Aliphatic hydrocarbons (n-alkanes) can be used as tracers to diagnose the sources of hydrocarbons because they originate from both anthropogenic (e.g., use of petroleum) and natural sources (e.g., terrestrial higher plant waxes, aquatic phytoplankton, and bacteria) 1–3. Linear alkylbenzenes (LABs) are raw materials for producing linear alkylbenzenesulfonate surfactants and can be discharged into the environment as unsulfonated residues 4. Therefore, LABs are highly related to and considered appropriate molecular markers of sewage pollution 5, 6. Polycyclic aromatic hydrocarbons (PAHs) are mainly generated from incomplete combustion of fossil fuel and biomass 7, which is typical of anthropogenic activities. The combined use of n-alkanes, LABs, and PAHs would allow one to examine the relative importance of the impacts from various terrestrially derived organic pollutants on a coastal zone, thus facilitating the implementation of appropriate environmental protection and management measures.
Coastal zones are characterized by abundant resources and beneficial climate conditions and thus are important recreational areas. Consequently, intensive anthropogenic activities may induce large amounts of terrestrial organic materials to deposit in coastal marine systems through riverine runoff 8, or atmospheric deposition 9, or both. For example, the annual riverine inputs of n-alkanes, LABs, and Σ15PAH (sum of the U.S. Environmental Protection Agency's 16 PAH priority pollutants minus naphthalene) from the Pearl River Delta (PRD) to the northern South China Sea (NSCS) were estimated to be 360 tons 10, 14 tons 11, and 33.9 tons 12, respectively. Because estuarine and coastal sediments are known to serve as both sinks and secondary sources of terrestrial organic materials 13, coastal marine systems are therefore dynamic areas connecting lands and oceans. Accordingly, carefully designed and executed investigations of estuarine and coastal sediments can lead to better understanding of the influences of terrestrial anthropogenic activities on coastal environments, providing information for the decision-makers who manage these vital coastal resources.
The coastal marine environment off South China includes the PRD, a well-known economic zone, as well as the economically less developed eastern and western coastal areas of Guangdong Province (Fig. 1). Pollution levels between the economically developed PRD region and the less developed eastern (and western) coastal areas of Guangdong Province are expected to differ, because industrialization and urbanization are at different stages in these two regions. Employment of multiple geochemical markers would help in characterizing this hypothesis. Characterization has been done of potential sources of n-alkanes 14, LABs 15, and PAHs 16–18 in sediments from the economically fastest growing areas, the PRD region and its adjacent areas, for example, the Pearl River Estuary (PRE) and the NSCS. However, only a few studies have been done on sediments from the rest of coastal Guangdong within a small range of specific sites 19, 20.
The present study was designed to analyze organic geomarkers in sediments from the economically less developed eastern and western coastal areas, and compare them with published data on sediments in the PRE coastal area of the PRD. Our aim was to determine potential spatial differences, and factors controlling them, in geomarker pollution between the PRD and the rest of coastal Guangdong, South China.
MATERIALS AND METHODS
Sample collection and extraction
From December 2006 to January 2007, surface sediment samples (n = 24; 0–5-cm layer) were collected using a stainless steel grab sampler off the eastern (labeled 1–17) and western (labeled 18–24) coast of Guangdong Province, South China (Fig. 1). Samples were sealed with polyethylene bags and transported to the laboratory, where they were preserved at −20°C until analysis.
Sediment samples were freeze-dried, homogenized with an 80-mesh sieve, and weighed before being packed into precleaned filters. After spiking with known amounts of surrogate standards (tetracosane-d50 for n-alkanes; 1-phenyldodecane-d30 for LABs; naphthalene-d8, acenaphthene-d10, phenanthrene-d10, chrysene-d12, and perylene-d12 for PAHs), a set of 15 field samples along with one procedural blank, spiked blank, matrix blank, and matrix-spiked blank was subjected to Soxhlet extraction with 200 ml of a hexane:acetone (1:1 in volume) mixture for 48 h. Copper sheets were added to the solvent mixture to remove elemental sulfur. Purification/fractionation procedures to collect n-alkanes 21, LABs 11, and PAHs 12 were followed as outlined in previous studies. Each fraction was concentrated to 0.5 ml under a gentle nitrogen flow. Known amounts of internal standards (triacontane-d62 for n-alkanes; 1-phenylpentadecane-d36 for LABs; 2-fluoro-1,1-biphenyl,p-terphenyl-d14 and dibenzo[a,h]anthracene-d14 for PAHs) were added before instrumental analysis.
A total of 23 n-alkanes (pristane [Pr], phytane [Ph], and n-C15 to n-C35, where the subscript indicates the carbon number), 19 LABs (i-CjLAB, where i represents the position of phenyl substitution and the subscript j indicates the carbon number [10–13] of the alkyl chain), and 28 PAH congeners, including 16 priority PAHs, were selected as target analytes (Supplemental Data, Table S1). A Shimadzu GCMS-QP2010 Plus gas chromatograph/mass spectrometer equipped with an AOC-20i auto injector was utilized to determine the concentrations of n-alkanes and PAHs. A Varian 3800 gas chromatograph interfaced with a Saturn 2000 ion trap mass spectrometer was employed for the quantification of LAB components. Chromatographic separation of n-alkanes was done with a 30 m × 0.25 mm i.d. (0.25-µm film thickness) DB-5 column, and LABs and PAHs were separated with a 60 × 0.25 mm i.d. (0.25-µm film thickness) DB-5 column (J&W Scientific). The mass spectra of n-alkanes and PAHs were generated in the full scan mode for peak confirmation, and those of LABs were acquired in the range of m/z 45 to 400.
Quality assurance and quality control
A standard solution was analyzed every 15 field samples as daily calibration. The difference between the measured and prepared concentrations of the target analytes must be less than 20% before sample analysis. For field samples, the recoveries of the n-alkane and LAB surrogate standards, tetracosane-d50 and 1-phenyldodecane-d30, were 70 ± 4% and 61 ± 12%, respectively, whereas those of the PAH surrogate standards, naphthalene-d8, acenaphthene-d10, phenanthrene-d10, and chrysene-d12, were 50 ± 15%, 65 ± 18%, 100 ± 21%, and 99 ± 15%, respectively. The recoveries of the same surrogate standards in quality assurance and quality control samples were 61 ± 5%, 75 ± 20%, 41 ± 5%, 53 ± 6%, 74 ± 12%, and 130 ± 9%, respectively. Quantitation of perylene-d12 was not possible given an unknown coeluting peak; thus, the recovery data were discarded. For those target analytes whose levels in procedural blanks exceeded the lowest calibration level, the average concentration in procedural blanks in a 0.5 ml volume was divided by the average dry sample weight (24, 25, and 15 g for n-alkanes, LABs, and PAHs, respectively) and defined as its reporting limit 22. Otherwise, the lowest calibration concentration in a 0.5 ml volume divided by the average dry sample weight was defined as the reporting limit (Supplemental Data, Table S1).
In the present study, ΣLAB is the sum of 19 LAB congeners (Supplemental Data, Table S1) unless otherwise noted. Among the 28 measured PAH compounds, naphthalene was frequently detected in procedural blanks and 9,10-diphenylanthracene was absent in nearly all samples; they were therefore excluded from the target list. The sum of the remaining 26 PAH compounds was designated as Σ26PAH. Σn-C15-35 is the sum of n-alkanes with carbon numbers of 15 to 35. Concentration data were presented on a sample dry weight basis. The quantification of any unresolved complex mixture (UCM) present in a sample was based on the UCM area and the average response factor of the individual n-alkane compounds with retention times within the UCM 23. All data were blank corrected but not surrogate recovery corrected. The nonparametric Mann–Whitney U and Kruskal–Wallis H tests were used to determine differences among data sets. Parameters concerning anthropogenic activities in the PRD and East and West Guangdong Province were summarized and correlated with the sediment concentrations of geochemical markers in the PRE and East and West Guangdong (three data points) to reveal factors affecting the pollution issues in the coastal areas of South China, and the actual significance level was reported. All statistical analysis was done with SPSS 13.0. Published data conducted in the coastal sediments of the PRD were combined with those obtained in the present study to interpret the potential differences of pollution levels and issues between the economically more developed PRD region and the less developed eastern and western regions.
RESULTS AND DISCUSSION
Occurrence and spatial distribution
Total concentrations of Σn-C15-35, ΣLAB, and Σ26PAH in coastal sediments off South China (Fig. 2) were in the ranges of 110 to 3,160, 11 to 162, and 26 to 600 ng/g, with median values of 730, 40, and 230 ng/g, respectively. The UCM concentrations, occurring in most samples, ranged from 2 to 205 µg/g (Supplemental Data, Table S2). The n-alkane concentrations in eastern and western costal sediments off South China were lower than those in Jiaozhou Bay and the Bohai Sea in North China and the Okinawa Trough in the East China Sea and were comparable to those in the Yangtze River Estuary in East China, the East China Sea, and the PRE and NSCS in South China (Supplemental Data; Table S3).
The sedimentary concentrations of ΣLAB in the eastern and western coast of South China were much lower than those in river sediments of Zhujiang and Dongjiangin, South China, Chaohu Lake, coastal sediments off South and Southeast Asia, Santa Monica Bay, California, USA; comparable to those in the Xijiang River, South China and São Sebastião, Brazil; and greater than those in the PRE and NSCS, as well as Santos, Brazil (Supplemental Data, Table S4). This comparison of sedimentary ΣLAB concentrations with those from around the world suggests that sewage pollution of the eastern and western coastal regions off South China was low to moderate.
The Σ26PAH concentrations in sediments from the eastern and western coast off South China were lower than those in Zhelin Bay, Daya Bay, the PRE, and the NSCS in South China, Jiaozhou Bay in North China, and the Yangtze River Estuary and the Minjiang Estuary in East China, but the concentrations were higher than those in Liaodong Bay, Bohai Bay, Laizhou Bay, and the Yellow River Estuary in North China (Supplemental Data, Table S5).
The sedimentary concentrations of Σ26PAH and Σn-C15-35 were statistically higher on the eastern coast (54–605 and 256–3,320 ng/g with median values of 257 and 900 ng/g, respectively) than on the western coast (26–178 and 130–817 ng/g with median values of 106 and 720 ng/g, respectively). Generally, sediment ΣLAB levels on the eastern coast (11–162 ng/g with a median of 55 ng/g; Fig. 2) were not statistically different from those on the western coast (19–53 ng/g with a median of 34.8 ng/g). A comparison with published data (Supplemental Data, Tables S3–S5) obtained for sediments from the PRE and NSCS indicates that the eastern and western coastal sediments off South China were loaded with comparable levels of Σn-C15-35, relatively higher concentrations of ΣLAB (Fig. 3A), and relatively lower occurrences of Σ25PAH (Σ26PAH minus perylene) (Fig. 3B). Possible reasons for the spatial distribution of sedimentary Σn-C15-35, ΣLAB, and Σ26PAH will be discussed in the following sections.
Poor to moderate correlations among the concentrations of Σn-C15-35, ΣLAB, and Σ26PAH (r2 = 0.31–0.48) suggest different sources of these three geochemical markers, which allow one to evaluate regional pollution issues from multiple perspectives.
The isoprenoid hydrocarbons pristane and phytane have been suggested as indicators of petroleum contamination 24 due to their common presence in crude oil. High Pr/n-C17 (0.85–4.76, mean 2.04; Supplemental Data, Table S2) and Ph/n-C18 (0.39–5.86, mean 1.81; Supplemental Data, Table S2) values suggest degraded oil, because isoprenoids are less susceptible to degradation than straight-chain alkanes 23. However, the use of Pr/n-C17 and Ph/n-C18 values might be somewhat questionable, because pristane and phytane could also come from biogenic sources such as zooplankton and bacteria 25. Low Pr/Ph values (0.2–0.6) were observed in sediments from Sfax, Tunisia 26, where petroleum hydrocarbons occur widely. In contrast, high Pr/Ph values (>3) suggest terrigenous organic matter 27. Therefore, the high Pr/Ph values (1.00–5.45 with an average of 2.20; Supplemental Data, Table S2) in the present study suggest biogenic sources for these two compounds.
In addition, Σn-C15-20/Σn-C21-35 (L/H) values < 1, approximately 1, and > 2 indicate sources of n-alkanes from higher plants, or bacteria, or both; from petroleum (and plankton); and from fresh oil, respectively 28, 29. Aliphatic hydrocarbons in the present study were of mainly biogenic origin, as suggested by low L/H (0.04–0.54 with a mean of 0.21) and high carbon preference index (CPI26-30) values (1.55–3.44 with a mean of 2.70; Supplemental Data, Table S2). Both values are different from those in Jiaozhou Bay, Qingdao (L/H > 3; CPI26–30 ∼1), where petroleum contamination was obvious 30. The CPI15-34 values (mean ± standard deviation: 2.54 ± 0.49; range: 1.4–3.48) in the present study were also higher than those in riverine runoff samples of the PRD (1.10 ± 0.34) 10, and urban aerosols of Guangzhou (1.03–1.16) 22, which were attributed to anthropogenic vehicular emissions. The compositional profiles of resolved n-alkanes (Fig. 4) exhibited an apparent odd–even preference, that is, highest relative concentration at n-C31. The CPI24-34 values (1.7–4.3; Supplemental Data, Table S2) in the present study were greater than 1 but lower than those of typical terrestrial higher plants (5–10) 31, suggesting additional contributions from other sources. In the present study, the addition of biogenic hydrocarbons from submerged or floating macrophytes to terrestrial plants was suggested by the aquatic macrophyte n-alkane proxy ((n-C23+n-C25)/(n-C23+n-C25+n-C29+n-C31)) values (0.16–0.26; Supplemental Data, Table S2). The CPI15-20 values were greater than 1 in approximately half of the samples (Supplemental Data, Table S2), suggesting biogenic sources, either from nonsiliceous plankton or a mixture of bacteria and plankton. Poor correlation (r2 = 0.28; excluding samples with markedly high and nondetectable UCM levels) between the concentrations of UCM and Σn-C15-35 suggested different sources of UCM and resolved n-alkanes. Taken together, these results all suggest that resolved aliphatic hydrocarbons under investigation in the present study were mainly derived biogenically, rather than anthropogenically. Although the presence of UCM in sediments has widely been taken as an indicator of petroleum contamination, its petroleum source could not be solidly confirmed due to the lack of specific biomarkers, such as hopane 32.
Biogenic hydrocarbons are not quite relevant for anthropogenic activities. Therefore, n-alkane concentrations in coastal sediments from the less developed eastern and western areas might not differ much from those in the PRE coastal area of the highly developed PRD region. This hypothesis is consistent with the observation of comparable sedimentary n-alkane levels between the PRE and the two coastal areas (Supplemental Data, Table S3).
Assessment of sewage pollution
The internal (I) isomers of LAB (i.e., those in which the phenyl group is in the middle of the alkyl chain) were demonstrated to be more recalcitrant to biodegradation than the external (E) ones 5, 33. The I/E ratio (6-C12LAB + 5-C12LAB)/(4-C12LAB + 3-C12LAB + 2-C12LAB) increased during LAB degradation under aerobic conditions 34. Thus, this ratio has been proposed 34 and applied 34–36 as an indicator of the extent of LAB degradation. Sedimentary I/E ratios in the present study were 0.54–1.21, similar to those in shampoo (0.5–1.6) 11, and in wastewater (0.8–1.4) 37. This possibly suggests that poorly treated municipal wastewater has been discharged into both the PRD and the eastern and western coastal areas. The sediment I/E ratios in the eastern and western coastal areas were slightly lower than those in the PRE (0.6–1.5) 15, which might suggest relatively greater discharge of untreated wastewater in eastern and western coastal areas than in the PRD regions 37.
Based on the results of a previous biodegradation experiment 34, I/E ratios are expected to increase with increasing distance from input sources. However, a seaward decreasing trend of I/E ratios was observed in Boston Harbor, Massachusetts, USA 38 and South China 15. The inconsistency between field observations and in-house biodegradation experiments suggests that the I/E ratio may not always be suitable for assessing the degradation of LABs under field conditions 38, or that the assessment of LAB degradation by using only the I/E ratio may be questionable 11. Another ratio, L/S ((5-C13LAB + 5-C12LAB)/(5-C11LAB +5-C10LAB)), was found to increase from sites adjacent to Portland Harbor, Maine, USA (0.69–0.82) to stations further away from the coast (1.6–4.0), and was therefore proposed as an indicator of LAB degradation 38. However, the L/S ratio varied among different commercial detergents, limiting the application of this ratio for assessing LAB degradation 11, 38. Luo et al. 15 observed a decrease in the relative abundance of the C12 homolog in sediment samples in the PRE and NSCS and proposed another ratio, C13/C12 ((6-, 5-, 4-, 3-, and 2-C13LAB)/(6-, 5-, 4-, 3-, and 2-C12LAB)) 15, as a more sensitive indicator.
In the present study, there was an apparent seaward decrease in the C13/C12 value (Supplemental Data, Fig. S1) but not the L/S value (Supplemental Data, Fig. S2). The C13/C12 values in suspended particles from wastewater influent (0.9–1.7; mean 1.3) 33 were similar to those from wastewater effluent (1.1–1.8, mean 1.3) 33, suggesting that municipal wastewater treatment did not change C13/C12 values. Therefore, the spatial distribution of C13/C12 ratios (Supplemental Data, Fig. S1) could not be caused solely by LAB degradation. The seaward decreasing trend of C13/C12 was consistent with the preferential vertical scavenging of the C13 homolog with increasing distance from the source, given its greater hydrophobicity and therefore higher partitioning into suspended particles compared with the C12 homolog 39.
Coastal marine sediments in eastern (and western) Guangdong should be more seriously affected by sewage pollution than those of the PRE and NSCS, coastal areas of the PRD, because the infrastructure is presumably less advanced in the less developed eastern (and western) Guangdong province. Indeed, sewage pollution in the eastern and western coastal areas off South China was more severe than that in the coastal areas of the PRD region (Fig. 3a). Although the PRD region is more populated (Supplemental Data, Fig. S3) 40 and consequently is expected to discharge larger amounts of domestic wastewater than the eastern and western regions, the much larger number of sewage treatment plants built and operated in the PRD region (Supplemental Data, Fig. S4; www.dowater.com) has probably reduced the amounts of LABs discharged with treated wastewater. Negative correlation between the per capita wastewater treatment capacity and LAB concentrations (Fig. 5) confirmed the hypothesis that economically developed PRD areas suffered less sewage pollution than the less developed eastern and western regions of Guangdong.
Perylene concentrations ranged from 5 to 260 ng/g, with a mean value of 70 ng/g. Principal component analysis, as detailed in the Supplemental Data, suggested that perylene was well correlated with the first principal component, which was heavily weighted with vehicular exhaust markers (IcdP and BghiP) 41, high molecular weight n-alkanes (n-C19–n-C35) and UCM, as well as component 2, which was highly correlated with parent PAH compounds (Supplemental Data, Table S6). The widespread occurrence of perylene in river and marine sediments 42–45, and even unpolluted sediments 46, was thought to stem from early diagenesis of aquatic precursors such as diatoms 47, 48 and terrestrial precursors such as photosynthetic pigments 49. Diagenetic perylene is normally most abundant in fine, anoxic sediments. Therefore, the prevailing anaerobic conditions for the same sampling areas in the present study 22 might suggest a diagenetic source of perylene. The seaward decline of perylene abundance in both sediments 50 and suspended particulate matter 51 from the PRE to the NSCS suggested that perylene was formed in situ in river sediments and was transported to marine sediments in association with terrestrially combustion-derived PAHs 50. In summary, the loading of perylene in the first factor with vehicular exhaust tracers (IcdP and BghiP) 41 and with high molecular weight n-alkanes (n-C19–n-C35), and the loading of perylene in the second factor with parent PAHs, may suggest that perylene in the present study was likely to be derived from diagenesis of terrestrial precursors under terrestrial environments and then underwent similar transport processes as anthropogenic PAHs.
Anthropogenic polycyclic aromatic hydrocarbons
The often cited indexes for the diagnostics of PAHs origins include An/(An + Phe), BaA/(BaA + Chr), Flu/(Flu + Pyr), IcdP/(IcdP + BghiP), and MP/P (sum of methylated phenanthrenes to phenanthrene) 52. An/(An + Phe) values < 0.1 and > 0.1 and BaA/(BaA + Chr) values < 0.2 and > 0.35 indicate petrogenic and pyrogenic sources, respectively 52. Flu/(Flu + Pyr) values > 0.5, between 0.4 and 0.5, and < 0.4 indicate coal or wood combustion (or both), petroleum combustion, and petroleum source of PAHs, respectively, as do IcdP/(IcdP + BghiP) ratios > 0.5, between 0.2 and 0.5, and < 0.2, respectively 52. The plots of An/(An + Phe) versus BaA/(BaA + Chr) (Supplemental Data, Fig. S5A) and Flu/(Flu + Pyr) versus IcdP/(IcdP + BghiP) (Supplemental Data, Fig. S5B) suggested mixed combustion sources for sedimentary PAHs in the present study. MP/P values for combustion-derived PAHs were typically < 1 and ranged from 2 to 6 in petroleum residues 16. The MP/P values under investigation were less than 1 (0.3–1.0), also suggesting the predominance of combustion-derived PAHs.
Higher population density and greater industrialization in the PRD region should have resulted in more serious pollution by combustion-derived PAHs in the PRD than in the eastern and western areas of the Guangdong Province. On the one hand, the number of registered vehicles in eastern Guangdong is greater than that in western Guangdong 40, possibly resulting in higher PAH emissions from petroleum combustion or unburned residues (or both) in eastern Guangdong. This is consistent with observations of generally higher UCM, a tracer of aged petroleum residues or degraded hydrocarbons, or both 25, in the eastern (<nd–205, mean 37 µg/g) compared with the western (<nd–22, mean 9 µg/g) coastal sediments. Furthermore, the greater population density in Guangdong's eastern cities, compared with the western ones 40, is probably an important cause for the statistically higher Σ25PAH (Σ26PAH minus perylene) concentrations in the eastern (38–340, mean 158 ng/g) than in the western (20–138, mean 83 ng/g) coastal sediments (Fig. 3B). This relation was previously demonstrated elsewhere: PAH concentrations in 11 inland lakes in Michigan, USA, were positively correlated with watershed population density 53. A positive correlation between the regionally average population density and PAH concentration (Fig. 6) was observed. Likewise, sedimentary pollution from anthropogenic activities (petroleum residues and combustion of fossil fuels) was also much greater in the coastal area of the economically developed PRD than in the less developed eastern and western Guangdong coasts (Fig. 3B).
The above application of n-alkanes, LABs, and PAHs in assessing the anthropogenic impacts on coastal marine sediments off South China, a dynamic coastal zone, indicated that the scopes and temporal trends of organic pollution were quite reflective of the extents of economic development, energy consumption, and wastewater treatment. This has somewhat validated the utility of multiple geochemical markers for identifying the key factors in dictating the patterns of organic pollution in rapidly evolving coastal regions. The present study has also pointed to the need for implementing rigorous but practical management measures to protect precious coastal resources in regions undergoing intensive economic growth and urbanization.
Figures S1–S5. (1.9 MB DOC).
The present study was financially supported by the National Natural Science Foundation of China (grant 41121063), the Natural Science Foundation of Guangdong Province (grant 9251064004000002), the Guangzhou Institute of Geochemistry, Chinese Academy of Sciences (GIGCAS 135 project Y234081001), and the Canada Research Chairs Program and the Natural Sciences and Engineering Research Council of Canada (to C.S. Wong). This is contribution IS-1576 from GIGCAS.