Introduction
Cetaceans are known to play several ecological roles in the ecosystems they inhabit, including the control of prey populations, and vertical and horizontal transport of nutrients (Roman et al., Reference Robbins, Felicetti and Sponheimer2014). To date, eight baleen whale and 27 odontocete species have been recorded in south-eastern Brazilian waters (Santos et al., Reference Santos, Siciliano, de Vicente, Alvarenga, Souza and Maranho2010; Miranda et al., Reference Minagawa and Wada2019; Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020). Except for the Bryde's whale (Balaenoptera edeni), for which movement knowledge is limited in south-eastern Brazil (Siciliano et al., Reference Siciliano, Brito and Azevedo2004; Santos et al., Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019), the other species of baleen whales are seasonal visitors during their migration to their reproductive areas in tropical regions (e.g. Lodi et al., Reference Liu, Chou and Chen1996; Zerbini et al., Reference Zerbini, Andriolo, Heide-Jorgensen, Pizzorno, Maia, VanBlaricom, DeMaster, Simões-Lopes, Moreira and Bethlem2006). Odontocete species include resident species such as franciscana (Pontoporia blainvillei) and Guiana dolphins (Sotalia guianensis), which occur in the inner part of the continental shelf and estuaries along the region (Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020). Other species, such as the Atlantic spotted dolphin (Stenella frontalis) and bottlenose dolphin (Tursiops truncatus) occur in coastal waters mainly in the 20–50 m isobaths (Santos et al., Reference Santos, Figueiredo and Van Bressem2017; Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020; Paschoalini & Santos, Reference Ott and Danilewicz2020). Some other species such as the common dolphin (Delphinus delphis) and orcas (Orcinus orca) are assumed to be occasional visitors (Santos et al., Reference Santos, Siciliano, de Vicente, Alvarenga, Souza and Maranho2010, Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019). Although their patterns of habitat use are poorly known in this region, these odontocetes seem to occur more frequently in shallow waters during the austral spring and summer (Tavares et al., Reference Siciliano, de Santos, Vicente, Alvarenga, Zampirolli, Laílson-Brito, Azevedo and Pizzorno2010; Santos et al., Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019).
Sympatric species with similar ecological niches compete when resources are limiting (Roughgarden, Reference Roman, Estes, Morissette, Smith, Costa, McCarthy, Nation, Nicol, Pershing and Smetacek1976). For that reason, they may develop niche segregation mechanisms to facilitate their coexistence and, consequently, minimize competition (Pianka, Reference Peterson and Fry1974). Such mechanisms are extremely important for maintaining the diversity of an environment (Chesson, Reference Chesson2000), and include foraging at different depths, time of the day or the consumption of distinct prey species (e.g. Gross et al., Reference Gross, Kiszka, Van Canneyt, Richard and Ridoux2009). Assessing the foraging areas and trophic ecology of co-occurring species may help in the understanding of the niche differentiation mechanisms developed among them (e.g. Browning et al., Reference Browning, CockCroft and Worthy2014a; Young et al., Reference Vighi, Borrel, Crespo, Oliveira, Simões-Lopes, Flores, García and Aguilar2017). Additionally, information on species feeding behaviour and habitat use helps to define their ecological roles in the ecosystems they inhabit, which can be used for conservation measures (Liu et al., Reference Lesage, Morin, Rioux, Pomerleau, Ferguson and Pelletier2015). The coastal waters along south-eastern Brazil are strongly influenced by anthropogenic pressures, such as harbours, oil and fishing industries (Santos et al., Reference Santos, Siciliano, de Vicente, Alvarenga, Souza and Maranho2010; Figueiredo et al., Reference Figueiredo, Santos, Siciliano and Moura2017), with several potential negative consequences for cetacean species. Therefore, investigating the ecological traits of local populations may help explain how these threats are affecting these species. Only a few studies have evaluated the feeding habits and trophic relationships of cetacean species in the region (e.g. Gross et al., Reference Gross, Kiszka, Van Canneyt, Richard and Ridoux2009; Bisi et al., Reference Bisi, Dorneles, Lailson-Brito, Lepoint, Azevedo, Flach, Malm and Das2013; Liu et al., Reference Lesage, Morin, Rioux, Pomerleau, Ferguson and Pelletier2015), and this information remains limited.
The analysis of carbon and nitrogen stable isotopes (expressed as δ 13C and δ 15N, respectively) provides a good complementary method for a better understanding of the ecological traits among living organisms (e.g. Kelly, Reference Kelly2000; Newsome et al., Reference Newsome, Martínez del Rio, Bearhop and Philips2010). This is because the δ 13C and δ 15N of a consumer reflect those of its prey (Peterson & Fry, Reference Passadore, Domingo and Secchi1987), with a small increase at each trophic level due to metabolic processes that discriminate against the light isotopes. Consequently, there is a mean increase of ~+1‰ and +3‰ in δ 13C and δ 15N values, respectively, for each trophic level increase (Minagawa & Wada, Reference Michener, Kaufman, Michener and Lajtha1984; Peterson & Fry, Reference Passadore, Domingo and Secchi1987). Therefore, δ 15N values are applied to estimate the relative trophic position of individuals within food webs (e.g. Vander Zanden & Rasmussen, Reference Valenzuela, Rowntree, Sironi and Serger2001), while δ 13C values are good indicators of the carbon source at the base of the food webs (Kelly, Reference Kelly2000). However, different studies have demonstrated that diet-to-tissue discriminations may vary largely, depending on the species and the tissue analysed (Borrell et al., Reference Borrell, Abad-Oliva, Gómez-Campos, Giménez and Aguilar2012; Browning et al., Reference Browning, Dold, I-Fran and Worthy2014b; Giménez et al., Reference Giménez, Ramírez, Almunia, Forero and de Stephanis2016), the quality of dietary protein (Robbins et al., 2005) and consumers' trophic position (Hussey et al., Reference Hussey, MacNeil, McMeans, Olin, Dudley, Cliff, Wintner, Fennessy and Fisk2014).
Carbon stable isotope ratios vary considerably among primary producers (e.g. Michener & Kaufman, Reference Melo, Santos, Bassoi and Araújo2007), allowing differentiation between terrestrial and marine, coastal and oceanic, or benthic and pelagic ecosystems (Peterson & Fry, Reference Passadore, Domingo and Secchi1987; France, Reference France1995). Moreover, there is a latitudinal gradient in δ 13C values, in which higher and lower δ 13C values are related with tropical and polar regions, respectively (Rau et al., Reference Polischuck, Hobson and Ramsay1982). Similarly, δ 15N values of primary producers vary according to the nitrogen sources available (N2, NH4+ and NO3−) and their concentration in the environment (Montoya, Reference Miranda, Luna, Sousa, Fruet and Zanoni2007), resulting in spatial and temporal gradients in δ 15N at the base of the food webs (Graham et al., Reference Graham, Newsome, Koch, McMahon, West, Bowen, Dawson and Tu2010). Therefore, δ 13C and δ 15N values provide useful information to investigate the feeding habits (e.g. relative trophic position) and foraging areas used by marine species (e.g. Gross et al., Reference Gross, Kiszka, Van Canneyt, Richard and Ridoux2009; Graham et al., Reference Graham, Newsome, Koch, McMahon, West, Bowen, Dawson and Tu2010; Chouvelon et al., Reference Chouvelon, Spitz, Caurant, Mèndez-Fernandez, Chappuis, Laugier, Le Goff and Bustamante2012).
Furthermore, δ 13C and δ 15N values can be used to quantify a population's isotopic niche, which is considered an indicator of their ecological niche dimensions (Newsome et al., Reference Moura, Tavares, Secco and Siciliano2007). Advanced statistical models, such as the Stable Isotope Bayesian Ellipses in R (SIBER, Jackson et al., Reference Jackson, Inger, Parnell and Bearhop2011), have been developed to estimate the isotopic niche width of populations. Such information can be applied to evaluate the level of trophic diversity (e.g. generalist vs specialist feeders) and overlap in isotopic niches among species or populations (e.g. Vighi et al., Reference Vander Zanden and Rasmussen2014; Liu et al., Reference Lesage, Morin, Rioux, Pomerleau, Ferguson and Pelletier2015; Kanaji et al., Reference Kanaji, Yoshida and Okazaki2017).
Studies of habitat use and trophic relationships among distinct cetacean species shed light to understand the ecological role of these mammals in aquatic food webs in the South-western Atlantic Ocean (e.g. Melo et al., Reference Melo, Santos, Bassoi and Araújo2010; Bisi et al., Reference Bisi, Dorneles, Lailson-Brito, Lepoint, Azevedo, Flach, Malm and Das2013; Di Beneditto & Monteiro, Reference Di Beneditto and Monteiro2016). Therefore, the main objective of the present study is to investigate the habitat use and feeding ecology of 10 cetacean species encountered in south-eastern Brazil. Carbon and nitrogen stable isotope analyses were applied to evaluate their foraging areas, identify trophic relationships, and niche segregation mechanisms among different species. Cetacean species investigated here include migratory baleen whales and resident toothed whales that occur in estuarine, coastal and oceanic environments along the study area.
Materials and methods
Study site
Cetacean samples were obtained along the coastal region of south-eastern Brazil, between Paraná state and the south of Rio de Janeiro state (24–26°S) (Figure 1). The region is influenced mainly by three water masses: the Coastal Water (CW), the Tropical Water (TW) and the South Atlantic Central Water (SACW) (Castro et al., Reference Castro, Lorenzzetti, Silveira, Miranda, Rossi-Wongtshowski, Del and Madureira2006). The CW is a low salinity water formed by the mixing between continental water and seawater and is encountered manly near the estuaries. The TW is a warm (>20°C), high salinity (>36 psu) water positioned between 0 and 200 m. The nutrient-rich SACW is positioned between depths of 200 and 500 m but surfaces along the coast of south-eastern Brazil mainly in summer months (Castro et al., Reference Castro, Lorenzzetti, Silveira, Miranda, Rossi-Wongtshowski, Del and Madureira2006), significantly influencing local productivity (Matsuura, Reference Magozzi, Yool, Vander Zanden, Wunder and Trueman1996). The south-eastern Brazilian coast presents a great diversity of oceanographic features, such as estuaries, bays and coastal islands (Tessler et al., Reference Tavares, Moreno, Siciliano, Rodriguéz, de Santos, Laílson-Brito and Fabían2006), which could influence the local biological productivity and prey abundance. The Cananéia–Iguape estuarine complex, for example, located in the southern portion of the study site, drains large amounts of organic matter to the coast and functions as a nursery area for several species of fish and invertebrates (Schaeffer-Novelli et al., Reference Sá Alves, Andriolo, Zerbini, Pizzonorno and Clapham1990). Similarly, the presence of islands along the coast may also contribute to increased primary productivity, by influencing physical processes that affect local nutrient concentrations (Gilmartin & Relevante, Reference Gilmartin and Relevante1974).
Study area in south-eastern Brazil. The dots indicate the locations where cetacean skin samples were collected between 2011 and 2016. The different states along the coast are indicated as SP (São Paulo), PR (Paraná) and RJ (Rio de Janeiro).
Sampling
Skin samples of cetacean species were collected from biopsied, bycaught or stranded individuals along the south-eastern coast of Brazil between 2011 and 2016 (Table 1). Biopsies were collected during oceanographic cruises using a crossbow (125 lb) with floating darts adapted to extract small portions of epithelial tissue. The biopsies were collected from the dorsal region of the cetacean body and were no more than the size of a pencil eraser. Bycaught cetaceans were accidentally killed in gillnet fisheries and skin samples were collected with the individuals still fresh (condition score 2, freshly dead). Lastly, skin samples from stranded cetaceans were collected during beach surveys along the coast when carcass condition scores were either 2 or 3 (moderately decomposed) (Geraci & Lounsbury, Reference Geraci and Lounsbury2005). All samples were stored at −20°C without the use of any preservative until analysis.
Cetacean species analysed in the present study
The number of skin samples obtained from stranded (STR), incidentally caught (IC) or biopsied (BP) individuals and the total number of samples (N) for each species; Mean (± SD) of C/N ratios, measured and mathematically corrected carbon (δ 13C and δ 13Ccor, respectively), and nitrogen stable isotopes (δ 15N).
Stable isotope analyses and lipid correction
After drying at 60°C for 72 h, ~0.3–0.7 mg of skin samples were weighed into tin capsules and sent to the Stable Isotope Core Laboratory (Washington State, USA) for analyses of carbon and nitrogen stable isotope ratios. Analyses were performed using an Isotope Ratio Mass Spectrometer (IRMS, Delta PlusXP, Thermofinnigan) connected to an Elemental Analyser (EA, ECS 4010, Costech Analytical, Valencia, CA). The analytical precision for both δ 13C and for δ 15N was < 0.1‰. Isotopic measurements are reported in the delta (δ) per mil (‰) notation, in relation to the international reference standards for δ 13C (Vienna Pee Dee Belemnite) and δ 15N (atmospheric nitrogen, N2).
Lipids are known to be depleted in 13C (DeNiro & Epstein, Reference DeNiro and Epstein1977), which can lead to an underestimation of δ 13C values in the tissues of an organism (Tieszen et al., Reference Tessler, Goya, Yosikawa, Hurtado and Muehe1983). The lipid content of a sample is correlated to its C/N ratio, and high values correspond to high quantities of lipids (McConnaughey & McRoy, Reference Matsuura1979). In this study, we corrected δ 13C values of all samples with C/N ratios higher than 3.0, as recommended for cetacean skin (Lesage et al., Reference Leatherwood2010). To promote the lipid correction, we used the equations proposed by McConnaughey & McRoy (Reference Matsuura1979) with adjustments of the constants suggested by Lesage et al. (Reference Leatherwood2010):
$$L = 93/\left[{1 + {\left({0.246 \times \displaystyle{C \over N}-0.775} \right)}^{{-}1}} \right]$$
$$\delta ^{13}{\rm Ccor} = \delta ^{13}{\rm C} + 6.386 \times \left[{0.004 + 3.90/\left({1 + \displaystyle{{287} \over L}} \right)} \right]\;$$where L is the quantity of lipids in the sample, C/N is the ratio of carbon to nitrogen, δ 13C and δ 13Ccor are sample measured and mathematically corrected δ 13C, respectively.
Data analyses
As data did not meet the assumptions of normality and homoscedasticity (tested using the Shapiro–Wilk and Bartlett test, respectively), δ 13C and δ 15N values were compared using Welch's Analyses of variance (Welch's ANOVA), followed by the Games–Howell post-hoc test for pair-wise comparisons between species. Only species with four or more samples were included in the statistical analyses to ensure the reliability of the results. A P value < 0.05 was used to indicate statistical significance. Additionally, hierarchical cluster analysis using the complete linkage method and Euclidean distances was carried out to detect isotopic similarities among the cetacean species. Due to the influence of milk on the isotopic ratios of nursing calves (Fogel et al., Reference Fogel, Tuross and Owsley1989), the calf of southern right whale (Eubalaena australis) was not included in this analysis. All statistical analyses were performed using R version 3.6.1 (R Core Team, Reference Rau, Sweeney and Kaplan2020).
To better understand the trophic relationships among species, their isotopic niches were estimated using the Stable Isotope Bayesian Ellipses package in R (SIBER) (Jackson et al., Reference Jackson, Inger, Parnell and Bearhop2011). Only species with five or more samples were considered for these analyses as SIBER requires at least five samples for reliable isotopic niche determination. Each species' isotopic niche width was estimated by calculating the standard ellipse area (SEA) corrected for small sample sizes (SEAC) and the Bayesian SEA based on the posterior distribution (SEAB). Isotopic niche widths were compared using the SEAB, for which we calculated the probabilities of real differences among niche widths based on 100,000 posterior draws. The overlap among the isotopic niches was computed from SEAC. All these estimations considered a prediction interval (p-interval) of 40% (Jackson et al., Reference Jackson, Inger, Parnell and Bearhop2011).
Results
Skin samples were obtained from 10 different cetacean species (seven toothed whales and three baleen whales) between 2011 and 2016 (Table 1). The mean δ 13C and δ 15N values for each species are displayed in Table 1 and Figure 2. The humpback whales (Megaptera novaeangliae) included one adult and three juveniles, while the E. australis included a mother and her calf. Juveniles M. novaeangliae had considerably higher δ 13C and δ 15N values in comparison to the adult. The calf E. australis had higher δ 15N and lower δ 13C values than the adult female. δ 13C and δ 15N varied significantly among the six species from which we had enough data to allow for statistical comparisons (Welch's ANOVA, P < 0.05, Table 2). The pair-wise comparisons were significant for all species regarding the δ 13C values, except between T. truncatus and both S. guianensis and P. blainvillei (Table 2). Nitrogen stable isotopes were significantly higher in T. truncatus than in the remaining species (Figure 2). In contrast, M. novaeangliae had significantly lower δ 15N than the other cetacean species. Lastly, δ 15N were significantly higher in D. delphis and S. frontalis than in S. guianensis but did not show significant differences between any of these species and P. blainvillei (Table 2).
Mean (±SD) stable carbon (δ 13C) and nitrogen (δ 15N) isotopes measured in (A) all cetacean species sampled in south-eastern Brazil between 2011 and 2016; and (B) focusing on Orcinus orca, Steno bredanensis, Sotalia guianensis and Pontoporia blainvillei. A – adult; C – calf; J – juvenile.
Results of the Games–Howell post hoc test for the pair-wise comparisons of mean carbon (δ 13C; lower-left) and nitrogen (δ 15N; upper-right) stable isotopes measured in skin samples of cetacean species from south-eastern Brazil, between 2011 and 2016
Significant differences (P < 0.05) are highlighted in bold. Only species with four or more samples available were considered: Tursiops truncatus (Tt); Delphinus delphis (Dd); Stenella frontalis (Sf); Pontoporia blainvillei (Pb); Sotalia guianensis (Sg); Megaptera novaeangliae (Mn).
Our hierarchical cluster analysis identified two main groups based on mean δ 13C and δ 15N values (node 1, Figure 3). One group included only baleen whales (M. novaeangliae and E. australis), with the lowest δ 13C and δ 15N values. Yet, these two species showed considerable dissimilarity in their isotopic values, as evidenced by the large Euclidean distance between them (node 4, Figure 3). The second main group included those species with higher δ 13C and δ 15N and was comprised mostly by toothed whales (aside from B. edeni and the juveniles M. novaeangliae). Within this group, T. truncatus (node 2) and the juveniles M. novaeangliae (node 3) were split from the other species, evidencing their large isotopic differences pertaining to their higher and lower δ 15N values, respectively. Further, our cluster analysis identified large isotopic dissimilarity between D. delphis and B. edeni (node 6, Figure 3), and between these species and the remaining toothed whales (node 5). The highest isotopic similarities were found between O. orca and S. bredanensis, and between P. blainvillei and S. guianensis (node 8).
Hierarchical cluster analysis based on mean carbon (δ 13C) and nitrogen (δ 15N) stable isotopes of the cetacean species sampled on the south-eastern coast of Brazil between 2011 and 2016. Tursiops truncatus (Tt); Pontoporia blainvillei (Pb); Sotalia guianensis (Sg); Stenella frontalis (Sf); Orcinus orca (Oo); Steno bredanensis (Sb); Delphinus delphis (Dd); Balaenoptera edeni (Be); Megaptera novaengliae (Mn); Eubalaena australis (Ea). A – adult; J – juvenile. The numbers indicate where the branches are separated based on their isotopic values.
Tursiops truncatus, followed by S. frontalis, had the largest isotopic niche areas estimated by the SIBER model, while the smallest niche width was observed in S. guianensis (Table 3, Figure 4). Intermediate niche widths were observed in D. delphis and in P. blainvillei. Furthermore, our analyses showed small isotopic niche overlap only between P. blainvillei and S. guianensis (Figure 4).
Standard ellipse area corrected for sample size (SEAC) estimated for species with total number of samples (N) > 5, and δ 13C – δ 15N biplots of individual cetaceans from which N < 5 in (A) all cetacean species sampled in south-eastern Brazil between 2011 and 2016; and (B) zoomed ellipses to show increased detail of (A). A – adult; C – calf; J – juvenile.
Standard ellipse area corrected for small sample sizes (SEAC; ‰2) and Bayesian standard ellipse area (SEAB, ‰2) with the 95% confidence interval (CI) calculated for the cetacean species sampled in south-eastern Brazil between 2011 and 2016
Only species with five samples or more were considered.
Discussion
In the present study we analysed carbon and nitrogen stable isotopes in 10 cetacean species sampled along the coastal waters of south-eastern Brazil. We discuss our results from isotopic data in relation to the patterns of habitat use and feeding habits, as well as the level of overlap in the use of resources among the co-occurring species.
Resident species
We considered T. truncatus, S. frontalis, S. guianensis and P. blainvillei as resident species, based on available information resulting from sightings, strandings and accidental catches along the south-eastern coast of Brazil (Santos et al., Reference Santos, Siciliano, de Vicente, Alvarenga, Souza and Maranho2010, Reference Santos, Figueiredo and Van Bressem2017, Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019), as well as recent results on local cetacean niche modelling (Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020). The large Euclidean distance between T. truncatus and the remaining odontocetes is related to its significantly higher δ 15N values in comparison to the remaining cetacean species. This was evident in the isotopic niche space, which showed niche segregation between T. truncatus and the other odontocete species, with complete lack of overlap in δ 15N values (Figure 4). The isotopic values estimated for T. truncatus indicate the consumption of coastal/benthic prey and relatively higher trophic position than the other species analysed. This is in agreement with previous studies on the same species in south-eastern Brazil, that also suggested the use of inshore areas along the 16 and 45 m isobaths (Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020; Paschoalini & Santos, Reference Ott and Danilewicz2020) and high trophic level (Bisi et al., Reference Bisi, Dorneles, Lailson-Brito, Lepoint, Azevedo, Flach, Malm and Das2013). Additionally, stomach content analyses of T. truncatus from south-eastern Brazil showed a great importance of demersal piscivorous fish, such as cutlass fish (Trichiurus lepturus), banned grunt (Conodon nobilis) and Atlantic midshipman (Porichthys porossimus) (Di Beneditto et al., Reference Di Beneditto, Ramos, Siciliano, Santos, Bastos and Fagundes-Netto2001; Melo et al., Reference Melo, Santos, Bassoi and Araújo2010; Moura et al., Reference Moreno, Zerbini, Danilewicz, de Santos, Simões-Lopes, Lailson-Brito and Azevedo2016), which could justify the high values estimated for both stable isotopes. Demersal prey have relatively higher δ 13C values than pelagic prey (e.g. Corbisier et al., Reference Corbisier, Soares, Petti, Muto, Silva, McClelland and Valiela2006), which are related to the greater 13C-enrichment that usually occurs in the coastal/benthic isotopic baselines (France, Reference France1995). In addition, feeding upon mesopredators (i.e. piscivorous fish) (Figueiredo et al., Reference Figueiredo, Santos, Yamaguti, Bernardes and Rossi Wongtschowski2002) would result in higher trophic positions, hence higher δ 15N values.
The isotopic niche area estimated for T. truncatus was considerably larger than those of the other cetacean species (Figure 4 and Table 3). Large niche area may indicate the use of a wide diversity of prey items, foraging on isotopically distinct prey and/or habitats (e.g. benthic vs pelagic), and feeding at different trophic positions. The large isotopic niche area observed in the present study is in agreement with the large spatial niche estimated for T. truncatus in a recent study using distributional data coupled with environmental variables in south-east Brazil (Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020). Additionally, T. truncatus have been described as generalist feeders (e.g. Leatherwood, Reference Leatherwood1975; Shane et al., Reference Secchi, Botta, Wiegand, Lopez, Fruet, Genoves and Di Tullio1986) showing large variation in the isotopic values of populations in other regions (e.g. Borrell et al., Reference Borrell, Aguilar, Tomero, Sequeira, Fernandez and Alis2006, 2021; Browning et al., Reference Browning, Dold, I-Fran and Worthy2014b). The ability to consume a large diversity of prey is especially beneficial for populations occurring in areas highly impacted by anthropogenic activities that may affect the availability of resources. For example, temporal changes in the most important prey species have been reported for a coastal population of T. truncatus from southern Brazil based on data from stomach contents coupled with stable isotope analyses (Secchi et al., Reference Schaeffer-Novelli, de Mesquita and Cintrón-Molero2016). This temporal change in their feeding habits was attributed to the overfishing of their main prey species in the 1970s (whitemouth croaker Micropogonias furnieri), resulting in increased consumption of other prey species in recent years. This was also observed for T. truncatus populations in other regions, where the great proportion of pelagic in comparison to demersal prey in their diet may be related to human-related eutrophication that is affecting the local populations of demersal species (Borrell et al., Reference Borrell, Vighi, Genov, Giovos and Gonzalo2021). This illustrates the importance of long-term studies assessing cetacean feeding habits to better understand the effects of anthropogenic actions on cetacean populations.
The second largest isotopic niche area was estimated for S. frontalis (Figure 4), which suggests foraging on isotopically variable prey and areas. This is in line with the large geographic niche observed for the species in the same region (Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020). Additionally, S. frontalis had relatively lower δ 13C values (Figure 2) in comparison to the other isotopically similar species identified by our cluster analysis (Figure 3), but its mean δ 15N was similar to those of S. bredanensis and O. orca (Table 2). The species has been recorded throughout the year in the shallow coastal waters (20–43 m) of south-eastern Brazil (Santos et al., Reference Santos, Figueiredo and Van Bressem2017; Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020) and dietary studies analysing stomach contents have identified great importance of pelagic prey, such as the tropical arrow squid (Dorytheuthis plei), the knobby (Argonauta nodosa) and the rough-scad (Trachurus lathami) (Melo et al., Reference Melo, Santos, Bassoi and Araújo2010; Lopes et al., Reference Lodi, Siciliano and Bellini2012a). The dominance of pelagic prey could explain their δ 13C values, since organisms occurring in the pelagic environments usually present lower δ 13C values than demersal/benthic organisms (France, Reference France1995).
Despite the statistical differences in δ 13C values, S. guianensis and P. blainvillei were placed together in our cluster analysis (Figure 3) and slightly overlapped in their isotopic niche areas (Figure 4). Similar δ 13C and δ 15N values have also been observed based on muscle tissues of these species in the central coast of Rio de Janeiro in south-eastern Brazil (Baptista et al., Reference Baptista, Kehrig, Di Beneditto, Hause-Davis, Almeida, Rezende, Siciliano, Moura and Moreira2016). This similarity in the isotopic ratios agrees with their local distribution, as both species are strongly associated with shallow waters and estuarine regions, where they feed mainly on small pelagic and demersal neritic prey (Di Beneditto et al., Reference Di Beneditto, Santos and Vidal2009; Campos et al., Reference Campos, Lopes, da Silva and Santos2020). However, S. guianensis and P. blainvillei from south-eastern Brazil differ in terms of prey preferences, as demonstrated in studies based on stomach contents. While the main prey of S. guianensis include the banded croaker (Paralonchurus brasiliensis) and the cutlass fish (Lopes et al., Reference Lopes, de Santos, da Silva, Bassoi and dos Santos2012b), the diet of P. blainvillei is dominated by the American coastal pellona (Pellona haroweri) and the bigtooth corvina (Isopisthus parvipinnis) (Campos et al., Reference Campos, Lopes, da Silva and Santos2020). Furthermore, the observed gradient in δ 13C values may be explained by small-scale differences in areas used (Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020), as S. guianensis were mainly sampled near the Cananéia estuary, while P. blainvillei samples were more uniformly distributed along the coast. Thus, the relatively small overlap in isotopic niche areas and the differences in δ 13C, even though both species are associated with coastal/estuarine waters, may suggest ecological and spatial segregation between these species. Nevertheless, our study counted only a small sample size for both species, thus our results should be interpreted with caution. Therefore, continued research is recommended for better understanding of the ecology and habitat use by these co-occurring odontocetes.
Occasional species
We considered occasional species those for which data are limited in the region, although they seem to occur more frequently and more abundantly at certain seasons of the year. This includes the O. orca, S. bredanensis, D. delphis and B. edeni (Tavares et al., Reference Siciliano, de Santos, Vicente, Alvarenga, Zampirolli, Laílson-Brito, Azevedo and Pizzorno2010; Santos et al., Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019). Orcinus orca and S. bredanensis were clustered together and had high isotopic similarity with S. guianensis and P. blainvillei (Figure 3). Their δ 15N values were higher than those of S. guianensis and P. blainvillei (Figure 2B), indicating relatively higher trophic levels. In south-eastern Brazil, S. bredanensis are mainly sighted in shallow coastal waters (Santos et al., Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019; Figueiredo et al., Reference Figueiredo, do Amaral and Santos2020) where cephalopods are the dominant prey item (Melo et al., Reference Melo, Santos, Bassoi and Araújo2010). The high δ 15N values observed here for S. bredanensis are similar to those observed for this species in other regions of south-eastern Brazil (Baptista et al., Reference Baptista, Kehrig, Di Beneditto, Hause-Davis, Almeida, Rezende, Siciliano, Moura and Moreira2016). The authors attributed S. bredanensis' relatively higher δ 15N to the use of larger prey due to their higher energy requirements in comparison to the smaller odontocetes. The high δ 15N observed in individuals of O. orca are in agreement with the species' high trophic position. In the South-western Atlantic Ocean, they have been reported feeding on swordfish Xiphias gladius, tunas Thunnus spp. (Dalla Rosa & Secchi, Reference Dalla Rosa and Secchi2007; Passadore et al., Reference Paschoalini and Santos2015), blue shark Prionace glauca (Passadore et al., Reference Paschoalini and Santos2015) and on small odontocetes, including P. blainvillei (Ott & Danilewicz, Reference Newsome, Clementz and Koch1998; Santos & Netto, Reference Santos and Netto2005). Thus, our isotopic data suggested that S. bredanensis and O. orca occupy relatively higher trophic positions than P. blainvillei and S. guianensis, although the former two species most likely differ considerably in terms of the main prey consumed.
The lower δ 15N values observed in O. orca in comparison to other cetacean species (e.g. T. truncatus, S. frontalis and D. delphis, Table 1) were unexpected, considering the prey types that could comprise their diet in the South-western Atlantic Ocean. Sightings of O. orca in the south-eastern Brazilian coast are more common between November and February (Siciliano et al., Reference Siciliano, Brito and Azevedo1999; Santos et al., Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019), the same period the samples in our study were obtained (December). As there is a seasonality of occurrence along the region, the stable isotopes measured in O. orca skin may be representing the isotopic values of prey consumed in different feeding areas. The time it takes for the isotopic values of prey to be reflected on toothed whales' skin tissue has an average half-life of ~24 and 47 days for δ 13C and δ 15N, respectively (Giménez et al., Reference Giménez, Ramírez, Almunia, Forero and de Stephanis2016). Accordingly, the nitrogen isotopic ratio observed in O. orca may be reflecting the isoscapes of other regions where they had been foraging during the previous months. If they had been foraging in offshore waters before approaching coastal regions, the lower δ 15N values from oceanic waters (Troina et al., Reference Tieszen, Boutton, Tesdahl and Slade2020a) could still be reflected in their skin isotopic values. Additionally, the diet-to-consumer discrimination could differ among these toothed whales, as diet-to-tissue (TDF) can be affected by consumers trophic level, the quality of dietary protein and prey isotopic values (Adams & Sterner, Reference Adams and Sterner2000; Robbins et al., 2005; Hussey et al., Reference Hussey, MacNeil, McMeans, Olin, Dudley, Cliff, Wintner, Fennessy and Fisk2014). Accordingly, if O. orca has lower TDF than the remaining toothed whales, the resulting δ 15N values observed in their skin could be relatively lower, even if they feed on higher trophic level prey. Further studies using complementary tools such as the analysis of δ 15N in individual compounds could help resolve their trophic position.
Our cluster analysis also identified a subgroup formed by D. delphis and B. edeni. It is important to note that, although they were grouped together, the isotopic differences between these two species were considerably large (node 6, Figure 3). This is most likely due to the large difference in δ 15N (~1.2‰, Table 1) between these two species, indicating relatively higher trophic position for D. delphis than for B. edeni. Individuals of both species seem to occur seasonally in the coastal waters of south-eastern Brazil. Although sightings of these species happen throughout the year in the shallow waters of the study area, frequency is higher during late spring, summer and early autumn (Siciliano et al., Reference Siciliano, Brito and Azevedo2004; Tavares et al., Reference Siciliano, de Santos, Vicente, Alvarenga, Zampirolli, Laílson-Brito, Azevedo and Pizzorno2010; Santos et al., Reference Santos, Laílson-Brito, Flach, Oshima, Figueiredo, Carvalho, Ventura, Molina and Azevedo2019). Such seasonality has been linked to the influence of the SACW during these months (Castro et al., Reference Castro, Lorenzzetti, Silveira, Miranda, Rossi-Wongtshowski, Del and Madureira2006), which results in increased productivity in coastal waters off south-eastern Brazil and, consequently, increased prey availability (Brandini et al., Reference Brandini, Tura and Santos2018). These physical events significantly affect baseline carbon isotopes, decreasing the δ 13C values in the organisms comprising the base of this food web (Troina et al., Reference Tieszen, Boutton, Tesdahl and Slade2020a). Among the organisms whose abundances are favoured by upwelling conditions of the SACW are small pelagic fishes and squids, such as the Brazilian sardine (Sardinella brasiliensis), the Argentine anchovy (Engraulis anchoita) (Bakun & Parrish, Reference Bakun and Parrish1991; Matsuura, Reference Magozzi, Yool, Vander Zanden, Wunder and Trueman1996) and the arrow squid (Dorytheuthis sanpaulensis) (Haimovici & Perez, Reference Haimovici and Perez1991). Balaenoptera edeni uses the south-eastern Brazilian coast mainly during warm months, apparently to forage upon large schools of Brazilian sardines (Siciliano et al., Reference Siciliano, Brito and Azevedo2004). Similarly, a diet rich on pelagic organisms has been suggested for D. delphis along its distribution in the Brazilian coast (Tavares et al., Reference Siciliano, de Santos, Vicente, Alvarenga, Zampirolli, Laílson-Brito, Azevedo and Pizzorno2010), although their diet may include relatively higher trophic level prey such as the longfine inshore squid (Melo et al., Reference Melo, Santos, Bassoi and Araújo2010). Thus, the lower δ 13C of D. delphis and B. edeni in comparison with the remaining species (Figure 2) could be reflecting foraging along the shelf breakwaters, where increased productivity due to upwelling events may provide important food sources for these species.
Despite the geographic overlap in occurrence between S. frontalis and D. delphis (Moreno et al., Reference Montoya, Michener and Lajtha2005; Tavares et al., Reference Siciliano, de Santos, Vicente, Alvarenga, Zampirolli, Laílson-Brito, Azevedo and Pizzorno2010), we found no evidence for overlap in the use of resources, as indicated by the lack of isotopic similarity between these two species. Similar isotopic values have been observed in S. frontalis and D. delphis from southern Brazil (31–34°S) nearshore (Botta et al., Reference Botta, Hohn, Macko and Secchi2012) and oceanic populations (Troina et al., Reference Troina, Dehairs, Botta, Di Tullio, Elskens and Secchi2020b). Nevertheless, the authors have concluded that despite the relatively high isotopic niche overlap between the oceanic populations of D. delphis and S. frontalis, the fact that they avoid one another (Di Tullio et al., Reference Di Tullio, Gandra, Zerbini and Secchi2016) is a strategy to minimize competition (Troina et al., Reference Troina, Dehairs, Botta, Di Tullio, Elskens and Secchi2020b). In south-eastern Brazil, data from stomach contents showed overlap in some prey species taken by these toothed whales, although they differed in terms of size-classes and the dominance of species (Melo et al., Reference Melo, Santos, Bassoi and Araújo2010). Therefore, segregation in the use of resources seems to be plausible, although further studies with larger sample sizes are recommended, as they would help improve our understanding about the trophic ecology and overlap between these enigmatic species.
Migratory species
The cluster analyses identified the largest isotopic differences between the adults M. novaeangliae and E. australis and the remaining cetaceans (node 1, Figure 3), with both baleen whale species having the lowest δ 13C and δ 15N values (Figure 2). These species occur along the Brazilian coastal waters during the austral winter and early spring, when they use the region for mating and parental care (Lodi et al., Reference Liu, Chou and Chen1996; Zerbini et al., Reference Zerbini, Andriolo, Heide-Jorgensen, Pizzorno, Maia, VanBlaricom, DeMaster, Simões-Lopes, Moreira and Bethlem2006). Their low δ 15N values were expected as these two species occupy relatively lower trophic positions: their main food item is krill, but copepods and small schooling fish are also important for E. australis (e.g. Hoffmeyer et al., Reference Hoffmeyer, Lindner, Carribero, Fulco, Menéndez, Fernandéz, Melissa, Diodato, Berasategui, Biancalana and Berrier2010) and M. novaeangliae (Clapham, Reference Clapham, Wursig, Thewissen and Kovacs2018), respectively. The low δ 13C most likely represents the isotopic baseline of Subantarctic and Antarctic waters (Brault et al., Reference Brault, Koch, McMahon, Broach, Rosenfield, Sauthoff, Loeb, Arrigo and Smith2018), their summer feeding grounds (Zerbini et al., Reference Zerbini, Andriolo, Heide-Jorgensen, Pizzorno, Maia, VanBlaricom, DeMaster, Simões-Lopes, Moreira and Bethlem2006). Considering estimated isotopic turnover of ~6 months in baleen whale skin (Busquets-Vass et al., Reference Busquets-Vass, Newsome, Calambokidis, Serra-Valente, Jacobsen, Aguiñiga-García and Gendron2017), this would coincide with the period these whales were feeding in polar waters prior to sampling off south-eastern Brazil. The large isotopic dissimilarity between these baleen whales (node 4, Figure 3) could be attributed to feeding in different areas in Subantarctic and Antarctic waters and the gradient in these regions' isoscapes (McMahon et al., Reference McConnaughey and McRoy2013). While the population of M. novaeangliae that breed off Brazil feed in South Georgia and Sandwich Islands (Zerbini et al., Reference Zerbini, Andriolo, Heide-Jorgensen, Pizzorno, Maia, VanBlaricom, DeMaster, Simões-Lopes, Moreira and Bethlem2006), a variety of feeding grounds in the South Atlantic Ocean have been suggested for E. australis, including the Georgia and Sandwich Islands, the Antarctic Peninsula and areas close to the Antarctic Convergence (IWC, 2001; Valenzuela et al., Reference Troina, Botta, Dehairs, Di Tullio, Elskens and Secchi2018).
The two E. australis sampled included an adult female and her calf. The calf's δ 13C and δ 15N values were, respectively, lower and higher than those of the mother (Table 1, Figure 2). This is probably due to the influence of the mother's milk on the isotopic values of the calf. As females produce milk by metabolizing their own tissues, the δ 15N in their nursing calves are relatively higher than in the mother (Fogel et al., Reference Fogel, Tuross and Owsley1989). On the other hand, the high fat content in aquatic mammals' milk (Costa, Reference Costa, Perrin, Wursig and Thewissen2002) result in lower δ 13C values in calves, due to the low 13C/12C in lipids (DeNiro & Epstein, Reference DeNiro and Epstein1977). Furthermore, lactation can also influence δ 13C and δ 15N in the mothers' tissues. An increase in δ 13C may be observed in the tissues of lactating females due to the production of lipid-rich 13C-depleted maternal milk (e.g. Borrell et al., Reference Borrell, Gómez-Campos and Aguilar2016; Gellippi et al., Reference Gellippi, Popp, Gauger and Caraveo-Patiño2020). Additionally, the protein synthesis during lactation reduces the δ 15N values in the maternal body protein (Kurle, Reference Kurle2002). Such patterns have been observed in northern elephant seals Mirounga angustirostris (Habran et al., Reference Habran, Debier, Crocker, Houser, Lepoint, Bouquegneau and Das2010) and in polar bears Ursus maritimus (Polischuck et al., Reference Pianka2001). Bearing this in mind, we could not discard the possibility that both δ 13C and δ 15N values in the adult female E. australis were influenced by metabolism during lactation, as we assume this female was breastfeeding due to the presence of a calf.
In the present study we analysed one adult and three juveniles M. novaeangliae. The adult was a biopsied female that had the lowest δ 13C and δ 15N values (Table 1), while the juveniles' isotopic values were notably higher, resulting in their clustering closer to the other toothed whales and B. edeni (Figure 3). The high δ 13C and δ 15N values observed in these juveniles may be indicative of foraging in an area with a different isotopic baseline. Juveniles of M. novaeangliae have already been reported opportunistically feeding during their migration to temperate and tropical zones (Danilewicz et al., Reference Danilewicz, Tavares, Moreno, Ott and Trigo2009; Sá Alves et al., Reference Roughgarden2009; Bortolotto et al., Reference Bortolotto, Kolesnikovas, Freire and Simões-Lopes2016). Baseline isoscapes are 13C-enriched in temperate and tropical zones in comparison with Subantarctic and Antarctic regions (McMahon et al., Reference McConnaughey and McRoy2013; Magozzi et al., Reference Lopes, Silva, Bassoi, Santos and de Santos2017; Troina et al., Reference Tieszen, Boutton, Tesdahl and Slade2020a). Additionally, the δ 13C and δ 15N values observed in juvenile M. novaeangliae were comparable to those of baseline organisms from the South-western Atlantic Ocean (McMahon et al., Reference McConnaughey and McRoy2013; Troina et al., Reference Tieszen, Boutton, Tesdahl and Slade2020a). Accordingly, our isotopic data suggest that the juveniles may be feeding during their displacement along the waters of the South-western Atlantic Ocean, which would consequently be influencing their skin isotopic values, averaging out the isotopic values from polar baselines. Such a pattern was also observed for M. novaeangliae specimens opportunistically feeding along the Australian coast, which had δ 13C and δ 15N values significantly higher than those of the same population that feed in Antarctica waters (Eisenmann et al., Reference Eisenmann, Fy, Holyoake, Coughran, Nicol and Nash2016). Additionally, higher δ 15N could indicate the consumption of other prey types such as small schooling fish that occupy a relatively higher trophic position than krill, which has been reported for the species in Brazilian waters (Sá Alves et al., Reference Roughgarden2009). Alternatively, nutritional stress could also have influenced the δ 15N values of these juveniles, as such conditions have been shown to increase δ 15N values due to the catabolism of tissues to balance energetic requirements (Hobson et al., Reference Hobson, Alisauskas and Clark1993; Doi et al., Reference Doi, Akamatsu and González2017).
Conclusion
In the present study we used δ 13C and δ 15N values in skin samples to assess the feeding ecology and habitat use by distinct cetacean species occurring along the south-eastern coast of Brazil. It is important to note that our ecological inferences were limited given our small sample size, especially for the baleen whales, and for O. orca and S. bredanensis. Nevertheless, stable isotopes allowed us to identify patterns of habitat use, feeding habits and trophic interactions that are useful to understand how these cetacean species explore and coexist in this area. Although information on the forage areas and feeding strategies of cetaceans is extremely important for management purposes, little is known about this subject for populations occurring in the South-west Atlantic Ocean. Therefore, continued investigation with increased sample size is recommended, perhaps coupled with other methodologies such as stomach content analysis or telemetry. Additionally, future studies in south-eastern Brazil should focus on sampling potential prey for isotopic measurements, as they would be extremely helpful to assess cetacean–prey relationships throughout the region.
Acknowledgements
The Coordination for the Improvement of Higher Education Personnel (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, CAPES) provided a post-doctoral fellowship to G.C.T. (Process number 88887.314453/2019-00 – PROANTAR). We also thank the Stable Isotope Core Laboratory of the Washington State University for the isotopic analyses.
Financial support
We thank Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP) for the financial support (Process number 10/51323-6 and 11/51543-9), which was essential for sampling cetaceans (beach and fishing fleet surveys for dead cetaceans, added to biopsy sampling in oceanographic cruises) and for the isotopic analyses.
