香京julia种子在线播放

A small section of the tip of the wing in the direction of growth was sampled from each cephalopod beak and rinsed with distilled water. The tip of the wing was selected as it signifies the most recent growth phase of the beak and therefore the most recent somatic growth (Cherel and Hobson, 2005). For isotopic analysis, beak wing tip samples were examined whole or cut in half for larger specimens. For squid species from pilot whale stomach contents, beaks were selected based on the mode of the LRL length for each prey species.

In whole prey specimens, approximately 1 cm² pieces of anterior dorsal white muscle was dissected from behind the head for fish and mid-anterior ventral mantle for cephalopods, cleaned of any skin or mesentery, and then rinsed with distilled water. A similar sized piece of pilot whale skin was also subsampled from larger frozen samples of skin and blubber. Any remaining subcutaneous adipose tissue was removed with a scalpel and the subsequent sample rinsed in distilled water. All samples were then frozen in preparation for isotopic analysis. Just prior to isotopic analysis, samples were freeze dried and then ground to a powder with a Wig–L-Bug^®. Since lipids are depleted in ¹³C, all tissue samples (except beaks) were delipidated using cyclohexane, which effectively adjusts δ¹³C values without affecting δ ¹⁵N (Chouvelon et al., 2014). The ground whale skin, fish muscle or cephalopod mantle samples, along with 3 ml of cyclohexane, were left to stand overnight in small, covered test tubes under a fume hood. After delipidating overnight, samples were then centrifuged, and the supernatant removed. Samples were subsequently subjected to two more cyclohexane rinses. Between rinses the samples were allowed to stand for another 1 - 2 hours. Following delipidation, samples were left to dry uncovered under a fume hood overnight.

Relative abundance of carbon (13C/12C) and nitrogen (15N/14N) stable isotopes was determined by a Finnigan Delta Plus Advantage stable isotope-ratio mass spectrometer at the University of Victoria, British Columbia, Canada. For all isotopic analysis there was a 10 percent replication measurement for each isotopic sample run. The results of the isotopic analysis are presented in the usual δ notation relative to Vienna PDP belemnite for δ13C and atmospheric N₂ (AIR) for δ15N. Replicate measurements of internal laboratory standards (DORM) indicated measurement errors of ± 0.1 ‰ and ± 0.2 ‰ for δ13C and δ15N, respectively. In addition, mean C: N mass ratios for pilot whales and prey species were within desirable ranges ≤ 3.5 ( Tables 1 , 2 ).

Since skin samples were examined from differing body sites, a nested analysis of similarities (ANOSIM) with 9999 permutations was used to determine if there were differences between standardized body sites (dorsal, lateral, and ventral) on separate δ¹³C and δ ¹⁵N values of skin from four individuals. Body site was treated as a fixed factor, and individuals as a random factor. ANOSIM was used because it is distribution free, robust for small sample sizes, and relies on permutations to determine significance. A permutational multivariate analysis of variance (PERMANOVA) model was then used to determine differences in pilot whale skin δ¹³C and δ ¹⁵N values simultaneously. The non-parametric PERMANOVA model was based on 9999 permutations using a Euclidean distance matrix with sex and stranding as fixed factors. PERMDISP was used to determine variation within factors, and any significant differences in the factors were examined using pairwise comparisons. Results of the PERMANOVA were visualized using a Principal Coordinates Analysis (PCO). Spearman’s correlations were also used to assess relationships between total length and δ ¹³C and δ ¹⁵N values. These correlations were conducted across all strandings, separately for males and females across all strandings, and within each stranding for males and females separately.

We employed the Stable Isotope Bayesian Ellipses in R (SIBER) package version 2.1.9 (Jackson et al., 2011) as a proxy for examining trophic diversity of the pilot whales. This method facilitated the quantification and comparison of isotopic niche widths and overlaps within and across different strandings, defined by stranding location and sex.

To quantify isotopic niches, we calculated three metrics: the Total Area (TA), the Standard Ellipse Area corrected for small sample sizes (SEAc), and the Bayesian Standard Ellipse Area (SEA _B ). The TA represents the whole isotopic niche, encompassing 100% of the data, but is sensitive to outliers, making it less reliable for analyzing the core niche. In comparison, the SEAc measures the core isotopic niche area, approximately 40% of the data, providing a robust estimate of central tendency while also correcting for small sample sizes. In contrast, the SEA _B , which also measures the area of the ellipse, accounts for uncertainty in the data using a Bayesian framework. Accounting for uncertainty is particularly important for small or uneven sample sizes, such as the male samples for Maria Island and King Island. Using Monte Carlo resampling, the SEA _B generates posterior distributions of ellipse areas. This allows estimation of the mode with 95% credible intervals, which reflect uncertainty in niche area estimates.

Furthermore, we also conducted a Bayesian analysis to evaluate the overlap and relative sizes of isotopic niches between different stranding groups and between sexes within those groups. This Bayesian approach considers the entire distribution of potential overlaps and size comparisons. Using the ‘bayesianOverlap’ function, we generated posterior distributions for the overlap, employing 4000 draws and a 95% probability interval to estimate and quantify the shared dietary isotopic space. We then calculated the bin mode of these distributions to identify the most probable range in overlap values. We also used the Bayesian posterior distribution of SEA _B to assess the probability as to whether the isotopic niche of one group was larger than another. The probabilities were calculated as the proportion of posterior draws where the SEA _B of one group exceeded that of another. Ellipses were also visualized using SEA _B.

Dietary proportions of pilot whales were estimated by implementing the MixSIAR package version 3.1.12 (Stock and Semmens, 2016), which allows for the incorporation of prior information and uses Bayesian inference to model the contributions of various prey sources to the diet. This approach allowed us to quantify the proportional consumption of prey items between the stranding groups, utilizing δ ¹⁵N and δ¹³C isotopic values of the whales (consumers) and their prey (sources).

Pilot whales consume a diverse range of cephalopods and fish. To reduce complexity caused by overlapping isotopic values among multiple species, prey items were categorized into five different groups: oceanic cephalopods (OC), oceanic fish (OF), near-shore cephalopods (NSC), near-shore fish (NSF), and ommastrephids represented by Todarodes filippovae (TF). Nearshore species included both shelf and neritic species. PERMANOVA was used to assess differences between the assigned prey groups for which we had raw isotopic data. These groups were then aggregated a posteriori to determine the proportion of cephalopods versus fish in the diet and whether the whales were more likely to feed nearshore or offshore.

Oceanic and near-shore cephalopod isotopic values were obtained from beaks retained in the pilot whale stomach contents as well as the mantle of whole specimens from local fishermen. As a result, we had both mantle (whole specimens) and beak tissues (PW stomach contents) from cephalopods for dietary analysis. Since isotopic discrimination varies according to the type of tissue (Teixeira et al., 2022), we applied a correction factor of -0.7 ‰ and 3.8 ‰ to the isotopic value of the beaks (wings) for δ¹³C and δ ¹⁵N, respectively (Jackson, unpublished data), which is similar to the difference in isotopic values between beaks and muscle observed in T. filippovae (Cherel et al., 2009). Isotopic data for oceanic fish, represented by a weighted mean and standard deviation, were sourced from Davenport and Bax (2002), while near-shore fish isotopic values were derived from the muscle of specimens collected by local fishermen (see Table 1 ).

We first ran a MixSIAR model with uninformed priors (equal contributions from all five sources). However, to refine the estimates of the dietary composition of pilot whales, we ran two other models, which incorporated prior information based on known global dietary preferences of long-finned pilot whales (G. melas) as compiled in the literature (see Becker et al., 2021, online resource 1). This sensitivity analysis allowed us to evaluate whether our choice of TEFs influenced the dietary estimates for each stranding.

While the global dietary preferences for G. melas were obtained from Becker et al. (2021), the cephalopod and fish prey categories were slightly modified according to the known distribution of species in Tasmanian waters (i.e. whether they were oceanic cephalopods, oceanic fish, near-shore cephalopods, near-shore fish or ommastrephids). For each cephalopod or fish species, percent contribution from multiple studies were averaged, then normalized so that all contributions summed to 100%. These normalized values were then scaled to represent the five sources used in the model. The informative priors for the five sources used in the models are as follows: OC = 1.64, OF = 0.07, NSC = 2.04, NSF = 0.2 and OMM = 1.04.

The trophic enrichment factors (TEFs) between the whale’s diet and its tissue were derived from feeding experiments on bottlenose dolphins (Tursiops truncatus). These experiments provided insights into the diet-to-skin enrichment factors (Browning et al., 2014; Giménez et al., 2016). Given the close taxonomic relationship between bottlenose dolphins and long-finned pilot whales, we assumed that the TEFs would be similar for pilot whales. Therefore, we used TEFs from Giménez et al. (2016) for both the uninformed model (Model 1) and one of the informative models (Model 2) with five source groups to assess how cephalopods and fish prey contributed to the diet of pilot whales ( Table 3 ). To address initial issues with model convergence, we also ran a third model comparison, using bottlenose dolphin TEFs from Browning et al. (2014) (see Table 3 ). While the dietary compositions are only reported for Models 1 and 2, Model 3 was used to compare convergence diagnostics and goodness of fit to ensure our models were robust.

Each of the three separate MixSIAR models was run with three MCMC chains, executing three million iterations, with a burn-in phase of 1.5 million iterations and a thinning interval of 500. To account for uncertainty in the models, we used both residual and process errors. The Gelman-Rubin diagnostic was used to determine if models converged, while the Deviance Information Criteria (DIC) was used to compare the goodness of fit between models.

All analyses were run in R version 4.3.3, except for PERMANOVA and ANOSIM which were undertaken using Primer v7 software (Primer-E). Values for δ¹³C, δ ¹⁵N, C: N mass ratio and length are all reported as mean ± SD. SEA _B is reported using the mode, along with 95% credible intervals. However, for dietary proportions we report medians, along with 95% credible intervals, which are a robust measure of central tendency but may not sum to one due to independent calculations for each dietary source.

Maria Island samples had the largest Bayesian Standard Ellipse Area (SEA _B ) (0.3; CI 0.2 - 0.4) among the three strandings. Based on Bayesian overlap function, it showed minimal overlap with Marion Bay and King Island (bin mode range: 0-10%), indicating a distinct isotopic niche. There was an 80% and 68% probability that Maria Island had a larger SEA _B than that of Marion Bay and King Island, respectively. King Island (SEA _B 0.2; CI 0.2 - 0.3) had a 65% probability of a larger ellipse area than Marion Bay (SEA _B 0.2; CI 0.2 - 0.3) with an overlap range of 50-60%, suggesting that at least half of the isotopic space is shared between these strandings ( Table 4 ; Figures 3A, B ).

The Bayesian overlap function also revealed a 70-80% overlap in the isotopic niches of Marion Bay females and males, indicating considerable similarity in their foraging habits or prey preferences. However, males exhibited a smaller isotopic niche (Total Area, TA: 0.8) than females (TA: 1.2), although both showed similar core niche metrics (SEA _C and SEA _B both at 0.2) indicating consistent niche utilization. Bayesian probability analysis revealed an 80% probability that the isotopic niche of females was larger than males, suggesting females had slightly greater dietary variation ( Table 4 ; Figures 3C, D ).

In contrast to Marion Bay, males from King Island and Maria Island were more likely to have larger isotopic niches than females (80% and 97%, respectively). The overlap between males and females was much lower at King Island (0-10%) and Maria Island (0-20%) compared to Marion Bay, indicating more distinct niches for females at these locations. Despite this, core niche metrics were similar between males and females across all strandings (SEA _C and SEA _B all at 0.2), males from King Island (CI 0.1 - 0.5) and Maria Island (CI 0.1 - 0.8) exhibited greater variability than females at both locations (CI 0.1 - 0.3), further highlighting the greater niche diversity among males ( Table 4 ; Figures 3C, D ).

A PERMANOVA on four of the five aggregated prey groups (nearshore cephalopods, offshore cephalopods, ommastrephids and nearshore fish) for which we had raw isotopic data showed a difference based on prey group (F _{3, 156} = 49.58, p = 0.0001). Pairwise comparisons revealed that all prey groups differed from one another (all p< 0.001). The fifth prey source (oceanic fish) was obtained from the literature (Davenport and Bax, 2002) ( Table 1 ). A visual inspection confirmed that the consumer’s isotopic data (pilot whales) fell within the expected mixing space defined by the five sources and adjusted for the TEFs sourced from Giménez et al. (2016) ( Figure 4 ) and Browning et al. (2014).

In our three MixSIAR models (see Table 3 ), we analyzed predicted dietary proportions using stranding as a fixed factor, as PCO analysis revealed that stranding was the most important factor accounting for variation in isotopic values. The Gelman-Rubin diagnostic for all parameters indicated strong convergence for all three models, with all R ^ below 1.05. A comparison of DIC values between Models 1 and 2, which used the same TEFs (Giménez et al., 2016) exhibited almost identical DIC values, indicating that informative priors resulted in a minimally stronger fit. Model 3, using TEFs from Browning et al. (2014) and informed priors, also converged well and fit within isotopic space.

When examining dietary contributions at each stranding using the informed priors model (Model 2), ommastrephids dominated the predicted diet of both Marion Bay (75%, CI: 59-86%) and King Island (87%, CI: 75-100%). Marion Bay included a smaller percentage of offshore fish (24%, CI: 14-37%), while King Island showed a smaller contribution from offshore cephalopods (11%, CI: 0-21%) ( Figure 5 ). Aggregated dietary proportions revealed that cephalopods dominated the diet at King Island (96%, CI: 92-100) and Marion Bay (75%, CI: 63-86), while Marion Bay also sourced 25% (CI: 13-37%) of its diet from fish. Both strandings predominantly sourced their prey offshore (King Island: 100%, CI: 94-100%; Marion Bay: 100%, CI: 96-100%) ( Figure 6 ). Overall, both aggregated and individual source contributions predicted from the generalist priors (Model 1) are almost identical to that of Model 2 ( Supplementary Tables S1 , S2 ).

In contrast, based on the informed model, the Maria Island stranding showed greater dietary variability, with offshore cephalopods contributing 38% (CI: 22-56%), ommastrephids 40% (CI: 0-54%), and nearshore cephalopods 16% (CI: 0-55%) ( Figure 5A ). Aggregated dietary proportions revealed that cephalopods accounted for 85% (CI: 45-100) of the diet, while fish contributed 15% (CI: 0-55%). Compared to the other strandings, Maria Island whales sourced prey predominantly from offshore (78%, CI: 45-90%) but also used nearshore habitats (22%, CI: 10-55%) ( Figure 6 ), suggesting greater plasticity in foraging strategies, as reflected by its distinctive mixing space. While the uninformed model also produced similar contributions at an aggregated level, individual sources revealed some variability. The model had a lower contribution of nearshore cephalopods (9%) and ommastrephids (25%) but higher contribution of nearshore fish (15%). Despite these differences, the diet of pilot whales across all strandings was dominated by cephalopods and offshore resources in both models, consistent with individual stranding events ( Supplementary Tables S1 , S2 ).

Variations in δ¹³C values often define differences in foraging strategies and habitat use (Cherel and Hobson, 2007). In our study, δ¹³C values showed a small but significant gradient among the three stranding locations, reflecting spatial variation in foraging habitats of marine predators due to δ¹³C differences at the base of the food web. These δ¹³C differences are influenced by factors such as latitudinal, nearshore/offshore, or pelagic/benthic gradients (Cherel and Hobson, 2007). Nearshore or demersal habitats, which tend to be more enriched in δ¹³C, generally have higher nutrient concentrations and greater productivity than offshore or pelagic habitats (France, 1995; McMahon et al., 2013).

Baseline δ¹³C values of the marine food web can also result from interannual changes in sea surface temperature (SST) and nutrient availability (McMahon et al., 2013). While all three whale strandings occurred in similar months (October and November), the King Island and Maria Island strandings occurred in late 2004, coinciding with rising SSTs resulting from weak to moderate El Niño conditions. In contrast, the Marion Bay stranding occurred late in 2005, when sea surface temperatures had begun to decline after the El Niño event (NOAA, 2006). Elevated SSTs during an El Niño event can reduce primary productivity, leading to more depleted baseline δ¹³C values due to slower carbon fixation rates. This variability in baseline δ¹³C values ascends through the food chain, influencing isotopic composition throughout the food web, including prey and predators (Rodríguez-Malagón et al., 2021). However, the distinct δ¹³C and δ ¹⁵N values observed between Maria Island and King Island, which both stranded in late 2004, suggests that spatial differences in foraging habitats, rather than any temporal variability were likely the primary driver of isotopic variation.

Spatial variability is further supported by the niche analysis, which revealed little to no overlap between the Marion Bay and Maria Island strandings, as well as the results from our mixing model. These findings suggest that pilot whales from both Marion Bay and King Island strandings primarily foraged in offshore habitats, consistent with their lower δ¹³C values. In contrast, the more enriched δ¹³C values of pilot whales from Maria Island suggest foraging in both nearshore and offshore habitats. This interpretation aligns with the PERMANOVA results, which revealed that the isotopic values differed between all locations. The slightly enriched δ ¹⁵N values for Maria Island whales further support this dual habitat use, as enriched δ ¹⁵N baselines are typically associated with inshore and demersal habitats (McMahon et al., 2013).

The differences in isotopic values observed in Tasmanian pilot whales also mirrors findings from New Zealand strandings. Isotopic differences were attributed to variations in primary productivity and baseline isotopic values across habitats rather than dietary shifts (Hinton et al., 2022). These findings suggest that spatially distinct foraging behavior is due to local ecological conditions and plasticity in foraging behavior, demonstrating ecological resilience. Together, these patterns highlight stranding location as a key driver in defining isotope niches and habitat preferences.

Other studies have similarly documented spatial differences in the foraging ecology of long-finned pilot whales based on stomach content analysis. Variability in diet has been documented due to seasonal, annual, or geographical differences (Gannon et al., 1997; Santos et al., 2014). For example, stranded pilot whales in New Zealand waters displayed evidence of both neritic and oceanic diets across different stranding events, though it remains unclear whether these patterns are a result of altered feeding behavior prior to stranding (Beatson et al., 2007a, b; Beatson and O’Shea, 2009).

Our isotopic dietary analysis reflects a preference for foraging in oceanic waters, with pilot whales inhabiting deeper waters off the continental shelf and edge (Gales et al., 1992; Monteiro et al., 2015). However, they exhibit plasticity in their foraging behavior by extending their foraging to nearshore and neritic waters at times, as the dietary pattern of Maria Island pilot whales suggests. Other researchers have also observed that both sub species of long-finned pilot whales, G. melas, can exploit both oceanic and neritic waters (Beatson et al., 2007a, b; Beatson and O’Shea, 2009; Spitz et al., 2011; Mèndez-Fernandez et al., 2012; Goetz et al., 2015; Becker et al., 2021). Isotopic mixing models confirmed a coastal or demersal foraging habitat for pilot whales (G. m. melas) stranded in Northwest Iberia, reinforcing findings from prior stomach content analyses (Monteiro et al., 2015). In contrast, pilot whales (G. m. melas) stranded in Scotland demonstrated a stronger preference for oceanic habitats and prey (Monteiro et al., 2015). Previous research also noted that pilot whales (G. m. edwardii) in the southwestern Atlantic Ocean primarily feed in oceanic waters, though they may sporadically exploit neritic or coastal prey when moving toward shore (Becker et al., 2021). These studies collectively highlight how regional differences in habitat use and prey distribution are key drivers of isotopic variation and differences in the foraging niche of long-finned pilot whale populations.

Some sexually dimorphic cetaceans display sex-specific foraging strategies due to differing physiological needs or habitat use (Evans and Hindell, 2004; Mendes et al., 2007; Louis et al., 2021, 2022) but evidence for such differences in pilot whales is less pronounced. In our study, sex played a lesser role than stranding site in explaining isotopic variability. This aligns with other studies that found no broad-scale differences in cephalopod prey selection between sexes. For example, a stomach content analysis of pilot whales from Tasmania, which included some of the same individuals used in this study for isotopic values from Marion Bay and Maria Island, reported no differences in the cephalopod assemblage between sexes. However, larger females generally consumed smaller Ommastrephid squid (Beasley et al., 2019). Similarly, in terms of fatty acid profiles, no difference was observed between male and female pilot whales stranded off southeast Tasmania (Walters, 2005). Similarly, research on G. melas species from the Strait of Gibraltar, the Atlantic, and New Zealand waters reported no significant sex-based differences in δ¹³C and δ ¹⁵N values (de Stephanis et al., 2008; Monteiro et al., 2015; Hinton et al., 2022). By contrast, male pilot whales in Kerguelen waters exhibited slightly higher skin δ ¹⁵N values than females (Fontaine et al., 2015), suggesting that finer-scale differences may be detectable between males and females.

Despite the lesser role of sex in isotopic variability in this study, finer-scale differences in isotopic values were observed. Bayesian overlap analysis documented minimal overlap between males and females from the King Island and Maria Island strandings. In contrast, the Marion Bay stranding revealed females were likely to have a larger niche than males, with niches between males and females exhibiting a 70-80% overlap. While the Marion Bay stranding shows significant dietary overlap, we also observed a negative correlation between male δ¹³C and δ ¹⁵N values and body length. The negative correlation between δ¹³C and body length suggests that larger males may take deeper dives than females while foraging, as δ¹³C values are typically more depleted in offshore waters and at greater depth due to reduced primary productivity and differences in nutrient cycling (Vanderklift and Ponsard, 2003; Cherel and Hobson, 2007). Diving deeper and longer would be consistent with larger individuals having some physiological advantages, such as a greater capacity for oxygen storage for prolonged dives (Schreer et al., 2001; Soto et al., 2008). A relationship between body size and depth has also been noted in the false killer whale (Pseudorca crassidens), where larger individuals exhibit similar depth-foraging behaviors (Riccialdelli and Goodall, 2015). In contrast, the negative correlation between δ ¹⁵N values and body length may reflect differences in prey selection or foraging strategies, rather than depth alone, as δ ¹⁵N values of particulate organic matter generally increase with depth (Mintenbeck et al., 2007). We suggest that larger males may be targeting different prey sizes or prey types located deeper in the water column than females. This facilitates reduced intra-group competition by accessing different prey resources, particularly for such a highly social species where shared foraging is common. In contrast, the broader isotopic niche observed in male pilot whales in the King Island and Maria Island strandings might reflect the ability to exploit a more extensive foraging range or adapt to changes in prey availability, providing an ecological advantage in a dynamic marine ecosystem.

Given their highly social nature, pilot whales within the same stranding event are expected to exhibit similar foraging patterns due to their strong cohesive social bonds (de Stephanis et al., 2008; Augusto et al., 2017). Nevertheless, observed variability in isotopic niches of males and females suggests that mass strandings may include multiple social groups (de Stephanis et al., 2008; Oremus et al., 2013), complicating the assessment of their dietary habits. Pilot whales typically form stable, matrilineal groups, social units made up of closely related individuals through maternal lines. These groups support strong kin-based social bonds, shaping the pod social structure and promoting coordinated foraging strategies (Oremus et al., 2013). However, pilot whales also display flexibility in their social organization, sometimes forming large, temporary aggregations that may exploit different foraging zones and depths (de Stephanis et al., 2008; Augusto et al., 2017). This social flexibility may contribute to the variability observed in the isotopic niches between males and females, as dynamic social units and changing foraging niches may result in distinct foraging adaptations among individuals.

Determining whether fish or squid are the major dietary component in the diet of pilot whales is essential for assessing their foraging strategies. Cephalopods are an important dietary component of many cetaceans, as approximately 80% of toothed whale species frequently consume cephalopods (Clarke, 1996; Rodhouse, 2013). Our mixing model suggested that these pilot whales also primarily consumed cephalopods in the months prior to stranding, with fish supplementing their diet to a lesser extent. For example, the model suggests that 15-25% of dietary intake was fish in the Maria Island and Marion Bay strandings. Stomach content analysis from these same strandings in a prior study also corroborates our findings of a preference for cephalopod consumption (Beasley et al., 2019). At Maria Island, 63% of pilot whale stomachs examined contained prey, primarily cephalopods (39%) with a small number of individuals with fish (3%). At Marion Bay, however, 71% of stomachs contained prey, exclusively cephalopods, with no fish recorded.

These findings generally align with our isotopic results, and are further supported by evidence of a cephalopod-dominated diet observed in pilot whales from New Zealand (Beatson et al., 2007a, b; Beatson and O’Shea, 2009), and Chile (Mansilla et al., 2012). However, cephalopod hard parts are preferentially retained in whale stomachs, which may overestimate cephalopod consumption (Bowen and Iverson, 2013). Nevertheless, other dietary tracers, such as fatty acid analysis, also support a cephalopod-dominant diet. Pilot whales from Tasmania exhibited a myctophid signature in their fatty acid profile (Walters, 2005), which was attributed to secondary effects from cephalopods, such as ommastrephids that eat myctophids (Pethybridge et al., 2012, 2013). This myctophid signature was retained in squid and subsequently transferred to its predator. Isotopic studies on pilot whales from the Southwest Atlantic also suggest a cephalopod-dominant diet, although there was a small proportion of fish consumption (Becker et al., 2021). Moreover, in the Kerguelen waters, heavy metal and isotopic analyses suggest a mixed diet of fish and squid, although no quantitative distinction between them was determined (Fontaine et al., 2015).

Off Tasmania, pilot whales are reported to target a diverse cephalopod assemblage. For example, a total of 26 cephalopod species across four strandings has been identified, with a broad range of 4 to 16 taxa recovered per stranding event (Beasley et al., 2019). A diverse cephalopod diet for the long-finned pilot whale in the Northern Hemisphere (G. m. melas) has also been documented (Santos et al., 2014). In contrast, diets in other regions in the Southern Hemisphere such as New Zealand (Beatson et al., 2007a, b; Beatson and O’Shea, 2009) and Chile (Mansilla et al., 2012) are reported to be dominated by three or fewer cephalopod species. Notably, the pilot whales stranded at Maria Island also consumed just three prey species, primarily nearshore, underscoring variability in foraging strategies across different locations.

The primary cephalopod species found in the stomachs of stranded individuals in Tasmania, Ommastrephidae sp., L. lorigera, and A. lesueurii, inhabit a range of marine environments (Beasley et al., 2019). Ommastrephids, which undertake both vertical and horizontal migrations (Gilly et al., 2006; Ji and Guo, 2024) were the main prey of pilot whales in the Marion Bay and King Island strandings and made up a significant proportion of the diet consumed at Maria Island, as well as in other regions (Spitz et al., 2011; Santos et al., 2014; Monteiro et al., 2015). Two important ommastrephid species commonly found in Tasmanian waters are N. gouldi and T.filippovae. N. gouldi is predominantly a shelf to nearshore and even estuarine species (Nowara and Walker, 1998) that aggregates at the bottom during the day but disperses in the water column in the evening (Winstanley et al., 1983). Conversely, T. filippovae is an epipelagic species commonly found in slope waters (Dunning, 1993). L. lorigera and A lesueurii, commonly encountered over slopes, seamounts and submarine ridges, also stay at depth during the day but migrate to the surface at night (Beasley et al., 2019). The deep-diving capabilities of long-finned pilot whales, with records of up to 800 m, facilitate their access to mesopelagic prey that migrate vertically at night (Baird et al., 2002; Heide-Jørgensen et al., 2002). This diving ability enables them to forage on a wide array of prey species across different marine environments.

Despite the worldwide dietary preference of pilot whales for cephalopods, fish also comprise a proportion of their diet in certain regions. This may be driven by the lower energy density of cephalopods in comparison to higher-energy fish, which may lead to supplementation with fish when nutritional needs increase or when cephalopod resources are low (Overholtz and Waring, 1991; Lockyer, 2007; Spitz et al., 2011). This kind of dietary plasticity highlights their ability to modify their foraging behavior in response to the diversity and prevalence of prey, thus reinforcing their role as a cephalopod specialist that exhibits opportunistic foraging when subjected to various ecological pressures (Santos et al., 2014).

Our findings enhance the understanding of the foraging strategies of pilot whales in Tasmania; however, they are accompanied by notable limitations. A lack of species-specific trophic enrichment factors (TEFs) introduces some uncertainty into our isotopic mixing models (Kadye et al., 2020). Species, trophic level, dietary source and tissue type are all critical considerations when selecting suitable TEFs for modeling the study species (Stephens et al., 2023). In closely related species, TEF values are relatively consistent (Giménez et al., 2016; Stephens et al., 2023) As a result, we chose TEFs obtained from bottlenose dolphins (Tursiops truncatus), a closely related species within the Delphinidae family (LeDuc et al., 1999). Both pilot whales and bottlenose dolphins share similar metabolic and physiological traits due to their close taxonomic relationship, and in Tasmanian waters, they occupy comparable trophic levels and consume similar food sources, including squid and fish (Davenport and Bax, 2002). Moreover, the TEFs were experimentally derived from skin, the same tissue analyzed in this study, ensuring compatibility in isotopic assimilation rates (Giménez et al., 2016). Monteiro et al. (2015) further validated the use of TEFs from bottlenose dolphins, demonstrating a similar diet to that of stomach contents examined in pilot whales from the Northeast Atlantic Ocean. By selecting TEFs from a closely related species, we have minimized uncertainty in our estimates of dietary composition, although a species-specific TEF would further refine our analysis.

Similarly, the selection of priors can also influence predictions of dietary contributions, particularly for whales in locations where foraging strategies are more flexible. Overall, aggregated dietary compositions between the informed and generalist models were consistent, as were the individual sources for Marion Bay and King Island. However, Maria Island displayed greater variability at the individual source level, highlighting the importance of using informed priors to reduce uncertainty. More ecologically plausible dietary estimates are achieved when using informed priors that reflect global dietary knowledge of pilot whales.

Additional uncertainties may be due to the exclusion of infrequent but ecologically important prey items in the mixing model, which may introduce error into dietary composition estimates. This is particularly relevant for highly mobile species, such as pilot whales, whose wide-ranging foraging habits can result in a diet that varies significantly across locations and seasons. In addition, data collected from stranded individuals may not accurately reflect natural foraging behavior, as strandings may result in altered prey intake. However, isotopic values of skin integrate multiple feeding events over time (Teixeira et al., 2022) reducing biases associated with stranding events.

Our study highlights the foraging plasticity of pilot whales in Tasmania, whereby they can potentially modify their foraging strategy based on the distribution or availability of prey. By confirming a dietary preference for a wide range of cephalopod species, we highlight their crucial ecological role as more generalist feeders when necessary. This ability for pilot whales to forage in both oceanic or offshore waters and nearshore waters increases their resilience to environmental changes and potential fluctuations in prey availability. Moreover, the ability to reduce intra-group competition via niche differentiation likely provides a buffer to pilot whale populations against environmental stressors, thus ensuring stable group dynamics over time. However, environmental changes, such as increasing sea surface temperatures, have resulted in many species of cetaceans demonstrating a poleward shift in their distribution or habitat (van Weelden et al., 2021). Similarly, increased ocean temperatures may also have a deleterious effect on their prey, with cephalopods likely to experience reduced survival rates, shorter development times and increased metabolic stress (Borges et al., 2023). Understanding the foraging ecology of cetaceans, including their interactions with prey, is therefore crucial, as they play an important role as sentinels of environmental change in the marine environment.