Fecal steroids in riverine runoff of the Pearl River Delta, South China: Levels, potential sources and inputs to the coastal ocean

Abstract

Domestic sewage is a potential source of contamination deteriorating water quality in rivers and coastal environments. The present study determined the concentrations of eight steroids in both filtrate and suspended particulate matter (SPM) samples collected monthly at the eight major riverine runoff outlets of the Pearl River Delta (PRD), South China, from March 2005 to February 2006. The concentration of the eight steroids (sum of which is defined as Σ₈steroid) in the filtrate and SPM samples ranged 16.7 – 1340 ng L⁻¹ and 0.44 – 240 µg g⁻¹, respectively. The filtrate samples contained the highest levels of cholesterol (CHOE) and stigmasterol (STIG) on average, whereas the SPM samples had the largest relative abundances of CHOE and coprostanol (COP). In general, larger proportions of the steroids were associated with SPM than with the filtrate phase. The combined (filtrate plus SPM) concentrations of COP in riverine runoff of the PRD were at the midpoint of the global range in surface water. The concentrations of Σ₈steroid and COP in SPM were not significantly correlated with particulate organic carbon (POC), riverine runoff, or SPM concentrations, which indicates the dominance of nonpoint input sources for the occurrence of steroids in the PRD. The annual riverine inputs of Σ₈steroid and COP were estimated at 360 and 31.6 tons yr⁻¹, respectively, and the monthly inputs of both Σ₈steroid and COP were positively correlated with monthly runoff discharge. Furthermore, the concentrations of Σ₈steroid, COP, and CHOE showed significant temporal and spatial variability. These results can be explained by a combination of climatic characteristics, socioeconomic conditions, and steroid sources. Analyses of the diagnostic indices of COP and background information suggested that riverine runoff from the PRD may have been affected by domestic wastewater, which appeared to be a major COP source to the Pearl River Estuary and South China Sea.

---

Environmental impact

Domestic sewage discharge has been suspected as the main source of anthropogenic contamination in China, particularly in regions with rich river systems. Eight fecal steroids of possibly human origin were measured in riverine runoff samples collected monthly from March 2005 to February 2006 at the eight major outlets in the Pearl River Delta of South China. A global comparison indicates that anthropogenic pollution in waters of the PRD, as characterized by the levels of coprostanol (COP), is moderate, but the main source of COP in the coastal waters of South China appeared to be raw sewage discharged via riverine runoff. Therefore, better wastewater treatment should be adopted to improve the water quality in the PRD.

Introduction

Domestic wastewater is one of the most important sources of environmental pollution. Contaminated waters are usually considered as a viable threat to both human health and the integrity of aquatic ecosystems, because they are likely to contain human-specific enteric pathogens.¹ Therefore, assessment of wastewater for possible human fecal contamination in rivers is a critical step towards water quality improvement and risk reduction of infecting diseases. Coprostanol (COP), which is a fecal sterol produced in the digestive tract of humans by microbial biohydrogenation of cholesterol (CHOE),² has been identified as a biomarker of human-derived sewage contamination due to its persistency and scarcity in the natural environment.^3–6 In fact, high levels of COP have been determined in human fecal materials,⁷ domestic waste water^8,9 and sediments contaminated with sewage.^7,10,11 Previous studies indicated that fecal sterol profiles varied significantly with different animal species.^3,12 For example, Shah et al.¹² investigated fecal sterols from humans, herbivores, carnivores, and birds, and found that the highest COP concentrations were found in human-derived fecal samples. Therefore, COP-dominated sterol profiles may be indicative of contamination of human origin.

The Pearl River Delta (PRD) is one of the most populated and economically fastest growing regions in China. Rapid economic development during the past 30 years in the PRD has resulted in accelerated urbanization and a substantial increase in population. On the other hand, existing wastewater treatment capacities have not been able to process the ever-increasing amount of domestic sewage. Statistical data showed that more than 50% of municipal waster generated in 2006 was discharged directly into rivers and streams in the PRD,¹³ and almost all domestic sewage in rural areas has been excreted without any treatment. Previous studies indicated that sediments from the Pearl River Estuary (PRE) and northern South China Sea (SCS) were significantly contaminated by fecal materials possibly originated from direct discharge of untreated wastewater locally, and the COP concentrations decreased with increasing distance from the PRD toward the open sea.¹⁴ These results indicate that domestic sewage discharge in the PRD is an important source of fecal contamination in waters of the PRE and SCS.

The present study was intended to assess the state of water quality in riverine runoff of the PRD with respect to human contamination. To accomplish this goal, samples of riverine runoff were collected monthly from March 2005 to February 2006 at the major eight outlets (to the coastal ocean), i.e., Humen (HM); Jiaomen (JM); Hongqilimen (HQ); Hengmen (HE); Modaomen (MD); Jitimen (JT); Hutiao (HT); and Yamen (YM) (Fig. 1), and analyzed for eight sterols. Levels and mass loadings of the steroids in the runoff samples were estimated, and their temporal and spatial variability in both the aqueous phase and suspended particulate matter (SPM) were examined to derive source information.

Fig. 1 Map showing the sampling sites and surrounding areas in the Pearl River Delta, China. HM–Humen, JM–Jiaomen, HQ–Hongqilimen, HE–Hengmen, MD–Modaomen, JT–Jitimen, HT–Hutiaomen, and YM–Yamen.

Methods

Sample collection and extraction

The sampling sites and procedures have been described in detail in a previous study.¹⁵ Briefly, 96 riverine runoff samples were collected at the eight runoff outlets (Fig. 1) from March 2005 to February 2006.¹⁶ Another one liter of runoff sample from each subsampling point was collected to measure the contents of suspended particulate matter (SPM), particulate organic carbon (POC), and dissolved organic carbon (DOC).¹⁷ The samples were cooled with ice during transport to the laboratory where they were processed immediately.

Water samples were filtered with a vermicular system, and SPM was retained by GF/F glass fiber filters (Whatman International Ltd., Maidstone, England) that were baked at 450 °C for at least 4 h prior to use. Filtrates were passed through a mixture of XAD-2:XAD-4 resin (1:1 in mass) in a glass column (25 mm i.d. and 400 mm length) to collect “dissolved-phase” organics. The glass column was washed three times with 50 mL of methanol each, followed by three consecutive extractions with 50 mL of a methylene chloride:methanol mixture (1:1 in volume) in an ultrasonic bath. The extracts were combined, spiked with a known amount of the surrogate standard (5-pregan-3-ol), and back extracted five times with 50 mL of methylene chloride each upon addition of 500 mL distilled water and an appropriate amount of NaCl. The organic fraction was further extracted with 10 mL of distilled water three times for removal of residual methanol. The filters loaded with SPM were freeze-dried, weighed, and Soxhlet-extracted for 72 h with 400 mL of a mixture of methylene chloride:acetone (1:1) upon addition of surrogate standards. After removal of the organic solvent layer, the sample was subject to further extraction for 36 h with 200 mL of methanol.

For filtrate samples, all extracts were concentrated, solvent-exchanged to hexane, and further reduced to 1 mL with a Zymark TurboVap 500 (Zymark Corporation, Hopkinton, MA, USA). For SPM samples, two fractions were condensed separately. They were fractionated on a glass column packed with alumina:silica gel (1:2, v:v). The fraction containing steroids was eluated with 100 mL of a mixture of methylene chloride:methanol (1:3 in volume) after other fractions (containing n-alkanes, polycyclic aromatic hydrocarbon, and organochlorine pesticides) were collected. The solvent was reduced by rotary evaporation to dryness and was taken up with methylene chlorine into a glass vial (about 2 mL volume). The extract was concentrated to dryness under gentle nitrogen flow and a certain amount of internal standard (cholesterol-d₆) and 0.1 mL of a derivatization reagent (N,O-Bis(trimethylsilyl)trifluoro acetamide, BSTFA) were added to each extract to derivatize the sterols to corresponding trimethylsilylethers under 60 °C for 2 h. Prior to instrumental analysis, the solvent volume was brought to 0.5 mL with hexane.

Instrumental analysis and quantification

A Hewlett-Packard 5890 series II gas chromatograph (GC) equipped with a mass spectrometry (HP 5973 MSD) operating in the electron impact and selective ion monitoring modes with an electron multiplier voltage of 1588 V was used. The separation of target analytes was conducted by a 30 m × 0.25 mm-i.d. (0.25 µm film thickness) DB-5MS column (J&W Scientific, Folsom, CA, USA). Column temperature was programmed for 80 °C to 250 °C at a rate of 20 °C min⁻¹ (held for 10 min), followed by a ramp to 290 °C at a rate of 5 °C min⁻¹ (held for 10 min). Injector temperature was set as 280 °C. Extract injection was conducted in the splitless mode. The target analytes, surrogate standard (5-pregan-3-ol), and internal standard (cholesterol-d₆) were monitored at m/z 370, 368, 386, 445, 231, 394, 473, 271, and 374 (Table S1).‡

Measurement of organic carbon

One liter of water sample was filtered immediately to collect SPM with GF/F glass fiber filters. The filters loaded with SPM were wrapped with aluminium foil and stored in plastic bags at −20 °C until analysis. A portion of filtrates for each sample was collected for DOC measurements within 48 h. The filtrate and filter samples were treated with 1 mol L⁻¹ HCl to remove carbonate residues prior to organic carbon measurement. POC was analyzed with an elemental analyzer (Elementar, Vario, EL III, Germany), and the filtrate samples were analyzed for DOC with a total organic carbon analyzer (Shimadzu TOC-VCPH, Kyoto, Japan). The blank values were subtracted from all measured samples values. The POC and DOC data were expressed as percent weight of total SPM (%) and normalized to total filtrate volumes (mg L⁻¹), respectively.

Quality assurance and quality control

A standard mixture of coprostanol, epicoprostanol, β-cholestanone, cholesterol, cholestanol, α-cholestanone, stigmasterol, and stigmastanol was obtained from AccuStandard (New Haven, CT, USA) for qualitative and quantitative analyses. Quantification was performed using the internal calibration method based on seven-point calibration for individual compounds (0.1, 0.2, 0.5 1, 2, 4, and 8 µg mL⁻¹). The linear regression coefficients for all target analytes and surrogate standard calibration curves were better than 0.99. The lowest concentration of the calibration standards was 0.1 µg mL⁻¹, corresponding to minimum quantitation limits of 1.3 ng L⁻¹ and 0.03 µg g⁻¹ (defined as reporting limits;^15,18 RLs, Table S1)‡ for filtrate sample (40 L) and average SPM dry weight of 1.9 g, respectively.

Results and discussion

Occurrence of steroids in riverine runoff of the Pearl River Delta

Concentrations of total eight steroids (defined as Σ₈steroid) in filtrates and SPM ranged from 16.7–1340 ng L⁻¹ (with an arithmetic average of 164 ng L⁻¹) and 0.44–240 µg g⁻¹ (with an arithmetic average of 31 µg g⁻¹), respectively (Fig. S1).‡ The total concentration of Σ₈steroid (filtrates SPM combined) varied from 199 to 4050 ng L⁻¹ with a mean value of 1120 ng L⁻¹ (the concentration unit in SPM was converted to ng L⁻¹). In the filtrate samples, CHOE and STIG were the most abundant components with average concentrations ± standard deviations of 79 ± 78 ng L⁻¹ and 46 ± 49 ng L⁻¹, respectively. In addition, concentrations of COP were lower than the RLs in 71% of the filtrate samples with a mean value of 1.9 ± 4.4 ng L⁻¹. ECOP, CHOA, bONE, aONE, and STAN were not detected or lower than RLs in more than 78% of the filtrate samples (Table 1). In the SPM samples, CHOE and COP were ubiquitous, and STIG was widely distributed (in 89 out of 96 samples). ECOP, bONE, aONE and CHOA were detected in more than 50% of the samples. However, STAN was not detectable in 83 samples. The most predominant steroids were CHOE, STIG and COP in the particulate phase with concentration levels of 14.7 ± 14.6, 8.4 ± 8.5 and 3.1 ± 5.0 µg g⁻¹, respectively. On the other hand, most steroids were associated with SPM due to their hydrophobicity (e.g., log K_ow of CHOE ∼7.28).¹⁹ For example, 83 ± 15% and 97 ± 7% of Σ₈steroid and COP were present in SPM (Fig. 2). The concentrations of Σ₈steroid and COP were not significantly correlated with POC (Fig. S2(a)),‡ riverine runoff (Fig. S2(b)),‡ and SPM contents (Fig. S2(c)).‡ This indicates that nonpoint input sources of fecal steroids are dominant in the watershed of the PRD.

Fig. 2 Phase distribution of individual steroids in riverine runoff of the Pearl River Delta. Solid bars and shaded bars represent suspended particulate matter (pore size > 0.7 µm) and filtrates (pore size < 0.7 µm), respectively. COP–coprostanol, ECOP–epicoprostanol, bONE–β-cholestanone, CHOE–cholesterol, CHOA–cholestanol, aONE–α-cholestanone, STIG–stigmasterol and STAN–stigmastanol (Tabel S1).‡

Comparison of fecal pollution levels in the Pearl River Delta and around the world

To compare the fecal pollution levels in waters of the PRD to those in other water bodies around the world, the highest concentrations of COP in both filtrates and SPM combined were employed because most previous studies provide only highest and lowest concentrations. In fact, a similar comparison has been conducted by Takada and Eganhouse.²⁰ In the present study, our data are compared with the date summarized by Takada and Eganhouse and recently found in the literature (Fig. 3) with details described in the Supplementary Information (Table S2).‡ Briefly, the maximum COP levels in the PRD were 3–4 times higher than those (from non-detectable to 26 ng L⁻¹) in lower Santa Ana River, southern California, USA,²¹ and were also significantly higher than the values found in the upper Wilson River, Australia,²² surface water of North America,²³ Tuggerah Lake of Australia,²⁴ seawater from Barcelona coastal of Spain,²⁵ and Scheldt estuary of Belguim.²⁶ The fecal pollution levels of the PRD were similar to those in some rivers of Vietnam,¹⁹ Solo River of Indonesia,²⁷ and Ebro River of Spain,²⁵ but the maximum value found in the present study was 1–2 orders of magnitude lower than the maximum concentrations reported for some estuaries, such as the Clyde estuary,²⁸ the Mersey estuary,²⁹ and the First of Clyde U.K.,²⁸ the Morlaix estuary, France,³⁰ the Fraser river estuary, Canada,³¹ the Krka River estuary, Croatia,³² and the Derwent estuary, Australia.³³ Furthermore, the levels are also lower than the levels in northern Italy,⁹ western Malaysia,¹⁹ and some rivers including Matsuyama, Tamagawa river, Minamiasakawa river, Fukuoka and Tsurumigawa River in Japan,²⁰ Ohio,³⁴ Narragansett,³⁵ and Great Lakes³⁶ of USA, Quebec of Canada, and grey water in northern Stockholm of Sweden.³⁷ On the other hand, COP concentrations in municipal sewage^19,25,38 and catch water²² were about 1 – 3 orders of magnitude higher than the levels obtained in the present study. The above comparison suggests that fecal pollution in riverine runoff of the PRD was at the midpoint of the global range found in surface water. Fig. 3 also indicates that the highest concentrations of COP were typically found in wastewater while relative low concentrations were measured in the freshwater and seawater, with the concentration levels following the sequence of raw wastewater > primary wastewater (activated sludge treatment) > septic tank water > trickling filter treated water ≥ catchment water≥ greywater > river ≥ estuary water≥ coastal water. This indicates that anthropogenic contributions were positively correlated to the concentrations of COP.

Fig. 3 Reported maximum concentration of coprostanol (µg L⁻¹) in different types of water around the world (detailed in Table S2).‡

Riverine inputs of steroids from the PRD to the coastal ocean

Monthly riverine inputs of individual steroids are computed by

F_i,j = (C_i,j,aq + C_i,j,SPM) × Q_i,j × 10⁻⁶ (1)

where F_i,j represents the flux of each steroid compound; C_i,j,aq and C_i,j,SPM (in ng L⁻¹) are the concentrations of each target compound in the aqueous and SPM phases, respectively; Q_i,j (in m³) is the runoff discharge; and the subscripts i and j indicate the ith outlet and the jth month, respectively. The total annual flux for the ith outlet was calculated by

The total riverine input of Σ₈steroid is estimated at 357 tons yr⁻¹ with CHOE constituting the largest proportion (162 tons yr⁻¹), followed by STIG (101 tons yr⁻¹) and COP (31.6 tons yr⁻¹) (Fig. 4). The riverine inputs of ECOP and STAN make up the smallest portions of the total, with only 4.8 and 5.9 tons yr⁻¹, respectively. The monthly fluxes of Σ₈steroid and COP were moderately correlated linearly with riverine runoff discharge (Fig. 5), with r² = 0.74 and 0.59. This suggests that riverine inputs of steroids from the PRD may have been controlled by hydrological conditions, which is similar to the findings associated with polycyclic aromatic hydrocarbons,¹⁵ polybrominated diphenyl ethers,¹⁸ and linear alkylbenzenes.³⁹

Fig. 4 Monthly (from March 2005 to February 2006) riverine inputs of individual steroids from the Pearl River Delta, China, to the coastal ocean. HM–Humen, JM–Jiaomen, HQ–Hongqilimen, HE–Hengmen, MD–Modaomen, JT–Jitimen, HT–Hutiaomen, and YM–Yamen. COP–coprostanol, ECOP–epicoprostanol, bONE–β-cholestanone, CHOE–cholesterol, CHOA–cholestanol, aONE–α-cholestanone, STIG–stigmasterol and STAN–stigmastanol (Table S1).‡

Fig. 5 Relationship between the monthly riverine inputs of (a) Σ₈steroid and (b) coprostanol (COP) and water flows (discharge amounts) from the eight runoff outlets of the Pearl River Delta, China.

Temporal and spatial variability of concentration levels and riverine inputs

An independent-samples t test (p < 0.01) indicated no significant temporal variability in the concentrations of Σ₈steroid, CHOE, and COP during the 12 month sampling period. However, the concentrations of STIG were considerably lower in the dry weather season (March, October–December 2005, and January–February 2006; 256 ± 156 ng L⁻¹) than in the wet weather season (from April to September 2005; 390 ± 329 ng L⁻¹). Conversely, the concentrations of Σ₈steroid, COP, and CHOE were significantly different spatially (p < 0.05), but the STIG concentration was not significantly different among the sampling sites (p < 0.05). In general, the concentrations of Σ₈steroid, COP, and CHOE were higher at the eastern four outlets (1251 ± 657, 111 ± 104, and 540 ± 308 ng L⁻¹, respectively) than at the western four outlets (990 ± 427, 60.8 ± 79.3, and 475 ± 224 ng L⁻¹, respectively). These results can be explained collectively by climatic characteristics, rates of social-economic development, and sources of steroids.

The PRD is located in a subtropical zone with annual precipitation of 1301–2279 mm, and more than 90% of the rainfalls occur during the wet weather season.⁴⁰ Moreover, the eastern region of the PRD has been better developed economically than the western region. It is known that STIG are mainly originated from plants⁴¹ and COP and CHOE are largely excreted from human beings and certain animals.⁴² In general, crops are grown in spring and summer as moisture and sunshine are abundant. At the same time, dense population in the eastern PRD region (e.g., Dongguan has a population density of ∼2700 residents km⁻¹ in 2005 and ranks second among all the cities of the Guangdong Province)⁴³ discharges huge amounts of fecal materials, and plentiful rainfall (1990 mm)⁴⁴ would wash off human or animal waste into the aquatic system by surface runoff and erosion. As a result, STIG concentrations are higher in the wet weather season than in the dry weather season with no apparent spatial variability, whereas the concentrations of COP and CHOE are higher in riverine runoff collected from the eastern outlets of the PRD than those from the western outlets but are not seasonally variable.

Seasonal and spatial trends for riverine inputs of individual steroids are mainly controlled by riverine runoff discharge as noted above. Large portions of Σ₈steroid, COP, CHOE, and STIG (85%, 84%, 85%, and 87%, respectively) were discharged via riverine runoff out of the PRD during the wet weather season, and were mostly discharged through the four eastern outlets (70%, 75%, 66%, and 71%). On the other hand, only 19%, 17%, 22% and 18% of the total riverine inputs of these steroids were discharged through Modaomen, a major runoff outlet in the western PRD region where runoff discharge is the largest among all the riverine runoff outlets.

Assessment of human-induced water pollution in the PRD using steroid compositional information

The occurrence of COP has often been used to indicate human-induced water pollution. There are three sets of COP-related indicators that have served the purpose. The first set is the plain COP concentration, because COP does not occur in natural or marine water and only small amounts of COP have been found in anaerobic sediments not previously contaminated by human fecal materials.³³ The second set of indicators includes organic carbon (OC) normalized COP concentration and the relative abundance of COP to total steroids (%COP).¹⁹ Because the COP concentration in sediment has been found to be dependent upon the OC content,^38,45 COP concentrations need to be normalized to OC to eliminate bias for evaluation of fecal pollution. Furthermore, %COP has been widely accepted as a useful diagnostic tool to assess fecal pollution because high %COP values were found to relate to large contributions from fecal materials of human origin.⁴⁶ The third set of indicators contains the ratios of some individual steroids to examine the relative contributions of different organism excreta.^4,12 In general, COP is microbially converted from CHOE in the human intestine and consequently is the predominant steroid in human feces.³³ Because, 5α sterols are thermodynamically more stable than their 5β epimers, the ratios of 5β to 5α epimers of stanols and stanones, i.e., COP/(COP + CHOA) and bONE/(aONE + bONE), can be used to indicate sewage pollution.⁶ Ratios of COP/CHOE and COP/ECOP are also often used as indicators of sewage pollution.^10,47

COP was found in all SPM samples from the present study with a total arithmetic average value of 87.5 ng L⁻¹ (filtrate and SPM samples combined), which exceeds the primary contact limit (60 ng L⁻¹) proposed by Leeming and Nichols.³³ Although herbivores may contribute COP to the environment,⁴ COP concentrations are much lower in herbivore feces than in human feces. In addition, the total number of herbivores is 26 million (including cattle, cow, goat and pig) in the PRD, less than the number (46 million) of registered residents.¹³ The OC normalized concentrations of COP from the present study ranged from 2 to 218 ng mg⁻¹ of OC with an arithmetic mean of 29 ng mg⁻¹ of OC, and %COP ranged from 0.6% to 59% with a mean of 8% (Fig. 6). These values are comparable to the values obtained from river and estuary sediments which had been known to be polluted by fecal matter.^11,19,48 Furthermore, COP concentrations were strongly correlated with OC-normalized COP concentrations and %COP (Fig. S3).‡ Therefore, all these indicators suggest that the aquatic system of the PRD is polluted by human-induced fecal matter.

Fig. 6 Compositional indices of steroids in the combined filtrate and particulate phases. COP/OC represents the coprostanol concentration normalized to organic carbon; %COP is the percent of COP concentration relative to the concentration of total steroids (eight steroids); COP/(COP + CHOA) is the ratio of the concentration of coprostanol versus the total concentration of coprostanol plus cholestanol; bONE/(aONE + bONE) is the ratio of the concentration of β-cholestanone versus the total concentration of α-cholestanone plus β-cholestanone; CHOE/COP is the ratio of the concentrations of coprostanol and cholesterol; and COP/ECOP is the ratio of the concentrations of coprostanol and epicoprostanol.

In the present study, both COP/(COP + CHOA) and bONE/(aONE + bONE) varied widely essentially in the range of 0–1 (Fig. 6). The values of COP/(COP + CHOA) in 40 out of 96 samples were higher than 0.7, whereas bONE/(aONE + bONE) in 49 of 96 samples were higher than 0.7. Grimalt et al. showed that a value of 5β/(5α + 5β) higher than 0.7 could indicate sewage pollution and less than 0.3 suggest no fecal contamination.²⁵ In addition, COP/ECOP could not be determined in 57% of the samples because ECOP was not detectable, but the concentrations of COP were about 12 times those of ECOP when ECOP was detectable. At the same time, CHOE was about 5 times more abundant than COP in all the samples. Therefore, sewage pollution is a key factor in influencing the water quality of the PRD.

A previous study determined the concentrations of COP in sediments of the Pearl River Estuary (PRE) and northern South China Sea,¹⁴ which directly receive riverine runoff from the eastern outlets sampled in the present study. The concentrations of COP in particulate samples normalized to OC from the present study were comparable to the levels found in the sediments of the PRE, where the concentrations of COP decreased with increasing distance from the PRE toward the open sea.¹⁴ However, the CHOE concentrations normalized to POC from the present study were significantly higher than those in sediment obtained from the PRE, resulting in higher values of COP/CHOE (0.25 ± 0.36) observed in the present study than those found in sediment of the PRE. A reasonable explanation is that CHOE is more easily dissolved in the aqueous phase and subsequently degraded than COP during sedimentation, as indicated by a lower log K_ow value for CHOE (∼7.28)¹⁹ than that for COP (∼8.82),¹⁴ leading to an increase of the COP/CHOE values from water to sediment. A previous study also suggested a higher degradation rate for CHOE than that for COP.⁴⁹ Therefore, it can be inferred that sewage transported by riverine runoff from the PRD was the main source of COP in the PRE and South China Sea.

Conclusions

Coprostanol was ubiquitous in riverine runoff from the PRD with concentrations at the midpoint of the global range. This, combined with the assessment of compositional indices of steroids, indicates that surface water in the PRD was considerably contaminated by human induced input sources. The total mass loadings of eight steroids were computed based on riverine runoff to estimate the potential contribution of Σ₈steroid to PRE and SCS. The annual riverine discharge of COP was approximately 32 tons yr⁻¹, the majority of which was discharged into the Pearl River Estuary and northern South China Sea. Finally, significant correlations were observed between the monthly fluxes of COP and Σ₈steroid and the corresponding monthly water discharge, indicating that hydrological conditions were an important factor controlling the mass loadings of steroids in riverine runoff from the PRD.

Acknowledgements

This research was financially supported by the National Natural Science Foundation of China (Nos. 40588001 and 40821003) and the Earmarked Fund of the State Key Laboratory of Organic Geochemistry (No. SKLOG2008A05). We thank Mr T. S. Xiang for assistance in the GC/MS analysis. The authors are also grateful to the sampling team consisted of mostly graduate students from the Guangzhou Institute of Geochemistry and Sun Yat-sen University for assistance in field work. This is contribution No. IS-1110 from GIGACS.

References

1. P. Tyagi, D. R. Edwards and M. S. Coyne, Chemosphere, 2007, 69, 1617–1624.
2. R. S. Rosenfield and T. F. Gallagher, Steroids, 1964, 4, 515–520.
3. R. Leeming, A. Ball, N. Ashbol and P. Nichols, Water Res., 1996, 30, 2893–2900.
4. V. G. Shah, R. Hugh Dunstan, P. M. Geary, P. Coombes, T. K. Roberts and T. Rothkirch, Water Res., 2007, 41, 3667–3674.
5. M. Nishimura, Geochim. Cosmochim. Acta, 1982, 46, 423–432.
6. M. Nishimura and T. Koyama, Geochim. Cosmochim. Acta, 1977, 41, 379–385.
7. A. D. Reeves and D. Patton, Phys. Chem. Earth, 2001, 26, 753–757.
8. M. M. Elhmmali, D. J. Roberts and R. P. Evershed, Environ. Sci. Technol., 2000, 34, 39–46.
9. G. Gilli, R. Rovere, D. Traversi, T. Schilirò and C. Pignata, J. Chromatogr., B: Anal. Technol. Biomed. Life Sci., 2006, 843, 120–124.
10. R. S. Carreira, A. L. R. Wagener and J. W. Readman, Estuar. Coast. Shelf Sci., 2004, 60, 587–598.
11. J. W. Readman, G. Fillmann, I. Tolosa, J. Bartocci and L. D. Mee, Mar. Pollut. Bull., 2005, 50, 310–318.
12. V. G. Shah, R. H. Dunstan, P. M. Geary, P. Coombes, T. K. Roberts and E. Von Nagy-Felsobuki, Water Res., 2007, 41, 3691–3700.
13. Statistical Bureau of Guangdong Province, Guangdong Statistical Yearbook, http://www.gdstats.gov.cn/tjnj/ml_c.htm, Accessed September 2007.
14. X.-Z. Peng, G. Zhang, B.-X. Mai, J.-F. Hu, K.-C. Li and Z.-D. Wang, Mar. Pollut. Bull., 2005, 50, 856–865.
15. J. Z. Wang, Y. F. Guan, H. G. Ni, X. L. Luo and E. Y. Zeng, Environ. Sci. Technol., 2007, 41, 5614–5619.
16. H.-G. Ni, F.-H. Lu, X.-L. Luo, H.-Y. Tian, J.-Z. Wang, Y.-F. Guan, S.-J. Chen, X.-J. Luo and E. Y. Zeng, Environ. Monit. Assess., 2008, 143, 291–301.
17. H.-G. Ni, F.-H. Lu, X.-L. Luo, H.-Y. Tian and E. Y. Zeng, Mar. Pollut. Bull., 2008, 56, 1150–1157.
18. Y. F. Guan, J. Z. Wang, H. G. Ni, X. L. Luo, B. X. Mai and E. Y. Zeng, Environ. Sci. Technol., 2007, 41, 6007–6013.
19. K. O. Isobe, M. Tarao, M. P. Zakaria, N. H. Chiem, L. Y. Minh and H. Takada, Environ. Sci. Technol., 2002, 36, 4497–4507.
20. H. Takada and B. P. Eganhouse, Molecular marker of anthropogenic waste, John Wiley & Sons, Inc., 1998.
21. J. A. Noblet, D. L. Young, E. Y. Zeng and S. Ensari, Environ. Sci. Technol., 2004, 38, 6002–6008.
22. V. G. Shah, R. Hugh Dunstan, P. M. Geary, P. Coombes, T. K. Roberts and T. Rothkirch, Water Res., 2007, 41, 3655–3666.
23. L. J. Standley, L. A. Kaplan and D. Smith, Environ. Sci. Technol., 2000, 34, 3124–3130.
24. R. Leeming, V. Latham, M. Rayner and P. Nichols, Detecting and distinguishing sources of sewage pollution in Austrialian inland and coastal water and sediments, Americian Chemical Society, Washingtong D.C., 1997.
25. J. Grimalt, P. Fernandez, J. M. Bayona and J. Albaiges, Environ. Sci. Technol., 1990, 24, 357–363.
26. M. Yde, E. De Wulf, S. De Maeyer-Cleempoel and D. Quaghebeur, Bull. Environ. Contam. Toxicol., 1982, 28, 129–134.
27. W. Li, J. Dagaut and A. Saliot, Biogeochemistry, 1995, 31, 139–154.
28. R. M. Goodfellow, J. Cardoso, G. Eglinton, J. P. Dawson and G. A. Best, Mar. Pollut. Bull., 1977, 8, 272–276.
29. M. F. Fitzsimons, M. K. Abdul Rashid, J. P. Riley and G. A. Wolff, Mar. Pollut. Bull., 1995, 30, 306–312.
30. M. Quémenéur and Y. Marty, Estuar. Coast. Shelf Sci., 1992, 34, 347–363.
31. L. M. Churchland, G. Kan and A. Ages, Can. J. Microbiol., 1982, 28, 239–247.
32. J. Laurcillard and A. Saliot, Mar. Chem., 1993, 43, 247–261.
33. R. Leeming and P. D. Nichols, Water Res., 1996, 30, 2997–3006.
34. J. J. Murtaugh and R. L. Bunch, J. Wat. Pollut. Control Fed., 1967, 39, 404–409.
35. L. A. LeBlanc, J. S. Latimer, J. T. Ellis and J. G. Quinn, Estuar. Coast. Shelf Sci., 1992, 34, 439–458.
36. B. J. Dutka, A. S. Y. Chau and J. Coburn, Water Res., 1974, 8, 1047–1055.
37. J. Ottoson and T. A. Stenström, Water Res., 2003, 37, 645–655.
38. K. A. Hughes and A. Thompson, Environ. Pollut., 2004, 127, 315–321.
39. H.-G. Ni, F.-H. Lu, J.-Z. Wang, Y.-F. Guan, X.-L. Luo and E. Y. Zeng, Environ. Pollut., 2008, 154, 348–355.
40. Zhong Shan Water Information Net, The Pearl River Delta, http://www.zswater.gov.cn/slbk_zssl/slzs/200503220059.htm, Accessed June, 2007.
41. W. De-Eknamkul and B. Potduang, Phytochemistry, 2003, 62, 389–398.
42. K. von Bergmann, T. Sudhop and D. Lütjohann, Am. J. Cardio., 2005, 96, 10–14.
43. Guangdong Statistical Yearbook, Principal aggregate indicators on national economic and social development and their related indices and growth rates, Guangzhou, 2005.
44. Guangdong Statistical Yearbook, Annual precipitation of various regions, 2005.
45. W.-L. Jeng and B.-C. Han, Estuar. Coast. Shelf Sci., 1996, 42, 727–735.
46. M. I. Venkatesan and I. R. Kaplan, Environ. Sci. Technol., 1990, 24, 208–214.
47. S. M. Mudge and C. E. Duce, Environ. Pollut., 2005, 136, 209–220.
48. J. H. Writer, J. A. B. Leenheer, B. Larry, G. L. Amy and S. C. Chapra, Water Res., 1995, 29, 1427–1436.
49. K. A. Sundin, R. L. Leeming and T. A. B. Stenström, Water Res., 1999, 33, 1975–1980.

---

Footnotes

† Part of a themed issue dealing with water and water related issues

‡ Electronic supplementary information (ESI) available: Tables S1 and S2 and Fig. S1–S3. See DOI: 10.1039/b909049p

---