The foodweb of El Tóbari Lagoon (central-east Gulf of California) was studied for four seasons through the carbon and nitrogen isotopic characterization of primary producers, invertebrates, fish, birds and potential food sources. The range of δ13C measured was much wider for potential food sources than for consumers. Many organisms presented enrichments of δ13C and δ15N values. There was a clear trend toward increasing δ13C and δ15N from base organisms to top-predators in the four seasons. The isotopic and percentage of contribution data confirmed that suspended particulate organic matter and phytoplankton are the main organic source supporting the foodweb. Our results also imply the occurrence of a nutrient transfer from zooplankton to some invertebrates and juvenile fishes. Consumers were composed in four trophic levels, with trophic level 2 occupied by zooplankton and filter-feeders and trophic level 4 occupied by carnivorous fish and most bird species. Carnivorous fish exhibited dietary similarities by a considerable sharing of resources, which could be related to the abundance of possible prey, between invertebrates and juvenile fishes. Crustaceans and fish represented the main food sources of birds, although some birds showed more dietary variation (marine and offshore prey).
Although coastal lagoons are considered among the most productive ecosystems, they are also some of the most threatened ecological systems in the world. They have high levels of primary production, are a source of nutrients via microbial loop and provide diverse habitats that offer optimal niches for numerous aquatic species that utilize these areas as refuge and/or breeding grounds (Abrantes and Sheaves, 2008).
The Gulf of California has a diversity of habitats that include mangrove forest, salt marshes, intertidal pools, swamps, freshwater inner lagoons, and brackish and sea water systems (Páez-Osuna et al., 2016). The need to account for the structure of ecosystems with respect to aspects such as the biological sustainability of living aquatic resources, the dynamics in the coupling between different ecosystems and the development of ecosystem management has been widely recognized (Martineau et al., 2004; Vizzini and Mazzola 2006). Information on energy sources and trophic structure is therefore fundamental for the understanding of the dynamics and persistence of these communities through time (Martineau et al., 2004; Carrasco et al., 2016). Moreover, coastal lagoons are complex environments that are connected to terrestrial and marine domains, and several organic matter sources are available to consumers. These ecosystems are composed of trophic diverse groups, encompassing species of different sizes and diverse feeding strategies (Martineau et al., 2004; Carrasco et al., 2016).
From an ecological point of view, the reconstruction of marine foodwebs has largely been constrained by methodological difficulties in obtaining the appropriate data, which are often derived from time-consuming analyses of gut contents of consumer species. Gut content studies have limitations, for example the large number of observations required to obtain the appropriate information over the time and space scales of interest and the fact that it is often not possible to identify all prey items (Zetina-Rejón et al., 2003; Carlier et al., 2007). The stable isotope analysis had been frequently used in the study of estuarine foodwebs to detail the trophic organization because differences in the natural abundance of isotopes of elements such as carbon (C) and nitrogen (N) between consumers and their diet are a consequence of nutrient and energy sources and exchanges, and they reflect trophic relationships (Post, 2002; Fry, 2006). In the subtropical lagoons there are diverse trophic groups, encompassing species of different sizes and diverse feeding strategies.
There are many detailed foodweb studies that have been published for coastal ecosystems in different areas (Carlier et al., 2007; Vander-Zanden and Fetzer, 2007) and some published studies from estuarine areas have aimed at identifying the sources of energy for one specie or a group of species, generally in relation to the importance of mangroves, salt marsh or seagrass habitats as a source of nutrients, and others to detailing relationships between species and among the estuarine community as a whole (Abrantes and Sheaves, 2008). Some studies (Nyunja et a., 2009; Mendoza-Carranza et al., 2010) found that the contribution of mangroves to foodwebs was confined only to mangroves areas, but a mixture of macroalgae and phytoplankton was a major carbon source for organisms in areas distant of mangroves or where they were scarce. Other studies confirmed that there are significant contributions of terrestrial organic matter if the rain regime is significantly high among seasons (Riera, 2007). Carlier et al. (2007) also reported the seasonal variations in the contribution of nutrients on coastal ecosystems. Both studies reported that detritus and phytoplankton supported the foodwebs.
In Mexico, foodweb studies are scarce for estuarine areas, which is important to determine the food sources and understand the structure and dynamics of biotic communities in these ecosystems. For example, Zetina-Rejón et al. (2003) used a mass-balance trophic model to describe the ecosystem structure and flows of energy in a subtropical lagoon of Gulf of California, and reported that detritus and phytoplankton supported the foodweb. Therefore, in this study, the stable isotope composition of organisms with different feeding habits collected in El Tóbari Lagoon, a subtropical coastal ecosystem of the central-east Gulf of California, was analyzed to provide a comprehensive picture of the trophic interactions occurring in the ecosystem. This ecosystem has areas bordered by mangroves in north and south zones but they are scarce in central area, with the rainy season during summer. The specific aims were to (1) identify the food sources supporting the dominant species in the foodweb, (2) investigate if there are seasonal variations of the organic matter sources. Our hypotheses were (1) phytoplankton and detritus are the main sources of food fueling the lagoon ecosystem foodweb; and (2) the stable isotopes signatures in the sources of food varies between seasons due to higher terrestrial run-off.
Materials and methods
The study was performed in El Tóbari Lagoon, located in the Northwestern Sonora state of Mexico, along the central-east coast of the Gulf of California (Figure 1). The coastal ecosystem has surface of 64.2 km2 and some areas bordered by two species of mangroves (Rhizophora mangle and Avicennia germinans). The predominant climate is semi-arid (warm-dry in spring, hot-rainy in summer, warm-rainy in autumn and cold-dry in winter), with an average annual precipitation of 259 mm. The ecosystem receives freshwater effluents from land runoff and some streams and is subject to regular effluents discharges from 10 drainage channels in the Yaqui Valley agricultural area. It also receives effluents from a nearby shrimp aquaculture farm (1190 ha) as well as untreated municipal sewage from surrounding towns. The ecosystem has an important ecological role, supporting a variety of endemic and migratory organisms including mangroves, macroalgae, mollusks, fish and birds. Thus, the lagoon has been included among the Conservation of Coastal and Oceanic Priority Zones, as determined by the National Commission for the Knowledge and Use of Biodiversity of Mexico (Aguilar et al., 2008).
Sampling was conducted during four seasons: summer (August 2011), winter (February 2012), spring (May 2012) and autumn (October 2012) in different sites of El Tóbari Lagoon (Figure 1). Seven sites with the most biodiversity (according to Aguilar et al. 2008) inside the lagoon were selected, where water, sediment and organism samples were collected. All samples were taken randomly to account for spatial heterogeneity effects. Water samples for the suspended particulate organic matter (SPOM) collection (duplicate per site, 14 samples for each season) were taken at the water surface using low density polyethylene bottles, stored in ice and transported to the laboratory. For stable isotope analysis, the SPOM was obtained by filtration on pre-combusted (400 °C for 4 h) Whatman GF/F glass fiber membranes. Subsequently, the membranes were washed with HCl 0.1N to remove carbonates, dried at 40 °C for 8 h and kept frozen until analysis. To characterize the organic matter in the sediments, duplicate surface sediment samples (14 samples for each season, at a 2 cm depth from the sediment surface) were taken with a Van Veen dredge and placed on ice in plastic containers for transport to the laboratory.
Different biological samples, representing producer and consumer compartments of the trophic network, were collected at ecosystem sites. Phytoplankton and zooplankton organisms were collected with 30 μm mesh and 270 μm mesh conical nets, respectively; towing of the plankton net was conducted 2 knots for ∼7 min, between the sites of the lagoon. Mangrove leaves (Rhizophora mangle and Avicennia germinans), macroalgae (Ulva intestinalis, Ulva lactuca, Gracilaria vermiculophylla and Spyridia filamentosa), bivalves (Chione gnidia, Chione fluctifraga, Crassostrea cortezienzes, Crassostrea gigas and Fistulobalanus dentivarians) and the snail Hexaplex erythrostoma were collected by hand.
Crustaceans (Litopenaeus vannamei and Callinectes arcuatus) and fishes (Mugil cephalus, Gerres cinereus, Haemulopsis leuciscus and Lutjanus argentiventris) were captured near selected sites using commercial gill nets. Two species of avifauna (Yellow-Crowned Night Heron Nyctanassa violacea and Clapper Rail Rallus longirostris) were captured using net traps, and blood samples, which were obtained from a wing vein using a commercial syringe, were stored in clean glass vials. Plasma was separated by centrifugation (3000 rpm by 10 min) and decanted. Naturally dead specimens of magnificent Frigate Bird (Fregata magnificens) and Cormorant (Phalacrocorax brasilianus) were also collected. Species identifications were made using the descriptions from Fischer et al. (1995a, b, c), and Sibley (2002). All samples were placed on ice for transport to the laboratory, washed with bi-distilled water and dissected in the laboratory. Mantle muscles of bivalves and snails, claw muscles of crabs, dorsal white muscles tissue of fish, chest muscles of magnificent Frigate Bird and Cormorant, and plasma-free blood of Yellow-Crowned Night Heron and Clapper Rail was taken for C and N stable isotope analyses. Surface sediment and biological samples were freeze-dried at -48 °C and 32 x10−3 mbar (Labconco freeze-dry system) for 72 h and treated with HCl 0.1N to remove carbonates, dried at 50 °C for 8 h, powdered, stored in clean low density polyethylene vials and kept frozen until C and N stable isotope analysis. White muscle for fishes is the best tissue for use in ecological work because of low isotopic variability than all other tissues. Samples of consumers were treated with methanol:chloroform (2:1 v/v) for 24 h to remove lipids, dried at 50 °C for 12 hand kept frozen until analysis. According to Logan et al. (2008), the effects of lipid remotion on the δ15N composition in tissues of organisms is minimal when the relation C:N is ≤4, and then no further corrections for invertebrates, fishes and birds were required.
The δ13C and δ15N data were tested for normality and variance homoscedasticity. As all stable isotope values follow a non-normal distribution according to the Shapiro-Wilk and Bartlett tests (p < 0.05), Kruskal-Wallis and Student-Newman-Keuls tests (Glantz 2002) were used to compare the differences in mean δ13C and δ15N values between the four sampling surveys in the lagoon ecosystem.
In order to quantify consumer diets in the ecosystem during the four sampling, a stable isotope mixing model of Phillips and Gregg (2003) was applied. The Isosource software package was used to run mixing equations using increments of 1 and 2.5%, and tolerances between 0.1 and 3‰. This is a linear model based on mass balance equations, where all possible combinations of each source contribution (0-100%) are examined in small increments (in this case 1-2.5%). According to authors, combinations that sum to the observed mixture isotopic signatures within a small tolerance (e.g. 0.1-3 ‰) are considered to be feasible solutions, from which the range of potential source contributions can be determined. Since each of these feasible source combinations are constrained to sum 100%, there are tradeoffs among the sources within their suitable ranges.
Stable isotopes compositions
Mean δ13C and δ15N of abiotic organic matter sources (sediment and SPOM), primary producers, and consumers (primary, secondary and tertiary) collected at El Tóbari Lagoon in the four seasons are presented in Figure 2. Although the number of samples varied among taxa, the δ13Cand δ15N ranges of each species were quite small with a coefficient of variation of less than 10% for most of the species analyzed. Sediments showed similar δ13C and δ15N values during the four seasons, while SPOM presented enrichment in δ13C signatures during the winter season and slightly enriched of δ15N signatures during summer and spring seasons. The highest δ13C values for phytoplankton were determined during the winter season with slight seasonal variations in their δ15N values.
The mangroves presented the lowest δ13C signatures in the four seasons (range from -27.24 to -25.74‰; Figure 2) with respect to other organisms and generated a dispersion of data; their δ15N values were also similar during the all seasons with slightly higher values for A. germinans (highest mean value of 12.10‰ in summer). Among macroalgae species, the lowest and highest δ13C values were found during summer for S. filamentosa (-21.98‰) and U. lactuca (-13.32‰), respectively. For δ15N, U. lactuca presented the highest (15.748‰ in autumn) and the lowest (10.40‰ in winter) values, while the other species did not present a seasonal range. Zooplankton were enriched with δ13C in winter (-17.51‰), and their δ15N values varied slightly between seasons ange from 9.95 to 11.78‰). The filter-feeder F. dentivarians exhibited a low δ13C value in the summer season (-20.03‰) and the other filter feeders showed homogeneous isotopic signatures of C (range from -16.9 to -16.21‰) and N (range from 12.12 to 14.24‰) during the four seasons. Among of the primary consumers, the omnivores presented slight variations in δ13C and δ15N composition, with the highest δ13C values belonging to H. erythrostoma (-13.18‰) and the highest δ15Nvalues belonging to the juveniles of C. arcuatus and G. cinereus (15.09‰ and 15.80‰, respectively) in summer. The H. leuciscus juveniles presented the highest and most homogeneous seasonal δ15N values (range from 15.62 to 17.22‰), while L. argentivensis had enriched δ13C values in winter (-13.95‰) and low δ15N values in spring (14.90‰). The tertiary consumers F. magnificens and P. brasilianus showed low δ13C isotopic composition (ranges from -18.92 to -18.01‰ and from -18.75 to -17.10‰, respectively) and the highest δ15N values (ranges from 18.19 to 19.44‰ and from 17.60 to 18.64‰, respectively) throughout the seasons.
Trophic groups and organic matter contributions
The distribution of δ13C and δ15N in the organic matter sources and the functional groups (primary producers and consumers) are presented in the Figure 3. Mangroves formed a separate group from other organisms and appeared as outliers, indicating a scarce carbon contribution to consumers. The abiotic organic matter sources had similar δ13C values of most primary producers, which may be derived from the incorporation of labile organic matter in sediments and water column to primary producers. The primary, secondary and tertiary consumers are grouped by their δ13C values, indicating flows of carbon through these organism communities. The pattern of δ15N values in functional groups were tertiary consumers > secondary consumers > primary consumers > primary producers, according to their trophic position.
The ranges of contribution to consumers in El Tóbari Lagoon were different among seasons. The filter-feeders presented seasonal variations in their dietary contributions. Phytoplankton has different ranges of contributions for C. gnidia, C. fluctifraga, C. cortezienzes and C. gigas during most seasons (from 0-53% to 93-100%). The balanus F. dentivarians exhibited similar seasonal contributions from sediments (from 0-4% to 0-53%) and SPOM (from 0-9% to 50-100%), and phytoplankton showed significant contributions during spring (86-100%) and autumn (67-100%) seasons. The sediments were the principal contributions for the crustaceans L. vannamei and H. erythrostoma, and for the juveniles and adults of the fish M. cephalus (from 0-78% to 38-100%); zooplankton also represented variable contributions for L. vannamei (0-98 to 0-100%) and juveniles M. cephalus (7-100 to 10-98%). The Shrimp L. vannamei was the main contributor for the juveniles of C. arcuatus during all seasons (from 8-95% to 93-100%). The juveniles of fishes G. cinereus, H. leuciscus and L. argentivensis presented slightly seasonal variations in their diet contributions, which were based on L. vannamei (from 0-100%), H. erythrostoma (from 0-78% to 0-94%), and juveniles of M. cephalus (from 3-91% to 0-100%) and C. arcuatus (from 0-100% to 7-100%).
Trophic levels in foodweb
The adults of the Crab C. arcuatus and the three species of fishes exhibited seasonal contributions from Shrimp L. vanamei (from 0-88% to 15-100%) and juveniles of fishes (from 0-76 to 0-100%). The bivalves C. gnidia and C. fluctifraga had similar seasonal contributions (from 0-48% to 0-85% and from 0-45% to 0-81%, respectively) for G. cinereus, H. leuciscus and L. argentivensis. The birds presented diets derived of juveniles of carnivore fishes (0-100%). However, shrimp and adults of Crab C. arcuatus had variable contributions (from 0-95% to 0-100%, and 0-100%, respectively) for the cormorant (P. brasilianus) and the magnificent Frigate Bird (F. magnificens).
Consumers were distributed into five trophic levels in El Tóbari Lagoon (Figure 4), considering the primary producers as the trophic level 1. Trophic level 2 was composed of Zooplankton, two Oysters (C. corteziensis and C. gigas), two Clams (C. gnidia and C. fluctifraga), the balanus F. dentivarians in two seasons (2.3 in summer and 2.7 in winter), the Snail H. erythrostoma and the Shrimp L. vannamei in three seasons (2.8 in summer, 2.7 in winter and 2.6 in spring). Trophic level 3 was composed by the balanus F. dentivarians in two seasons (3.2 in spring and 3.0 in autumn), the Shrimp L. vannamei in one season (3.1 in autumn), the juvenile of three fish species (M. cephalus, G. cinereusand H. leuciscus), the adult of M. cephalus in two seasons (3.4 in winter and 3.5 in spring) and the juvenile of C. arcuatus. Trophic level 4 was composed by the adult of M. cephalus in two seasons (4.1 in summer and 4.0 in autumn), the adult of the Crab C. arcuatus, the adult of carnivore fishes and most of bird species (from 4.0 for N.violacea in summer to 4.9 for P. brasilianus in spring and F. magnificens in summer). Trophic level 5 was occupied by F. magnificens (5.2 in spring and 5.5 in autumn).
The δ13C values of sediments and SPOM (ranges of -21.91to -19.86‰ and of -21.07 to -16.63‰, respectively) were close to the values reported in different coastal lagoons of temperate (Riera, 2007; Carlier et al., 2008) and subtropical latitude regions (Serrano-Grijalva et al., 2011) and reflected seasonal fluctuations. The enrichment of δ15N values in SPOM during summer and spring seasons may reflect anthropogenic inputs, as increments that were reported in aquaculture (from 1‰ to 17‰) and sewage domestic (from 1 to 3‰) effluents impacted ecosystems (Vizzini and Mazzola, 2006; Serrano-Grijalva et al., 2011). The isotopic values of SPOM suggest that are important part of the base of foodweb in El Tóbari Lagoon, in contrast with our hypotheses, considering that sediment contributions were mainly for secondary consumers.
The mangrove detritus exhibited a carbon isotope signal in the same range as the measured δ13C values for mangrove leaves (Nyunja et al., 2009; Serrano-Grijalva et al., 2011). In this sense, the differences in δ13C between mangroves and primary consumers, and the relative high δ13C values of sediment close to mangrove vegetation, imply the low contribution of mangroves to El Tóbari Lagoon foodweb, which was confirmed by the ranges of percentages determined for both mangrove species (from 0-12% to 0-43%). This low contribution is in accordance with previous studies conducted in subtropical lagoons (Nyunja et al., 2009; Serrano-Grijalva et al., 2011) and in contrast to those conducted in tropical lagoons (Abrantes and Sheaves, 2008; Mendoza-Carranza et al., 2010). Bouillon et al. (2000) explain these differences by the selective mode of feed by plankton organisms on bulk suspended matter, the abundance of nutritious material and the predominance of certain plankton species.
The δ13C composition of most macroalgae species (from -19.59 to -13.32‰) was similar to that reported in different ecosystems (Dittel et al., 2006; Vizzini and Mazzola, 2006; Serrano-Grijalva et al., 2011) and U. lactuca presented temporal variations (significant enrichment in summer, -13.3‰). Serrano-Grijalva et al. (2011) reported the increments in δ13C signatures of macroalgae as disturbances produced by organic matter from shrimp aquiculture wastes in a subtropical lagoon, considering that phytoplankton uptake the isotopically light dissolved inorganic carbon resulting from the bacterial loop (Bouillon et al., 2000) and macroalgae use of heavier carbon isotope of the water column. Piñón-Gimate et al. (2009) also reported the use preferential of heavier nutrients isotopes by macroalgae in sites impacted by agricultural and sewage effluents. In the present study, these sources in El Tóbari Lagoon were associated with effluents from shrimp farm and agricultural drains. The shift of natural to anthropogenic sources in the ecosystem was likely aggravated by the construction of the stone barrier in 1973. This barrier changed the tidal regime inside the lagoon and reduced the flow of water in and out of the lagoon, which decreased current velocities; these changes led to an increase in sedimentation and the consequential limitation of the exportation of organic matter to adjacent areas. However, further studies are necessary to explore these possibilities, because the other macroalgae species of El Tóbari Lagoon do not showed temporal variations.
The δ13C composition of phytoplankton (from -19.40 to -17.62‰) is within the range of that reported in other lagoon ecosystems (from -18 to -22‰) (Hsieh et al., 2002; Dittel et al., 2006). The isotopic contributions determined by Isosource program infer that SPOM and phytoplankton are the main organic source supporting the foodweb of El Tóbari Lagoon, confirming partially the hypotheses that phytoplankton and detritus are the bases of foodweb in the studied area. These data are similar to the results found by Hsieh et al. (2002), but in contrast with the findings of Serrano-Grijalva et al. (2011), that reported the sediment detritus as the base of the foodweb. The ranges of feasible contributions for zooplankton reflect that its diets are supported by the phytoplankton and SPOM, as reported by other authors (Bouillon et al., 2000; Maguire and Jonathan, 2006; Carlier et al., 2006). The δ13C composition of zooplankton exhibited a seasonal variation (with significant enrichment in winter, -17.51‰). These changes are indicative that there are seasonal variabilities in the sources of organic matter in the ecosystem or by the selective mode of feed by zooplankton organisms on bulk suspended matter and the predominance of certain zooplankton species, as was reported by Bouillon et al. (2000). Previous research has shown that the relative importance of allochthonous sources of organic carbon decreases with increasing autochthonous nutrients in the column (Maguire and Jonathan, 2006); however, the impact of anthropogenic organic matter is related to the increment of nutrients caused by bacterial activity, which produce changes in the δ13C signatures of primary producers and consumers (Bouillon et al., 2000; Carlier et al., 2007). The main dietary contributors of filter feeders were phytoplankton, SPOM and zooplankton. Martineau et al. (2004) and Carlier et al. (2007) found that filter feeders may preferentially feed on a 13C-enriched fraction within sediment and SPOM pools, through selective feeding and/or absorption, which could account for some differences between the δ13Cvalues of filter feeders and those of sediment and SPOM.
Omnivores also exhibited a preferential use of food sources. Their δ13C composition did not present significant seasonal variations. The snail H. erythrostoma presented the most homogeneous δ13C values, which indicated a consistent diet, with the main dietary contributions from sediments and, in less degree, from macroalgae and some filter feeders. Fischer et al. (1995a) describe their food habits as omnivorous, with preference for large clams (e.g. Megapitaria squalida), and it is possible that the snail is selectively feeding on benthic components from the sediment. This selective food mode also seems to be used by the Shrimp L. vannamei because its δ13C signature is different from that of most other collected prey, and their main dietary contributors were sediment, macroalgae and zooplankton. Shrimp is zooplanktivorous until its juvenile stage, and seston represent their principal food source during their adult stage (Martinez-Cordova et al., 2003; Gamboa-Delgado et al., 2011).
The δ13C values of juveniles and adults of the fish M.cephalus were similar to those found in macroalgae, but their principal food source were sediment and macroalgae. Several studies have found that benthic algae contribute heavily to M. cephalus diets in estuary ecosystems from other regions (Cardona, 2000) and that the seston contribution may be higher when the sources change (Fischer et al., 1995c). In this study these changes may be associated to the seasonal variations of macroalgae blooms previously reported in other regions of Gulf of California (Piñón-Gimate et al., 2009). Among primary consumers, juveniles of C. arcuatus showed seasonal variations in their resources, with a decrement in the zooplankton contributions during spring and autumn. The juveniles of G. cinereus, H. leuciscus and L. argentivensis presented slightly seasonal variations in δ13C signature, despite their diverse food resources (bivalves, crustaceans and juveniles of M. cephalus). The variations in δ13C values in some primary consumers may be related to anthropogenic effluents in El Tóbari Lagoon, considering that wide isotopic signatures may be produced by shrimp aquaculture and sewage domestic effluents (Vizzini and Mazzola, 2006; Serrano-Grijalva et al., 2011).
The carnivores in the secondary consumer group (adults of C. arcuatus, G. cinereus, H. leuciscus and L. argentivensis) presented a δ13C signature and percentages of contributions consistent with two bivalves (C. gnidia and C. fluctifraga), crustaceans (H. erythrostoma, L. vannamei and C. arcuatus) and juveniles of fishes as their food sources. Carnivores are opportunistic depredators and may shift their prey selection according to abundance and competition for food resources with other species (Fischer et al., 1995a, 1995c). In this study, the observed small seasonal variations in fish δ13C values and in the percentages of contributions most likely occurred because similar food resources or/and their prey were available across seasons, which is frequently observed in subtropical and tropical ecosystems (Serrano-Grijalva et al., 2011). This pattern seems to occur for most tertiary consumer (birds) species because they have also exhibit slightly seasonal variations in their δ13C signatures and in the percentages of contributions. However, F. magnificens and P. brasilianus in all seasons, as well as R. longirostris during summer, showed low δ13C values. This finding may provide evidence that these species have preference for pelagic preys and/or that their diets are more variable and composed of marine and offshore prey. Hobson et al. (1993) found that birds with 13C enrichment were consistent with an inshore/benthic feeding preference pattern. Seabirds are migratory and may move seasonally or daily among continental and ocean ecosystems of differing carbon isotopic signatures. Further isotopic investigations in El Tóbari Lagoon are required to establish how patterns of δ13C abundance in seabird tissues are linked with a preference for pelagic, inshore or off shore feeding, as several factors may explain the δ13C gradient shown in seabirds and their prey (Hobson, 1993; Hobson et al., 1994).
Our δ15N results indicated that the foodweb of El Tóbari Lagoon is composed of 5 trophic levels, which is consistent with previous studies where different top predators were considered (Hobson and Welch, 1992; Carlier et al., 2007; Serrano-Grijalva et al., 2011). According to the δ15N values, and similarly to proposal with contributions data, phytoplankton organisms are part of the base of the foodweb of El Tóbari Lagoon and are the main food source for zooplankton (trophic level 2). The degree of dietary support of phytoplankton to foodwebs is variable in coastal environments and frequently is driven by abiotic parameters (Bouillon et al., 2000; Abrantes and Sheaves, 2008). Mangroves are considered carbon suppliers of many species in subtropical lagoon, which was not reflected in this study. Most primary consumers (filter feeders, snails and shrimp) occupied the intermediate trophic level. Some studies have been reported that filter feeders use the more refractory and degraded organic matter of sediment and SPOM (Carlier et al., 2007; Riera et al., 2007). These derived δ15N changes of deposited organic matter during diagenetic process are due to the selective removal of components or fractions with different isotopic ratios; microorganisms had a preferential use of 14N, producing an enrichment in organic substrates in 15N (Carlier et al., 2007), which generated a significant 15N-enrichment of organic matter in the sediment and then in the 15N-enrichment of deposit and filter feeders.
Omnivorous species in estuary and lagoon environments are often characterized by sharing common resources, and they are flexible, which allows them to exploit temporary abundances in prey populations. Thus, the ability of certain species to feed at more than one trophic level can also increase the intraspecific variability of δ15N values. In this study, H. erythrostoma, M. cephalus and L. vannamei presented selective feeding on sediments, while juveniles of C. arcuatus and G. cinereus exhibited carnivorous habits. L. vannamei is capable of selecting specific prey items from the mixture of components available in the water column and sediments (Gamboa-Delgado et al., 2011). In fact, studies of microscopic analyses of shrimp gut contents have revealed the presence of a large variety of food items, including detritus, Nematodes, Copepods, Amphipods, Polychaetes, Bivalves, and Diatoms (Martinez-Cordova et al., 2003; Gamboa-Delgado et al., 2011). An alternative explanation is that these benthic organisms are dependent on organic matter supplied from shrimp-pond effluents and/or agricultural discharges, which may produce significant increases of δ15N values by the nitrification of NH+4 (Serrano-Grijalva et al., 2011).
The fish of El Tóbari Lagoon occupied trophic level 4 and their δ15N presented a slightly intra-specific variability in signals (Figure 2). This finding can be explains by the dietary overlap of the fish species considered. This finding implies a considerable sharing of resources, which may be related to an abundance of prey items, available to juveniles and adults. The relative size of a species plays an important role in the structure of foodwebs. Fish predators are often larger than their prey and attain a higher trophic position as their size increases (Pasquaud et al., 2008). However, this relationship between size and δ15N was not obvious for predatory fish species in the study, and some juvenile species (L. argentivensis) occupied similar trophic levels as adults. Estuarine fish populations are generally dominated by specimens that are small in size (juveniles and young adults), which is related to a nursery function attributed to estuarine habitats for juvenile fish (Woodward and Hildrew, 2002). In this study, the epibenthic fish M. cephalus had a relatively high trophic position (trophic levels ranging between 3.4 and 4.1), which is surprising, given its reported dietary ecology (Fischer et al., 1995c). This finding is consistent with a fish predator of polychaete annelids and benthic isopods, as reported by Pasquaud et al. (2008). This may be also related to the presence of anthropogenic organic matter in the sediments, from shrimp aquaculture and/or agricultural origins, with high δ15N signatures. We previously reported high δ15N values for M. cephalus (16.19 ± 0.31‰) in Estero de Urias Lagoon, an impacted by shrimp aquaculture and sewage sludge coastal ecosystem from southeast coast of the Gulf of California (Jara-Marini et al., 2009; 2014). These data contrast with values (δ15N = 5.1 ± 0.5‰) reported for this fish in non-impacted by anthropogenic activities lagoon of Kenya (Nyunja et al., 2009). However, further and detailed studies are needed for elucidate this possibility.
The highest trophic level was occupied by birds as secondary carnivores. The trophic levels of birds were similar to those reported for birds of temperate and other coastal areas (Hobson, 1993; Hobson et al., 1994; Hobson, 2009). F. magnificens and P. brasilianus showed the highest δ15N compositions, which may be a result of their narrow prey selectivity, as was reported by Hobson (2009) for P. auritus. Other birds exhibits a varied range in δ15N signatures, which could be evidence of an inshore/offshore or benthic/pelagic gradient. Most bird’s food sources were the crustaceans L. vannamei and C. arcuatus and the juveniles of fishes M. cephalus, G. cinereus, H. leuciscus and L. argentivensis. The use of isotopic composition in bird ecology has limitations. Birds are usually migratory and exhibit a wide variety of possible prey (Hobson, 1993). Ainley et al. (2018) suggested that the availability of prey and the means and rate by which prey become available, rather than the type of prey, have a strong bearing on the structure of pelagic seabird faunas. The tissue analyzed may represent another limitation. Hobson et al. (1994) provided estimates for isotopic turnover in birds and proposed that isotopic analyses of muscle tissue represent an integration over 3-4 weeks. Dietary information derived from the isotopic analysis of liver integrates diet over a week, and that derived from bone collagen integrates diet over the life time of birds. The isotopic measurement of several tissues from the same individual will thus yield information on changes in diet through time. However, in this study, we were not able to collect bone and liver tissues. Despite the enormous potential of the isotope technique in seabird dietary studies, it is important to stress the need to establish isotopic inventories of prey of as many species as possible in appropriate study areas.
A clear trend toward increasing δ13C and δ 15N from base organisms to top-predators in the studied foodweb in the four seasons was detected. The isotopic and contribution data confirmed that the SPOM and phytoplankton are the main organic source supporting the foodweb in El Tobari Lagoon. Consumers were distributed into four trophic levels, with trophic level 2 occupied by zooplankton and filter-feeders and trophic level 4 occupied by carnivorous fish and most bird species. The different carnivorous fish presented dietary similarities implying a considerable sharing of resources, which could be related to the abundance of possible prey, between juvenile, adults and the species. Crustaceans and fish represented the main food sources of birds, although some birds exhibited more variable food sources (marine and offshore prey).
The authors thank G. Leyva-García and D. Aguilera-Márquez for their technical support in sample collection and preparation. We also thank Dr. Chris Eastoe and his work group of the Environmental Isotope Laboratory of the Department of Geosciences of the University of Arizona at Tucson, AZ, USA, for the use of their facilities for the stable isotope analyses.
This research was supported by a research grant from the Consejo Nacional de Ciencia y Tecnología de México (Project CONACYT CB-2008-C01-103522).