Trophic structures in tropical marine ecosystems: a comparative investigation using three different ecological tracers

Ecological Indicators 81 (2017) 315–324

Contents lists available at ScienceDirect

Ecological Indicators journal homepage: www.elsevier.com/locate/ecolind

Original Article

Trophic structures in tropical marine ecosystems: a comparative investigation using three different ecological tracers

MARK

Fany Sardennea, Stephanie Hollandab, Sabrena Lawrenceb, Rona Albert-Arrisolb, ⁎ Maxime Degrootea, Nathalie Bodinb,c, a b c

Research Institute for Sustainable Development (IRD), UMR MARBEC, Avenue Jean Monnet, 34200 Sète, France Seychelles Fishing Authority (SFA), Fishing Port, Victoria, Seychelles Research Institute for Sustainable Development (IRD), UMR MARBEC, Fishing Port, Victoria, Seychelles

A R T I C L E I N F O

A B S T R A C T

Keywords: Stable isotopes Fatty acids Mercury Tropical food web Indian ocean

We looked at how three ecological tracers may influence the characterization and interpretation of trophic structures in a tropical marine system, with a view to informing tracer(s) selection in future trophic ecology studies. We compared the trophic structures described by stable isotope compositions (carbon and nitrogen), the total mercury concentration (THg) and levels of essential fatty acids (EFA) at both the individual and species level. Analyses were undertaken on muscle tissue samples from fish and crustacean species caught in the waters surrounding the Seychelles. The carbon isotope composition (δ13C) correlated to the proportion of arachidonic acid (ARA), whereas the nitrogen isotope composition (δ15N) correlated to the proportion of docosahexaenoic acid (DHA) and THg. At the individual level, trophic position obtained with these three last tracers are similar. In contrast, the eicosapentaenoic acid (EPA) was not clearly correlated to any of the tracers. At the species level, the use of EFA (ARA and DHA), as compared to stable isotopes, resulted in slight structural modifications, mainly in the middle trophic levels. For example, the EFA overestimated the trophic positions of Thunnus alalunga and Etelis coruscans but underestimated the trophic positions of other snappers and groupers. While ARA mainly originates from coastal/benthic areas, DHA is conserved throughout the food web and may be used as a proxy indicator of trophic position. However, metabolic disparities can affect ecological tracers and in turn, distort the trophic structures derived from their results. This is especially true for species with close trophic ecologies. Despite these caveats, we think that analysing at the individual level the wealth of ARA, DHA and THg data that has already been obtained through earlier nutrition or food security studies would enhance our understanding of trophic structures.

1. Introduction Understanding the intricate flows of energy and nutrients through food webs is an essential component of both fundamental and applied ecology. In marine systems, comparisons of these flows are made to assess the effects of anthropogenic pressures such as fishing, to trace the transfer of pollutants through food webs and to quantify the impacts of global climate change (e.g., Litzow et al., 2006; Hebert et al., 2008). In tropical marine systems, however, this type of quantitative information is often lacking, despite their sensitivity to climate change and exposure to high levels of fishing pressure and coastal habitat destruction (Munday et al., 2009). Trophic ecology studies on marine species commonly use intrinsic ecological tracers. These tracers are biogeochemical compounds found within organisms, including stable isotopes, biomagnifying pollutants ⁎

and fatty acids (Ramos and González-Solís, 2012). Nitrogen and carbon are two elements commonly used as stable isotope tracers. With nitrogen, the 15N/14N ratio generally increases in fish species from one trophic level to the next and consequently, it can be used to indicate trophic position (Vander Zanden and Cabana Rasmussen, 1997). With carbon, the 13C/12C ratio can be used to determine the sources of primary production in a food web (i.e., benthic vs. pelagic inputs) (France, 1995). Examples of biomagnifying pollutants include methyl-mercury (MeHg) and organic polychlorobiphenyls. Chemically stable and persistent, these compounds accumulate in the tissues of marine organisms and their concentrations are regulated by factors such as exposure to chemical compounds, including dietary exposure (Kelly et al., 2007). As these compounds do not tend to degrade, their concentrations generally increase with each trophic level (Lavoie et al., 2013; Chouvelon et al., 2014). Finally, fatty acids are long carbon chains constituting lipids that

Corresponding author at: Seychelles Fishing Authority (SFA), Fishing Port, Victoria, Seychelles. E-mail address: [email protected] (N. Bodin).

http://dx.doi.org/10.1016/j.ecolind.2017.06.001 Received 28 February 2017; Received in revised form 26 May 2017; Accepted 1 June 2017 1470-160X/ © 2017 Elsevier Ltd. All rights reserved.

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

three tracers and comparing the trophic structures (i.e., relative proximities among individuals and species on two dimensions) described by each tracing method. In the absence of information on baselines compositions of the Seychelles’ marine ecosystem, the present study focused on the relative trophic position of individuals/species in the food web instead of absolute trophic level.

are necessary for a variety of physiological functions. A subset of these, known as ‘essential fatty acids’ (EFA), govern key functions such as growth, development and immune response (Arts et al., 2001). EFA and their precursors are mainly found in marine plants and algae and cannot be readily synthesized by all consumers. Thus, these EFA are incorporated into consumers in a conservative manner and can therefore be used to track different primary production sources and predators-prey relations throughout food webs (Dalsgaard et al., 2003; Piché et al., 2010). Three EFA are recognized as being essential for consumers: docosahexaenoic acid (DHA; 22:6n-3), eicosapentaenoic acid (EPA; 20:5n-3) and arachidonic acid (ARA; 20:4n-6) (Arts et al., 2001). In marine ecosystems, these three EFA are commonly used as biomarkers to understand trophic structure and called ‘fatty acids trophic markers’ (e.g., Müller-Navarra et al., 2000; Budge et al., 2006; Iverson, 2009). For example, DHA increases with the trophic position of Coral Sea-caught albacore tuna Thunnus alalunga (Parrish et al., 2015) and with nitrogen stable isotope values across a Mediterranean lagoon food web, from sediment organic matter to fish (Koussoroplis et al., 2011). Significant positive correlations were also highlighted between DHA/EPA ratios, nitrogen stable isotope signatures and degree of carnivory in coastal calanoid copepods (El-Sabaawi et al., 2008). An advantage of ecological tracers is their ability to reflect the integration of consumers’ diets over a relatively long timeframe. However, their main limitation is that they are influenced by both environmental (abiotic) and biological (metabolic) processes (Ramos and González-Solís, 2012). For example, stable isotope compositions are affected by growth rates and baseline changes; local pollution sources and the transfer of contaminants to specific organs (for accumulation or excretion) can affect outputs from pollutant tracers; and the turn-over and modifications that EFA may undergo within consumers remain largely unknown (Iverson, 2009; Ramos and González-Solís, 2012). To overcome these limitations and provide complementary insights, the combined use of different ecological tracers and/or tissues is a growing area of research in the field of trophic ecology. This combined approach allow researchers to better characterize contaminant and nutrient flows within food webs and better detect differences in dietary habits among sympatric species (e.g., Hebert et al., 2009; Le Croizier et al., 2016; Sardenne et al., 2016). However, although informative, this approach multiplies the number of expensive and timeconsuming analyses to be undertaken, requires adapted statistical approaches, and can lead to interpretation challenges when each tracer leads to a different conclusion. For example, the trophic level attributed to the outcomes of a stable isotope analysis can be uncorrelated with the mercury concentrations detected (Bond, 2010), as mercury bioaccumulation also depends on dietary exposure and size/age of organisms. In this study, we looked at how the choice of ecological tracer can influence the characterization of trophic structure (i.e. on two dimensions) and trophic positions in the food chain (i.e. on one dimension) in a tropical marine ecosystem, with a view to improving tracer(s) selection in future studies. Drawing on the fundamental theories associated with each tracer, as previously outlined, we hypothesized that independent ecological tracers should yield similar results. For example, nitrogen isotope composition, DHA and mercury contents theoretically increase with the trophic position and thus should be correlated and provide the same trophic information. To test this hypothesis, we compared the results of three independent ecological tracers tested on 17 species of market fish and crustaceans collected within the Seychelles’ surrounding waters. These species, all of which belonged to trophic level 3 or above, were selected to capture a wide range of phylogenetic differences (i.e., different families), vertical distributions (i.e., pelagic, demersal and benthic) and habitat type (i.e., oceanic and coastal) (Table 1). The three tracers used were the stable isotope compositions of carbon (δ13C) and nitrogen (δ15N), total mercury concentrations (THg) and EFA levels (DHA, EPA and ARA). Our investigation focused on identifying the correlations present between the

2. Material and methods 2.1. Fish and crustacean sampling A total of 92 individuals (17 fish and crustacean species; Table 1) were collected from the Plateau of Mahé, Seychelles (3.68°S to 6.51°S and 53.93°E to 57.14°E) between April 2014 and February 2015 using either purse-seine, hand-line, long-line, snorkelling, trap or tangle net during scientific surveys of the Seychelles Fishing Authority. To limit the influence of size on the ecological tracers (Dang and Wang, 2012), similarly sized individuals were selected for each species (i.e., intraspecies coefficient of variation < 25%). For each individual, fork length (fish) or total length (crustaceans) was recorded using callipers and a sample of white muscle tissue (∼20 g) was taken from the front dorsal region and stored at −80 °C shortly after collection. For the stable isotope analysis, a subsample of each muscle sample was freeze dried and ground into a fine homogeneous powder using a Retsch Mixer Mill MM200. Another subsample was keep frozen for the THg and EFA analyses. 2.2. Stable isotope analysis The stable isotope analysis was performed on samples of lipid-free powder (0.4 ± 0.1 mg) that were packed into tin capsules. The lipids were removed, as they are known to influence δ13C values (Post et al., 2007; Bodin et al., 2009). They were extracted from the powdered samples (approximately 350 ± 100 mg) following the method set out in Bodin et al. (2009) (12 ml of dichloromethane at 100 °C under 1900 psi for 20 min using an Dionex Accelerated Solvent Extractor 200) that allows simultaneous C and N stable isotope analysis without an unwanted effect on δ15N values. Lipid extracts were evaporated and weighed to the nearest 0.1 mg to determine the total lipid content, expressed as a percentage of the dry weight (dw). The lipid-free powders were analyzed using an Elemental Analyser (Flash EA 1112; Thermo Scientific) coupled to an Isotope Ratio Mass Spectrometer (Delta V Advantage with a Conflo IV interface; Thermo Scientific) at the LIENSs Stable Isotope facility (La Rochelle, France). Results were reported in the δ unit notation and expressed as parts per thousand (‰), relative to international standards (atmospheric N2 for nitrogen and Vienna-Pee Dee Belemnite for carbon). Calibration was completed using reference materials (IAEA-N2, −NO−3, −600 for nitrogen; USGS24, IAEA-CHE, −600 for carbon). Analytical precision, based on replicate measurements of acetanilide (Thermo Scientific) was < 0.15‰ for both δ15N and δ13C. The effectiveness of the chemical extraction was checked by examining the C:N ratio from the percent element weight (C:N < 3.5; Post et al., 2007). 2.3. Total mercury concentration The biomagnifying form of Hg (MeHg) constitutes most of the THg in the upper trophic levels (Cai et al., 2007; McMeans et al., 2015). Thus, for analytical convenience, we only measured THg. The analysis was performed on fresh white muscle samples (10–50 mg) using a Direct Mercury Analyser DMA-80 Dual Cell (Milestone) at the Seychelles Fishing Authority (Victoria, Seychelles). The results were reported in parts per million (ppm), which equates to μg g−1 in wet weight. Calibration blanks were run in between each sample to ensure Hg levels were reset to 0.1 ng. Analytical performance was checked every 15–20 samples against two laboratory control analyses (performed on large 316

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

Table 1 Phylogenic and general ecological information on the 17 species. Data are derived from SeaLifeBase (Palomares and Pauly, 2017) with trophic levels based on previous diet studies or on food items.

Fish

Family

Species

Code

Vertical distribution

Habitat

Movement patterns

Trophic level

Scombridae

Albacore (Thunnus alalunga) Bigeye tuna (Thunnus obesus) Skipjack tuna (Katsuwonus pelamis) Yellowfin tuna (Thunnus albacares) Wahoo (Acanthocybium solandri) Indian mackerel (Rastrelliger kanagurta) Common dolphinfish (Coryphaena hippurus) Brownspotted grouper (Epinephelus chlorostigma) White-blotched grouper (Epinephelus multinotatus) Tomato hind (Cephalopholis sonnerati) Deepwater longtail red snapper (Etelis coruscans) Emperor red snapper (Lutjanus sebae) Two-spot red snapper (Lutjanus bohar) Green jobfish (Aprion virescens)

ALB BET SKJ YFT WAH RAG DOL EFH EWU EFT ETC LUB LJB AVR

Pelagic (epipelagic) Pelagic (epipelagic) Pelagic (epipelagic) Pelagic (epipelagic) Pelagic (epipelagic) Pelagic (epipelagic) Pelagic (neritic) Demersal Demersal Demersal Pelagic (mesopelagic) Demersal Demersal Demersal

oceanic oceanic oceanic oceanic oceanic & coastal oceanic & coastal oceanic & coastal coastal (reef) coastal (reef) coastal (reef/sand interface) coastal (reef) coastal (reef) coastal (reef) coastal (reef/sand interface)

Migratory Migratory Migratory Migratory Migratory Migratory Migratory Resident Resident Resident Resident Resident Resident Resident

4.3 4.5 4.4 4.4 4.3 3.2 4.4 4.0 3.9 3.8 4.4 4.1 4.3 4.3

Longlegged spiny lobster (Panulirus longipes) Pronghorn spiny lobster (Panulirus penicillatus) Spanner crab (Ranina ranina)

LOJ NUP RAQ

Benthic Benthic Benthic

coastal (reef) coastal (reef) coastal (sandy bottom)

Resident Resident Resident

3.6 3.6 3.9

Coryphaenidae Serranidae

Lutjanidae

Crustacean

Palinuridae Raninidae

derived from two of the ecological tracers, at both the individual and species levels. EPA was excluded from this analysis because of the absence of relationship of this compound with the other tracer (see results paragraph 3.1). Agreement at the individual level was tested using Procrustes analysis and ProTests. Procrustes is a shape analysis method that compares two ordination shapes by superimposing them and maximizing the fit between the corresponding observations of the two configurations (Peres-Neto and Jackson, 2001; Oksanen et al., 2007). The method is based on the least squares criterion that minimizes the sum of the squared residuals between the two configurations. The r2 value (between 0 and 1) measures the degree of association between the two ordinations: high values indicate a strong concordance. The ProTest provided a permutation test to assess the statistical significance of the Procrustes fit. We used ProTest with 2000 permutations of individuals to examine the degree of consistency between the trophic structures, as described by the results of the stable isotope and EFA analyses. At the species level, a hierarchical clustering approach, based on Euclidean distance matrices and the Ward method, was used to identify different species groupings using each set of results, one by one and combined (after scaling to overcome unit differences). We also investigated the general differences observed between the three ecological tracers among species, families, vertical distribution, or habitat/ movement (the two latest factors giving similar grouping in the present study). However, the best models (as selected by the AIC) only contained the species covariate. Consequently, F-tests and post hoc t-tests were used to detect differences between species. Statistical analyses were performed using R 3.1.2 software (R Core Team, 2013) and the ‘vegan’ package.

homogenized wet samples of white muscle and liver from Thunnus obesus) and the certified reference material, tuna fish flesh homogenate (IAEA-436) (Bodin et al., 2017). Satisfactory accuracy (97–104%) was calculated with an analytical variability below 5% (n = 29). Quantification limits were calculated from blank measurements with Hg values of 0.0016 ppm. 2.4. EFA levels EFA levels were determined by analysing the total lipids present in fresh white muscle samples (i.e., there was no separation of lipid classes). This method is easier to implement than a separation method and has been used in several ecological studies (e.g., Müller-Navarra et al., 2000; Parrish et al., 2015). It assumes that differences in total lipid contents among species mainly correspond to differences in storage lipids, while structural lipids remain fairly constant (Litzow et al., 2006). Lipids were extracted from the samples (∼100 mg) using a mixture of chloroform and methanol (2:1, v/v), following the method in (Folch et al., 1957). The samples were then transesterified with 10% wt boron trifluoride–methanol. The fatty acids methyl esters (FAME) were separated and quantified using gas chromatography (a TRACE 1310 gas chromatograph equipped with a split/splitless injector operated in the splitless mode) using a FAMEWAX™ column (30 m in length, 0.32 mm internal diameter, 0.25 μm film thickness; Restek) and a flame ionisation detector at the analytical platform MetaToul-LIPIDOMIQUE (Toulouse, France). The FAME were identified by comparing sample retention times to those of commercial standard mixtures (Menhaden oil and Food Industry FAME Mix; Restek) using Xcalibur 2.2 software (Thermo Scientific). EFA levels were expressed as a percentage of the total identified FAME.

3. Results A summary of the results obtained for the three tracers are presented in Table 2.

2.5. Data analysis Prior to analysis, the THg data were transformed into log(x + 1) to reduce the influence of outliers. Correlations among the ecological tracers (regardless of species) were investigated using several approaches. Firstly, we considered the simple relationships among the ecological tracers using Spearman’s rank correlation (where rs is the correlation coefficient). When a correlation was detected, the model (linear or not) selected by the Akaike Information Criterion (AIC) was adjusted and the normality of residuals was checked using Q–Q plots. Secondly, we considered the consistency between trophic structures

3.1. Correlations between ecological tracers, trophic structures and positions Rather strong correlations were found between δ13C and ARA, between δ15N and DHA (rs > 0.6, P < 0.05) and, to a lesser extent, between THg, δ15N and EPA (−0.3 > rs > −0.45, P < 0.05; Fig. 1). However, the only strong linear relationship we identified was between δ13C and ARA (r2 = 0.72; Fig. 2A). The linear relationship between δ15N and DHA was weaker (r2 = 0.38). This is mainly due to the tuna 317

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

Table 2 Muscle composition (mean ± standard deviation) in stable isotopes of nitrogen (δ15N) and carbon after lipid removal (δ13C), total mercury (THg) and essential fatty acids (docosahexaenoic acid (DHA), arachidonic acid (ARA), and eicosapentaenoic acid (EPA)) of 14 fish and three crustacean species collected in the water surrounding the Seychelles (western Indian Ocean) from April 2014 to February 2015. N = number of individuals analyzed. Size corresponds to fork length for fish and total length for crustaceans. stable isotopes (‰) δ15N

Species

Code

N

size (cm)

δ13C

Fish

T. alalunga T. obesus K. pelamis T. albacares A. solandri R. kanagurta C. hippurus E. chlorostigma E. multinotatus C. sonnerati E. coruscans L. sebae L. bohar A. virescens

ALB BET SKJ YFT WAH RAG DOL EFH EWU EFT ETC LUB LJB AVR

2 8 9 4 2 7 7 6 7 6 7 7 7 7

99.5 ± 2.1 48.6 ± 5.7 59.5 ± 10.2 132.8 ± 13.2 109.4 ± 15.1 25.9 ± 0.5 99.3 ± 6.4 37.7 ± 2.4 62.9 ± 7.2 40.3 ± 5.0 55.5 ± 12.9 56.5 ± 6.9 66.2 ± 4.1 52.1 ± 1.8

−17.3 −17.2 −16.9 −16.6 −16.1 −17.4 −16.6 −16.6 −16.3 −16.9 −18.1 −16.2 −15.9 −16.1

Crustacean

P. longipes P. penicillatus R. ranina

LOJ NUP RAQ

2 2 2

8.2 ± 0.2 7.8 ± 0.2 8.6 ± 0.1

−13.8 ± 0.1 −13.8 ± 0.6 −14.9 ± 0.0

± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.1 0.2 0.3 0.2 0.2 0.2 0.1 0.4 0.3 0.1 0.2 0.3 0.4 0.3

12.3 12.3 11.5 12.3 14.5 11.9 13.3 13.7 13.0 14.0 13.6 14.2 13.4 14.1

± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.1 0.6 0.6 0.5 0.6 0.3 0.3 0.3 0.2 0.3 0.3 0.1 0.3 0.2

11.2 ± 0.0 11.9 ± 0.5 11.5 ± 0.3

mercury (ppm ww)

essential fatty acids (% of total FA)

THg

ARA

0.394 0.104 0.195 0.361 1.292 0.033 0.196 0.195 0.274 0.158 0.240 0.192 0.395 0.116

± ± ± ± ± ± ± ± ± ± ± ± ± ±

0.056 0.037 0.098 0.187 0.898 0.008 0.048 0.036 0.146 0.059 0.163 0.105 0.130 0.028

0.025 ± 0.002 0.023 ± 0.000 0.035 ± 0.008

4.1 4.6 5.1 5.1 7.2 2.6 8.7 6.5 8.1 4.6 3.9 7.5 8.9 7.1

± ± ± ± ± ± ± ± ± ± ± ± ± ±

EPA 0.2 1.1 1.0 0.9 1.4 0.7 1.9 1.3 1.6 1.0 0.2 0.5 1.7 0.8

27.3 ± 3.2 16.7 ± 1.3 16.0 ± 1.8

3.8 3.1 3.0 3.0 2.8 5.8 3.8 2.5 2.7 2.4 4.2 3.4 2.7 3.6

DHA ± ± ± ± ± ± ± ± ± ± ± ± ± ±

1.0 1.2 1.2 1.3 0.6 2.2 0.4 0.5 0.3 0.3 0.3 0.3 0.3 0.6

15.5 ± 0.1 15.3 ± 2.6 16.1 ± 0.1

38.8 30.0 22.7 22.2 33.9 15.9 38.5 33.3 27.5 29.0 42.9 36.7 30.1 41.7

± ± ± ± ± ± ± ± ± ± ± ± ± ±

3.7 7.2 6.8 4.7 0.5 5.4 3.2 3.1 2.2 4.9 3.3 2.8 4.5 2.1

7.5 ± 0.1 9.4 ± 2.9 18.7 ± 0.4

performed on the stable isotope and two EFA (ARA and DHA) results indicated that trophic structure concordance was consistently significant and relatively strong (P < 0.001; r2 = 0.45). This consistency was strengthened by the high compatibility between δ13C values and ARA levels. For species, the stable isotope and EFA trophic structures based on clustering were relatively similar. Here, crustaceans were clearly differentiated from fish species due to their high δ13C values and ARA levels (Fig. 3). The main differences observed between the two trophic structures related to five fish species whose trophic levels ranged between 3.8 and 4.4. These differences were driven by the low correlation between their δ15N values and DHA levels. T. alalunga and Etelis coruscans had mean isotopic compositions that were close to the medium-low δ15N values exhibited by the tropical tunas (i.e., T. obesus, K. pelamis and T. albacares) and R. Kanagurta, and were gathered in Group I (Fig. 4A). However, in the EFA clustering, their high DHA level (> 25% of total FA) put them closer to the demersal fish species (Group II; Fig. 4A). The opposite was observed for the cluster comprising Cephalopholis sonnerati, Epinephelus multinotatus and Lutjanus bohar. This was due to their high δ15N values and medium-low DHA levels. Finally, the combination of the three datasets (stable isotope, EFA and THg; Fig. 4B) led to different merging for the Groups I and II that respectively contained 10 fish species and the three tropical tunas plus R. kanagurta. The crustaceans (Group III) were clearly distinguished from fish and occupied the lowest trophic positions together with the Indian mackerel. Among them, the DHA assigned the spanner crab Ranina ranina to a slightly higher trophic position than either species of spiny lobster Panulirus longipes and P. penicillatus (t = −2.5 and −2.7, both P < 0.05); however, no differences were observed in the δ15N values (F = 2.6, P = 0.23; Fig. 5A) and the THg levels (F = 5.6, P = 0.10; Fig. 5C). For the high trophic positions, both δ15N values and THg levels assigned the Wahoo Acanthocybium solandri to the highest position while the DHA assigned the snapper Etelis coruscans (Fig. 5).

Fig. 1. Correlation plots among the three ecological tracers studied (Spearman’s test): lipid-free stable isotope ratios (carbon: δ13C and nitrogen: δ15N, ‰), log-transformed total mercury concentrations (THg, ppm) and essential fatty acids (docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA) and arachidonic acid (ARA), expressed as a percentage of total lipid content). The colored cells indicate a significant correlation between the two tracers (P < 0.05) while the uncolored cells indicate a non-significant correlation. The numbers in the cells are the associated Spearman’s correlation coefficients calculated for each pair (rs; note that these values can be negative or positive) and the cell’s colour intensity is proportional to rs.

3.2. Species-specific compositions

species analyzed (i.e., T. alalunga, T. obesus and Katsuwonus pelamis), all of which produced relatively low δ15N values but high DHA levels (Fig. 2b). The weak relationships between THg and δ15N and THg and EPA were a result of the low THg levels detected in the three crustacean species and Indian mackerel Rastrelliger kanagurta (Fig. 2CD). No correlation was observed between THg and total lipid content (rs = −0.16, P = 0.14), while the correlation between THg and size was strong (rs = 0.75, P < 0.001). For trophic structures based on individuals, the ProTest analyses

δ13C values were found to be higher in crustaceans than fish (mean values:−14.2‰ vs. −16.6‰; t = 8.9, P < 0.001) but the opposite was true for δ15N (mean values: 11.5‰ vs. 13.0‰; t = −7.1, P < 0.001). Among the fish species analyzed, a wide range of isotopic values were obtained (Fig. 3). For example, the ranges of δ13C values for two different snapper species were −18.4‰ to −17.8 ‰ (E. coruscans) and −16.6‰ to −15.1‰ (L. bohar) while δ15N ranges were 318

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

Fig. 2. Relationships between (A) lipid-free stable carbon isotope (δ13C) and arachidonic acid (ARA), (B) lipid-free stable nitrogen isotope (δ15N) and docosahexaenoic acid (DHA), (C) THg (log scale) and δ15N and (D) THg (log scale) and eicosapentaenoic acid (EPA) identified from the muscle samples taken from 17 species (see Table 1 for species codes) caught in the waters around the Seychelles.

P. penicillatus and R. ranina (t = 9.2, P < 0.001). Large differences were also observed among the fish (F = 19.3, P < 0.001). For example, R. kanagurta contained relatively low levels of all three EFA, particularly DHA (6.1–21.0%; according to species, 2.3 < t < 11.2, P < 0.05). The highest DHA levels were found in the snappers E. coruscans (36.2%–45.4%) and Aprion virescens (38.7%–44.2%), the tuna T. alalunga (36.1%–41.4%) and the common dolphinfish Coryphaena hippurus (33.8%–41.1%) (−1.8 < t < −0.5, all P > 0.08). The highest ARA levels were found (5.9%–12.2%) in C. hippurus, A. solandri, the grouper Epinephelus multinotatus and two snapper species (L. sebae and L. bohar).

10.5‰ to 12.6‰ (K. pelamis) and 14.0‰ to 14.9‰ (Acanthocybium solandri). The THg concentrations varied considerably among the different species (F = 22.7, P < 0.001) although generally, they were lower in crustaceans (0.022–0.041 ppm; t = −6.6, −7.1 and −6.0, P < 0.001) than in fish (0.022–1.927 ppm) (Table 2). Among the fish, large differences were found. For example, the Indian mackerel R. kanagurta had THg values that were similar to those found in crustaceans (0.024–0.045 ppm; t = −0.3, P = 0.76). In general, A. solandri was found to have the highest THg levels (0.657–1.927 ppm; −11.1 < t < −2.6, all P < 0.001), followed by L. bohar (0.269–0.648 ppm) and T. alalunga (0.354–0.434 ppm). A. solandri also had the highest inter-individual variability in THg (coefficient of variation = 69.5%). Levels of ARA, EPA and DHA in crustaceans were found to be ∼20%, ∼16% and ∼12%, and ∼7%, ∼4% and ∼30% in fish (Table 2). Among the crustaceans, P. longipes contained more ARA than

4. Discussion Three different ecological tracers were investigated and their outputs compared with a view to explaining the trophic structures of the Seychelles tropical marine ecosystems. Trophic structures determined 319

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

Fig. 3. Trophic structures with17 species (see Table 1 for species codes) caught in the waters around the Seychelles (mean ± standard deviation by species), using (A) lipid-free stable isotopes values: carbon (δ13C) and nitrogen (δ15N); and (B) two essential fatty acid compositions: docosahexaenoic acid (DHA) and arachidonic acid (ARA).

Fig. 4. Hierarchical clustering of the 17 species (see Table 1 for species codes) caught in the waters around the Seychelles, based on Euclidean distance matrices and the Ward clustering method, according to their (A) mean lipid-free stable isotope compositions (carbon (δ13C) and nitrogen (δ15N); left panel) and essential fatty acids (docosahexaenoic acid (DHA) and arachidonic acid (ARA); right panel) considered separately with arrows indicating positions changes; and (B) combined mean compositions in δ13C, δ15N, DHA, ARA and total mercury (log-transformed THg).

320

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

Fig. 5. Relative trophic positions of 17 species (see Table 1 for details) caught in the waters around the Seychelles according to their muscle composition in (A) lipid-free stable nitrogen isotope (δ15N), (B) essential docosahexaenoic acid (DHA), and (C) total mercury (THg). These relative trophic positions should not be directly converted into absolute trophic levels because of the species movement behaviour among habitats with potentially different baseline values.

producers also have higher δ13C values than their offshore counterparts because they are exposed to different CO2 diffusion resistances (i.e., low/high diffusion resistance and 13C depletion in turbulent/stagnant waters) (France, 1995). Thus, this specific combination of high δ13C values and ARA level that characterize coastal primary producers can be transferred to upper trophic levels via primary consumers (e.g., sea urchins and or coral fish species) feeding on coralline algae or corals. For example, spiny lobsters feed on both coralline algae and sea urchins (Palomares and Pauly, 2017) and in our study, they showed the highest ARA level and δ13C values (Fig. 3A). Tunas are unable to synthesize EFA because they lack the necessary desaturase enzymes (Tocher, 2003) but ARA is particularly important for their reproduction (Mourente et al., 2001; Sardenne et al., 2017). Interestingly, T. albacares migrates to the Seychelles during its breeding season (Zudaire et al., 2015). We suggest that in part, this migration may be triggered by a need to find coastal, ARA-rich prey species. A similar hypothesis has been proposed to explain the breeding migration of the gray whale Eschrichtius robustus from polar waters (rich in DHA) to tropical lagoons (rich in ARA) (Caraveo-Patiño et al., 2009). Thus, as already shown in marine mammal predators (Spitz et al., 2012), prey quality may affect feeding strategies in tunas.

using carbon and nitrogen stable isotopes share some similarities with the structures described by the ARA and DHA levels, while δ15N and THg values indicated similar trophic levels. Among the 17 species, the results of all three ecological tracers varied greatly, even between species that were closely related in both phylogeny and ecology. 4.1. Ecological interpretation of observed correlations between the ecological tracers In the waters around the Seychelles, ARA levels increase with δ13C values, suggesting that here, ARA has a coastal or benthic origin. A similar observation has been made in a number of other ecosystems including Senegal’s up-welling (Le Croizier et al., 2016), the North Eastern Atlantic (Stowasser et al., 2009) and the tropical waters of Australia and Mozambique (Couturier et al., 2013). ARA is synthesized from linoleic acid (18:2n-6) via elongase and desaturase enzymes systems (Tocher, 2003). This synthesis occurs in numerous primary producers, especially the red algae belonging to the phylum Rhodophyta (Gerwick and Bernart, 1993). In the fringing reefs of the Caribbean, van Duyl et al. (2011) observed high ARA levels in coralline red algae (ARA ∼ 32%) and some coral mucus (ARA ∼ 15%). Coastal primary 321

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

The relationship between δ15N values and DHA level (Fig. 2B) has been observed in a number of ecosystems (e.g., Rooker et al., 2006; Koussoroplis et al., 2011). Evolutionary processes appear to be the primary driver of this relationship: most marine consumers are unable to synthesize this EFA but acquire it through their diets (Tocher, 2003). DHA is mainly synthesized by dinoflagellates, preserved in food webs and tends to increase with each step up in trophic position (Dalsgaard et al., 2003; Parrish et al., 2015). In parallel, δ15N values also increase with trophic levels due to the preferential metabolization and excretion of 14N, relative to 15N, during amino acid transamination (Macko et al., 1986). Regardless of any additional age or growth influences on nitrogen metabolism, δ15N values should increase between 2.5–3.4‰ from trophic level n to n + 1 (mean trophic enrichment factors proposed for fish by Caut et al. (2009) and Minagawa and Wada (1984), respectively). In this study, this would equate to an increase in DHA between the prey and consumers levels that corresponded to 40–50 % of the total fatty acid contents, which seems unrealistic. Thus, this relationship is probably not linear across the entire food web, as observed in a Mediterranean lagoon (Koussoroplis et al., 2011), maybe because all species have a high content of DHA subsistence-level regardless of the trophic position. In contrast to the other EFA, EPA level was poorly correlated with values produced by the other ecological tracers. A weak negative correlation between EPA and THg was driven by the low levels of both tracers found in the three crustaceans and R. kanagurta (Fig. 2D). The low THg levels observed in these four species is likely due to their relatively low trophic levels. THg increases with trophic level (a similar pattern is also seen in δ15N), a characteristic that has been observed in several different food webs (e.g., Cai et al., 2007; McMeans et al., 2015). In general, THg levels tend to increase with age and body size (Kojadinovic et al., 2006), as both detoxification capabilities and growth rates reduce (Dang and Wang, 2012). Age can also influence δ15N values: in species with indeterminate growth such as fish and crustaceans, older individuals are generally larger, tend to feed on larger prey and consequently, have higher δ15N values than their younger congeners. Regardless of trophic level, THg levels in consumers are also related to the presence of local sources of contamination and levels of contamination in prey species (Loseto et al., 2008). In pelagic ecosystems, MeHg is mainly found in the poorly oxygenated waters of the mesopelagic zone (around 200–1000 m deep) (Kojadinovic et al., 2007; Choy et al., 2009). Consumers that feed at such depths are more exposed to mercury contamination, regardless of their trophic level. It is also interesting to note that mercury is transferred more readily between trophic levels in pelagic food webs than benthic webs (McMeans et al., 2015).

information on isotopic baseline composition in our study that most likely differ between sandy and reef habitats on which spanner crab and spiny lobster depend on, respectively (Table 1). Differences in trophic positioning were also observed for five mid-to-high trophic level species: for instance, the temperate tuna species T. alalunga and the snapper E. coruscans were characterized by relatively low stable isotope-derived trophic positions when compared to their EFA- and THgderived trophic positions, indicating that perhaps the first were underestimated (Fig. 5). The three tracers, however, agreed on the trophic proximity of these species despite their different habitats and vertical distribution (oceanic vs. coastal and surface vs. deep waters) (Fig. 4). Species’ migration between ecosystems with different baseline values can also affect the trophic tracers’ outputs. In the case of the migratory Indian mackerel and the tropical tuna species, EFA and isotopes yielded a similar pattern of results, a combination which set them to a lower trophic position than most of the demersal species. Such unexpected trophic positioning might result from differences in baseline values between epipelagic and bathydemersal zones, and highlight these two communities rely on different resources in the Seychelles EEZ. Similarly, EFA and isotope outputs assigned A. virescens and L. sebae to the highest trophic positions while placing the tropical tunas at intermediate ones. This result was at odds with THg outcomes, which assigned similar values to these fish species regardless of habitat but varied with size. Thus, the isotope-derived grouping appears more driven by habitat and vertical distribution than the EFA-derived group (e.g., Group I only contains pelagic species when it is derived from the stable isotope outputs but pelagic and demersal species when it is derived from EFA outputs) and the THg-derived trophic position. Differences in baseline values between habitats might be a stronger constraint for the stable isotope analysis. Collectively, these differences may also be drive by differences in metabolism. EFA retention is regulated to optimize physiological performances, especially in lipids constituting cellular membranes (Kainz et al., 2006). The physiological status (e.g., age, reproductive status) of individuals were unknown but these factors may have affected EFA levels and inter-individual variability. Stable isotope compositions can also be affected by species-specific physiological processes, including tissue turn-over and growth rates (Ramos and González-Solís, 2012). Thus, the sensitivity of ecological tools to physiological processes should be evaluated using experimental approaches that are able to control for diet and physiological status. In situ, the inclusion of covariates such as age, growth rates or reproductive status may assist with these considerations.

4.2. Trophic structures as defined by the stable isotope, EFA and THg outputs

The trophic positions derived for most of the studied species using the stable isotope outputs (Fig. 3A) were similar to those obtained by Trystram et al. (2015) around Réunion Island (20.88°S; 55.45°E). The low δ13C values found for the snapper E. coruscans in the waters around both the Seychelles and Réunion (around −18‰) are probably linked to its diet of deep prey species such as myctophids. The low δ15N values observed for the four tuna species (< 13‰) is likely related to their opportunistic diets which include a relatively lower component of juvenile fish, as compared to other epipelagic marine predators (e.g., dolphinfish C. hippurus and wahoo A. solandri) (Trystram et al., 2015). These lower values may also be driven by recent and significant periods of residency in the Mozambique Channel, an area characterized by a lower nitrogen baseline and shorter marine food webs (as compared to other high-sea areas in the western Indian Ocean) (Kojadinovic et al., 2006; Sardenne et al., 2016). Species-specific THg concentrations were also in line with previously reported values for tropical marine fish in the western Indian Ocean (Matthews, 1983; Kojadinovic et al., 2006, 2007; Bodin et al., 2017). An exception to this was the extremely high value we obtained for A. solandri. However, this result was predominantly driven by the low number of individuals analyzed (n = 2)

4.3. Species-specific patterns in the three ecological tracer outputs

In general, the stable isotope and EFA tracers both described similar trophic structures, an outcome driven by the correlations we have outlined above. At the individual level, the trophic structures were particularly similar and strong correlations were observed among tracers (i.e. trophic structures similarly shaped). At the species level, the method selected did affect details in the trophic structure, with differences observed between close trophic positions in relation to the species’ habitat and movements. For example, while both methods clearly distinguished the three crustacean species from the fish species, the EFA outputs assigned R. ranina a slightly higher trophic position than either species of spiny lobsters, while the stable isotope outputs did not. EFA outputs are in line with the literature as the diet of R. ranina includes small DHA-rich fish (estimated tropic level of 3.9; Table 1) oppositely to spiny lobsters’ diet based on molluscs and echinoderms (estimated tropic level of 3.6; Table 1) (Palomares and Pauly, 2017). Relative isotopic-derived trophic positions for Seychelles’ crustaceans may be biased due to lack of 322

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

of active tissues of the Gray whale. Mar. Ecol. 30, 437–447. http://dx.doi.org/10. 1111/j. 1439-0485.2009.00289.X. Caut, S., Angulo, E., Courchamp, F., 2009. Variation in discrimination factors (Δ 15 N and Δ13C): the effect of diet isotopic values and applications for diet reconstruction. J. Appl. Ecol. 46, 443–453. http://dx.doi.org/10.1111/j.1365-2664.2009.01620.X. Chouvelon, T., Caurant, F., Cherel, Y., Simon-Bouhet, B., Spitz, J., Bustamante, P., 2014. Species- and size-related patterns in stable isotopes and mercury concentrations in fish help refine marine ecosystem indicators and provide evidence for distinct management units for hake in the Northeast Atlantic. ICES J Mar Sci 71, 1073–1087. http://dx.doi.org/10.1093/icesjms/fst199. Choy, C.A., Popp, B.N., Kaneko, J.J., Drazen, J.C., 2009. The influence of depth on mercury levels in pelagic fishes and their prey. Proc. Natl. Acad. Sci. 106, 13865–13869. http://dx.doi.org/10.1073/pnas.0900711106. Couturier, L.I.E., Rohner, C.A., Richardson, A.J., Marshall, A.D., Jaine, F.R.A., Bennett, M.B., Townsend, K.A., Weeks, S.J., Nichols, P.D., 2013. Stable isotope and signature fatty acid analyses suggest reef manta rays feed on demersal zooplankton. PLoS One 8, e77152. http://dx.doi.org/10.1371/journal.pone.0077152. Dalsgaard, J., John, M.S., Kattner, G., Müller-Navarra, D., Hagen, W., 2003. Fatty acid trophic markers in the pelagic marine environment. Adv. Mar. Biol. 46, 225–340. Dang, F., Wang, W.-X., 2012. Why mercury concentration increases with fish size? Biokinetic explanation. Environ. Pollut. 163, 192–198. http://dx.doi.org/10.1016/j. envpol.2011.12.026. El-Sabaawi, R., Dower, J.F., Kainz, M., Mazumder, A., 2008. Characterizing dietary variability and trophic positions of coastal calanoid copepods: insight from stable isotopes and fatty acids. Mar. Biol. 156, 225–237. http://dx.doi.org/10.1007/ s00227-008-1073-1. Folch, J., Lees, M., Sloane-Stanley, G.H., 1957. A simple method for the isolation and purification of total lipids from animal tissues. J. Biol. Chem. 226, 497–509. France, R.L., 1995. Carbon-13 enrichment in benthic compared to planktonic algae: foodweb implications. Mar. Ecol. Prog. Ser. 124, 307–312. http://dx.doi.org/10. 3354/meps124307. Gerwick, W.H., Bernart, M.W., 1993. Eicosanoids and related compounds from marine algae. In: Attaway, D.H., Zaborsky, O.R. (Eds.), Pharmaceutical and Bioactive Natural Products. Springer, US, pp. 101–152. http://dx.doi.org/10.1007/978-1-4899-23912_4. Hebert, C.E., Weseloh, D.V.C., Idrissi, A., Arts, M.T., O’Gorman, R., Gorman, O.T., Locke, B., Madenjian, C.P., Roseman, E.F., 2008. Restoring piscivorous fish populations in the laurentian great lakes causes seabird dietary change. Ecology 89, 891–897. http://dx.doi.org/10.1890/07-1603.1. Hebert, C.E., Weseloh, D.V.C., Gauthier, L.T., Arts, M.T., Letcher, R.J., 2009. Biochemical tracers reveal intra-specific differences in the food webs utilized by individual seabirds. Oecologia 160, 15–23. http://dx.doi.org/10.1007/s00442-009-1285-1. Iverson, S.J., 2009. Tracing aquatic food webs using fatty acids: from qualitative indicators to quantitative determination. In: Kainz, M.J., Brett, M.T., Arts, M.T. (Eds.), Lipids in Aquatic Ecosystems. Springer, pp. 281–308. Kainz, M., Telmer, K., Mazumder, A., 2006. Bioaccumulation patterns of methyl mercury and essential fatty acids in lacustrine planktonic food webs and fish. Sci. Total Environ. 368, 271–282. http://dx.doi.org/10.1016/j.scitotenv.2005.09.035. Selected papers from the 7th International Conference on Mercury as a Global Pollutant, Ljubljana, Slovenia June 27–July 2, 2004. Kelly, B.C., Ikonomou, M.G., Blair, J.D., Morin, A.E., Gobas, F.A.P.C., 2007. Food webspecific biomagnification of persistent organic pollutants. Science 317, 236–239. http://dx.doi.org/10.1126/science.1138275. Kojadinovic, J., Potier, M., Le Corre, M., Cosson, R.P., Bustamante, P., 2006. Mercury content in commercial pelagic fish and its risk assessment in the Western Indian Ocean. Sci. Total Environ. 366, 688–700. http://dx.doi.org/10.1016/j.scitotenv. 2006.02.006. Kojadinovic, J., Potier, M., Le Corre, M., Cosson, R.P., Bustamante, P., 2007. Bioaccumulation of trace elements in pelagic fish from the Western Indian Ocean. Environ. Pollut. 146, 548–566. http://dx.doi.org/10.1016/j.envpol.2006.07.015. Lichens in a Changing Pollution Environment. Koussoroplis, A.-M., Bec, A., Perga, M.-E., Koutrakis, E., Bourdier, G., Desvilettes, C., 2011. Fatty acid transfer in the food web of a coastal Mediterranean lagoon: evidence for high arachidonic acid retention in fish. Estuar. Coast. Shelf Sci. 91, 450–461. http://dx.doi.org/10.1016/j.ecss.2010.11.010. Lavoie, R.A., Jardine, T.D., Chumchal, M.M., Kidd, K.A., Campbell, L.M., 2013. Biomagnification of mercury in aquatic food webs: a worldwide meta-analysis. Environ. Sci. Technol. 47, 13385–13394. Le Croizier, G., Schaal, G., Gallon, R., Fall, M., Le Grand, F., Munaron, J.-M., Rouget, M.L., Machu, E., Le Loc’h, F., Laë, R., De Morais, L.T., 2016. Trophic ecology influence on metal bioaccumulation in marine fish: inference from stable isotope and fatty acid analyses. Sci. Total Environ. 573, 83–95. http://dx.doi.org/10.1016/j.scitotenv. 2016.08.035. Litzow, M.A., Bailey, K.M., Prahl, F.G., Heintz, R., 2006. Climate regime shifts and reorganization of fish communities: the essential fatty acid limitation hypothesis. Mar. Ecol. Prog. Ser. 315, 1–11. http://dx.doi.org/10.3354/meps315001. Loseto, L.L., Stern, G.A., Deibel, D., Connelly, T.L., Prokopowicz, A., Lean, D.R.S., Fortier, L., Ferguson, S.H., 2008. Linking mercury exposure to habitat and feeding behaviour in Beaufort Sea beluga whales. J. Mar. Syst. 74, 1012–1024. http://dx.doi.org/10. 1016/j.jmarsys.2007.10.004. Sea ice and life in a river-influenced arctic shelf ecosystem. Müller-Navarra, D.C., Brett, M.T., Liston, A.M., Goldman, C.R., 2000. A highly unsaturated fatty acid predicts carbon transfer between primary producers and consumers. Nature 403, 74–77. http://dx.doi.org/10.1038/47469. Macko, S.A., Estep, M.L.F., Engel, M.H., Hare, P.E., 1986. Kinetic fractionation of stable nitrogen isotopes during amino acid transamination. Geochim. Cosmochim. Acta 50,

and the inclusion of a large wahoo (120 cm fork length) that had a high level of contamination (1.93 ppm ww). Nonetheless, it confirms that A. solandri has a higher trophic level than other pelagic species such as tunas, as suggested by the other two tracing methods. Among the EFA, DHA was present most often, with the highest levels observed in the two species of snapper E. coruscans and A. virescens, the temperate tuna T. alalunga and the dolphinfish C. hippurus (DHA > 35%). Studies from the Pacific Ocean have also reported the latter two species having high DHA levels (∼25%; Ostrowski and Divakaran, 1989; Parrish et al., 2015). In comparison to most of the other fish considered, the Indian mackerel R. kanagurta returned low overall EFA values despite having a relatively high EPA level (∼6%). This result is consistent with the findings of Sahena et al. (2010) in Malaysia that reported ∼11% for EPA and DHA and ∼2% for ARA. In the crustacean species, ARA and EPA were the main EFA detected (> 15% for both). 5. Conclusion The trophic scenarios generated by the three ecological tracers allowed us to evaluate how method choice can influence our general understanding of trophic ecology in the study area. Correlations among the tracers can strengthen the ecological interpretations that can be made when looking at data that was originally obtained for another purpose e.g., for nutritional or food safety reasons (i.e., EFA and THg), a tool which can be useful in ecological meta-analyses. Depending on the ecological tracer chosen, differences in trophic structures can be observed at the species level, especially for migratory species, as shown here. We recommend (i) to preferentially work at the individual level and (ii) that each ecosystem or food web that is studied should be clearly delimited before analysing the ecological tracers, especially as linear relationships may not be consistent between higher and lower trophic levels due to the species migrations between ecosystems. Acknowledgments This work will contribute to the SEYFISH project, co–funded by the Institute for Research and Development (IRD) and the Seychelles Fishing Authority (SFA). We thank all the SFA staff and fishermen who helped throughout the sampling period. Particular thanks go to Natifa Pillay and Maria Cedras. We are also grateful to the team of UMR TOXALIM (Aurélien Amiel, Edwin Fouché, Laurent Debrauwer and Hervé Guillou) for hosting us and helping with the fatty acid analyses and to Jane Alpine for professional job of editing the manuscript. Finally, we warmly thank the two anonymous reviewers for their comments and wise suggestions that clarified and improved our manuscript. References Arts, M.T., Ackman, R.G., Holub, B.J., 2001. Essential fatty acids in aquatic ecosystems: a crucial link between diet and human health and evolution. Can. J. Fish. Aquat. Sci. 58, 122–137. http://dx.doi.org/10.1139/f00-224. Bodin, N., Budzinski, H., Le Ménach, K., Tapie, N., 2009. ASE extraction method for simultaneous carbon and nitrogen stable isotope analysis in soft tissues of aquatic organisms. Anal. Chim. Acta 643, 54–60. http://dx.doi.org/10.1016/j.aca.2009.03. 048. Bodin, N., Lesperance, D., Albert, R., Hollanda, S.J., Michaud, P., Degroote, M., Churlaud, C., Bustamante, P., 2017. Trace elements in oceanic pelagic communities in the western Indian Ocean. Chemosphere 174, 354–362. http://dx.doi.org/10.1016/j. chemosphere.2017.01.099. Bond, A.L., 2010. Relationships between stable isotopes and metal contaminants in feathers are spurious and biologically uninformative. Environ. Pollut. 158, 1182–1184. http://dx.doi.org/10.1016/j.envpol.2010.01.004. Budge, S.M., Iverson, S.J., Koopman, H.N., 2006. Studying trophic ecology in marine ecosystems using fatty acids: a primer on analysis and interpretation. Mar. Mammal Sci. 22, 759–801. http://dx.doi.org/10.1111/j.1748-7692.2006.00079.X. Cai, Y., Rooker, J.R., Gill, G.A., Turner, J.P., 2007. Bioaccumulation of mercury in pelagic fishes from the northern Gulf of Mexico. Can. J. Fish. Aquat. Sci. 64, 458–469. http:// dx.doi.org/10.1139/f07-017. Caraveo-Patiño, J., Wang, Y., Soto, L.A., Ghebremeskel, K., Lehane, C., Crawford, M.A., 2009. Eco-physiological repercussions of dietary arachidonic acid in cell membranes

323

Ecological Indicators 81 (2017) 315–324

F. Sardenne et al.

dx.doi.org/10.1890/110140. Rooker, J.R., Turner, J.P., Holt, S.A., 2006. Trophic ecology of Sargassum-associated fishes in the Gulf of Mexico determined from stable isotopes and fatty acids. Mar. Ecol. Prog. Ser. 313, 249–259. Sahena, F., Zaidul, I.S.M., Jinap, S., Yazid, A.M., Khatib, A., Norulaini, N.A.N., 2010. Fatty acid compositions of fish oil extracted from different parts of Indian mackerel (Rastrelliger kanagurta) using various techniques of supercritical CO2 extraction. Food Chem. 120, 879–885. http://dx.doi.org/10.1016/j.foodchem.2009.10.055. Sardenne, F., Bodin, N., Chassot, E., Amiel, A., Fouché, E., Degroote, M., Hollanda, S., Pethybridge, H., Lebreton, B., Guillou, G., Ménard, F., 2016. Trophic niches of sympatric tropical tuna in the Western Indian Ocean inferred by stable isotopes and neutral fatty acids. Prog. Oceanogr. 146, 75–88. http://dx.doi.org/10.1016/j.pocean. 2016.06.001. Sardenne, F., Kraffe, E., Amiel, A., Fouché, E., Debrauwer, L., Ménard, F., Bodin, N., 2017. Biological and environmental influence on tissue fatty acid compositions in wild tropical tunas. Comp. Biochem. Physiol. A Mol. Integr. Physiol. 204, 17–27. http:// dx.doi.org/10.1016/j.cbpa.2016.11.007. Spitz, J., Trites, A.W., Becquet, V., Brind’Amour, A., Cherel, Y., Galois, R., Ridoux, V., 2012. Cost of living dictates what whales, dolphins and porpoises eat: the importance of prey quality on predator foraging strategies. PLoS One 7, e50096. http://dx.doi. org/10.1371/journal.pone.0050096. Stowasser, G., McAllen, R., Pierce, G.J., Collins, M.A., Moffat, C.F., Priede, I.G., Pond, D.W., 2009. Trophic position of deep-sea fish—assessment through fatty acid and stable isotope analyses. Deep Sea Res. Part Oceanogr. Res. Pap. 56, 812–826. http:// dx.doi.org/10.1016/j.dsr.2008.12.016. Tocher, D.R., 2003. Metabolism and functions of lipids and fatty acids in teleost fish. Rev. Fish. Sci. 11, 107–184. http://dx.doi.org/10.1080/713610925. Trystram, C., Roos, D., Guyomard, D., Jaquemet, S., 2015. Mechanisms of trophic partitioning within two fish communities associated with a tropical oceanic island. West. Indian Ocean J. Mar. Sci. 14, 93–111. van Duyl, F.C., Moodley, L., Nieuwland, G., Ijzerloo, L., van Soest, R.W.M., van Houtekamer, M., Meesters, E.H., Middelburg, J.J., 2011. Coral cavity sponges depend on reef-derived food resources: stable isotope and fatty acid constraints. Mar. Biol. 158, 1653–1666. http://dx.doi.org/10.1007/s00227-011-1681-z. Vander Zanden, M.J., Cabana Rasmussen, G.J.B., 1997. Comparing trophic position of freshwater fish calculated using stable nitrogen isotope ratios (δ15N) and literature dietary data. Can. J. Fish. Aquat. Sci. 54, 1142–1158. http://dx.doi.org/10.1139/f97016. Zudaire, I., Murua, H., Grande, M., Goñi, N., Potier, M., Ménard, F., Chassot, E., Bodin, N., 2015. Variations in the diet and stable isotope ratios during the ovarian development of female yellowfin tuna (Thunnus albacares) in the Western Indian Ocean. Mar. Biol. 1–15. http://dx.doi.org/10.1007/s00227-015-2763-0.

2143–2146. Matthews, A.D., 1983. Mercury content of commercially important fish of the Seychelles, and hair mercury levels of a selected part of the population. Environ. Res. 30, 305–312. http://dx.doi.org/10.1016/0013-9351(83)90216-5. McMeans, B.C., Arts, M.T., Fisk, A.T., 2015. Impacts of food web structure and feeding behavior on mercury exposure in Greenland Sharks (Somniosus microcephalus). Sci. Total Environ. 509–510, 216–225. http://dx.doi.org/10.1016/j.scitotenv.2014.01. 128. (Special Issue: Mercury in Canada’s North). Minagawa, M., Wada, E., 1984. Stepwise enrichment of 15N along food chains: further evidence and the relation between δ15N and animal age. Geochim. Cosmochim. Acta 48, 1135–1140. http://dx.doi.org/10.1016/0016-7037(84)90204-7. Mourente, G., Megina, C., Díaz-Salvago, E., 2001. Lipids in female northern bluefin tuna (Thunnus thynnus thynnus L.) during sexual maturation. Fish Physiol. Biochem. 24, 351–363. http://dx.doi.org/10.1023/A. 1015011609017. Munday, P.L., Cheal, A.J., Graham, N.A.J., Meekan, M., Pratchett, M.S., Sheaves, M., Sweatman, H., Wilson, S.K., 2009. Tropical coastal fish. In: Poloczanska, E.S., Hobday, A.J., Richardson, A.J. (Eds.), Marine Climate Change Impacts and Adaptation Report Card for Australia 2009. National Climate Change Adaptation Research Plans. Oksanen, J., Kindt, R., Legendre, P., O’Hara, B., Stevens, M.H.H., Oksanen, M.J., Suggests, M., 2007. The vegan package. Community Ecol. Package 10, 631–637. Ostrowski, A.C., Divakaran, S., 1989. The amino acid and fatty acid compositions of selected tissues of the dolphin fish (Coryphaena hippurus) and their nutritional implications. Aquaculture 80, 285–299. http://dx.doi.org/10.1016/0044-8486(89) 90176-2. Palomares, M.L.D., Pauly, D., 2017. SeaLifeBase [WWW Document]. www.sealifebase. org. URL http://www.sealifebase.org/search.php (Accessed 19 May 2017). Parrish, C.C., Pethybridge, H., Young, J.W., Nichols, P.D., 2015. Spatial variation in fatty acid trophic markers in albacore tuna from the southwestern Pacific Ocean—a potential ‘tropicalization’ signal. Deep Sea Res. Part II Top. Stud. Oceanogr. 113, 199–207. http://dx.doi.org/10.1016/j.dsr2.2013.12.003. (Impacts of climate on marine top predators). Peres-Neto, P.R., Jackson, D.A., 2001. How well do multivariate data sets match? The advantages of a Procrustean superimposition approach over the Mantel test. Oecologia 129, 169–178. http://dx.doi.org/10.1007/s004420100720. Post, D.M., Layman, C.A., Arrington, D.A., Takimoto, G., Quattrochi, J., Montaña, C.G., 2007. Getting to the fat of the matter: models, methods and assumptions for dealing with lipids in stable isotope analyses. Oecologia 152, 179–189. http://dx.doi.org/10. 1007/s00442-006-0630-X. R Core Team, 2013. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Ramos, R., González-Solís, J., 2012. Trace me if you can: the use of intrinsic biogeochemical markers in marine top predators. Front. Ecol. Environ. 10, 258–266. http://

324