Analysis of sediments from Shetland Island voes for polycyclic aromatic hydrocarbons, steranes and triterpanes

Abstract

A few days after the grounding of the oil tanker Braer on 5 January 1993, an Exclusion Zone was designated by Order under the Food and Environment Protection Act 1985, prohibiting the harvesting of farmed or wild shellfish within the Zone to prevent contaminated products reaching the market place. The order was progressively lifted for species that were found to be free of petrogenic taint and for which the polycyclic aromatic hydrocarbon (PAH) levels were within the range for reference samples. This Order, however, still remains in place for mussels (Mytilus edulis) as the PAH levels are higher than in reference mussels. To investigate the possible source of PAHs found in these mussels, sediments were collected from three reference and three Zone sites and their hydrocarbon compositions studied using the n-alkane composition and concentration, PAH composition and concentration and the sterane and triterpane composition. The reference site at Olna Firth was found to have the highest levels of 2–6-ring parent and branched PAHs, the highest concentration in one of the pooled sediments being 4530 ng g⁻¹ dry weight. Values in the other two reference sites (Vaila Sound and Mangaster Voe) ranged from 248.7 to 902.2 ng g⁻¹ dry weight. PAH concentrations at the Zone sites (Sandsound Voe, Stromness Voe and Punds Voe) ranged from 641.0 to 2766 ng g⁻¹ dry weight. The PAH data were normalised to the percentage of organic carbon and log-transformed prior to being analysed using principal component analysis. The mean total PAH concentrations for Zone sites were found not to be significantly different from the reference sites. The PAH concentration ratios were consistent with the main source of PAHs being pyrolysis. However, there was a petrogenic contribution, suggested by the presence of alkylated PAHs, with Punds Voe having the largest petrogenic hydrocarbon content. This was supported by the triterpane profiles and the presence of a UCM in the aliphatic chromatograms from Punds Voe sediments.

---

Introduction

The tanker MV Braer grounded on Garth Ness in south Shetland on 5 January 1993. Over the following 12 days, the entire cargo of 84 700 t of Norwegian Gulfaks light crude oil, together with a quantity of bunker fuel oil, was released from the ship. An Exclusion Zone was designated on 8 January 1993 by Order under the Food and Environment Protection Act 1985 (Fig. 1). The Order prohibited the harvesting of farmed or wild shellfish within the Zone to prevent contaminated products reaching the market place.¹ The restriction was lifted for fish and shellfish over the following 2.5 years under the criteria that the fish or shellfish contained no petrogenic taint and that the concentration of polycyclic aromatic hydrocarbons (PAHs) was within the range for reference fish and shellfish.² Since 9 February 1995 the Zone has remained effective for only Norway lobster (Nephropsnorvegicus) and blue mussels (Mytilus edulis). PAHs were originally selected for analysis owing to concern over their toxic and carcinogenic properties and because it was clear that contaminated and reference samples collected as a result of the Braer oil spill could easily be distinguished by their PAH profiles. Hydrophobic compounds such as PAHs typically become associated with particulate material and are deposited on the sediment. PAHs can come from two main sources: (1) petrogenic, including Gulfaks and other crude oils, bituminous deposits or petroleum products; and (2) pyrogenic, those formed in natural combustion processes, mainly forest fires, by the combustion of fossil fuels, coal and peat, from the incineration of agricultural, industrial and municipal wastes or from domestic wood burning giving rise to emissions, spillages or effluents. It is possible to distinguish between petrogenic and pyrogenic PAHs by studying a variety of PAH concentration ratios; a number of different indices have been developed to assess the different origins of these compounds.^3–8 Lower temperature generation of PAHs yields abundant alkyl-substituted compounds and thermodynamically favoured isomers, as found for petrogenic sources, whereas high temperature processes (pyrolytic sources) generate mainly parent compounds. Isomer ratios such as the phenanthrene/anthracene (P/A) and the fluoranthene/pyrene (Fl/Py) ratio can help identify pyrogenic sources. Comparison of alkylated PAHs with the parent compound, using the methylphenanthrene/phenanthrene (MP/P) and fluoranthene + pyrene/methylfluoranthene + methylpyrene (Fl + Py/MFl + MPy) indices, can be used to help identify petrogenic contamination (Table 1).

Fig. 1 Map of Shetland showing the FEPA Exclusion Zone, together with the location of the reference sites at Olna Firth (R1), Mangaster Voe (R2) and Vaila Sound (R3), and the sites within the Zone, Sandsound Voe (Z1), Stromness Voe (Z2) and Punds Voe (Z3). An expanded scale map is shown for each voe and in these maps the shaded area represents the land. Thirty-six samples (•) were collected at random from each voe.

Table 1 Typical PAH concentration ratios for pyrolytic and petrogenic origins: phenanthrene/anthracene (P/A); fluoranthene/pyrene (Fl/Py); sum of methylphenanthrenes/phenanthrene (MP/P); fluoranthene + pyrene/methylfluoranthene + methylpyrene (Fl + Py/MFl + MPy)

Origin	P/A	Fl/Py	MP/P	Fl + Py/MFl + MPy
Pyrolytic	<10	>1	<1	∼3
Petrogenic	>10	<1	>1	<3
Ref.	3–5	4, 5	5–7	8

---

If petrogenic contamination of a sediment is suspected, then it is useful to try to identify the source of the oil. The pentacyclic triterpanes and the tetracyclic steranes, also known as biomarker compounds, are not thought to be toxic but determination of these compounds in sediment can be useful in identification of spilled oils.^9–12 Triterpanes and steranes are derived from once living organisms. These compounds are more stable to biodegradation than n-alkanes, isoprenoids and PAHs and are therefore useful for the identification of oils. Hopanes (pentacyclic triterpanes) are derived from bacteriohopanetetrol found in lipid membranes of prokaryotic organisms. Hopanes with more than 30 carbon atoms are called homohopanes, and norhopanes (C₂₉ hopanes) are hopanes that have lost a methyl group. The biological configuration of homohopanes is [17β,21β,22R]. This configuration is thermodynamically unstable and hence diagenesis and catagenesis convert the biological 17β,21β configuration to the αβ configuration. Similarly, the 22R configuration will eventually convert to an equilibrium mixture of 22S and 22R diastereoisomers, as found in crude oils. This results in two peaks due to the 22S and 22R diastereoisomers in the m/z 191 mass chromatogram of all crude oils. The 22S/(22S + 22R) ratio in most crude oils is between 0.57 and 0.62.¹³

Steranes are derived from sterols in eukaryotic organisms and, unlike hopanes, are not found in unpolluted recent surface sediment and are therefore especially useful for identification purposes.¹⁴ Steranes are converted from sterols by diagenetic and catagenic changes and are abundant in mature sediments and petroleum. Highly weathered oils, such as Gulfaks, are often abundant in biomarker compounds and give a characteristic profile which can be used to identify oil contamination.

To investigate the source of the PAHs in sediments from Shetland voes, samples were collected from within and outwith the FEPA Exclusion Zone and analysed for PAHs, steranes and triterpanes (geochemical biomarkers) and other aliphatic hydrocarbons, specifically n-alkanes. PAH concentrations in sediment are related to the particle size and total organic carbon content; muddy sediments (high total organic carbon content) with a smaller particle size and a larger surface area to volume ratio will accumulate PAHs to a greater extent than the coarser, sandy sediments.^15,16 As such, these parameters were determined for each sediment prior to statistical analysis.

Experimental

Sediment sampling

For each site 36 random numbers, between 0 and 15, were generated. The vessel sampled the first station and then moved at 4 knots, over a pre-determined track. Samples were collected at the time intervals, in minutes, taken from the random number table. This resulted in sediment being collected from pseudo-randomly generated points (Fig. 1). The 36 sediment samples from each site were later pooled into six pools of six (see below). This sampling procedure was chosen based on previous data to ensure a 90% statistical power for detecting a 65% difference in total PAH concentration between reference and Zone sites.

The individual sediment samples were collected from five of the six voes from the FRV Clupea, using a 0.1 m² modified Day grab. The presence of a sill at the entrance to the sixth voe, Stromness Voe, meant that a smaller vessel, the Moder Di from the North Atlantic Fisheries College, was used. A hand held Van Veen grab was used for the collection of the 36 sediments from this voe. All sediments were collected in March 1997. Each sediment was mixed before transferring (∼200 g) to a solvent washed aluminium can which was labelled and stored at −20 °C until required for analysis. All sediments from one site were allowed to thaw at room temperature overnight and thoroughly mixed before pooling. Pooled samples were generated by random selection (Tables 2 and 4). Each pooled sample consisted of aliquots (25 ± 1 g) from six individual samples. Each pooled sample was treated as an individual sample for determination of particle size, total organic carbon and moisture content. The six pooled samples from the six sites were analysed for PAHs and n-alkanes while selected samples for were analysed for steranes and triterpanes. Selected individual sediments were analysed in triplicate for PAHs. The same individual samples were freeze-dried prior to the extraction and analysed in triplicate for PAHs.

The sediment particle size was determined, after freeze-drying, by laser granulometry employing a Malvern Mastersizer E particle size analyser. The total organic carbon content was determined on freeze-dried sediments, following acid treatment to remove carbonate, by combustion in a Perkin-Elmer (Beaconsfield, Bucks., UK) CHN elemental analyser.

Isolation of hydrocarbons from sediment

Each sediment sample was thoroughly mixed and an aliquot (approximately 10 g) removed for determination of water content by oven drying at 80 °C for 22 ± 2 h.¹⁷ To a second aliquot of sediment (10–20 g) were added the aliphatic hydrocarbon internal standards, heptamethylnonane and squalane, and the deuterated aromatic standards (naphthalene, biphenyl, dibenzothiophene, anthracene, pyrene and benzo[a]pyrene). The hydrocarbons were extracted using dichloromethane–methanol with sonication. The halogenated solvent was isolated and dried over Na₂SO₄. Solvent exchange to isohexane was performed and the concentrated isohexane solution fractionated by isocratic normal phase HPLC. Two fractions were collected and concentrated prior to chromatographic analysis. Procedural blanks were performed with each batch of sediment samples, and the hydrocarbon concentration was adjusted to take this in to account.¹⁷

Determination of polycyclic aromatic hydrocarbons by GC-MS

The concentration and composition of the PAHs were determined by GC-MS using an HP 6890 Series gas chromatograph interfaced with an HP 5973 mass-selective detector (MSD) and fitted with a cool on-column injector (Hewlett-Packard, Stockport, UK). A non-polar methylsilicone column was used for the analyses (Ultra 1, 25 m × 0.2 mm id, 0.33 µm film thickness; Hewlett-Packard). The carrier gas was helium set at a constant flow rate of 0.7 ml min⁻¹. Injections were made at 50 °C and the oven temperature held at this temperature for 3 min. Thereafter the temperature was raised at 20 °C min⁻¹ to 100 °C, followed by a slower ramp of 4 °C min⁻¹ to a final temperature of 270 °C. The MSD was set for selective ion monitoring (SIM) with a dwell time of 50 ms. A total of 25 ions plus the six internal standard ions were measured over the period of the analysis.^1,2 Thus the analysis incorporated 2–6-ring, parent and branched PAHs. The limit of detection, calculated as three times the standard deviation of the mean value from six procedural blanks, was found to be <0.1 ng g⁻¹ for benzo[b]fluoranthene and chrysene and less than 0.3 ng g⁻¹ for benzo[a]pyrene. Recoveries of >70% were achieved for sediments spiked to give a concentration of 1 ng g⁻¹ dry weight of individual PAHs. Good reproducibility was achieved for individual PAHs.¹⁸ Further quality control was assured through successful participation in the PAH programme of QUASIMEME (Quality Assurance of Information for Marine Environmental Monitoring in Europe).

Determination of aliphatic hydrocarbons (n-alkanes)

The n-alkane distribution was determined by GC-FID using a HP 5890 Series II gas chromatograph (Hewlett-Packard) equipped with an HP 7673 automated, cool on-column injector and fitted with a non-polar, Ultra 1 column (25 m × 0.2 mm id, film thickness 0.33 µm; Hewlett-Packard). The carrier gas was ECD grade nitrogen (16 psi). Injections were made at 60 °C and the oven temperature was held at this temperature for 3 min. Thereafter the temperature was raised at 4 °C min⁻¹ to 280 °C and held at this temperature until the end of the run. The detector was maintained at 300 °C throughout. Data were collected using a PE Nelson 600 series link box and processed using a Turbochrom 3 data station (Perkin-Elmer).

Determination of steranes and triterpanes

The sterane and triterpane composition was determined by GC-MS using an HP 5890 Series II gas chromatograph (Hewlett-Packard) interfaced with an Autospec high resolution mass spectrometer (Micromass UK, Manchester, UK) and fitted with a cool on-column injector. Perfluorokerosene was used as the calibrant. A methylsilicone (Ultra 1, 25 m × 0.2 mm id, 0.33 µm film thickness; Hewlett-Packard) capillary column was used for the separation. Injections were made at 60 °C and the oven temperature was held constant for 0.5 min, after which it was increased at 40 °C min⁻¹ to 150 °C, followed by a slower ramp of 5 °C min⁻¹ to a final temperature of 300 °C, this temperature being held for 22 min. Helium (15 psi) was used as the carrier gas. Biomarker analysis was carried out using the SIM mode. Triterpanes were monitored using m/z 177 and 191 and steranes using m/z 217 and 218 with a dwell time of 80 ms and a delay of 10 ms.

Results and discussion

Gulfaks crude oil

Gulfaks is a naturally degraded North Sea crude oil. The biodegradation has removed most of the n-alkanes, resulting in a heavier crude with a higher concentration of PAHs. The PAH profile is dominated by the C1–C4 naphthalenes with lesser amounts of alkylated phenanthrenes and anthracenes and relatively small amounts of the parent compounds. However, weathering of the oil with time will result in an increasing proportion of 4–6-ring compounds as the 2–3-ring compounds are more volatile and have a greater solubility in water.

Gulfaks crude oil contains the triterpane bisnorhopane (two fewer methyls than hopane), which is characteristic of North Sea oils, C₂₉ hopane and hopane. The homohopane C₃₁–C₃₅ doublets, the 22S and 22R diastereoisomers, typical of crude oils, can also be observed in the m/z 191 mass chromatogram (Fig. 2). These doublet peaks decrease in size with increasing carbon number.

Fig. 2 (a) Triterpane profile for Gulfaks crude oil, showing the North Sea oil specific marker bisnorhopane (BNH), and containing C₂₉ hopane (NH) and hopane along with the doublet peaks due to the C₃₁–C₃₅ homohopane diastereoisomers (C₃₁-H to C₃₅-H). (b) The sterane profile of Gulfaks oil is typical of a heavily biodegraded oil, having a high diasterane/sterane ratio (C₂₇-DiaS/C₂₇-S). The 20S/20R ethylcholestane ratio (S-EC/R-EC) is >1 and is also an indication of a biodegraded oil.

The m/z 217 mass chromatogram of Gulfaks shows a characteristic and complex sterane distribution. The 20S/20R ethyl-5α(H),14α(H),17α(H)-cholestane ratio is >1. In most crude oils, the 20S and 20R isomers are present in equal amounts, and a high 20S/20R ratio is an indication of thermal maturity.¹⁰ Diasteranes are rearranged steranes and are more stable to biodegradation than steranes and a high diasterane/sterane ratio can indicate a heavily biodegraded oil.¹² The C₂₇ diasterane/C₂₇ sterane ratio in Gulfaks is high, ∼0.7, and can be used as an indicator of Gulfaks oil (Fig. 2).

Analysis of reference site sediments

Particle size analysis (PSA). The PSA data on pooled sediments from the reference sites (Olna Firth, Mangaster Voe and Vaila Sound) are given in Table 2. The distributions for the <63 µm fraction ranged from 14.0 to 52.0%, both samples coming from Vaila Sound. However, one sample from Olna Firth gave a value of 4.6% for the <63 µm fraction, due to the large amount of stones in the sediment. The mean for the <63 µm fraction at Olna Firth, Mangaster Voe and Vaila Sound were 34.6% (s 6.31, n = 5), 44.4% (s 5.52, n = 6) and 27.2% (s 13.72, n = 6), respectively. Sediments with a smaller particle size (higher percentage for the <63 µm fraction) tend to accumulate PAHs to a greater extent.

Table 2 PAH concentrations, particle size and carbon (%) from pooled sediments taken from the reference sites: Olna Firth (R1), Mangaster Voe (R2) and Vaila Sound (R3). The first two letters of the voe are used to identify samples, e.g., VA for Vaila, and P1–P6 refer to the pool numbers. The numbers in parentheses indicate the individual sample numbers that make up the pool

Sample No.	[PAHs]/ng g^−1 dry weight	Median particle size/µm	Particle size <63 µm (%)	Carbon (%)
^a Analysed in triplicate for PAHs. ^bNot included in calculation of mean.
OLP1 (34, 35, 19, 21, 31, 8)	4530	93	42.5	6.384
OLP2 (28, 30, 23, 25,16, 7)	2034	1835	24.1	6.086
OLP3 (22,11, 26, 5, 14, 27)	3400	397	32.3	6.440
OLP4^a (20, 17, 3, 12, 13, 1)	2902 (s 702.6)	338	34.7	5.716
OLP5 (18, 10, 9, 6, 33, 36)	3488	4585^b	4.6^b	7.160
OLP6 (4, 32, 15, 24, 2, 9)	3378	122	39.2	5.181
Mean	3289 (s 744.4)	557 (s 650)	34.6 (s 6.31)	6.161 (s 0.618)
MAP1 (11, 12, 10, 18, 21, 9)	478.1	103	36.9	2.056
MAP2 (29, 17, 7, 35, 15, 33)	444.3	62	50.4	1.963
MAP3 (8, 16, 32, 3, 31, 19)	429.7	103	39.9	2.093
MAP4 (24, 13, 2, 6, 5, 22)	740.0	59	51.8	2.816
MAP5 (36, 27, 26, 30, 4, 23)	376.7	94	41.0	1.5
MAP6^a (20, 14, 34, 25, 1, 28)	588.2 (s 177.9)	72	46.4	1.865
Mean	509.5 (s 121.5)	82 (s 18)	44.4 (s 5.52)	2.052 (s 0.389)
VAP1 (31, 1, 9, 36, 4, 21)	902.2	107	37.3	0.718
VAP2 (30, 22, 12, 6, 26, 14)	543.8	167	27.0	1.410
VAP3^a (17, 28, 33, 10, 16, 13)	603.0 (s 281.4)	190	16.7	0.994
VAP4 (27, 20, 15, 11, 8, 5)	248.7	227	14.0	0.720
VAP5 (23, 25, 32, 3, 24, 35)	313.9	258	15.9	0.532
VAP6 (2, 29, 7, 34, 19, 18)	517.8	59	52.0	0.814
Mean	521.6 (s 212.0)	168 (s 68)	27.2 (s 13.72)	0.865 (s 0.280)

---

Total organic carbon. Total organic carbon on pooled sediments from the reference sites (Olna Firth, Mangaster Voe and Vaila Sound) are given in Table 2. The total organic carbon ranged from 0.532% in a sediment from Vaila Sound to 7.160% in a sediment from Olna Firth. The mean values, for the six pools at each site, were 6.161% (s 0.618, n = 6), 2.052% (s 0.389, n = 6) and 0.865% (s 0.280, n = 6) at Olna, Mangaster, and Vaila, respectively. Olna pooled sediments all had a higher total organic carbon content than the other reference site sediments and therefore would be expected to accumulate hydrophobic compounds such as PAHs to a greater extent than sandy sediments.³

Polycyclic aromatic hydrocarbons. The total PAH concentration (Table 2) in the pooled reference samples ranged from 248.7 ng g⁻¹ dry weight in a sample from Vaila Sound, to 4530 ng g⁻¹ dry weight in a pooled sediment from Olna Firth. The mean total PAH concentration was 509.5 ng g⁻¹ dry weight (s 121.5, n = 6) at Mangaster Voe and 521.6 ng g⁻¹ dry weight (s 212.0, n = 6) at Vaila Sound. Triplicate analysis of pooled sediment MAP6 gave a mean PAH concentration of 588.2 ng g⁻¹ dry weight (s 177.9, n = 3). Triplicate analysis of an individual sediment, MA35, gave a mean PAH concentration of 571.3 ng g⁻¹ dry weight (s 243.5, n = 3) for the wet extraction method and 594.4 ng g⁻¹ dry weight (s 96.8, n = 3) when the sample was freeze-dried prior to extraction. The total PAH concentrations of pooled sediments showed a high variability when analysed in triplicate. Triplicate analysis of individual sediments and freeze-dried sediments also showed a high variability, suggesting that pooling the samples did not introduce an error and freeze-drying did not improve the method. Repeated analysis of a QUASIMEME sediment has shown that the method employed for PAH analysis is precise and accurate. Hence, the high variability in the replicate analysis must be associated with the considerable heterogeneity within the sediments. At Mangaster Voe the four- and five-ring compounds gave the largest contribution to the total PAH concentration and DBT (parent and branched) the smallest (Fig. 3).

Fig. 3 Mean PAH concentration (ng g⁻¹ dry weight) by group for reference (R) and Zone (Z) voes. The error bars represent one standard deviation from the mean. Nap, naphthalenes (parent and C₁–C₄); 178, phenanthrene/anthracene (parent and C₁–C₃); DBT, dibenzothiophene (parent and C₁–C₃); 202, fluoranthene/pyrene (parent and C₁–C₃); 228, benzanthracenes/benzophenanthrenes/chrysene/triphenylenes (parent and C₁–C₂); 252, benzofluoranthene/benzopyrene/perylene (parent and C₁–C₂); 278, indenopyrene/benzoperylene (parent and C₁–C₂). A full description of these PAH groups has been presented previously.¹⁸

One pooled sample from Vaila Sound (VAP3) was analysed in triplicate and gave a mean of 603.0 ng g⁻¹ dry weight (s 281.4, n = 3). Triplicate analysis of an individual sediment gave a mean of 700.5 ng g⁻¹ dry weight (s 360.5, n = 3). The same individual sample freeze-dried before analysis gave a mean of 557.8 ng g⁻¹ dry weight (s 210.9, n = 3). The PAH distributions in both pooled and individual sediments from Vaila Sound were similar, showing a fairly even distribution between the 2–5-ring compounds but with lesser amounts of DBT (parent and branched) and six-ring compounds (Fig. 3). The PAH levels at both Mangaster Voe and Vaila Sound are elevated in comparison with a marine sediment, but the contamination is low to moderate for a coastal sediment,⁵ and could partly be a result of the small boat activity in these areas.

The total PAH concentrations in pooled samples from Olna Firth were considerably higher than at the other two reference sites (Table 2) and ranged from 2034 to 4530 ng g⁻¹ dry weight with a mean of 3289 ng g⁻¹ dry weight (s 744.4, n = 6). Triplicate analysis of pooled sample OLP4 gave a mean PAH concentration of 2902 ng g⁻¹ dry weight (s 702.6, n = 3). Triplicate analysis of an individual sediment, OL30, gave a mean concentration of 1626 ng g⁻¹ dry weight (s 375.3, n = 3). The same sample freeze-dried before extraction gave a mean of 2071 ng g⁻¹ dry weight (s 201.6, n = 3). Again, the variability was high for sediments analysed in triplicate, which suggests that sediments from Olna were not homogeneous. The PAH concentration was highest for the four- and five-ring compounds and lowest for branched and parent DBT (Fig. 3). All ring size groups, including alkylated PAHs, except for DBT, showed higher concentrations than any other site. The high PAH levels observed in these sediments were unexpected for a reference site and were also considerably higher than for the other two reference sites and the sites from within the Zone. Olna sediments, however, contained a high percentage of total organic carbon and would be expected to accumulate PAHs more readily than the other reference site sediments. Olna is located in a discrete Voe, remote from any urban and industrial activity, and would have only small boat traffic. However, a main road (A970) runs alongside Olna Firth. This could result in an accumulation of PAHs in the sediments due to road runoff.

PAH concentration ratios. PAH concentration ratios are often used to distinguish between pyrolytic and petrogenic sources. A summary of indices commonly used is shown in Table 1. Clean sediments in the Shetland area tend to be dominated by 3–5-ring parent compounds resulting from pyrolysis.¹⁹ This profile is typical of most areas in the North Sea. Alkylated PAHs dominate unburned fossil fuels, and therefore the presence of alkylated PAHs in the reference site sediments suggests possible petrogenic contamination. However, at all three reference sites the proportion of parent PAHs (not including DBT) was high (>36%), indicating that there could also be a pyrolytic source of PAHs (Table 3). In contrast, a sediment collected in 1994 from an area severely impacted by the oil spill (Burra Haaf-S94/228) contained only 3.4% parent compounds. The methylphenanthrene/phenanthrene ratio (MP/P) is commonly used to ascertain emission sources by comparing alkylated and parent compounds.^5–7 Alkylated homologues are deficient in combustion-generated PAHs, giving an MP/P ratio of <1. However this ratio was >1 in all reference sites, ranging from 1.0 to 2.2, which suggests there is some petrogenic contamination (Table 3). The index for fluoranthene + pyrene/methylfluoranthene + methylpyrene (Fl + Py/MFl + MPy) again compares alkylated and parent PAHs, and has been reported as an indication of pyrogenic inputs. Values of near 3 have been found in sediments where the main source of contamination is pyrolysis, with lower values indicating a smaller pyrogenic input and greater petrogenic input.⁸ Values found in reference site sediments ranged from 1.6 to 2.9. Other indicators of pyrolytic sources compare isomer ratios. The one most often used is the phenanthrene/anthracene ratio (P/A).^3–5 Phenanthrene is the thermodynamically more stable isomer. The P/A ratio is temperature dependent and decreases with increasing temperature, therefore high temperature processes can be characterised by low P/A values (<10). The slow thermal maturation of organic matter in petroleum is governed by thermodynamic properties and leads to much higher P/A values (>10). The P/A ratio for crude oils is normally close to 50. At all reference sites the P/A values were <10, indicating that the major source of PAHs at the reference sites was pyrolysis. Similarly, the fluoranthene/pyrene ratio (Fl/Py) is often used to distinguish between pyrolytic and petrogenic sources.^4,5 Values of >1 are associated with pyrolytic origins. All reference sites displayed a Fl/Py value of >1, again indicating that the major source of PAHs was pyrolysis. These results suggest that pyrolysis was the main source of PAHs in the reference site sediments, but there was a limited petrogenic input. The concentration ratios for the contaminated sediment from Burra Haaf (S94/228) are consistent with petrogenic contamination (Table 3).

Table 3 Percentage of parent compounds (excluding DBT), phenanthrene/anthracene (P/A), fluoranthene/pyrene (Fl/Py), methylphenanthrene/phenanthrene (MP/P) and fluoranthene + pyrene/methylfluoranthene + methylpyrene (Fl +Py/MFl + MPy) ratios in selected reference site sediments. The concentration ratios for S94/228 (a sediment from Burra Haaf known to be contaminated with oil) all indicate petrogenic contamination

	S94/228	VAP1	VAP2	OLP4	OLP1	MAP5	MAP6
^a Excludes DBT. ^bNap, naphthalene.
Naphthalene (%)	28.9	17.9	15.9	12.6	7.8	6.7	4.1
178 (%)	35.0	21.8	18.0	15.4	12.2	12.4	12.9
DBT (%)	4.3	2.1	2.2	1.8	1.5	2.0	2.7
202 (%)	15.4	20.6	20.2	17.8	14.6	14.7	19.1
228 (%)	9.2	16.0	13.3	18.0	17.2	14.4	17.1
252 (%)	5.6	14.9	21.5	25.7	34.7	32.5	31.5
276 (%)	1.6	6.6	9.0	8.6	12.0	17.2	12.5
Parent^a/ng g^−1	1379.7	402.6	227.4	1502.3	2029.0	178.7	170.2
Parent^a (%)	3.4	45.6	42.7	41.3	45.5	48.4	36.5
Parent^a−Nap^b	1375.0	382.8	218.9	1466.1	2002.7	175.8	168.3
P/A	32.4	6.2	6.1	6.0	5.2	7.1	5.3
Fl/Py	0.6	1.2	1.3	1.1	1.1	1.3	1.2
MP/P	8.9	1.0	1.5	2.1	1.9	1.5	2.2
Fl + Py/MFl + Mpy	0.2	2.9	2.3	2.3	2.0	2.4	1.7

---

Total aliphatic hydrocarbons. The n-alkane profiles of all reference site sediments showed no evidence of petrogenic contamination [Fig. 4(a)]. There was no unresolved complex mixture (UCM) and the profile showed an odd-numbered long-chain predominance, typical of biogenic material. The total n-alkane concentrations were higher at Olna Firth (27.3–41.3 µg g⁻¹) than the other two reference sites (1.4–9.7 µg g⁻¹), but this is to be expected at a site with a high total organic carbon content.

Fig. 4 (a) Aliphatic hydrocarbon profile of a typical reference site sediment. The profile is dominated by the odd-numbered long-chain components, typical of biogenic material. (b) Aliphatic hydrocarbon profile of a typical Punds Voe sediment. The unresolved complex mixture (UCM) shows a bimodal distribution, suggesting petrogenic contamination. The internal standard used contained squalane (Sq) and heptamethylnonane (HMN). Squalane was used for the quantification.

Steranes and triterpanes. Vaila Sound pooled sediment samples VAP3 and VAP6 were analysed for steranes and triterpanes. The m/z 191 mass chromatograms for both samples contained mainly natural triterpanes, but there was limited evidence for petrogenic contamination [Fig. 5(a), VAP3]. The largest peaks came from hopane, the biogenic triterpane (22R)-17α, 21β-homohopane and the naturally occurring triterpene diploptene. There was also a C₂₉ hopane peak but no bisnorhopane, indicating that there was no North Sea oil contamination. There were, however, small doublet peaks decreasing in size with increasing carbon number, for the C₃₂–C₃₅ homohopane 22S and 22R diastereoisomers, which indicates petrogenic contamination. The petrogenic C₃₁ homohopane doublet was masked by the naturally occurring isomer (22R)-17α,21β-homohopane. There was an unusually high proportion of C₂₉ hopane, which is often associated with oils from carbonate source rocks, which includes most oils from the Middle East. The presence of steranes in the m/z 217 mass chromatogram indicated petrogenic contamination, the 20S/20R ethylcholestane ratio was >1, but higher than would be found in an oil such as Gulfaks crude oil. There was little evidence of the C₂₇ diasteranes, indicating the oil was not heavily biodegraded.

Fig. 5 (a) Triterpane profile of a sediment suspected to contain a limited amount of Middle Eastern oil. This profile was typical of sediments from the reference sites, Vaila Sound and Mangaster Voe, and Zone sites Stromness and Sandsound Voes. The chromatograms showed both natural [(22R)-17α,21β-homohopane and diploptene] and petrogenic biomarkers. The small doublet peaks due to the C₃₂–C₃₅ homohopane diastereoisomers (C₃₂-H to C₃₅-H) are indicative of crude oil contamination. The high ratio of C₂₉ hopane (NH) to hopane and the lack of a bisnorhopane peak indicate that the contamination is due to a Middle Eastern oil. (b) Triterpane profiles of a sediment from Olna Firth. The profile is dominated by natural biomarkers such as (22R)-17α,21β-homohopane and the triterpene diploptene. There is little evidence of any petrogenic contamination; the hopane and C₂₉ hopane peaks are small, and there is a lack of any homohopane doublet peaks.

Mangaster Voe pooled sediment samples MAP6 and MAP3 were analysed for triterpanes and steranes. The m/z 191 mass chromatograms showed clearly both natural and petrogenic triterpanes, and both sediments had profiles similar to those from Vaila Sound [Fig. 5(a)]. The sterane profile was also similar to the Vaila Sound profiles, showing no evidence of diasteranes. The petrogenic contamination at both Vaila Sound and Mangaster Voe was not due to a North Sea oil but could have been from a Middle Eastern oil. Shipping activity could result in the accumulation of Middle Eastern oil in the sediments, but in these areas there would only be small boat traffic, which use mainly marine diesel, and would not give a crude oil triterpane profile. Transportation of oil from areas where there is shipping activity into Mangaster Voe and Vaila Sound could result in petrogenic contamination in the sediments.

Olna pooled sediment samples OLP6 and OLP4 were analysed for biomarker compounds. Neither samples contained petrogenic biomarkers. The m/z 191 mass chromatogram [Fig. 5(b), OLP2] shows a distribution which is dominated by components of very recent biogenic origin, (22R)-17α,21β-homohopane and the natural triterpene diploptene, with very little contribution from petrogenic hopanes. The C₃₁–C₃₅ homohopane doublet peaks were not present, suggesting that there was no crude oil contamination at this site. The m/z 217 mass chromatograms showed no steranes, again indicating no crude oil contamination. An individual sediment, OL30, was also analysed and showed similar m/z 191 and 217 mass chromatograms. The unusually high PAH levels found in sediments from this reference site were therefore not due to crude oil. The PAH ratios were consistent with pyrolytic sources.

Analysis of zone site sediments

Particle size analysis. PSA on pooled sediments from Sandsound Voe, Stromness Voe and Punds Voe (Table 4) showed a distribution ranging, for the <63 µm fraction, from 12.1% for a sample from Sandsound Voe to 72.6% from a Stromness Voe sediment. The mean for the <63 µm fraction at Sandsound, Stromness and Punds Voes were 29.6% (s 10.19, n = 6), 57.7% (s 16.54, n = 6) and 27.4% (s 4.03, n = 6), respectively. Stromness Voe sediments, on average, had a smaller particle size than the other Zone sites and are likely to accumulate PAHs to a greater extent.

Table 4 PAH concentrations, particle size and carbon (%) from pooled sediments taken from the Zone sites: Sandsound Voe (Z1), Stromness Voe (Z2) and Punds Voe (Z3). The first two letters of the voe are used to identify samples, e.g., SA for Sandsound, and P1–P6 refer to the pool numbers. The numbers in parentheses indicate the individual sample numbers that make up the pool

Sample No.	[PAHs]/ng g^−1 dry weight	Median particle size/µm	Particle size <63 µm (%)	Carbon (%)
^a Analysed in triplicate for PAHs.
SAP1 (4, 36, 8, 35, 28, 10)	1218	2754	25.0	4.484
SAP2 (16, 29, 6, 5, 17, 12)	1448	226	40.0	5.832
SAP3 (23, 15, 26, 20, 11, 30)	1353	1260	31.7	4.580
SAP4 (34, 21, 32, 18, 1, 19)	2043	109	42.7	6.342
SAP5^a (2, 33, 24, 7, 31, 13)	1276 (s 182.2)	1148	26.2	4.430
SAP6 (14, 3, 25, 9, 22, 27)	1228	583	12.1	4.814
Mean	1428 (s 286.2)	1013 (s 888)	29.6 (s 10.19)	5.080 (s 0.737)
STP1 (11, 20, 30, 35, 3,24)	1241	54	55.4	4.322
STP2 (26, 32, 6, 29, 12, 7)	970.8	2724	22.9	4.791
STP3^a (16, 2, 22, 1, 8, 13)	1430 (s 292.3)	38	69.7	6.418
STP4 (18, 19, 10, 14, 34, 17)	1586	39	72.6	6.597
STP5 (33, 15, 28, 25, 5, 27)	1084	38	61.2	4.120
STP6 (9, 23, 4, 36, 31, 21)	1122	46	64.5	4.619
Mean	1239 (s 210.8)	490 (s 999)	57.7 (s 16.54)	5.144 (s 0.988)
PUP1^a (18, 19, 1, 7, 4, 20)	641.0 (s 197.2)	221	26.2	1.522
PUP2 ( 17, 5, 30, 11, 35, 31)	2548	244	24.3	1.233
PUP3 (25, 34, 9, 33, 14, 22)	2400	136	31.0	1.471
PUP4 (29, 15, 3, 13, 26, 10)	2345	153	28.6	1.265
PUP5 (21, 16, 32, 8, 36, 2)	904.8	400	21.2	1.328
PUP6 (28, 23, 27, 24, 6, 12)	2766	117	33.2	1.486
Mean	1934 (s 835.3)	212 (s 96)	27.4 (s 4.03)	1.384 (s 0.113)

---

Total organic carbon. The total organic carbon content (Table 4) of the pooled sediments ranged from 1.233% from a Punds Voe sediment to 6.597% for a sediment from Stromness Voe. The mean values for Sandsound, Stromness and Punds Voes were 5.080% (s 0.737, n = 6), 5.144% (s 0.988, n = 6) and 1.384% (s 0.113, n = 6), respectively. Punds Voe might be expected to accumulate PAHs to a lesser extent than the other Zone sites, having a smaller mean total organic carbon.

Polycyclic aromatic hydrocarbons. The total PAH concentration (Table 4) in the pooled Zone sediments ranged from 641.0 to 2766 ng g⁻¹ dry weight, both samples coming from Punds Voe. The mean PAH concentration at Sandsound Voe was 1428 ng g⁻¹ dry weight (s 286.2, n = 6) and that at Stromness Voe was 1239 ng g⁻¹ dry weight (s 210.8, n = 6). Pooled sediment SAP5 was analysed in triplicate and gave a mean total PAH concentration of 1276 ng g⁻¹ dry weight (s 182.2, n = 3). An individual sediment, SA5, was also analysed in triplicate and gave a mean total PAH concentration of 1100 ng g⁻¹ dry weight (s 44.5, n = 3). Triplicate analysis was also carried out on this sample after freeze-drying the sediment and gave a mean total PAH concentration of 1454 ng g⁻¹ dry weight (s 32.7, n = 3). The PAH concentration at Sandsound Voe was highest for the four- and five-ring compounds and lowest for the parent and branched naphthalenes and DBTs (Fig. 3). Pooled sediment STP3 was analysed in triplicate and gave a mean total PAH value of 1430 ng g⁻¹ dry weight (s 292.3, n = 3) . Individual sediment ST5 was also analysed in triplicate and gave a mean total PAH concentration of 1869 ng g⁻¹ (s 136.0, n = 3). Extraction of this sample in triplicate was carried out after freeze-drying. Analysis gave a mean total PAH content of 1513 ng g⁻¹ dry weight (s 39.2, n = 3). The PAH distribution in sediments from Stromness Voe was highest for the four- and five-ring compounds and lowest for naphthalenes and DBTs which had a smaller contribution to the total (Fig. 3). The variability for the total PAH concentrations was generally lower for sediments from Stromness and Sandsound Voes (RSD 7.3 and 2.6%, respectively), suggesting that the sediments at these sites were more homogeneous.

At Punds Voe the mean total PAH concentration was 1934 ng g⁻¹ dry weight (s 835.3, n = 6). This value was higher than the other two Zone sites, despite two of the values being lower than any other Zone sediments. One pooled samples (PUP1) was analysed in triplicate and gave a mean total PAH content of 641.0 ng g⁻¹ dry weight (s 197.2, n = 3). Individual sediment (PU17) was also analysed in triplicate and gave a mean total PAH content of 791.3 ng g⁻¹ dry weight (s 66.1, n = 3). The same individual sediment was freeze-dried prior to analysis. The mean total PAH content using this method was 931.1 ng g⁻¹ dry weight (s 115.2, n = 3). In sediments from Punds Voe the four-ring compounds gave the biggest contribution to the total PAH concentration and DBT (parent and branched) the smallest (Fig. 3). Punds Voe is close to the entrance of Scalloway harbour, and is associated with shipping activity; elevated PAH levels might be predicted at this site.

PAH concentration ratios. The proportion of parent compounds (not including DBT) was >36% at Stromness and Sandsound Voes (Table 5). Although equivalent values were noted for some of the sediments from Punds Voe, others contained a smaller proportion of parent PAHs, in one case as low as 14.1%. The higher proportion of alkylated compounds at Punds Voe were indicative of a petrogenic source for the PAHs. MP/P ratios ranged from 1.4 to 3.6, indicating that there was possibly petrogenic contamination at the Zone sites (Table 5). The Fl + Py/MFl + MPy index in the Exclusion Zone sediments ranged from 1.5 to 3.3 at Sandsound and Stromness Voes. The values for the Punds Voe sediments were lower, ranging from 0.5 to 2.2, indicating that there was a greater petrogenic input at Punds Voe. Similarly to the reference sites, the P/A values for all Zone site sediments were <10, indicating that there was a pyrolytic source of PAHs in Zone site sediments. The Fl/Py ratio was >1 for all Zone site sediments except PUP2, again suggesting that the main source of PAHs was pyrolysis. These results suggest that PAHs in Zone site sediments were of both petrogenic and pyrogenic origins, with Punds Voe sediments having the larger petrogenic contribution.

Table 5 Percentage of parent compounds (excluding DBT), phenanthrene/anthracene (P/A), fluoranthene/pyrene (Fl/Py), methylphenanthrene/phenanthrene (MP/P) and fluoranthene + pyrene/methylfluoranthene + methylpyrene (Fl +Py/MFl + MPy) ratios in selected Exclusion Zone site sediments

	SAP1	SAP2	STP3	STP4	PUP2	PUP6
^a Excludes DBT. ^bNap, naphthalene.
Naphthalene (%)	4.1	3.5	4.6	3.5	10.0	5.2
178 (%)	12.9	9.9	14.4	15.8	20.5	13.1
DBT (%)	2.7	2.3	2.9	6.9	6.4	4.0
202 (%)	19.1	16.7	16.8	15.2	21.1	34.1
228 (%)	17.1	18.7	17.6	16.1	19.5	22.3
252 (%)	31.5	35.0	27.3	27.1	18.0	16.4
276 (%)	12.5	13.8	16.4	15.4	4.5	5.0
Parent^a/ng g^−1	170.2	884.3	509.5	565.4	335.7	1054.3
Parent^a (%)	36.5	44.3	42.9	38.3	14.1	39.7
Parent^a−Nap^b	168.3	875.5	504.5	564.3	320.6	1045.5
P/A	7.3	6.9	3.3	4.0	7.2	2.2
Fl/Py	1.1	1.1	1.1	1.1	0.8	1.1
MP/P	1.6	1.4	2.8	3.6	2.1	2.1
Fl + Py/MFl + MPy	2.0	2.2	1.6	1.4	1.3	2.2

---

Total aliphatic hydrocarbons. The n-alkane profiles of sediments from Stromness and Sandsound Voes showed no evidence of petrogenic contamination. The profiles were similar to those of the reference site sediments, showing an odd-numbered long-chain predominance and without a UCM. Punds Voe sediments, however, were characterised by a bimodal UCM, which is indicative of the presence of petrogenic hydrocarbons (Fig. 4).

Steranes and triterpanes. Sandsound pooled samples (SAP4 and SAP5) and Stromness sediments (STP3 and STP6) showed similar triterpane profile to the pooled Vaila sediments [Fig. 5(a), VAP3]; they comprised both natural [diploptene and (22R)-17α,21β-homohopane] and petrogenic biomarkers (hopane and C₂₉ hopane). Again, any petrogenic contamination was mainly due to a Middle Eastern oil as there were only very small bisnorhopane peaks present in the m/z 191 mass chromatograms. The high proportion of C₂₉ hopane to hopane also indicated possible contamination from Middle Eastern oil. The m/z 217 chromatogram gave no evidence of Gulfaks crude oil, as no C₂₇ diasterane peaks could be seen. This confirmed that the petrogenic contamination at these sites was mainly from a Middle Eastern oil, with possibly a small contribution from a North Sea oil. Similar to the reference sites, Mangaster and Vaila, there would be only small boat activity present in these voes. Hence Middle Eastern oil contamination must be transported into these sites from areas where there is shipping activity.

Punds pooled samples analysed for biomarker compounds (PUP4, PUP5, PUP2, PUP3) and an individual sample (PU17) all showed a similar triterpane profile, dominated by petrogenic components [Fig. 6(a)]. The C₃₁–C₃₅ homohopane doublet peaks could be clearly seen. The profiles were fairly typical of a North Sea oil; the largest peak was hopane and C₂₉ hopane was present. In addition, the North sea oil specific marker bisnorhopane was present in the chromatogram. The proportion of C₂₉ hopane to hopane oils is normally about 35% in North Sea oils. However, the proportion found in Punds Voe sediments was higher than this (∼50%), suggesting there was possibly some Middle Eastern oil present. The sterane profile [Fig. 6(b)] was not identical with that of Gulfaks oil [Fig. 2(b)]. The C₂₇ diasterane/C₂₇ sterane ratio was less than normally found in Gulfaks crude oil (∼0.3). The 20S/20R ethylcholestane ratio was >1, but was higher than in Gulfaks. It is possible that Gulfaks was present in Punds Voe, but contamination from other oils could have reduced the diasterane/sterane ratio, giving a profile that was not typical of Gulfaks oil.

Fig. 6 (a) The m/z 191 mass chromatogram of a Punds Voe sediment is fairly typical of a North Sea oil, containing the North sea oil specific biomarker bisnorhopane (BNH). The doublet peaks due to the C₃₁–C₃₅ homohopane diastereoisomers (C₃₁-H to C₃₅-H) are clearly observed and suggest crude oil contamination. The ratio of C₂₉ hopane (NH) to hopane, however, is higher than found in most North Sea oils, suggesting that there may be Middle Eastern oil also present. (b) The m/z 217 mass chromatogram shows a 20S/20R ethylcholestane ratio (S-EC/R-EC) of >1, but higher than normally found in Gulfaks crude oil. C₂₇ diasteranes (C₂₇-DiaS) are present, but the ratio to C₂₇ steranes (C₂₇-S) is less than normally found in Gulfaks oil.

Differences in PAH concentration between voes. Differences in PAH concentration between voes and regions (reference/Exclusion Zone) were investigated using a combination of multivariate and regression techniques. First, principal component analysis was used to explore the relationship between the concentrations of the 36 PAH compounds measured in each pool. Rather than use the raw data, however, the concentrations were normalised to the percentage organic carbon content of the pool, to account for the increased accumulation of PAHs by sediments with high carbon content.³ The normalised concentrations were then log-transformed, to reduce skewness and homogenise variances.

The first three principal components explained 89% of the variation in the data (70, 11 and 8%, respectively), and are the only components that need to be considered. The first, and most important, component gives positive weight to nearly all the compounds, and so is a measure of ‘overall' PAH concentration. The second component contrasts the naphthalenes and some parent 3–6-ring compounds with the alkylated 3–6-ring compounds, particularly those with two or three methyl groups. The third component contrasts the naphthalenes with the parent 3–6-ring compounds, particularly those with 4–6 rings.

The principal components were used to suggest meaningful combinations of the original compound concentrations for use as response variables in further analysis. The first component clearly suggests total PAH concentration. The second and third components are contrasts, and suggest concentration ratios involving the naphthalenes, the parent 3–6-ring compounds and the alkylated 3–6-ring compounds. Letting N, A and P denote the total concentration of the naphthalenes, the alkylated 3–6-ring compounds and the parent 3–6-ring compounds, respectively, all normalised for organic carbon, the following variables were considered: total concentration, i.e., N + A + P; concentration ratio of alkylated vs. parent 3–6-ring compounds, i.e. A/P; and concentration ratio of 3–6-ring compounds vs. naphthalenes, i.e., (A + P)/N.

Fig. 7 shows a strip-plot of these variables by voe. The data are plotted on a logarithmic scale, and all subsequent analyses are on a logarithmic scale with results back-transformed for presentation.

Fig. 7 Strip-plot of the first three principal component scores by voe. The three components used were the total PAH concentration, alkylated vs. parent (3–6-ring) and 3–6-ring vs. naphthalenes. The data are plotted on a logarithmic scale. The means for the Exclusion Zone and reference site sediments and the means for each voe, along with ±2 standard error bars, are shown.

The effects of voe and region on the three response variables were investigated using various linear mixed models. First, a model focusing on specific differences between voes was considered. This model can be written as:

to indicate that each observation depends on the voe from which it came plus an error term. Here voe is treated as a fixed effect, and it is of no concern whether a voe is a reference or is in the Exclusion Zone. The error is the residual random variation between pools within voes. The model is just a standard one-way analysis of variance, except that the residual variance was allowed to vary between voes, as preliminary analysis showed this to be necessary (cf., Fig. 7). A second model focused on differences between regions: This model treats region as a fixed effect, measuring the difference between reference voes and voes in the Exclusion Zone, and region.voe as a random effect, allowing for random variation between voes within regions. More complicated models involving particle size were also considered. All models were fitted by restricted maximum likelihood using the statistical package Genstat for Windows, with the significance of the various effects being assessed by likelihood-based tests.

There were significant differences between voes for all three response variables (total concentration, p < 0.001; A/P, p < 0.005; (A + P)/N, p < 0.001). The estimated mean levels for each voe, with ±2 standard error bars, are shown in Fig. 7. In particular, total PAH concentration is greater in Punds Voe, Vaila Sound and Olna Firth than in Mangaster, Sandsound and Stromness Voes. In general, Punds Voe stands out as more variable than the other voes, and very different to Sandsound and Stromness Voes, the other two voes in the Exclusion Zone. This is not surprising, since Punds Voe is close to the entrance to Scalloway Harbour. When Punds Voe was excluded from the analysis, voe had no significant effect on A/P, although it was still significant for total concentration and (A + P)/N.

There were no significant differences between region for any of the response variables, regardless of whether Punds Voe was included or not. However, too much should not be read into this result; there were only three voes in each region, and the test for differences had low statistical power. To demonstrate, Fig. 7 shows the estimated mean level, with ±2 standard error bars, for each region. The standard errors are large and, for instance, the mean total concentration in the Exclusion Zone could be anywhere between one third and three times that in the Reference region, with 95% confidence.

There was also no evidence of any effect of particle size on any of the response variables. In particular, this suggests that normalising for organic carbon is adequate for explaining the increased uptake of PAHs by sediment with high carbon content and correspondingly small particle size.

Conclusions

The mean total PAH concentrations were found not to be significantly different for reference and Zone site sediments. The reference site Olna Firth was found to have the highest levels of PAHs. However, the PAH concentration ratios (P/A, Fl/Py) indicated that pyrolysis was the main source of PAHs. The n-alkane profile and the biomarker profiles were dominated by biogenic components and gave no evidence of petrogenic contamination. The reference sites at Vaila Sound and Mangaster Voe and the Zone sites at Stromness Voe and Sandsound Voe showed some evidence of petrogenic contamination from the PAH concentration ratios (MP/P and Fl + Py/MFl + MPy). This was confirmed by the triterpane profile which, although dominated by natural compounds, did show some evidence of petrogenic contamination by Middle Eastern oil. Only the Zone site, Punds Voe, showed clear evidence of petrogenic contamination, from the PAH concentration ratios (MP/P and Fl + Py/MFl + MPy), the UCM in the n-alkane profile and the triterpane profile which was dominated by petrogenic biomarkers. The m/z 191 mass chromatogram indicated that the petrogenic contamination was due to a mixture of Middle Eastern and North sea oils. As Punds Voe is situated close to Scalloway harbour, this is not unexpected. In cases of high petrogenic contamination, the n-alkane profiles, the biomarker profiles and to some extent the PAH concentration ratios all prove to be useful indicators of oil contamination. However, PAH concentration ratios do not always give conclusive evidence of petrogenic contamination. When there is a mixed source of PAHs the P/A and Fl/Py ratios tend to indicate pyrolysis as being the main input and the MP/P and Fl + Py/MFl + MPy ratios suggest a possible petrogenic input. In cases where pyrolysis is the main source of PAHs but with a small petrogenic input, only biomarker analysis can identify the crude oil contamination.

References

1. K. J. Whittle D. A. Anderson P. R. Mackie C. F. Moffat N. J. Shepherd A. H. McVicar, in The Impact of an Oil Spill in Turbulent Waters: the Braer, ed. J. M. Davies and G. Topping, The Stationery Office, Edinburgh, 1997, pp. 144–160..
2. G. Topping J. M. Davies P. R. Mackie C. F. Moffat, in The Impact of an Oil Spill in Turbulent Waters: the Braer, ed. J. M. Davies and G. Topping, The Stationery Office, Edinburgh, 1997, pp. 121–143..
3. P. Baumard, H. Budzinski and P. Garrigues, Mar. Pollut. Bull., 1998, 36, 577.
4. P. Baumard, H. Budzinski, Q. Michon, P. Garrigues, T. Burgeot and J. Bellocq, Estuarine Coastal Shelf Sci., 1998, 47, 77.
5. H. Budzinski, I. Jones, J. Bellocq, C. Pierard and P. Garrigues, Mar. Chem., 1997, 58, 85.
6. M. B. Fernandes, M. A. Sicre, A. Boireau and J. Tronczynski, Mar. Pollut. Bull., 1997, 34, 857.
7. P. Garrigues, H. Budzinski, M. P. Manitz and S. A. Wise, Polycyclic Aromatic Compd., 1995, 7, 275.
8. J. C. Colombo, E. Pelletier, C. Brochu and M. Khalil, Environ. Sci. Technol., 1989, 23, 888.
9. D. M. Jones, S. J. Rowland and G. Douglas, Mar. Pollut. Bull., 1986, 17, 24.
10. F. Brakstad and O. Grahl-Nielsen, Mar. Pollut. Bull., 1988, 19, 319.
11. D. Munoz, M. Guiliano, P. Doumenq, F. Jacquot, P. Scherrer and G. Mille, Mar. Pollut. Bull., 1997, 34, 868.
12. G. A. Wolff, M. R. Preston, G. Harriman and S. J. Rowland, Mar. Pollut. Bull., 1993, 26, 567.
13. K. E. Peters J. M. Moldowan, in The Biomarker Guide: Interpretating Fossils in Petroleum and Ancient Sediment, Prentice-Hall, Englewood Cliffs, NJ, 1993, pp. 110–265..
14. A. S. MacKenzie S. C. Brassell G. Eglington J. R. Maxwell, Science, 217, 491..
15. S. Mitra, T. M. Dellapenna and R. M. Dickhut, Estuarine Coastal Shelf Sci., 1999, 49, 311.
16. R. Law and E. Andrulewicz, Mar. Pollut. Bull., 1983, 14, 289.
17. L. Webster, P. R. Mackie, S. J. Hird, P. D. Munro, N. A. Brown and C. F. Moffat, Analyst, 1997, 122, 1485.
18. L. Webster, L. Angus, G. Topping, E. J. Dalgarno and C. F. Moffat, Analyst, 1997, 122, 1491.
19. L. C. Massie, A. P. Ward, J. M. Davies and P. R. Mackie, Mar. Environ. Res., 1985, 15, 165.

---