sixteen variables (PAHs) were classified into three major groups. The first group is composed of low molecular weight PAHs with 2–3 rings and alkyl substituted PAHs, including Nap, PhA and MNap. The second group includes DBahA, BeP, An and Pery. Detectable frequencies of the four compositions were less than 70% (Table 1). In this group, PAH compositions with similar detectable
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
Fig. 2. Hierarchical dendrogram for PAHs using average linkage between groups and Pearson correlation as a measure interval.
frequencies were clustered together, e.g., DBahA and BeP. The third group, including most 4–6 rings PAHs, was subdivided into two subgroups, namely 4 ring PAHs (FlA, Py, Chy and BaA) and 5–6 ring PAHs (BbF, BkF, BaP, BghiP and IP). These PAHs are usually detected in pyrogenic sources, e.g., combustion of coal, wood, vehicle fuel and waste tire (Levendis et al., 1998; Wang et al., 2007). 3.4.2. Principal components analysis Principal components analysis (PCA) was performed to further understand the relationship between PAH compositions and possible chemical sources for each factor. Factor loadings of the normalized dataset are listed in Table S1 and plotted in Fig. 3a. Most of the variance (76.2%) of the normalized dataset was explained by the first four factors. Factor 1 explains 34.0% of the total variance. High negative loadings of MNap, Nap and PhA and high positive loadings of BbF, BkF, BaP, IP and BghiP were observed in this factor (Table S1). These PAH compositions are clustered on the left and right sides of the x-axis in Fig. 3a, respectively. MNap, Nap and PhA belongs to the low molecular weight PAHs with 2–3 rings or alkyl substituted PAHs, which are abundant in petrogenic source mainly caused by petroleum spills (Marr et al., 1999; Dobbins
359
et al., 2006). Samples with high negative score have a dominance of 2–3 rings PAHs, which originated mainly from petrogenic source. BghiP is identified as a tracer of auto emissions because it was found to be enriched in a traffic tunnel along with BaP (Harrison et al., 1996; Larsen and Baker, 2003; Boonyatumanond et al., 2007). IP is found in petroleum as a tracer of diesel combustion (Kavouras et al., 2001). The high level of BkF relative to other PAHs is suggested to indicate diesel vehicles (Larsen and Baker, 2003). BkF and its isomers (e.g., BbF and BaP) are dominant compounds in particulate samples of roadside air (Boonyatumanond et al., 2007). Samples with high positive score have a dominance of 5–6 rings PAHs, and vehicular traffic pollution is a major contributor of PAHs contamination in the samples. To sum up, the loadings for Factor 1 are primarily due to the separation between petrogenic source and vehicular traffic pollution. Factor 2, accounting for 16.5% of total variance, is negatively dominated by MNap and Nap and positively by FlA, Py, Chy and BaA (Table S1). In Fig. 3a, MNap and Nap are projected in the lower left quadrant, and they are related to petrogenic source. FlA, Py, Chy and BaA are projected in the upper right quadrant. The four compositions are high molecular weight PAHs with 4 rings as markers of coal combustion (Duval and Friedlander, 1981). The furnace effluents from combustion of pulverized coals contained FlA and Py (Levendis et al., 1998). Likewise, the semi-volatile PAHs of FlA and Py are attributed to unburned fossil fuels (Kavouras et al., 2001). Contents of the four PAHs in soils mainly reflect the contribution of combustion of coal and biomass. Therefore, the loadings for Factor 2 are primarily due to the separation between petrogenic source and combustion of coal and biomass. Factors 3 and 4 are responsible for 14.5% and 11.2% of the variance, respectively (Table S1). Factor 3 is heavily weighted by BeP and DBahA, while Factor 4 is dominated by An and Pery. However, no chemical sources with such a PAH profile have been reported until now. They might be likely to represent ‘‘noise” due to their low detectable ratios (50–65%). It is interesting that the PCA result in this paper is compared with the PCA result of sedimentary PAHs in our previous work, and PAH compounds with high positive loadings in extracted factors are parallel between in soil and in sediment (Liu et al., 2009). This indicates that soil PAHs and sedimentary PAHs in Shanghai have common sources. Fig. 3b shows a plot of the scores of samples for the first two principal components and stack column graphs of the percentages that 2–6 ring PAHs account for in the samples. In Fig. 3b, the urban influenced samples (I) project in the upper right quadrant, indicat-
Fig. 3. Loading plot (a) of the first two principal components and score plot (b) of each sample. The variance accounted for by each PC is shown after the axis label. Stack column graphs show percentages accounted for by 2–6 ring PAHs in the samples.
360
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
ing that these samples suffered greatly from contamination by vehicular traffic and combustion of coal and biomass. Likewise, PAHs from vehicular traffic and combustion of coal and biomass belong to pyrogenic PAHs and should have higher proportions of 4–6 rings PAHs, which agrees with their stack column graphs (I). Furthermore, these samples (I) were collected from heavy PAH contaminated regions in Fig. 1b. The regions are close to power plants and located in urban or industrial areas where vehicular traffic pollution is obvious. We note that Site 901 is situated in a petrochemical industrial region, where vehicular traffic pollution is not prominent in theory, but there is a similar stack column as other sites, e.g., Sites 604 and 405. According to photos pictured near Site 901 (Fig. S3), exhaust from petrochemical processing usually includes combustion products from hydrocarbons gases. The treatment of the exhaust by direct combustion produces high molecular weight PAHs and contaminates the surrounding soils, as does combustion of gasoline or diesel fuel in vehicles. Therefore, it is reasonable that the high PAH contamination in the Shanghai soil is due to the contribution of vehicular traffic and combustion of coal and petroleum.
In Fig. 3b, samples from Chongmingdao Island (II) are projected in the upper left quadrant, illustrating a weak contribution of vehicular traffic and a strong contribution of combustion of coal and biomass. Chongmingdao Island is an isolated region currently without linkage of a bridge or tunnel to the mainland of Shanghai. The isolation limits development of vehicular traffic, leading to pollution from vehicular traffic on the island being less than that on the mainland. Nowadays, Chongmingdao Island supports large scale agriculture, especially green agriculture and industrial development polluting the environment is limited. Here, the strong contribution of combustion of coal and biomass is attributed to straw combustion. In local agricultural activity, rice straw is burned in situ to release nutrients for the next growing season. The concentrations of atmospheric PAHs with lighter molecular weights, e.g., Py, FlA, PhA, Fl and Nap (2–4 ring PAHs), are obviously higher on rice straw burning days than on non-burning days, but no concentration difference is observed for higher molecular weight PAHs, e.g., 5–6 ring PAHs (Yang et al., 2006). The stack column graphs of Site A02 and Site A04 (Fig. 3bII) also reflect the dominance of 2–4 ring PAHs. In the samples from III (Sites 201, 301
Fig. 4. Isopleth maps of source diagnostic ratios of PAHs in surface soils of the Shanghai area.
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
361
and 302), the PAH concentration is low, even less than the MDLs and the total concentrations of PAHs are <150 ng/g dw. Overall, multivariate statistical analysis based on the centered log ratio procedure unveils the relationships of PAH compositions and identifies vehicular traffic and combustion of coal and biomass as dominant contributors of PAH contamination in soils. The spatial distribution contour of total PAH concentrations and stack column graphs of 2–6 ring PAH percentages further confirm the results of the multivariate statistic analysis.
Acknowledgements
3.5. Source diagnostic ratios of PAHs
Appendix A. Supplementary material
The relative abundances or diagnostic ratios are valid indicators of PAH sources because isomer pairs are diluted to a similar extent upon mixing with natural particulate matter and are distributed similarly to other phases as they have comparable thermodynamic partitioning and kinetic mass transfer coefficients (Dickhut et al., 2000). Diagnostic ratios of PAHs, such as ratios of An/(PhA + An), FlA/(FlA + Py), BaA/(BaA + Py) and IP/(IP + BghiP), can be applied to identify the possible emission sources, as noted in Fig. 4 (Yunker et al., 2002). Isopleth maps of source diagnostic ratios of PAHs are shown Fig. 4. The ratio of An/(An + PhA) ranges from 0.01–0.26, FlA/(FlA + Py) from 0.11–0.77, BaA/(BaA + Chy) from 0.30–0.60 and IP/(IP + BghiP) from 0.30–0.56. The results of these ratios indicate that the combustion of biomass, coal and petroleum are the major sources of PAHs in Shanghai. The region with higher PAH contamination is companied with a higher ratio of An/(An + PhA) (Fig. 4a). The ratios of FlA/(FlA + Py) and BaA/(BaA + Chy) show that combustion of coal and biomass are the predominant sources (Fig. 4b and c), but that of IP/(IP + BghiP) shows petroleum combustion is the major source (Fig. 4d). This may be because FlA, Py, BaA and Chy originate mainly from the contribution of combustion of coal and biomass, while IP and BghiP source chiefly from vehicular traffic pollution (e.g. combustion of fuel oils), according to PCA results. Therefore, the results from source diagnostic ratios of PAHs can reflect basically the contribution of combustion sources of PAHs.
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.orggeochem.2009.12.009.
4. Conclusions Although the total PAH concentrations ranged from 94– 37,300 ng/g dw, most surface soils in Shanghai have low PAH contamination. The concentrations of total PAH and most individual PAHs follow log normal distributions except for the Nap and MNap components. Isopleth maps are a valid approach to identify local hot spots in PAHs contamination and three regions with high PAH contamination were delimited using the ordinary Kriging method with log transformation. The surface soils with high PAHs concentrations are found in the urban area and petrochemical industrial area and anthropological influence is the major contributor of PAHs contamination in Shanghai surface soil. The ecological risk from PAH contamination should be considered for the persons living in the three regions. A centered log ratio transformation is an effective normalization procedure for PAH concentration data. Two multivariate statistic analysis methods based on this data normalization procedure have been successfully applied to explore the characteristics and to identify PAH sources. Vehicular traffic and combustion of coal and biomass are the main sources of PAHs in Shanghai surface soil. The results from multivariate statistical analysis are combined with PAH percentages by ring number and isopleth maps of total concentrations of 18 PAHs and source diagnostic ratios of PAHs, which further confirm the sources of PAHs. Such map based approaches are one of valid path in identification of PAH sources.
This work was supported by the National Natural Science Foundation of China (No. 20907034), Program for Young Excellent Talents in Tongji University (No. 2008KJ022), the Foundation of the State Key Laboratory of Pollution Control and Resource Reuse (No. PCRRK09001), and the National Key Technology R&D Program (No. 2008BAC46B02). We also thank Mark Bernard Yunker, Lorenz Schwark, Xue-Tong Wang and an anonymous reviewer for helpful comments and suggestions.
Associate Editor—Mark Yunker References Bonn, B.A., 1998. Polychlorinated dibenzo-p-dioxin and dibenzofuran concentration profiles in sediment and fish tissue of the Willamette Basin, Oregon. Environmental Science and Technology 32 (6), 729–735. Boonyatumanond, R., Murakami, M., Wattayakorn, G., Togo, A., Takada, H., 2007. Sources of polycyclic aromatic hydrocarbons (PAHs) in street dust in a tropical Asian mega-city, Bangkok, Thailand. Science of the Total Environment 384 (1– 3), 420–432. Bouloubassi, I., Fillaux, J., Saliot, A., 2001. Hydrocarbons in surface sediments from the Changjiang (Yangtze River) estuary, East China Sea. Marine Pollution Bulletin 42 (12), 1335–1346. Cai, Q.-Y., Mo, C.-H., Wu, Q.-T., Katsoyiannis, A., Zeng, Q.-Y., 2008. The status of soil contamination by semivolatile organic chemicals (SVOCs) in China: a review. Science of the Total Environment 389 (2–3), 209–224. Cheng, J., Yuan, T., Wu, Q., Zhao, W., Xie, H., Ma, Y., Ma, J., Wang, W., 2007. PM10bound polycyclic aromatic hydrocarbons (PAHs) and cancer risk estimation in the atmosphere surrounding an industrial area of shanghai, China. Water, Air, and Soil Pollution 183 (1–4), 437–446. Dickhut, R.M., Canuel, E.A., Gustafson, K.E., Liu, K., Arzayus, K.M., Walker, S.E., Edgecombe, G., Gaylor, M.O., MacDonald, E.H., 2000. Automotive sources of carcinogenic polycyclic aromatic hydrocarbons associated with particulate matter in the Chesapeake Bay region. Environmental Science and Technology 34 (21), 4635–4640. Dobbins, R.A., Fletcher, R.A., Benner Jr., B.A., Hoeft, S., 2006. Polycyclic aromatic hydrocarbons in flames, in diesel fuels, and in diesel emissions. Combustion and Flame 144 (4), 773–781. Duval, M.M., Friedlander, S.K., 1981. Source resolution of polycyclic aromatic hydrocarbons in the Los Angeles atmosphere: application of a CMB with first order decay. US EPA Report EPA-600/2-81-161, US Government Printing Office, Washington, DC. Feng, J., Chan, C.K., Fang, M., Hu, M., He, L., Tang, X., 2006. Characteristics of organic matter in PM2.5 in Shanghai. Chemosphere 64 (8), 1393–1400. Harrison, R.M., Smith, D.J.T., Luhana, L., 1996. Source apportionment of atmospheric polycyclic aromatic hydrocarbons collected from an urban location in Birmingham, UK. Environmental Science and Technology 30 (3), 825–832. Jiang, Y.F., Wang, X.T., Wang, F., Jia, Y., Wu, M.H., Sheng, G.Y., Fu, J.M., 2009. Levels, composition profiles and sources of polycyclic aromatic hydrocarbons in urban soil of Shanghai, China. Chemosphere 75 (8), 1112–1118. Johansson, E., Wold, S., Sjoedin, K., 1984. Minimizing effects of closure on analytical data. Analytical Chemistry 56 (9), 1685–1688. Johnsen, A.R., Wick, L.Y., Harms, H., 2005. Principles of microbial PAH-degradation in soil. Environmental Pollution 133 (1), 71–84. Johnston, K., Hoef, J.M.V., Krivoruchko, K., Lucas, N., 2004. Using ArcGIS geostatistical analyst (GIS by ESRI). In: ArcGIS, ESRI, p. 300. Kavouras, I.G., Koutrakis, P., Tsapakis, M., Lagoudaki, E., Stephanou, E.G., Von Baer, D., Oyola, P., 2001. Source apportionment of urban particulate aliphatic and polynuclear aromatic hydrocarbons (PAHs) using multivariate methods. Environmental Science and Technology 35 (11), 2288–2294. Kipopoulou, A.M., Manoli, E., Samara, C., 1999. Bioconcentration of polycyclic aromatic hydrocarbons in vegetables grown in an industrial area. Environmental Pollution 106 (3), 369–380. Kucera, M., Malmgren, B.A., 1998. Log ratio transformation of compositional data – a resolution of the constant sum constraint. Marine Micropaleontology 34, 117– 120. Larsen III, R.K., Baker, J.E., 2003. Source apportionment of polycyclic aromatic hydrocarbons in the urban atmosphere: a comparison of three methods. Environmental Science and Technology 37 (9), 1873–1881. Lehndorff, E., Schwark, L., 2009. Biomonitoring airborne parent and alkylated threering PAHs in the Greater Cologne Conurbation II: regional distribution patterns. Environmental Pollution 157 (5), 1706–1713.
362
Y. Liu et al. / Organic Geochemistry 41 (2010) 355–362
Levendis, Y.A., Atal, A., Carlson, J.B., 1998. On the correlation of CO and PAH emissions from the combustion of pulverized coal and waste tires. Environmental Science and Technology 32 (23), 3767–3777. Li, Y.T., Li, F.B., Chen, J.J., Yang, G.Y., Wan, H.F., Zhang, T.B., Zeng, X.D., Liu, J.M., 2008. The concentrations, distribution and sources of PAHs in agricultural soils and vegetables from Shunde, Guangdong, China. Environmental Monitoring and Assessment 139 (1–3), 61–76. Liu, M., Cheng, S.B., Ou, D.N., Hou, L.J., Gao, L., Wang, L.L., Xie, Y.S., Yang, Y., Xu, S.Y., 2007a. Characterization, identification of road dust PAHs in central Shanghai areas, China. Atmospheric Environment 41 (38), 8785–8795. Liu, Y., Chen, L., Tang, Y., Huang, Q., Zhao, J., 2007b. Determination of trace polycyclic aromatic hydrocarbons in surface sediments of Huangpu River by high performance liquid chromatography. Chinese Journal of Chromatography 25 (3), 356–361. Liu, Y., Chen, L., Zhao, J.F., Huang, Q.H., Zhu, Z.L., Gao, H.W., 2008. Distribution and source of polycyclic aromatic hydrocarbons in surface sediments of rivers and an estuary in Shanghai, China. Environmental Pollution 154 (2), 298–305. Liu, Y., Chen, L., Huang, Q.H., Li, W.Y., Tang, Y.J., Zhao, J.F., 2009. Source apportionment of polycyclic aromatic hydrocarbons (PAHs) in surface sediments of the Huangpu River, Shanghai, China. Science of the Total Environment 407 (8), 2931–2938. Marr, L.C., Kirchstetter, T.W., Harley, R.A., Miguel, A.H., Hering, S.V., Hammond, S.K., 1999. Characterization of polycyclic aromatic hydrocarbons in motor vehicle fuels and exhaust emissions. Environmental Science and Technology 33 (18), 3091–3099. Murakami, M., Nakajima, F., Furumai, H., 2005. Size- and density-distributions and sources of polycyclic aromatic hydrocarbons in urban road dust. Chemosphere 61 (6), 783–791. Pollard, A.M., Blockley, S.P.E., Lane, C.S., 2006. Some numerical considerations in the geochemical analysis of distal microtephra. Applied Geochemistry 21 (10), 1692–1714. NRC, 1983. Polycyclic Aromatic Hydrocarbons: Evaluation of Sources and Effects. National Academy Press, Washington, DC. Ren, Y., Zhang, Q., Chen, J., 2006. Distribution and source of polycyclic aromatic hydrocarbons (PAHs) on dust collected in Shanghai, People’s Republic of China. Bulletin of Environmental Contamination and Toxicology 76 (3), 442–449. Tang, L., Tang, X.Y., Zhu, Y.G., Zheng, M.H., Miao, Q.L., 2005. Contamination of polycyclic aromatic hydrocarbons (PAHs) in urban soils in Beijing, China. Environment International 31 (6), 822–828.
UNEP, 2002. Europe regional report. Global Environment Facility, Switzerland. USEPA, 1996. Test Methods for Evaluating Solid Waste, Physical/Chemical Methods SW-846. Office of Solid Waste and Emergency Response, Washington, DC, USA. Wang, X.J., Zheng, Y., Liu, R.M., Li, B.G., Cao, J., Tao, S., 2003. Kriging and PAH pollution assessment on the topsoil of Tianjin area. Bulletin of Environmental Contamination and Toxicology 71 (1), 189–195. Wang, Z., Li, K., Lambert, P., Yang, C., 2007. Identification, characterization and quantitation of pyrogenic polycyclic aromatic hydrocarbons and other organic compounds in tire fire products. Journal of Chromatography A 1139 (1), 14–26. Yang, H.-H., Tsai, C.-H., Chao, M.-R., Su, Y.-L., Chien, S.-M., 2006. Source identification and size distribution of atmospheric polycyclic aromatic hydrocarbons during rice straw burning period. Atmospheric Environment 40 (7), 1266–1274. Ye, B., Zhang, Z., Mao, T., 2006. Pollution sources identification of polycyclic aromatic hydrocarbons of soils in Tianjin area, China. Chemosphere 64 (4), 525– 534. Yunker, M.B., Macdonald, R.W., 2003. Alkane and PAH depositional history, sources and fluxes in sediments from the Fraser River Basin and Strait of Georgia, Canada. Organic Geochemistry 34 (10), 1429–1454. Yunker, M.B., Macdonald, R.W., Vingarzan, R., Mitchell, R.H., Goyette, D., Sylvestre, S., 2002. PAHs in the Fraser River basin: a critical appraisal of PAH ratios as indicators of PAH source and composition. Organic Geochemistry 33 (4), 489– 515. Zedeck, M.S., 1980. Polycyclic aromatic hydrocarbons: a review. Journal of Environmental Pathology and Toxicology 3 (5–6), 537–567. Zhang, C., Selinus, O., 1997. Spatial analyses for copper, lead and zinc contents in sediments of the Yangtze River basin. Science of the Total Environment 204 (3), 251–262. Zhang, C., Selinus, O., 1998. Statistics and GIS in environmental geochemistry – some problems and solutions. Journal of Geochemical Exploration 64 (1–3), 339–354. Zhang, H.B., Luo, Y.M., Wong, M.H., Zhao, Q.G., Zhang, G.L., 2006. Distributions and concentrations of PAHs in Hong Kong soils. Environmental Pollution 141 (1), 107–114. Zhang, Y.J., Zhu, S.Q., Xiao, R., Wang, J., Li, F.S., 2008. Vertical transport of polycyclic aromatic hydrocarbons in different particle-size fractions of sandy soils. Environmental Geology 53 (6), 1165–1172.