- Email: [email protected]
A case study on fuel oil contamination in a mangrove swamp in Hong Kong
Marine Pollution Bulletin 51 (2005) 1092–1100 www.elsevier.com/locate/marpolbul
A case study on fuel oil contamination in a mangrove swamp in Hong Kong Nora F.Y. Tam, Teresa W.Y. Wong, Y.S. Wong
*
Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong
Abstract Mangroves commonly found along tropical and subtropical coastlines are susceptible to oil pollution. In December 2000, around 500 1 m tall Kandelia candel saplings at the age of 3–5 years old located at the foreshore region of Sheung Pak Nai swamp, Hong Kong SAR, were found to be damaged by oil pollution. More than 80% of the saplings were either dead or washed away and leaving less than 5% healthy saplings with dense green leaves. Elevated concentrations of light n-alkanes (ranging from n-C14 to n-C20), pristane and phytane were recorded in surface sediments collected in December 2000. The ratio between light and total n-alkanes was 0.4. The total petroleum hydrocarbons (60–80 lg g1 TPH) and unresolved complex mixtures (60–70 lg g1 UCM) were higher than the background values of other mangrove sediments in Hong Kong, which were 40 and 20 lg g1, respectively. In certain root zone sediments, TPH concentrations were above 1000 lg g1. These results suggest that surface sediments in Sheung Pak Nai were contaminated by petroleum oil, most likely by illegal discharge of fuel oil which occurred between 1998 and 2002. One year later, in December 2001, unhealthy saplings had recovered and re-grown. The concentrations of TPH and UCM in sediments declined to around 40 lg g1, pristane and phytane dropped by 80%, and the ratio of light to total n-alkanes was 0.15, suggesting that residual oil in sediments was weathered leading to a remarkable recovery of the unhealthy saplings. 2005 Elsevier Ltd. All rights reserved. Keywords: Kandelia candel; Petroleum hydrocarbons; Fuel oil; Alkane; UCM
1. Introduction Oil contamination, irrespective of its scale, has been recognized as one of the most serious threats to marine environments, particularly in tidal flats. Mangrove wetlands commonly found along most low wave energy shorelines in tropical and subtropical regions are well known for their high vulnerability to oil spills. The most serious oil spill accidents which affected mangrove habitats was the Galeta oil spill took place at the east of the Caribbean entrance to the Panama Canal in 1986, where about 8 million litres of crude oil were leaked out from a
*
Corresponding author. Tel.: +86 852 2788 9377; fax: +86 852 2788 9002. E-mail address: [email protected] (Y.S. Wong). 0025-326X/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2005.06.005
storage tank of a nearby oil refinery (Garrity et al., 1994). Mangrove trees were found to be killed or with growth impaired due to physical and chemical effects of the spilled oil. Floating oil settles and smothers the breathing and feeder roots. Oil dispersed on tree roots did not only cause death of some trees but also reduced growth of the survivors including defoliation and losses in canopy density, malformation of fruits, sub-lethal biochemical and molecular damages (Da Silva et al., 1997; Duke et al., 1997). Previous oil spill studies showed that oil residues in sediments at concentrations above 100 lg g1 dry weight would cause sub-lethal effects to mangrove trees (Burns et al., 1994; Levings and Garrity, 1997). The unique anoxic and waterlogged features of mangrove sediments reduced oil degradation, especially the residual aromatic hydrocarbons, whereas crude oil residues accumulated and persisted in deep
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
mud for up to a period of 20 years (Burns et al., 1994, 2000). The mortality and/or damages due to oil contamination depend on the type, quantity and weathered state of the oil, the mangrove species, and the prevailing climatic and tidal conditions. The scale and type of oil spill also affect the rate of dissipation and degradation of the residual oil in sediments (Burns and Codi, 1998). Previous studies on oil contamination in mangroves were mostly focussed on large scale crude oil spills as these were often perceived to be major threats. The contamination by leakage of fuel oil, improper disposal of used motor oil and illegal discharge of spent lubricating oil have received less attention although they account for a significant percentage of all oil inputs in harbours and other coastal waterways. Most oil spill studies were carried out in Central America, South America and Australia, while little information is available in mangroves around South China Sea especially the Pearl River Estuary. Only one crude oil spill accident in a mangrove swamp has been reported in Hong Kong SAR (Ke et al., 2002; Wong et al., 2002) despite mangroves in these areas are under all kinds of risks of oil pollution such as rapid industrialization and urbanization, heavy shipping transportation, and illegal oil smuggling operations. The present study therefore aims to investigate the oil contamination and its damages on young saplings of Kandelia candel, a dominant mangrove plant species in Sheung Pak Nai swamp. The degree of oil weathering and the recovery of mangrove saplings were also examined. Sheung Pak Nai mangrove swamp (1137.3 0 E, 216.9 0 N) situates in North Western part of the New Territories, Hong Kong SAR. The stand spreads over 4 km of coastlines along Deep Bay and the total mangrove area is around 6.34 ha (Tam et al., 1997). The substrate is made up of soft and deep fluidized mud especially near the foreshore region. The swamp is dominated by Kandelia candel, which has been planted since 1980s by local farmers who rely on fisheries from their oyster farms and aquaculture ponds as well as pigsties for their living. The mature trees with an average height of 4–5 m are healthy and produce large quantities of propagules. Saplings of 3–5 years old have also been planted at the low tidal mudflat region in recent years. Across Deep Bay, the opposite shore is Shekou, a densely populated and highly industrialized district in Shenzhen, PeopleÕs Republic of China. In December 2000, the saplings planted at the mudflat region were found either missing, dead or in unhealthy condition. Patches of oil film were seen on the mudflat at low tides with a strong oil smell. Illegal smuggling of diesel and fuel oil in the Pearl River Estuary Region, in particular Deep Bay was often reported from 1998 to 2000. The merchant ships and fishing boats unloaded the oil from larger ships at the border between Hong Kong SAR and
1093
Mainland China, and oil was discharged directly to the sea if they were detected by the Marine Police. The scale was around thousand tons of fuel oil in each smuggling operation. The smuggling activities have been under control from year 2001 onwards because of the joint efforts between the Hong Kong SAR and the Mainland administrations.
2. Materials and methods 2.1. Vegetation survey and sediment sampling Three 10 m · 10 m quadrats, located at different positions at the low tidal mudflat planted with saplings of K. candel in Sheung Pak Nai mangrove swamp, were sampled in December 2000 and December 2001. The saplings were planted in quadrats with regular distance of 0.5 m, each containing 400 saplings. Quadrat 1 was at the fringe of the mudflat, nearest to the sea with muddy sediments. Quadrat 2 was 8 m away from the fringe and sediments were less silty than that in Q1. Quadrat 3 was at a tidal position similar to that of Q2 but was near to a sandy path paved by local villages, and its sediments were sandy with lower water content than Q1 and Q2 (Table 1). In each quadrat, the number of saplings remained was counted and the health condition of each sapling was recorded according to a standard scaling system for assessing oil spill impact on a brackish marsh plant community (Hester and Mendelssohn, 2000). Surface sediment samples (1–10 cm) were randomly collected using a solvent-cleaned stainless steel spatula, and three replicate samples (each sample had more than 1 kg fresh sediments) were collected from each quadrat. In March 2001, triplicate root zone sediments of healthy and unhealthy saplings were sampled. This was done by digging up three individuals of each type of saplings from the mudflat. Bulk sediments were removed by shaking the roots, and sediments around the roots were then harvested. Dead saplings were not sampled because most roots were rotten and root zone sediments were difficult to obtain. Sediments were freeze-dried, ground
Table 1 Particle size distribution (%) and water content (% dry weight) of sediments in three quadrats in Sheung Pak Nai mangrove swamp Measurements
Quadrat 1
Quadrat 2
Quadrat 3
Particle sizes > 2 mm 1 mm–2 mm 500 lm–1 mm 250 lm–500 lm 125 lm–250 lm 63 lm–125 lm <63 lm Water content
4.29 6.56 6.53 16.08 25.11 7.46 33.97 40.81
9.78 8.43 7.09 18.26 16.48 15.55 24.41 40.60
23.23 19.04 12.71 12.28 11.08 4.93 16.73 34.85
1094
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
into powder and sieved through a 2 mm sieve, then stored in dark glass bottles at room temperature prior to analyses for petroleum hydrocarbons. Sub-samples of the fresh sediments were air-dried and analysed for pH, soil texture (wet sieving technique) and organic matter concentrations (loss on ignition at 550 C for 6 h). 2.2. Determination of petroleum hydrocarbons in sediments Sediment samples were extracted following the protocol described by Wang et al. (1994) with minor modification. Five to 20 g freeze-dried sediments, depending on the level of contamination, were transferred to a 250 ml pre-ashed conical flask. Around 1 g of activated copper (copper powder purchased from BDH was activated by 2 M hydrochloric acid and subsequently washed by distilled water and solvents) was added to remove sulphur. A surrogate standard, o-terphenyl (ChemService) at a concentration of 69.5 lg l1, was spiked prior to extraction. The sediment was then sonicated in an ultrasonic bath with a 60 ml mixture of n-hexane and dichloromethane (1:1 v/v) for 30 min. The ultrasonic extraction step was repeated twice with 60 ml dichloromethane each time. The extract was combined and filtered through a Whatman no. 541 filter paper to remove large particles. The extract was then concentrated to around 1 ml by rotary evaporation and dried to 0.2–0.4 ml under a pure nitrogen stream. The cleanup and fractionation of the extract was done by a self-packed silica gel column, a 25 ml acetone-rinsed burette (1 cm · 45 cm) contained 3 g firmly packed activated silica gel (72–230 mesh, Sigma) and a layer of 1–2 cm Na2SO4 at the surface to absorb water. Twenty milliliters hexane was added to condition the column. The concentrated extract was then applied to the column with an additional 3 ml hexane for complete transfer. All the solvent eluted up to this point was discarded. When the Na2SO4 layer was nearly exposed to air, 12 ml n-hexane was used to elute the aliphatic fraction (F1), 15 ml 1:1 benzene: n-hexane was used to elute the aromatic fraction (F2). Half of the F1 and half of the F2 fractions were combined to obtain a mixture of aliphatic and aromatic fraction (F3). One hundred microliters of an internal standard, 5-a-androstane (100 lg l1, Sigma) was spiked to F1 and F3 fractions, and the volume in each fraction was concentrated to 0.5 ml under a pure nitrogen stream. The concentrations of petroleum hydrocarbons in F1 and F3 fractions were quantified by GC–FID (HewlettPackard 5890 Gas Chromatograph with a flame ionization detector). The Rtx-5 fused silica capillary column (Restek, Bellefonte, PA with a dimension of 30 m long, 0.32 mm internal diameter and 0.25 lm film thick) was used. The column flow was maintained at a flow rate of 1.5 ml min1. The oven temperature program was as
follows: hold at 50 C for 2 min, increased to 300 C at a rate of 6 C min1 and a final hold at 300 C for 16 min. The injector and detector temperatures were set at 290 and 300 C, respectively. Sample extract (1 ll) was injected in a splitless mode with a 1 min purge off. Before analysing the sample extract, a mixture of standards including n-alkanes (n-hexadecane n-C16, n-octadecane n-C18, n-eicosane n-C20, n-docosane n-C22, n-tetracosane n-C24, n-hexacosane n-C26, n-octacosane n-C28 and n-triacontane n-C30, Alltech) and an internal standard 5-a-androstane were used for calibration. Five points calibration curves using peak areas were obtained and the response factors were used to determine the concentrations of various hydrocarbons in the sample extract, including (i) n-alkanes from n-C14 to n-C33 in F1 (aliphatic) fraction; (ii) isoprenoid alkanes (pristane and phytane); (iii) unresolved complex mixture (UCM), appeared as the hump between the lower baseline and the base of the resolved peaks subtracting the internal and surrogate standards, in F1 and F3 (mixture of aliphatic and aromatic) fractions; and (iv) total petroleum hydrocarbons (TPH), the sum of (i) and (iii), in F1 and F3 fractions. The concentrations of TPH and UCM in F2 (aromatic) fraction were estimated by subtracting the respective values in F1 from that in F3. 2.3. Quality control and statistical analyses The recovery and precision of the analytical procedures were checked by extraction of three replicates of 15 g clean sand (BDH Laboratory Supplies) spiked with eight n-alkane standards described above, each at 7.5 lg. The recoveries for n-alkanes from n-C16 to n-C18 fell between 70% and 90% while the values for n-C20 to n-C30 were between 90% and 96%, comparable to those reported in literature (Colombo et al., 1989; Doskey, 2001). The precision for n-alkanes, in terms of relative standard deviation values, ranging from 0.23% to 2.95% was satisfactory. The method detection limits for n-alkanes were 0.003 lg g1 freeze-dried weight. The efficacy of the analytical procedure for mangrove sediments was also checked by spiking a surrogate standard, o-terphenyl (ChemService) to the sediment prior to extraction. The recovery of the surrogate standard was 102.83% ± 13.56 (n = 70), ranging from 68.02% to 135.07%, which were within the standard range of 40– 140% set by Massachusetts Department of Environmental Protection (1998). To ensure the response factors remained at a similar level throughout GC analyses, a mixture of standards at a mid-point concentration (25 lg l1) was repeatedly measured at each batch of 15–20 samples injected. The relative percent differences between two response factors were kept within 25%, otherwise, a new calibration curve would be prepared. The mean and standard deviation values of the triplicates were calculated. The differences in petroleum
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
hydrocarbon concentrations and diagnostic ratios among three quadrats in each sampling time were determined by a parametric one-way analysis of variance (ANOVA). The differences between healthy and unhealthy saplingsÕ root zone sediments were also compared by one-way ANOVA. Prior to ANOVA test, data were checked for homogeneity of variances by LeveneÕs test and normality by Kolmogorov–Smirnov (Zar, 1999). If the data did not fulfil the above two criteria, they would be transformed either by natural logarithm or squared root. If the ANOVA test was significant at p 6 0.05 level, the least significant differences were calculated. All statistical analyses were done by SPSS 10.0 package (SPSS Inc., USA).
3. Results and discussion 3.1. Vegetation survey In December 2000, more than 80% saplings were either missing (probably washed away) or dead (Table 2). The percentages of missing saplings were high in Q1 while more dead saplings were still found in Q3, suggesting that saplings in Q1, closest to the sea, might have died for some time and the dead saplings were washed away by tides. Most unhealthy individuals were either leafless or with dry brown leaves hanging on the branches. Leaf defoliation was considered as one of the sub-lethal damages by oil spill contamination. Duke et al. (1997) reported that in addition to the 69 ha of mangroves (6% of the total mangrove forests in the Galeta Oil Spill area) died off, an area of 307 ha of the mangroves (34% of the total area) were found to have unusually ‘‘open’’ canopy with reduced density of leaves.
1095
When the dead and unhealthy saplings in Sheung Pak Nai were dug, the roots were black in colour in contrast to the whitish root system in healthy plants, and a strong oil smell was noted. Among saplings, a thin layer of oil was often found floating on the mudflat at low tides. Nevertheless, there was no sign of oil coverage on leaves or branches, indicating that lethal and sub-lethal damages were probably due to oil residues accumulated in sediments, in particular around root zones. The sublethal effects including leaf losses, reduction in shoot and leaf biomass of red mangroves were recorded when the concentrations of oil residues in mangrove sediments were between 100 and 1000 lg g1 dry weight (Levings and Garrity, 1997). The toxic doses of oil residues vary among plant species. Levings and Garrity (1995) reported that most young saplings and mature trees of red mangrove (Rhizophora mangle) and black mangrove (Avicennia germinans) died and defoliated when the area was oiled by No. 6 fuel oil but such oil pollution had less effect on the white mangrove (Laguncularia racemosa). Young Kandelia candel saplings may be more vulnerable to oil damage. One year later, in December 2001, signs of floating oil were not detected on the mudflat. The damaged saplings of Kandelia candel had recovered from their unhealthy condition, at scales between 0 and 0.5, suggesting that re-growth occurred naturally once pollution from illegal oil smuggling activities were stopped and the accumulated oil residues, in particular, the lighter and more toxic fraction were weathered. Jernelov and Linden (1983) also reported that the previously defoliated mangrove area, due to pollution by the Saint Peter oil tanker accident, had recovered with new leaves, flowers and seedlings 3–4 months after the oil spill. The effects of oil contamination on mangroves could be divided into four stages, namely (i) ‘‘initial effect’’: propagules and
Table 2 Survey of Kandelia candel saplings planted on mudflats in three quadrats (Q1, Q2 and Q3) in Sheung Pak Nai mangrove swamp in December 2000 (total number of saplings planted was 400 in each quadrat) Ratings
Conditions
0 0.5
Healthy dense green leaves Dense leaves on the branches but some turn yellow and are unhealthy, a few branches are leafless The density of leaves is fewer than above, some turn yellow, some are wilting, some branches are leafless Leaves are wilting, most branches are leafless but have more leaves than that in rating 3 Dying, very few wilting leaves are still ‘‘hanging’’ on a few branches Plants are dead, all leaves have been shed, branches are dried
1 2 3 Death
Saplings remained (dead + live) Saplings washed away
Number of saplings
% to saplings remained
% of total saplings planted
Q1
Q2
Q3
Q1
Q2
Q3
13 31
16 23
11 29
15 36
12 17
3.9 10.2
Q1 3.25 7.75
Q2 4.0 5.75
Q3 2.75 7.25
15
5
5
18
4
1.8
3.75
1.25
1.25
2
3
2
2
2
0.7
0.5
0.75
0.5
3
1
1
4
1
0.4
0.75
0.25
0.25
21
84
237
25
64
83.2
5.25
21.0
59.25
85 315
132 268
285 115
21.25 78.75
33.0 67.0
71.25 28.75
1096
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
young plants are most likely to die; (ii) ‘‘structural damages’’: trees begin to die; (iii) ‘‘stabilization’’: deterioration of mangroves ceases but no improvement can be detected; and (iv) ‘‘recovery’’: mangroves show signs of improvement in leaf density, colonization and survival rates (Da Silva et al., 1997). The present study shows that the young saplings in Sheung Pak Nai mangrove swamp had gone through these four stages in one yearÕs time.
(Fig. 1). These results suggest that contamination in Sheung Pak Nai mudflat was most likely by fuel oil as their alkane carbon numbers were within the similar ranges of fuel oil (n-C14–n-C25). The n-alkane range for diesel is n-C8–n-C21, No. 2 fuel oil is n-C8–n-C21, and No. 6 fuel oil is n-C12–beyond n-C34 (Bruce and Schmidt, 1994). The concentrations of light n-alkanes ranging from n-C14 to n-C20 were high in all surface sediment samples collected in December 2000 (Fig. 2). The two isoprenoids, pristane and phytane, common indicators for petroleum pollution and dominant saturated hydrocarbons in partially weathered petroleum hydrocarbons (Sauer and Uhler, 1994), also present at elevated concentrations in surface mangrove sediments with values of 0.2 and 0.3 lg g1, respectively. In all surface sediments, LALK/TALK (light to total n-alkanes) ratios varied from 0.36 to 0.51, around 40% of total n-alkanes were made up of light n-alkanes (Table 3). This is another evidence for fuel oil contamination (Snedaker et al., 1995). The CPI (carbon
3.2. Petroleum hydrocarbon concentrations in surface sediments The profiles of aliphatic petroleum hydrocarbons among three quadrats were comparable despite Q3 had more sandy surface sediments. The surface sediments collected from three mudflat areas in December 2000 had a prominent UCM (unresolved complex matrix) hump in the range of n-alkanes between n-C14 and n-C26 with a peak in the range of n-C19 to n-C21
FID1 A, (A:\01092117.D) counts
17
8000
phy 19
I.S. 21
Quadrat 1 23
7000
25
6000 5000
27 29
15
4000
33
UCM
14
3000
31
2000 1000 0 0
10
20
30
40
17
7000
Quadrat 2
phy 19
6000
min
I.S.
FID1 A, (A:\01112881.D) counts
50
21 23 25
5000
27 29
15
4000
31
UCM
14
3000
33
2000 1000 0 0
10
20
30
FID1 A, (A:\01112882.D)
17
counts
phy 19
5000
40
I.S.
Quadrat 3 23
3000
25
27 29
31
UCM
14
min
21
4000
15
50
33
2000 1000 0 0
10
20
30
40
50
min
Fig. 1. GC–FID chromatograms of aliphatic fractions (F1) of surface sediments collected in three mudflat areas (Quadrats 1–3) in December 2000 (the number on top of the peak represents the number of carbon in n-alkanes; phy = phytane; UCM = unresolved complex mixture; IS = internal standard).
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100 December 2000
December 2001 0.5
Quadrat 1
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
0.5
0.5 0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
0.5
0.5
Quadrat 3
Quadrat 3
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Concentrations (µg / g sediment)
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Quadrat 2
0.4
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Concentrations (µg / g sediment)
Quadrat 2
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Concentrations (µg / g sediment)
Quadrat 1
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
0.5
1097
Fig. 2. Profiles of n-alkanes and isoprenoid alkanes, pristane and phytane, in surface sediments collected in three mudflat areas (Quadrats 1–3) in December 2000 and 2001 (mean and standard deviations of triplicates are shown).
preference index) of the two ranges from n-C14 to n-C21 and n-C24 to n-C33 were closer 1 in December 2000 than that in 2001 (Table 3), indicating the dominance of n-alkanes by petrogenic inputs. In December 2000, concentrations of total petroleum hydrocarbons (TPH-F3) and unresolved complex matrix (UCM-F3) in surface sediments collected from three quadrats were not significantly different (Table 4). The average TPH-F3 (70 lg g1 dry weight) and UCM-F3 (63 lg g1 dry weight) values were comparable to the TPH values (75 ± 10 lg g1) in mangrove sediments in Bulwer Island at the mouth of Brisbane River, near heavy industry, refineries and major port area, which caused chlorophyll-deficiency mutations in Avicennia
marina (Duke and Watkinson, 2002). The TPH-F3 and UCM-F3 concentrations in sediments collected in 2000 were higher than the background levels of other muddy mangrove sediments in Hong Kong, which were 26.8 ± 3.3 and 20.7 ± 2.9 lg g1 dry weight, respectively (Wong et al., 2002). The hydrocarbon concentrations recorded in the present study were also higher than that in a mangrove habitat in the Caribbean Island contaminated by oil from shipping activities (Bernard et al., 1996). Tolosa et al. (1996) proposed that sediment UCM concentrations below 10 lg g1 are common in coastal environments distant from hydrocarbon inputs. All these results illustrate that Sheung Pak Nai mudflat had been contaminated by fuel oil, probably from
1098
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
Table 3 Diagnostic ratios of CPI14–21, CPI24–33, and LALK/TALK in surface sediments collected in December 2000 and 2001 (mean and standard deviations of triplicates are shown) Quadrats
Quadrat 1 Quadrat 2 Quadrat 3
LALK/TALK
CPI14–21
CPI24–33
2000
2001
2000
2001
2000
2001
0.43 ± 0.07 0.40 ± 0.04 0.44 ± 0.07
0.15 ± 0.01 0.15 ± 0.02 0.21 ± 0.12
1.20 ± 0.21 1.56 ± 0.17 1.68 ± 0.74
1.58 ± 0.23 2.06 ± 0.21 4.93 ± 1.46
1.44 ± 0.15 2.10 ± 0.44 2.18 ± 0.11
2.50 ± 0.18 3.29 ± 0.21 3.52 ± 0.01
LALK/TALK (light alkanes/total alkanes) = (sum of C14–20)/(sum of C14–33); CPI14–21 (carbon preference index in the range of n-C14 to n-C21) = (sum of C15, C17, C19 and C21)/(sum of C14, C16, C18 and C20); CPI24-33 (carbon preference index in the range of n-C24 to n-C33) = (sum of C25, C27, C29, C31and C33)/(sum of C24, C26, C28 C30, and C32).
Table 4 Concentrations of petroleum hydrocarbons in aliphatic (F1), aromatic (F2) and total (F3) fractions of the surface sediments collected from three quadrats in December 2000 and 2001 (mean and standard deviations of triplicates are shown) Quadrats
F1 (aliphatic) 2000
F2 (aromatic) 2001
F3 (total = aliphatic + aromatic)
2000
2001
2000
2001
Total petroleum hydrocarbon (TPH = resolved + unresolved) Quadrat 1 50.9 ± 6.7 30.5 ± 6.4 Quadrat 2 57.3 ± 14.0 37.1 ± 7.0 Quadrat 3 44.6 ± 10.8 36.9 ± 5.9
15.2 ± 6.2 21.9 ± 5.4 18.3 ± 6.0
12.9 ± 3.3 10.6 ± 3.4 12.2 ± 2.4
66.2 ± 3.3 79.2 ± 17.4 62.8 ± 16.8
43.4 ± 8.6 47.7 ± 10.2 49.1 ± 6.4
Unresolved complex matrix (UCM) Quadrat 1 43.7 ± 6.5 Quadrat 2 50.8 ± 12.2 Quadrat 3 39.3 ± 9.4
14.9 ± 6.3 21.4 ± 5.5 17.4 ± 6.1
12.2 ± 2.8 9.8 ± 3.1 10.6 ± 2.3
58.6 ± 2.2 72.2 ± 15.9 56.7 ± 15.3
39.8 ± 7.8 42.9 ± 8.9 42.9 ± 5.2
27.6 ± 5.9 33.1 ± 6.2 32.2 ± 4.9
leakage and discharge of illegal smuggling operations in the Pearl River Estuary region which was active between 1998 and 2000. In 2001, concentrations of petroleum hydrocarbons and UCM in both aliphatic and aromatic fractions declined significantly (Table 4), suggesting that oil residues in surface sediments were weathered by physical/chemical processes and biodegradation. Fuel oil with a higher proportion of light hydrocarbons than that in crude oil would be easier to degrade. The loss of fuel oil in mudflat sediments at Sheung Pak Nai mangrove swamp was faster than the weathering of spilled crude oil at Yi O, another mangrove swamp in Hong Kong (Wong et al., 2002). Zhang and Lin (1996) also found that up to 70% diesel oil applied to mangrove sediments were degraded in one month under laboratory conditions. Although concentrations of total n-alkanes maintained at around 2–3 lg g1 in these two years, n-alkanes had shifted from petrogenic to natural biogenic sources as reflected in their profiles (Fig. 2): (i) most light n-alkanes except n-C17 disappeared and concentrations of light n-alkanes declined, (ii) the two isoprenoids (pristane and phytane) dropped by 80% or more; and (iii) the odd number n-alkanes from n-C25 to n-C33 were dominant in all surface sediments. The n-C17 peak indicates the petroleum hydrocarbons from phytoplanktons (Doskey, 2001) while the mangrove debris contribute to n-C25, n-C27, n-C29, n-C31 and n-C33 (Nishigima et al.,
2001). Quadrat 3 had an extremely large peak at nC17, suggesting a significant input from phytoplankton. During the 2001Õs sampling, a thick mat of Halophilia ovata (a local seagrass species) with healthy leaves and well-developed roots in the top 1 cm of the sediment was found in Q3 while the amounts of H. ovata were less in the other two quadrats. The ratios of LALK/TALK dropped from 0.4 in 2000 to 0.15 in 2001 while CPI values increased significantly (Table 3). These results also indicate that the petrogenic light n-alkanes had been lost by weathering and n-alkanes was dominant by the odd carbon number with large molecular sizes. 3.3. Petroleum hydrocarbon concentrations in root zone sediments Total petroleum hydrocarbons concentrations in root zone sediments of the unhealthy saplings were higher than that in healthy saplings although variations between saplings of the same category were large (Table 5). For most of the root zone sediments, petroleum hydrocarbon concentrations were lower than that in surface sediments, suggesting that some degradation had already taken place between December 2000 and March 2001. One unhealthy sapling had exceptionally high concentrations of petroleum hydrocarbons in its root zone sediments, greater than 1000 lg g1 dry weight (Table 5). A very strong fuel oil smell was noted when this
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
1099
Table 5 Concentrations of petroleum hydrocarbons and diagnostic ratios in root zone sediments of healthy and unhealthy saplings in March 2001 Healthy saplings Sapling 1
Unhealthy saplings
Sapling 2
Sapling 3
Petroleum hydrocarbon concentrations (lg g1 freeze-dry wt) TPH-F1 17.83 51.43 21.27 TPH-F2 5.53 8.17 11.42 TPH-F3 23.36 59.61 32.69 Diagnostic ratios CPI14–21 CPI24–33 LALK/TALK
0.91 2.20 0.23
0.95 2.38 0.30
1.12 2.26 0.32
Mean ± SD
Sapling 1
Sapling 2
Sapling 3
Mean ± SD
30.18 ± 15.09 8.37 ± 2.41 38.15 ± 15.37
48.27 19.14 67.41
33.85 12.61 46.46
505.57 604.13 1111.7
195.9 ± 219.1 211.9 ± 277.3 408.5 ± 497.3
0.99 ± 0.11 2.28 ± 0.07 0.28 ± 0.04
1.10 2.03 0.59
1.26 2.42 0.31
0.93 1.35 0.90
1.09 ± 0.14 1.93 ± 0.44 0.60 ± 0.24
TPH-F1: aliphatic petroleum hydrocarbons; TPH-F2: aromatic petroleum hydrocarbons; TPH-F3: total petroleum hydrocarbons (sum of aliphatic and aromatic); CPI14–21 (carbon preference index in the range of n-C14 to n-C21) = (sum of C15, C17, C19 and C21)/(sum of C14, C16, C18 and C20); CPI24–33 (carbon preference index in the range of n-C24 to n-C33) = (sum of C25, C27, C29, C31and C33)/(sum of C24, C26, C28 C30, and C32); LALK/ TALK (light alkanes/total alkanes) = (sum of C14–20)/(sum of C14–33).
counts
FID1 A, (A:\01111728.D)
16
70000
17 18 19
15
60000
20
50000 40000
14
21
30000
22
20000
UCM
10000 0 0
10
20
30
40
50
min
Fig. 3. GC–FID chromatograms of aliphatic fractions (F1) of root zone sediments collected from an unhealthy sapling in Quadrat 3 in March 2001. (The number on top of the peak represents the number of carbon in n-alkanes; UCM = unresolved complex mixture; the petroleum hydrocarbon level was exceptionally high with a clear pattern of weathered fuel oil.)
unhealthy sapling was dug up from the mudflat. This particular root zone sediment was mainly composed of partially weathered fuel oil as it had a large UCM in the range of n-C14 to n-C26 (Fig. 3), very high peaks of n-alkanes from n-C14 to n-C22 and the two isoprenoids, n-alkanes were dominant by short chain n-alkanes (LALK/TALK index was 0.9) and CPI values were close to 1 (Table 5). These strongly prove that Sheung Pak Nai mangrove swamp was suffered from fuel oil contamination and some oil was migrated into deeper sediments around the roots leading to severe death of young K. candel saplings.
4. Conclusions The present study demonstrates a mangrove swamp in Hong Kong SAR was contaminated by fuel oil due to illegal smuggling activities and discharge in Pearl River Estuary. The saplings of Kandelia candel, a dominant mangrove species planted by local villagers were suffered from oil pollution with massive mortality and defoliation in December 2000 although the plant surfaces were not covered by oil and the contamination
was not in a large scale. The concentrations of total petroleum hydrocarbons in surface sediments were around 60–80 lg g1 dry weight in December 2000 with one root zone sediment sample reached an exceptionally high concentration of >1000 lg g1 dry weight. In December 2001, the unhealthy saplings had recovered with new leaves, the total petroleum hydrocarbon levels in sediments also dropped to 30–40 lg g1 dry weight due to weathering processes and the control of anthropogenic fuel oil contamination by local authorities in early 2001. The GC–FID chromatograms, the profiles of n-alkanes and the diagnostic ratios in sediments all reveal that the petroleum hydrocarbons had shifted from petrogenic to biogenic sources in December 2001.
Acknowledgements The work described in this paper was fully supported by a grant from the Research Grant Council, HKSAR (Ref. no. CityU 1110/02M). The authors gratefully acknowledge Mr. Benz Chan for his assistance in the GC analyses.
1100
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
References Bernard, D., Pascaline, H., Jeremie, J.J., 1996. Distribution and origin of hydrocarbons in sediments from lagoons with fringing mangrove communities. Marine Pollution Bulletin 32, 734–739. Bruce, L.G., Schmidt, G.W., 1994. Hydrocarbon fingerprinting for application in forensic geology: review with case studies. American Association of Petroleum Geologists Bulletin 78, 1692– 1710. Burns, K.A., Codi, S., 1998. Contrasting impacts of localised versus catastrophic oil spills in mangrove sediments. Mangroves and Salt Marshes 2, 63–74. Burns, K.A., Garrity, S.D., Jorissen, D., MacPherson, J., Stoelting, M., Tierney, J., Yelle-Simmons, L., 1994. The Galeta oil spill. II. Unexpected persistence of oil trapped in mangrove sediments. Estuarine, Coastal and Shelf Science 38, 349–364. Burns, K.A., Codi, S., Duke, N.C., 2000. Gladstone, Australia field studies: weathering and degradation of hydrocarbons in oiled mangrove and salt marsh sediments with and without the application of an experimental bioremediation protocol. Marine Pollution Bulletin 41, 392–402. Colombo, J.C., Pelletier, E., Brochu, C., Khalil, M., 1989. Determination of hydrocarbon sources using n-alkane and polyaromatic hydrocarbon distribution indices. Case study: Rio de la Plata Estuary, Argentina. Environmental Science and Technology 23, 888–894. Da Silva, E.M., Peso-Aguiar, M.C., Navarro, M.F.T., Chastinet, C.B.A., 1997. Impact of petroleum pollution on aquatic coastal ecosystems in Brazil. Environmental Toxicology and Chemistry 16, 112–118. Doskey, P.V., 2001. Spatial variation and chronologies of aliphatic hydrocarbons in Lake Michigan sediments. Environmental Science and Technology 35, 247–254. Duke, N.C., Watkinson, A.J., 2002. Chlorophyll-deficiency propagules of Avicennia marina and apparent longer term deterioration of mangrove fitness in oil-polluted sediments. Marine Pollution Bulletin 44, 1269–1276. Duke, N.C., Pinzon, Z.S., Prada, M.C., 1997. Large-scale damage to mangrove forests following two large oil spills in Panama. Biotropica 29, 2–14. Garrity, S.D., Levings, S.C., Burns, K.A., 1994. The Galeta oil spill. I. Long-term effects on the physical structure of the mangrove fringe. Estuarine, Coastal and Shelf Science 38, 327–348. Hester, H.W., Mendelssohn, I.A., 2000. Long-term recovery of a Louisiana brackish marsh plant community from oil-spill impact: vegetation response and mitigation effects of marsh surface elevation. Marine Environmental Research 49, 233–254. Jernelov, A., Linden, O., 1983. The effect of oil pollution on mangroves and fisheries in Ecuador and Colombia. In: Teas, H.J. (Ed.), Tasks for Vegetation Science, Biology and Ecology of
Mangroves, vol. 8. Dr. W. Junk Publisher, The Hague, pp. 185– 188. Ke, L., Wong, T.W.Y., Wong, Y.S., Tam, N.F.Y., 2002. Fate of polycyclic aromatic hydrocarbon (PAH) contamination in a mangrove swamp in Hong Kong following an oil spill. Marine Pollution Bulletin 45, 339–347. Levings, S.C., Garrity, S.D., 1995. Oiling of mangrove keys in the 1993 Tampa Bay oil spill. In: Proceedings of the 1995 International Oil Spill Conference. American Petroleum Institute, Washington, DC. Levings, S.C., Garrity, S.D., 1997. Sublethal injury to red mangroves two years after oiling. In: Proceedings of the 1997 International Oil Spill Conference. American Petroleum Institute, Washington, DC, pp. 1040–1041. Massachusetts Department of Environmental Protection, 1998. Method for the Determination of Extractable Petroleum Hydrocarbons (EPH), Commonwealth of Massachusetts, Massachusetts, 41pp. Nishigima, F.N., Weber, R.R., Bicego, M.C., 2001. Aliphatic and aromatic hydrocarbons in sediments of Santos and Cananeia, SP, Brazil. Marine Pollution Bulletin 42, 1064–1072. Sauer, T.C., Uhler, A.D., 1994. Pollutant source identification and allocation: advances in hydrocarbon fingerprinting. Remediation 4, 25–45. Snedaker, S.C., Glynn, P.W., Rumbold, D.G., Corcoran, E.F., 1995. Distribution of n-alkanes in marine samples from Southeast Florida. Marine Pollution Bulletin 30, 83–89. Tam, N.F.Y., Wong, Y.S., Lu, C.Y., Berry, R., 1997. Mapping and characterization of mangrove plant communities in Hong Kong. Hydrobiologia 352, 25–37. Tolosa, I., Bayona, J.M., Albaiges, J., 1996. Aliphatic and polycyclic aromatic hydrocarbons and sulphur/oxygen derivatives in northwest Mediterranean sediments: spatial and temporal variability, fluxes and budgets. Environmental Science and Technology 30, 2495–2503. Wang, Z., Fingas, M., Li, K., 1994. Fractionation of a light crude oil and identification and quantitation of aliphatic, aromatic, and biomarker compounds by GC–FID and GC–MS. Journal of Chromatographic Science 32, 361–366. Wong, T.W.Y., Ke, L., Wong, Y.S., Tam, N.F.Y., 2002. Study of the sediment contamination levels in a mangrove swamp polluted by marine oil spill. In: Proceedings of the 25th Arctic and Marine Oil Spill Technical Seminar, Environment Canada, Ottawa, Canada, pp. 73–90. Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice Hall, Inc., NJ, USA. Zhang, T.C., Lin, P., 1996. The soil microbial function of Kandelia candel mangroves: degradation of diesel oil. In: Morton, B. (Ed.), Proceedings of the Third International Conference on the Marine Biology of the South China Sea. Hong Kong University Press, Hong Kong, pp. 389–395.
A case study on fuel oil contamination in a mangrove swamp in Hong Kong Nora F.Y. Tam, Teresa W.Y. Wong, Y.S. Wong
*
Department of Biology and Chemistry, City University of Hong Kong, Tat Chee Avenue, Kowloon, Hong Kong
Abstract Mangroves commonly found along tropical and subtropical coastlines are susceptible to oil pollution. In December 2000, around 500 1 m tall Kandelia candel saplings at the age of 3–5 years old located at the foreshore region of Sheung Pak Nai swamp, Hong Kong SAR, were found to be damaged by oil pollution. More than 80% of the saplings were either dead or washed away and leaving less than 5% healthy saplings with dense green leaves. Elevated concentrations of light n-alkanes (ranging from n-C14 to n-C20), pristane and phytane were recorded in surface sediments collected in December 2000. The ratio between light and total n-alkanes was 0.4. The total petroleum hydrocarbons (60–80 lg g1 TPH) and unresolved complex mixtures (60–70 lg g1 UCM) were higher than the background values of other mangrove sediments in Hong Kong, which were 40 and 20 lg g1, respectively. In certain root zone sediments, TPH concentrations were above 1000 lg g1. These results suggest that surface sediments in Sheung Pak Nai were contaminated by petroleum oil, most likely by illegal discharge of fuel oil which occurred between 1998 and 2002. One year later, in December 2001, unhealthy saplings had recovered and re-grown. The concentrations of TPH and UCM in sediments declined to around 40 lg g1, pristane and phytane dropped by 80%, and the ratio of light to total n-alkanes was 0.15, suggesting that residual oil in sediments was weathered leading to a remarkable recovery of the unhealthy saplings. 2005 Elsevier Ltd. All rights reserved. Keywords: Kandelia candel; Petroleum hydrocarbons; Fuel oil; Alkane; UCM
1. Introduction Oil contamination, irrespective of its scale, has been recognized as one of the most serious threats to marine environments, particularly in tidal flats. Mangrove wetlands commonly found along most low wave energy shorelines in tropical and subtropical regions are well known for their high vulnerability to oil spills. The most serious oil spill accidents which affected mangrove habitats was the Galeta oil spill took place at the east of the Caribbean entrance to the Panama Canal in 1986, where about 8 million litres of crude oil were leaked out from a
*
Corresponding author. Tel.: +86 852 2788 9377; fax: +86 852 2788 9002. E-mail address: [email protected] (Y.S. Wong). 0025-326X/$ - see front matter 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.marpolbul.2005.06.005
storage tank of a nearby oil refinery (Garrity et al., 1994). Mangrove trees were found to be killed or with growth impaired due to physical and chemical effects of the spilled oil. Floating oil settles and smothers the breathing and feeder roots. Oil dispersed on tree roots did not only cause death of some trees but also reduced growth of the survivors including defoliation and losses in canopy density, malformation of fruits, sub-lethal biochemical and molecular damages (Da Silva et al., 1997; Duke et al., 1997). Previous oil spill studies showed that oil residues in sediments at concentrations above 100 lg g1 dry weight would cause sub-lethal effects to mangrove trees (Burns et al., 1994; Levings and Garrity, 1997). The unique anoxic and waterlogged features of mangrove sediments reduced oil degradation, especially the residual aromatic hydrocarbons, whereas crude oil residues accumulated and persisted in deep
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
mud for up to a period of 20 years (Burns et al., 1994, 2000). The mortality and/or damages due to oil contamination depend on the type, quantity and weathered state of the oil, the mangrove species, and the prevailing climatic and tidal conditions. The scale and type of oil spill also affect the rate of dissipation and degradation of the residual oil in sediments (Burns and Codi, 1998). Previous studies on oil contamination in mangroves were mostly focussed on large scale crude oil spills as these were often perceived to be major threats. The contamination by leakage of fuel oil, improper disposal of used motor oil and illegal discharge of spent lubricating oil have received less attention although they account for a significant percentage of all oil inputs in harbours and other coastal waterways. Most oil spill studies were carried out in Central America, South America and Australia, while little information is available in mangroves around South China Sea especially the Pearl River Estuary. Only one crude oil spill accident in a mangrove swamp has been reported in Hong Kong SAR (Ke et al., 2002; Wong et al., 2002) despite mangroves in these areas are under all kinds of risks of oil pollution such as rapid industrialization and urbanization, heavy shipping transportation, and illegal oil smuggling operations. The present study therefore aims to investigate the oil contamination and its damages on young saplings of Kandelia candel, a dominant mangrove plant species in Sheung Pak Nai swamp. The degree of oil weathering and the recovery of mangrove saplings were also examined. Sheung Pak Nai mangrove swamp (1137.3 0 E, 216.9 0 N) situates in North Western part of the New Territories, Hong Kong SAR. The stand spreads over 4 km of coastlines along Deep Bay and the total mangrove area is around 6.34 ha (Tam et al., 1997). The substrate is made up of soft and deep fluidized mud especially near the foreshore region. The swamp is dominated by Kandelia candel, which has been planted since 1980s by local farmers who rely on fisheries from their oyster farms and aquaculture ponds as well as pigsties for their living. The mature trees with an average height of 4–5 m are healthy and produce large quantities of propagules. Saplings of 3–5 years old have also been planted at the low tidal mudflat region in recent years. Across Deep Bay, the opposite shore is Shekou, a densely populated and highly industrialized district in Shenzhen, PeopleÕs Republic of China. In December 2000, the saplings planted at the mudflat region were found either missing, dead or in unhealthy condition. Patches of oil film were seen on the mudflat at low tides with a strong oil smell. Illegal smuggling of diesel and fuel oil in the Pearl River Estuary Region, in particular Deep Bay was often reported from 1998 to 2000. The merchant ships and fishing boats unloaded the oil from larger ships at the border between Hong Kong SAR and
1093
Mainland China, and oil was discharged directly to the sea if they were detected by the Marine Police. The scale was around thousand tons of fuel oil in each smuggling operation. The smuggling activities have been under control from year 2001 onwards because of the joint efforts between the Hong Kong SAR and the Mainland administrations.
2. Materials and methods 2.1. Vegetation survey and sediment sampling Three 10 m · 10 m quadrats, located at different positions at the low tidal mudflat planted with saplings of K. candel in Sheung Pak Nai mangrove swamp, were sampled in December 2000 and December 2001. The saplings were planted in quadrats with regular distance of 0.5 m, each containing 400 saplings. Quadrat 1 was at the fringe of the mudflat, nearest to the sea with muddy sediments. Quadrat 2 was 8 m away from the fringe and sediments were less silty than that in Q1. Quadrat 3 was at a tidal position similar to that of Q2 but was near to a sandy path paved by local villages, and its sediments were sandy with lower water content than Q1 and Q2 (Table 1). In each quadrat, the number of saplings remained was counted and the health condition of each sapling was recorded according to a standard scaling system for assessing oil spill impact on a brackish marsh plant community (Hester and Mendelssohn, 2000). Surface sediment samples (1–10 cm) were randomly collected using a solvent-cleaned stainless steel spatula, and three replicate samples (each sample had more than 1 kg fresh sediments) were collected from each quadrat. In March 2001, triplicate root zone sediments of healthy and unhealthy saplings were sampled. This was done by digging up three individuals of each type of saplings from the mudflat. Bulk sediments were removed by shaking the roots, and sediments around the roots were then harvested. Dead saplings were not sampled because most roots were rotten and root zone sediments were difficult to obtain. Sediments were freeze-dried, ground
Table 1 Particle size distribution (%) and water content (% dry weight) of sediments in three quadrats in Sheung Pak Nai mangrove swamp Measurements
Quadrat 1
Quadrat 2
Quadrat 3
Particle sizes > 2 mm 1 mm–2 mm 500 lm–1 mm 250 lm–500 lm 125 lm–250 lm 63 lm–125 lm <63 lm Water content
4.29 6.56 6.53 16.08 25.11 7.46 33.97 40.81
9.78 8.43 7.09 18.26 16.48 15.55 24.41 40.60
23.23 19.04 12.71 12.28 11.08 4.93 16.73 34.85
1094
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
into powder and sieved through a 2 mm sieve, then stored in dark glass bottles at room temperature prior to analyses for petroleum hydrocarbons. Sub-samples of the fresh sediments were air-dried and analysed for pH, soil texture (wet sieving technique) and organic matter concentrations (loss on ignition at 550 C for 6 h). 2.2. Determination of petroleum hydrocarbons in sediments Sediment samples were extracted following the protocol described by Wang et al. (1994) with minor modification. Five to 20 g freeze-dried sediments, depending on the level of contamination, were transferred to a 250 ml pre-ashed conical flask. Around 1 g of activated copper (copper powder purchased from BDH was activated by 2 M hydrochloric acid and subsequently washed by distilled water and solvents) was added to remove sulphur. A surrogate standard, o-terphenyl (ChemService) at a concentration of 69.5 lg l1, was spiked prior to extraction. The sediment was then sonicated in an ultrasonic bath with a 60 ml mixture of n-hexane and dichloromethane (1:1 v/v) for 30 min. The ultrasonic extraction step was repeated twice with 60 ml dichloromethane each time. The extract was combined and filtered through a Whatman no. 541 filter paper to remove large particles. The extract was then concentrated to around 1 ml by rotary evaporation and dried to 0.2–0.4 ml under a pure nitrogen stream. The cleanup and fractionation of the extract was done by a self-packed silica gel column, a 25 ml acetone-rinsed burette (1 cm · 45 cm) contained 3 g firmly packed activated silica gel (72–230 mesh, Sigma) and a layer of 1–2 cm Na2SO4 at the surface to absorb water. Twenty milliliters hexane was added to condition the column. The concentrated extract was then applied to the column with an additional 3 ml hexane for complete transfer. All the solvent eluted up to this point was discarded. When the Na2SO4 layer was nearly exposed to air, 12 ml n-hexane was used to elute the aliphatic fraction (F1), 15 ml 1:1 benzene: n-hexane was used to elute the aromatic fraction (F2). Half of the F1 and half of the F2 fractions were combined to obtain a mixture of aliphatic and aromatic fraction (F3). One hundred microliters of an internal standard, 5-a-androstane (100 lg l1, Sigma) was spiked to F1 and F3 fractions, and the volume in each fraction was concentrated to 0.5 ml under a pure nitrogen stream. The concentrations of petroleum hydrocarbons in F1 and F3 fractions were quantified by GC–FID (HewlettPackard 5890 Gas Chromatograph with a flame ionization detector). The Rtx-5 fused silica capillary column (Restek, Bellefonte, PA with a dimension of 30 m long, 0.32 mm internal diameter and 0.25 lm film thick) was used. The column flow was maintained at a flow rate of 1.5 ml min1. The oven temperature program was as
follows: hold at 50 C for 2 min, increased to 300 C at a rate of 6 C min1 and a final hold at 300 C for 16 min. The injector and detector temperatures were set at 290 and 300 C, respectively. Sample extract (1 ll) was injected in a splitless mode with a 1 min purge off. Before analysing the sample extract, a mixture of standards including n-alkanes (n-hexadecane n-C16, n-octadecane n-C18, n-eicosane n-C20, n-docosane n-C22, n-tetracosane n-C24, n-hexacosane n-C26, n-octacosane n-C28 and n-triacontane n-C30, Alltech) and an internal standard 5-a-androstane were used for calibration. Five points calibration curves using peak areas were obtained and the response factors were used to determine the concentrations of various hydrocarbons in the sample extract, including (i) n-alkanes from n-C14 to n-C33 in F1 (aliphatic) fraction; (ii) isoprenoid alkanes (pristane and phytane); (iii) unresolved complex mixture (UCM), appeared as the hump between the lower baseline and the base of the resolved peaks subtracting the internal and surrogate standards, in F1 and F3 (mixture of aliphatic and aromatic) fractions; and (iv) total petroleum hydrocarbons (TPH), the sum of (i) and (iii), in F1 and F3 fractions. The concentrations of TPH and UCM in F2 (aromatic) fraction were estimated by subtracting the respective values in F1 from that in F3. 2.3. Quality control and statistical analyses The recovery and precision of the analytical procedures were checked by extraction of three replicates of 15 g clean sand (BDH Laboratory Supplies) spiked with eight n-alkane standards described above, each at 7.5 lg. The recoveries for n-alkanes from n-C16 to n-C18 fell between 70% and 90% while the values for n-C20 to n-C30 were between 90% and 96%, comparable to those reported in literature (Colombo et al., 1989; Doskey, 2001). The precision for n-alkanes, in terms of relative standard deviation values, ranging from 0.23% to 2.95% was satisfactory. The method detection limits for n-alkanes were 0.003 lg g1 freeze-dried weight. The efficacy of the analytical procedure for mangrove sediments was also checked by spiking a surrogate standard, o-terphenyl (ChemService) to the sediment prior to extraction. The recovery of the surrogate standard was 102.83% ± 13.56 (n = 70), ranging from 68.02% to 135.07%, which were within the standard range of 40– 140% set by Massachusetts Department of Environmental Protection (1998). To ensure the response factors remained at a similar level throughout GC analyses, a mixture of standards at a mid-point concentration (25 lg l1) was repeatedly measured at each batch of 15–20 samples injected. The relative percent differences between two response factors were kept within 25%, otherwise, a new calibration curve would be prepared. The mean and standard deviation values of the triplicates were calculated. The differences in petroleum
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
hydrocarbon concentrations and diagnostic ratios among three quadrats in each sampling time were determined by a parametric one-way analysis of variance (ANOVA). The differences between healthy and unhealthy saplingsÕ root zone sediments were also compared by one-way ANOVA. Prior to ANOVA test, data were checked for homogeneity of variances by LeveneÕs test and normality by Kolmogorov–Smirnov (Zar, 1999). If the data did not fulfil the above two criteria, they would be transformed either by natural logarithm or squared root. If the ANOVA test was significant at p 6 0.05 level, the least significant differences were calculated. All statistical analyses were done by SPSS 10.0 package (SPSS Inc., USA).
3. Results and discussion 3.1. Vegetation survey In December 2000, more than 80% saplings were either missing (probably washed away) or dead (Table 2). The percentages of missing saplings were high in Q1 while more dead saplings were still found in Q3, suggesting that saplings in Q1, closest to the sea, might have died for some time and the dead saplings were washed away by tides. Most unhealthy individuals were either leafless or with dry brown leaves hanging on the branches. Leaf defoliation was considered as one of the sub-lethal damages by oil spill contamination. Duke et al. (1997) reported that in addition to the 69 ha of mangroves (6% of the total mangrove forests in the Galeta Oil Spill area) died off, an area of 307 ha of the mangroves (34% of the total area) were found to have unusually ‘‘open’’ canopy with reduced density of leaves.
1095
When the dead and unhealthy saplings in Sheung Pak Nai were dug, the roots were black in colour in contrast to the whitish root system in healthy plants, and a strong oil smell was noted. Among saplings, a thin layer of oil was often found floating on the mudflat at low tides. Nevertheless, there was no sign of oil coverage on leaves or branches, indicating that lethal and sub-lethal damages were probably due to oil residues accumulated in sediments, in particular around root zones. The sublethal effects including leaf losses, reduction in shoot and leaf biomass of red mangroves were recorded when the concentrations of oil residues in mangrove sediments were between 100 and 1000 lg g1 dry weight (Levings and Garrity, 1997). The toxic doses of oil residues vary among plant species. Levings and Garrity (1995) reported that most young saplings and mature trees of red mangrove (Rhizophora mangle) and black mangrove (Avicennia germinans) died and defoliated when the area was oiled by No. 6 fuel oil but such oil pollution had less effect on the white mangrove (Laguncularia racemosa). Young Kandelia candel saplings may be more vulnerable to oil damage. One year later, in December 2001, signs of floating oil were not detected on the mudflat. The damaged saplings of Kandelia candel had recovered from their unhealthy condition, at scales between 0 and 0.5, suggesting that re-growth occurred naturally once pollution from illegal oil smuggling activities were stopped and the accumulated oil residues, in particular, the lighter and more toxic fraction were weathered. Jernelov and Linden (1983) also reported that the previously defoliated mangrove area, due to pollution by the Saint Peter oil tanker accident, had recovered with new leaves, flowers and seedlings 3–4 months after the oil spill. The effects of oil contamination on mangroves could be divided into four stages, namely (i) ‘‘initial effect’’: propagules and
Table 2 Survey of Kandelia candel saplings planted on mudflats in three quadrats (Q1, Q2 and Q3) in Sheung Pak Nai mangrove swamp in December 2000 (total number of saplings planted was 400 in each quadrat) Ratings
Conditions
0 0.5
Healthy dense green leaves Dense leaves on the branches but some turn yellow and are unhealthy, a few branches are leafless The density of leaves is fewer than above, some turn yellow, some are wilting, some branches are leafless Leaves are wilting, most branches are leafless but have more leaves than that in rating 3 Dying, very few wilting leaves are still ‘‘hanging’’ on a few branches Plants are dead, all leaves have been shed, branches are dried
1 2 3 Death
Saplings remained (dead + live) Saplings washed away
Number of saplings
% to saplings remained
% of total saplings planted
Q1
Q2
Q3
Q1
Q2
Q3
13 31
16 23
11 29
15 36
12 17
3.9 10.2
Q1 3.25 7.75
Q2 4.0 5.75
Q3 2.75 7.25
15
5
5
18
4
1.8
3.75
1.25
1.25
2
3
2
2
2
0.7
0.5
0.75
0.5
3
1
1
4
1
0.4
0.75
0.25
0.25
21
84
237
25
64
83.2
5.25
21.0
59.25
85 315
132 268
285 115
21.25 78.75
33.0 67.0
71.25 28.75
1096
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
young plants are most likely to die; (ii) ‘‘structural damages’’: trees begin to die; (iii) ‘‘stabilization’’: deterioration of mangroves ceases but no improvement can be detected; and (iv) ‘‘recovery’’: mangroves show signs of improvement in leaf density, colonization and survival rates (Da Silva et al., 1997). The present study shows that the young saplings in Sheung Pak Nai mangrove swamp had gone through these four stages in one yearÕs time.
(Fig. 1). These results suggest that contamination in Sheung Pak Nai mudflat was most likely by fuel oil as their alkane carbon numbers were within the similar ranges of fuel oil (n-C14–n-C25). The n-alkane range for diesel is n-C8–n-C21, No. 2 fuel oil is n-C8–n-C21, and No. 6 fuel oil is n-C12–beyond n-C34 (Bruce and Schmidt, 1994). The concentrations of light n-alkanes ranging from n-C14 to n-C20 were high in all surface sediment samples collected in December 2000 (Fig. 2). The two isoprenoids, pristane and phytane, common indicators for petroleum pollution and dominant saturated hydrocarbons in partially weathered petroleum hydrocarbons (Sauer and Uhler, 1994), also present at elevated concentrations in surface mangrove sediments with values of 0.2 and 0.3 lg g1, respectively. In all surface sediments, LALK/TALK (light to total n-alkanes) ratios varied from 0.36 to 0.51, around 40% of total n-alkanes were made up of light n-alkanes (Table 3). This is another evidence for fuel oil contamination (Snedaker et al., 1995). The CPI (carbon
3.2. Petroleum hydrocarbon concentrations in surface sediments The profiles of aliphatic petroleum hydrocarbons among three quadrats were comparable despite Q3 had more sandy surface sediments. The surface sediments collected from three mudflat areas in December 2000 had a prominent UCM (unresolved complex matrix) hump in the range of n-alkanes between n-C14 and n-C26 with a peak in the range of n-C19 to n-C21
FID1 A, (A:\01092117.D) counts
17
8000
phy 19
I.S. 21
Quadrat 1 23
7000
25
6000 5000
27 29
15
4000
33
UCM
14
3000
31
2000 1000 0 0
10
20
30
40
17
7000
Quadrat 2
phy 19
6000
min
I.S.
FID1 A, (A:\01112881.D) counts
50
21 23 25
5000
27 29
15
4000
31
UCM
14
3000
33
2000 1000 0 0
10
20
30
FID1 A, (A:\01112882.D)
17
counts
phy 19
5000
40
I.S.
Quadrat 3 23
3000
25
27 29
31
UCM
14
min
21
4000
15
50
33
2000 1000 0 0
10
20
30
40
50
min
Fig. 1. GC–FID chromatograms of aliphatic fractions (F1) of surface sediments collected in three mudflat areas (Quadrats 1–3) in December 2000 (the number on top of the peak represents the number of carbon in n-alkanes; phy = phytane; UCM = unresolved complex mixture; IS = internal standard).
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100 December 2000
December 2001 0.5
Quadrat 1
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
0.5
0.5 0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
0.5
0.5
Quadrat 3
Quadrat 3
0.4
0.4
0.3
0.3
0.2
0.2
0.1
0.1
0.0
0.0
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Concentrations (µg / g sediment)
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Quadrat 2
0.4
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Concentrations (µg / g sediment)
Quadrat 2
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
Concentrations (µg / g sediment)
Quadrat 1
C14 C15 C16 C17 pristane C18 phytane C19 C20 C21 C22 C23 C24 C25 C26 C27 C28 C29 C30 C31 C32 C33
0.5
1097
Fig. 2. Profiles of n-alkanes and isoprenoid alkanes, pristane and phytane, in surface sediments collected in three mudflat areas (Quadrats 1–3) in December 2000 and 2001 (mean and standard deviations of triplicates are shown).
preference index) of the two ranges from n-C14 to n-C21 and n-C24 to n-C33 were closer 1 in December 2000 than that in 2001 (Table 3), indicating the dominance of n-alkanes by petrogenic inputs. In December 2000, concentrations of total petroleum hydrocarbons (TPH-F3) and unresolved complex matrix (UCM-F3) in surface sediments collected from three quadrats were not significantly different (Table 4). The average TPH-F3 (70 lg g1 dry weight) and UCM-F3 (63 lg g1 dry weight) values were comparable to the TPH values (75 ± 10 lg g1) in mangrove sediments in Bulwer Island at the mouth of Brisbane River, near heavy industry, refineries and major port area, which caused chlorophyll-deficiency mutations in Avicennia
marina (Duke and Watkinson, 2002). The TPH-F3 and UCM-F3 concentrations in sediments collected in 2000 were higher than the background levels of other muddy mangrove sediments in Hong Kong, which were 26.8 ± 3.3 and 20.7 ± 2.9 lg g1 dry weight, respectively (Wong et al., 2002). The hydrocarbon concentrations recorded in the present study were also higher than that in a mangrove habitat in the Caribbean Island contaminated by oil from shipping activities (Bernard et al., 1996). Tolosa et al. (1996) proposed that sediment UCM concentrations below 10 lg g1 are common in coastal environments distant from hydrocarbon inputs. All these results illustrate that Sheung Pak Nai mudflat had been contaminated by fuel oil, probably from
1098
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
Table 3 Diagnostic ratios of CPI14–21, CPI24–33, and LALK/TALK in surface sediments collected in December 2000 and 2001 (mean and standard deviations of triplicates are shown) Quadrats
Quadrat 1 Quadrat 2 Quadrat 3
LALK/TALK
CPI14–21
CPI24–33
2000
2001
2000
2001
2000
2001
0.43 ± 0.07 0.40 ± 0.04 0.44 ± 0.07
0.15 ± 0.01 0.15 ± 0.02 0.21 ± 0.12
1.20 ± 0.21 1.56 ± 0.17 1.68 ± 0.74
1.58 ± 0.23 2.06 ± 0.21 4.93 ± 1.46
1.44 ± 0.15 2.10 ± 0.44 2.18 ± 0.11
2.50 ± 0.18 3.29 ± 0.21 3.52 ± 0.01
LALK/TALK (light alkanes/total alkanes) = (sum of C14–20)/(sum of C14–33); CPI14–21 (carbon preference index in the range of n-C14 to n-C21) = (sum of C15, C17, C19 and C21)/(sum of C14, C16, C18 and C20); CPI24-33 (carbon preference index in the range of n-C24 to n-C33) = (sum of C25, C27, C29, C31and C33)/(sum of C24, C26, C28 C30, and C32).
Table 4 Concentrations of petroleum hydrocarbons in aliphatic (F1), aromatic (F2) and total (F3) fractions of the surface sediments collected from three quadrats in December 2000 and 2001 (mean and standard deviations of triplicates are shown) Quadrats
F1 (aliphatic) 2000
F2 (aromatic) 2001
F3 (total = aliphatic + aromatic)
2000
2001
2000
2001
Total petroleum hydrocarbon (TPH = resolved + unresolved) Quadrat 1 50.9 ± 6.7 30.5 ± 6.4 Quadrat 2 57.3 ± 14.0 37.1 ± 7.0 Quadrat 3 44.6 ± 10.8 36.9 ± 5.9
15.2 ± 6.2 21.9 ± 5.4 18.3 ± 6.0
12.9 ± 3.3 10.6 ± 3.4 12.2 ± 2.4
66.2 ± 3.3 79.2 ± 17.4 62.8 ± 16.8
43.4 ± 8.6 47.7 ± 10.2 49.1 ± 6.4
Unresolved complex matrix (UCM) Quadrat 1 43.7 ± 6.5 Quadrat 2 50.8 ± 12.2 Quadrat 3 39.3 ± 9.4
14.9 ± 6.3 21.4 ± 5.5 17.4 ± 6.1
12.2 ± 2.8 9.8 ± 3.1 10.6 ± 2.3
58.6 ± 2.2 72.2 ± 15.9 56.7 ± 15.3
39.8 ± 7.8 42.9 ± 8.9 42.9 ± 5.2
27.6 ± 5.9 33.1 ± 6.2 32.2 ± 4.9
leakage and discharge of illegal smuggling operations in the Pearl River Estuary region which was active between 1998 and 2000. In 2001, concentrations of petroleum hydrocarbons and UCM in both aliphatic and aromatic fractions declined significantly (Table 4), suggesting that oil residues in surface sediments were weathered by physical/chemical processes and biodegradation. Fuel oil with a higher proportion of light hydrocarbons than that in crude oil would be easier to degrade. The loss of fuel oil in mudflat sediments at Sheung Pak Nai mangrove swamp was faster than the weathering of spilled crude oil at Yi O, another mangrove swamp in Hong Kong (Wong et al., 2002). Zhang and Lin (1996) also found that up to 70% diesel oil applied to mangrove sediments were degraded in one month under laboratory conditions. Although concentrations of total n-alkanes maintained at around 2–3 lg g1 in these two years, n-alkanes had shifted from petrogenic to natural biogenic sources as reflected in their profiles (Fig. 2): (i) most light n-alkanes except n-C17 disappeared and concentrations of light n-alkanes declined, (ii) the two isoprenoids (pristane and phytane) dropped by 80% or more; and (iii) the odd number n-alkanes from n-C25 to n-C33 were dominant in all surface sediments. The n-C17 peak indicates the petroleum hydrocarbons from phytoplanktons (Doskey, 2001) while the mangrove debris contribute to n-C25, n-C27, n-C29, n-C31 and n-C33 (Nishigima et al.,
2001). Quadrat 3 had an extremely large peak at nC17, suggesting a significant input from phytoplankton. During the 2001Õs sampling, a thick mat of Halophilia ovata (a local seagrass species) with healthy leaves and well-developed roots in the top 1 cm of the sediment was found in Q3 while the amounts of H. ovata were less in the other two quadrats. The ratios of LALK/TALK dropped from 0.4 in 2000 to 0.15 in 2001 while CPI values increased significantly (Table 3). These results also indicate that the petrogenic light n-alkanes had been lost by weathering and n-alkanes was dominant by the odd carbon number with large molecular sizes. 3.3. Petroleum hydrocarbon concentrations in root zone sediments Total petroleum hydrocarbons concentrations in root zone sediments of the unhealthy saplings were higher than that in healthy saplings although variations between saplings of the same category were large (Table 5). For most of the root zone sediments, petroleum hydrocarbon concentrations were lower than that in surface sediments, suggesting that some degradation had already taken place between December 2000 and March 2001. One unhealthy sapling had exceptionally high concentrations of petroleum hydrocarbons in its root zone sediments, greater than 1000 lg g1 dry weight (Table 5). A very strong fuel oil smell was noted when this
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
1099
Table 5 Concentrations of petroleum hydrocarbons and diagnostic ratios in root zone sediments of healthy and unhealthy saplings in March 2001 Healthy saplings Sapling 1
Unhealthy saplings
Sapling 2
Sapling 3
Petroleum hydrocarbon concentrations (lg g1 freeze-dry wt) TPH-F1 17.83 51.43 21.27 TPH-F2 5.53 8.17 11.42 TPH-F3 23.36 59.61 32.69 Diagnostic ratios CPI14–21 CPI24–33 LALK/TALK
0.91 2.20 0.23
0.95 2.38 0.30
1.12 2.26 0.32
Mean ± SD
Sapling 1
Sapling 2
Sapling 3
Mean ± SD
30.18 ± 15.09 8.37 ± 2.41 38.15 ± 15.37
48.27 19.14 67.41
33.85 12.61 46.46
505.57 604.13 1111.7
195.9 ± 219.1 211.9 ± 277.3 408.5 ± 497.3
0.99 ± 0.11 2.28 ± 0.07 0.28 ± 0.04
1.10 2.03 0.59
1.26 2.42 0.31
0.93 1.35 0.90
1.09 ± 0.14 1.93 ± 0.44 0.60 ± 0.24
TPH-F1: aliphatic petroleum hydrocarbons; TPH-F2: aromatic petroleum hydrocarbons; TPH-F3: total petroleum hydrocarbons (sum of aliphatic and aromatic); CPI14–21 (carbon preference index in the range of n-C14 to n-C21) = (sum of C15, C17, C19 and C21)/(sum of C14, C16, C18 and C20); CPI24–33 (carbon preference index in the range of n-C24 to n-C33) = (sum of C25, C27, C29, C31and C33)/(sum of C24, C26, C28 C30, and C32); LALK/ TALK (light alkanes/total alkanes) = (sum of C14–20)/(sum of C14–33).
counts
FID1 A, (A:\01111728.D)
16
70000
17 18 19
15
60000
20
50000 40000
14
21
30000
22
20000
UCM
10000 0 0
10
20
30
40
50
min
Fig. 3. GC–FID chromatograms of aliphatic fractions (F1) of root zone sediments collected from an unhealthy sapling in Quadrat 3 in March 2001. (The number on top of the peak represents the number of carbon in n-alkanes; UCM = unresolved complex mixture; the petroleum hydrocarbon level was exceptionally high with a clear pattern of weathered fuel oil.)
unhealthy sapling was dug up from the mudflat. This particular root zone sediment was mainly composed of partially weathered fuel oil as it had a large UCM in the range of n-C14 to n-C26 (Fig. 3), very high peaks of n-alkanes from n-C14 to n-C22 and the two isoprenoids, n-alkanes were dominant by short chain n-alkanes (LALK/TALK index was 0.9) and CPI values were close to 1 (Table 5). These strongly prove that Sheung Pak Nai mangrove swamp was suffered from fuel oil contamination and some oil was migrated into deeper sediments around the roots leading to severe death of young K. candel saplings.
4. Conclusions The present study demonstrates a mangrove swamp in Hong Kong SAR was contaminated by fuel oil due to illegal smuggling activities and discharge in Pearl River Estuary. The saplings of Kandelia candel, a dominant mangrove species planted by local villagers were suffered from oil pollution with massive mortality and defoliation in December 2000 although the plant surfaces were not covered by oil and the contamination
was not in a large scale. The concentrations of total petroleum hydrocarbons in surface sediments were around 60–80 lg g1 dry weight in December 2000 with one root zone sediment sample reached an exceptionally high concentration of >1000 lg g1 dry weight. In December 2001, the unhealthy saplings had recovered with new leaves, the total petroleum hydrocarbon levels in sediments also dropped to 30–40 lg g1 dry weight due to weathering processes and the control of anthropogenic fuel oil contamination by local authorities in early 2001. The GC–FID chromatograms, the profiles of n-alkanes and the diagnostic ratios in sediments all reveal that the petroleum hydrocarbons had shifted from petrogenic to biogenic sources in December 2001.
Acknowledgements The work described in this paper was fully supported by a grant from the Research Grant Council, HKSAR (Ref. no. CityU 1110/02M). The authors gratefully acknowledge Mr. Benz Chan for his assistance in the GC analyses.
1100
N.F.Y. Tam et al. / Marine Pollution Bulletin 51 (2005) 1092–1100
References Bernard, D., Pascaline, H., Jeremie, J.J., 1996. Distribution and origin of hydrocarbons in sediments from lagoons with fringing mangrove communities. Marine Pollution Bulletin 32, 734–739. Bruce, L.G., Schmidt, G.W., 1994. Hydrocarbon fingerprinting for application in forensic geology: review with case studies. American Association of Petroleum Geologists Bulletin 78, 1692– 1710. Burns, K.A., Codi, S., 1998. Contrasting impacts of localised versus catastrophic oil spills in mangrove sediments. Mangroves and Salt Marshes 2, 63–74. Burns, K.A., Garrity, S.D., Jorissen, D., MacPherson, J., Stoelting, M., Tierney, J., Yelle-Simmons, L., 1994. The Galeta oil spill. II. Unexpected persistence of oil trapped in mangrove sediments. Estuarine, Coastal and Shelf Science 38, 349–364. Burns, K.A., Codi, S., Duke, N.C., 2000. Gladstone, Australia field studies: weathering and degradation of hydrocarbons in oiled mangrove and salt marsh sediments with and without the application of an experimental bioremediation protocol. Marine Pollution Bulletin 41, 392–402. Colombo, J.C., Pelletier, E., Brochu, C., Khalil, M., 1989. Determination of hydrocarbon sources using n-alkane and polyaromatic hydrocarbon distribution indices. Case study: Rio de la Plata Estuary, Argentina. Environmental Science and Technology 23, 888–894. Da Silva, E.M., Peso-Aguiar, M.C., Navarro, M.F.T., Chastinet, C.B.A., 1997. Impact of petroleum pollution on aquatic coastal ecosystems in Brazil. Environmental Toxicology and Chemistry 16, 112–118. Doskey, P.V., 2001. Spatial variation and chronologies of aliphatic hydrocarbons in Lake Michigan sediments. Environmental Science and Technology 35, 247–254. Duke, N.C., Watkinson, A.J., 2002. Chlorophyll-deficiency propagules of Avicennia marina and apparent longer term deterioration of mangrove fitness in oil-polluted sediments. Marine Pollution Bulletin 44, 1269–1276. Duke, N.C., Pinzon, Z.S., Prada, M.C., 1997. Large-scale damage to mangrove forests following two large oil spills in Panama. Biotropica 29, 2–14. Garrity, S.D., Levings, S.C., Burns, K.A., 1994. The Galeta oil spill. I. Long-term effects on the physical structure of the mangrove fringe. Estuarine, Coastal and Shelf Science 38, 327–348. Hester, H.W., Mendelssohn, I.A., 2000. Long-term recovery of a Louisiana brackish marsh plant community from oil-spill impact: vegetation response and mitigation effects of marsh surface elevation. Marine Environmental Research 49, 233–254. Jernelov, A., Linden, O., 1983. The effect of oil pollution on mangroves and fisheries in Ecuador and Colombia. In: Teas, H.J. (Ed.), Tasks for Vegetation Science, Biology and Ecology of
Mangroves, vol. 8. Dr. W. Junk Publisher, The Hague, pp. 185– 188. Ke, L., Wong, T.W.Y., Wong, Y.S., Tam, N.F.Y., 2002. Fate of polycyclic aromatic hydrocarbon (PAH) contamination in a mangrove swamp in Hong Kong following an oil spill. Marine Pollution Bulletin 45, 339–347. Levings, S.C., Garrity, S.D., 1995. Oiling of mangrove keys in the 1993 Tampa Bay oil spill. In: Proceedings of the 1995 International Oil Spill Conference. American Petroleum Institute, Washington, DC. Levings, S.C., Garrity, S.D., 1997. Sublethal injury to red mangroves two years after oiling. In: Proceedings of the 1997 International Oil Spill Conference. American Petroleum Institute, Washington, DC, pp. 1040–1041. Massachusetts Department of Environmental Protection, 1998. Method for the Determination of Extractable Petroleum Hydrocarbons (EPH), Commonwealth of Massachusetts, Massachusetts, 41pp. Nishigima, F.N., Weber, R.R., Bicego, M.C., 2001. Aliphatic and aromatic hydrocarbons in sediments of Santos and Cananeia, SP, Brazil. Marine Pollution Bulletin 42, 1064–1072. Sauer, T.C., Uhler, A.D., 1994. Pollutant source identification and allocation: advances in hydrocarbon fingerprinting. Remediation 4, 25–45. Snedaker, S.C., Glynn, P.W., Rumbold, D.G., Corcoran, E.F., 1995. Distribution of n-alkanes in marine samples from Southeast Florida. Marine Pollution Bulletin 30, 83–89. Tam, N.F.Y., Wong, Y.S., Lu, C.Y., Berry, R., 1997. Mapping and characterization of mangrove plant communities in Hong Kong. Hydrobiologia 352, 25–37. Tolosa, I., Bayona, J.M., Albaiges, J., 1996. Aliphatic and polycyclic aromatic hydrocarbons and sulphur/oxygen derivatives in northwest Mediterranean sediments: spatial and temporal variability, fluxes and budgets. Environmental Science and Technology 30, 2495–2503. Wang, Z., Fingas, M., Li, K., 1994. Fractionation of a light crude oil and identification and quantitation of aliphatic, aromatic, and biomarker compounds by GC–FID and GC–MS. Journal of Chromatographic Science 32, 361–366. Wong, T.W.Y., Ke, L., Wong, Y.S., Tam, N.F.Y., 2002. Study of the sediment contamination levels in a mangrove swamp polluted by marine oil spill. In: Proceedings of the 25th Arctic and Marine Oil Spill Technical Seminar, Environment Canada, Ottawa, Canada, pp. 73–90. Zar, J.H., 1999. Biostatistical Analysis, fourth ed. Prentice Hall, Inc., NJ, USA. Zhang, T.C., Lin, P., 1996. The soil microbial function of Kandelia candel mangroves: degradation of diesel oil. In: Morton, B. (Ed.), Proceedings of the Third International Conference on the Marine Biology of the South China Sea. Hong Kong University Press, Hong Kong, pp. 389–395.