香京julia种子在线播放

Single baleen plates from 13 adult and 1 subadult PRWs were sampled for stable isotope analysis, including: 6 males; 6 females; and 2 individuals of unknown sex ( Table 1 ), all registered into the South Australian Museum collection and loaned to us for stable isotope analysis. Baleen plates were originally obtained from beach-washed or stranded PRWs along the South Australian and Tasmanian coasts ( Figure 2 ). If a complete set was collected, the longest plate was chosen for analysis, taken from ¾ back from the front of the mouth (Williamson, 1973; Pivorunas, 1979), and where possible, baleen plates included the unerupted portion within the gum, i.e., to include the most recent growth. Juveniles were excluded from the study to ensure younger whales, that may still be suckling and thus return elevated isotopic values, were not included (Borrell et al., 2016). Additionally, this was done to avoid ontogenetic changes in baleen growth rate found in juveniles (Schell et al., 1989a; Best & Schell, 1996; Kemper & Leppard, 1999; Eisenmann et al., 2016; Lysiak et al., 2018).

Baleen plates were first scrubbed with Milli-Q water using steel wool, then cleaned with a 2:1 chloroform to methanol solution twice to remove surface lipids and contaminants prior to sampling. Starting at the proximal end (most recent growth beneath the gum), approximately 0.5 mg of baleen powder was extracted every 1 cm along the longitudinal growth axis 0.5 cm from the labial edge using a Dremel engraving tool. Powdered samples were transferred to tin capsules and compressed airtight for processing and analysis. Samples were processed using the Flash 2,000 organic elemental analyser, interfaced with a Delta V Advantage Isotope Ratio Mass Spectrometer via a ConFlo IV interface (Bioanalytical Mass Spectrometry Facility, Mark Wainwright Analytical Centre, UNSW Sydney). Nitrogen and carbon reference standards USGS40 and USGS41A were used to correct potential instrumental drift prior to isotope abundance calculation. Ratios of C:N ranged from 3.4 ± 0.1 and were within the established range of pure keratin (3.4 ± 0.5) (Hobson & Schell, 1998; Trueman et al., 2019).

PRW baleen isotope values were compared to literature-derived bulk δ ¹⁵N and δ ¹³C isotope values of their potential prey species ( Table 2 ). Voucher prey samples were selected from Australian waters at temperate latitudes as well as euphausiid species from Antarctica (to rule out feeding within the Southern Ocean). Due to no available bulk stable isotope values of prey species within the two upwelling regions (eastern GAB and Bonney Upwelling), we selected the only and closest available prey samples to these two regions, including off southeast Australia, southeast Tasmania, and eastern Australia (northern NSW) ( Figure 2 ). For comparison with PRW stable isotope values, prey estimates were adjusted for trophic fractionation. These estimates were calculated as per Eisenmann et al. (2016):

Where, δX = stable isotope values of either δ ¹³C or δ ¹⁵N, B = baleen, p = prey, TF = trophic enrichment factor, SD = standard deviation, x = ¹³C or ¹⁵N. Here, we used the TF of krill-feeding fin whales (Balaenoptera physalus) from the Northern Hemisphere, known to be +2.77 ‰ for δ ¹⁵N and +2.26 ‰ for δ ¹³C (Borrell et al., 2012). This TF was also more accurate for Southern Hemisphere humpback whales (Eisenmann et al., 2016) compared to that of generic estimates (DeNiro & Epstein, 1981; Minagawa & Wada, 1984; Peterson & Fry, 1987), however blue whale TF aligned more closely with the established TF of bottlenose dolphins (Busquets-Vass et al., 2017). As we could not directly measure the amount of isotopic enrichment of PRWs, relative to known, local prey sources (due to the lack of data), we assumed they align with the majority of baleen whale species with estimated TFs. We also assume that the prey isotope values (mean and standard deviation) capture the natural variation within the prey species’ isotopic range that may vary over time and space (Rodríguez-Malagón et al., 2021). We used previously averaged values for Antarctic krill (E. superba) and Australian fish by (Eisenmann et al., 2016) to best account for temporal variability where available. Copepod δ ¹⁵N and δ ¹³C values were averaged as they did not significantly vary across different water types off southeast Australia (Henschke et al., 2015). Literature-derived prey samples from southeast Tasmania (N. australis) and Antarctica (E. superba) were lipid extracted/corrected, while non-lipid extracted/corrected prey samples included those from eastern Australia (T. gregaria, E. similis) and southeast Australia (copepods and Australian pelagic fish species) ( Table 2 ). Given that lipid-corrected prey samples of euphausiids show a slight increase in δ ¹⁵N values (Logan et al., 2008), it is possible N. australis and E. superba true range is slightly lower. We believe the impact of this on our conclusions is minimal, given E. superba is beyond the range of PRW δ ¹⁵N values and N. australis is enriched, relative to most PRW values, by a whole trophic level.

The rate of baleen growth for PRWs has not yet been estimated, and due to the uncertainty in their migration and feeding patterns we were unable to use predictable annual shifts in movements (and associated physiological changes triggered by fasting) to calculate their baleen growth rate. Therefore, we assume the cyclical nature of δ ¹⁵N values in PRW baleen reflects the fluctuations in seasonal prey availability and annual changes in food web dynamics (driven by predictable upwelling) within waters off southern Australia; a likely feeding ground.

Based on the assumption that stable isotope oscillations are formed due to seasonal food web changes, we estimated individual-specific baleen growth rates by measuring the distance between two adjacent δ ¹⁵N minimums. Depending on the length of the baleen plate and the number of clear, measurable oscillations, this resulted in between one and three growth rate measures per individual. When there were two or more oscillations measured within the baleen of one individual, the average of these became their individual-specific average. The species average was calculated by taking the overall mean of all baleen growth measures, resulting in 18.5 ± 4.8 cm/year (see Supplementary Table 1 ). Using species average baleen growth rates is a more robust calculation as it: (a) partially accounts for the high variability seen within and across individuals, likely representing changes in behaviour across years rather than an explicit change in tissue growth rate; and (b) not all individuals have clear, measurable oscillations.

Dates were assigned to baleen segments for analysis with oceanographic variables following the protocol outlined by Dedden and Rogers (2022). Briefly, we used the species-specific baleen growth rate, stranding date (i.e., closest date available to when the whale died), and sampling interval to calculate the number of days between each data point. A date was then assigned to each data point based on the cumulative days since death (i.e., last known baleen growth). PRW individuals M5589 and M18943 were excluded as they did not have an available stranding date, a necessity for attributing time to baleen segments. Like most studies involving bulk stable isotopes, time attribution assumes that stable isotope assimilation is constant in all individuals despite potential differences in physiology (e.g., reproductive state).

To account for the ~1 cm of fragile keratin at the proximal end (within the gum), which was too thin to be sampled, the date of each data point was shifted forward (towards the proximal end) by 1 cm (~19 days) so that the stranding date would match the point of most recent growth at the proximal edge.

Two regions were selected that reflect possible PRW feeding locations off mainland southern Australia. These were two of the seven hotspots previously identified by Kemper et al. (2013) and reflect: (a) where the majority of samples were sourced, and (b) two important upwelling areas that many other species of marine wildlife within the region rely on (Gill, 2002; Rogers et al., 2013). These included two upwelling regions, one within the eastern GAB (-37.167 S 132.94 E to -34.167 S 136.917 E) and the other further southeast known as the Bonney Upwelling (-38.376 S 137.521 E to -36.223 S 141.378 E). While both upwelling regions are within proximity of each other off southern Australia, combining regions would result in an area so large it would underestimate localised upwelling events due to averaging across surrounding regions. By selecting smaller areas of known productivity, we avoid underestimating these events.

Linear mixed-effects models were selected as the most appropriate for analysis as they allow for nested random effects and correlation structures. Using the ‘lme’ function from the ‘nlme’ package (Pinheiro et al., 2007) in R version 4.1.2 (R Core Team, 2021), we investigated whether δ ¹⁵N values from PRWs were associated with SST and chlorophyll-a concentrations at the eastern GAB and Bonney Upwelling. SST and chlorophyl-a concentrations at both locations were included as fixed effects. All δ ¹⁵N values for each individual were included as the response. Individual was incorporated as a nested random effect due to the replicate samples along the baleen for each whale. Autocorrelation in PRW data was accounted for via the inclusion of an order 1 autocorrelation structure ‘corAR1’ in the lme models, assuming the repeated measurements are equally spaced. Here, the addition of the ‘corAR1’ term was included after a comparison against the same model without this correlation structure, assessed via Akaike Information Criterion (AIC). Both chlorophyll-a concentrations and SST were analysed at a zero-lag period, however we also tested lag periods of 1–3 months to account for any delay between the upwelling event and prey abundance that PRWs rely on (resulting in 16 models – 8 for each fixed effect). As the different lag periods were directly correlated, all linear mixed-effect models were reported based on their statistical significance rather than compared. Both marginal and conditional R² values were reported to show the influence of fixed effects alone (R²m) and combined with the random effects (R²c). To access whether relationships were consistent between male and female PRWs, sex was added to significant models as a fixed effect. We tested the homogeneity of variance in δ ¹⁵N and δ ¹³C values across months via the Bartlett’s test to determine whether the isotopic range for each month was equal. Due to the lack of intra-annual variation observed in δ ¹³C values, we did not analyse the relationship between PRW δ ¹³C values and chlorophyll-a concentrations or SST. We did not adjust δ ¹³C values for changes due to the Suess effect (Young et al., 2013), instead the original δ ¹³C values were used. This is because our focus was to assess whether changes in intra-annual variability in stable isotope values were associated with the timing of intra-annual changes in oceanographic features. The shifts in δ ¹³C values due to the Suess effect will be slight within any year, which is not true across years, but here we do not examine long-term (inter-annual) temporal changes in δ ¹³C values. The original δ ¹³C values were also used to infer potential prey types. This is because the shift in δ ¹³C values, due to the Suess effect, over the time of our study would be relatively low compared to the difference in δ ¹³C values that identify whether the PRWs remain relatively sedentary and feed within temperate waters or whether they annually migrate south to feed on Antarctic krill. In addition, the difference in δ ¹³C values due to the Suess effect would be relatively small compared to the likely difference in the prey stable isotope values used here and the ‘true’ prey values. The stable isotope prey values available were not closely tied to the PRWs in this study. Our prey stable isotope values were not from prey collected from the region the whales were feeding, nor the year they stranded. With the data available we infer the broad region and prey types the PRW’s feed on rather than identify their foraging to particular prey species.

The data presented suggests that PRWs remain at mid latitudes year-round and lack north-south movements across distinct stable isotope gradients into Antarctic waters, supporting previous speculations (Kemper, 2002). Their carbon stable isotope record ranges between -20.32 and -13.15 ‰, which coincides with values well outside the Antarctic range (St John Glew et al., 2021). Modelled carbon values of particulate organic matter (POM) gradually decrease at approximately -40°S from -20 to -24 ‰ to around -28 to -31 ‰ towards high latitudes in the Southern Ocean (St John Glew et al., 2021). The absence of PRW carbon values below approximately -20 ‰ (i.e., values within the Antarctic range) suggest these whales are likely absent from the region entirely. There is a lack of baseline stable isotope data for mid-latitude waters between the Subtropical Convergence and coastal regions off southern Australia, making movement patterns of PRWs hard to distinguish. However, the gradual increase in carbon values from Antarctica to temperate waters and the comparable values off New Zealand (-22 to -19 ‰) (St John Glew et al., 2021), suggest the range of PRW carbon values may sit between the Subtropical Convergence and coastal waters off southern Australia, which support the limited sighting records of the whales within these regions (Kemper et al., 2013). Individual whales M23548 and M23863 show spikes in their carbon stable isotope values, as seen along their carbon oscillation patterns. However, as those few enriched carbon values are not associated with potential prey items it is difficult to ascertain to cause of such changes.

Compared to the clustered carbon stable isotope values (suggesting a primarily mid-latitudinal distribution), PRW nitrogen stable isotope values exhibit a large range, where the overlap between them and Australian krill and copepod species suggest a likely food source. Generally, higher nitrogen values (i.e., oscillation peaks) were associated with Australian krill species, where most of the overlap occurred with Australian krill species T. gregaria and E. similis. Copepods from the southeast Australian coast also appear to be included in PRW diet. It is possible the overlap between PRW stable isotope values and these prey species is not a true representation of their primary diet, and that other euphausiid species such as N. australis may be important as previously suggested (Kemper, 2002). This is because the samples for T. gregaria and E. similis were collected from eastern Australia (northern NSW), likely outside the PRW distribution (most northerly occurrence of PRWs is off southern NSW (Pavey, 1992)), and that the trophic enrichment factor of PRWs can not be accurately tested here. However, in the absence of prey stable isotope values from waters off southern Australia and the Subtropical Convergence, we utilised the best available prey data for similar mid-latitude regions and assumed PRW trophic enrichment factor based on other baleen whale species to infer their diet and habitat use. Furthermore, the similarity in available nitrogen and carbon stable isotope values of krill and copepod species between Australia, Tasmania, and New Zealand make it difficult to distinguish exact feeding patterns and movements between these areas, highlighting the need for additional prey samples. Australian pelagic fish species appear to be absent from the diet of Australian PRWs, as expected due to the small diameter of baleen fringes that indicate a diet of small euphausiids, copepods, and other small plankton, as identified through stomach content analysis (Budylenko et al., 1973; Sekiguchi et al., 1992; Kemper, 2002). Thus, the baleen structure and previous stomach content analysis support the isotopic overlap between PRWs and copepod and euphausiid species found in this study.

We found that nitrogen and carbon stable isotope oscillations were synchronous within individual whales, however, the interpretation of these results is challenging. This is because changes in stable isotope values, particularly for nitrogen, may be due to various factors including direct changes in diet or changes in the animal’s physiology (DeNiro & Epstein, 1981; Fry, 1988). As a result, we suggest the changes in stable isotope values between nitrogen and carbon may represent either: (a) active changes in feeding due to oceanographic changes and shifts in food web dynamics; (b) metabolic changes due to seasonal variation in prey abundance; or (c) inshore/offshore movements between waters off southern Australia and the Subtropical Convergence.

Firstly, synchronous cycles in nitrogen and carbon stable isotopes may reflect changes in active feeding due to seasonal fluctuations in food web dynamics. In marine environments, the abundance of nutrients in surrounding waters can trigger fluctuations in food web dynamics that affect the efficiency of energy transfer from the bottom to the top of the food web (van Ruth et al., 2018). For example, high-nutrient water during summer upwelling is dominated by diatoms, facilitating a short classical food web that provides enhanced productivity (Sommer et al., 2002). In contrast, nutrient-poor water in winter typically supports a longer, less efficient microbial food web that is generally unproductive and dominated by picophytoplankton (Azam et al., 1983; Stockner, 1988). Food web dynamics of the eastern GAB are likely underpinned by variations in nutrient availability and changes in the number of trophic links from the bottom to the top of the food web (van Ruth et al., 2018). In the present study, we found that higher nitrogen values of PRWs (i.e., the enriched portion of the synchronous cycles) overlap with the nitrogen values of temperate prey species, being Australian krill and copepods. We also found that these higher values displayed more variation and generally occurred during the summer months when both the eastern GAB and Bonney Upwelling regions are underpinned by a classical food web during the productive upwelling season. The high variability likely reflects moderate bursts of productivity during this time, with a microbial food web persisting in the background (van Ruth et al., 2018). In contrast, we found that lower nitrogen values of PRWs (i.e., the depleted portion of the synchronous cycles) did not overlap with any available prey data. These slightly lower, less-varied nitrogen values correspond to the presence of a microbial food web during winter. During these times, when larger zooplankton is patchy and/or not available, it may be possible PRWs are consuming other small plankton available to them, a potential food item suggested by Kemper (2009). While this positive relationship between PRW nitrogen values and SST (to indicate the food web present) is weak, it provides support for the overlap seen between the nitrogen values of PRWs and Australian krill and copepod species. Typically, periods of upwelling in summer are associated with cooler temperatures (van Ruth et al., 2010a; van Ruth et al., 2018), however, here, summer corresponds to slightly warmer SST. Generally, this region has a small SST difference between seasons, however, by taking the monthly SST average, to match with PRW isotope values, upwelling events may be underestimated.

Alternatively, it is possible the low (and less variable) nitrogen stable isotope values observed in PRW baleen, that fall below the range of Australian krill and copepod species, reflects periods of fasting during winter when resources are sparce, however, we believe this explanation is unlikely. As found in fin and bowhead whales, fasting may drive a decrease in nitrogen without subsequent enrichment triggered by starvation (Aguilar et al., 2014; Matthews & Ferguson, 2015). If PRWs were to experience metabolic changes due to fasting, it could be expected that changes in nitrogen would not match those in carbon. For example, when an animal switches from diet-derived food sources to stored body reserves, while nitrogen generally changes due to fractionation associated with fasting (Cherel et al., 2005a; Lübcker et al., 2020), carbon does not. This is due to the carbon stable isotope values of the prey’s geographical location and the whales’ stored energy remaining the same (Eisenmann et al., 2016). In this study, we observed synchronous cycles in PRW nitrogen and carbon stable isotope values, suggesting that metabolic changes associated with fasting is unlikely. Rather, a decrease in nitrogen for PRWs more likely reflects a change in seasonal prey (i.e., feeding on smaller plankton) or a shift in movement to another feeding area.

Lastly, it is possible that the synchronicity in nitrogen and carbon stable isotope oscillations reflect movements of PRWs between coastal waters off southern Australia (higher nitrogen and carbon values) and the Subtropical Convergence (lower nitrogen and carbon values). Whilst the stranding and sighting records of PRWs suggest they can occur in coastal waters year-round, sightings at the Subtropical Convergence (Kemper et al., 2013) suggest some may move between these two regions. Once summer upwelling around the eastern GAB and Bonney Coast weakens, and waters become nutrient poor, some PRWs may move south to exploit resources at the Subtropical Convergence during austral winter. Oceanic fronts at the Subtropical Convergence have been linked with the foraging behaviour for many species of marine mammals (Beauplet et al., 2004; Baylis et al., 2008) and seabirds (Charrassin & Bost, 2001; Spear et al., 2001), and while the Subtropical Front is relatively weak to the south of Australia, the region is characterised by high levels of productivity (James et al., 2002). Baylis et al. (2008) found a seasonal shift in foraging behaviour for New Zealand fur seals, where females moved between the Bonney Upwelling in autumn to regions associated with the Subtropical Front during winter in response to the decline in coastal productivity. They further suggested the strength and duration of upwelling and productivity in coastal areas, associated with the Bonney Upwelling during summer and autumn, may influence the time females can sustain themselves in this habitat before southward movements to alternate foraging grounds (Baylis et al., 2008). This has also been observed for other otariid seal species (Goldsworthy, 1995; McCafferty et al., 1998; Thompson et al., 2003; Beauplet et al., 2004; Goldsworthy, 2006), where females increased foraging-trip duration and/or distance towards the Subtropical Front during pup-rearing period, likely reflecting increased demands and/or changes in prey availability. While we found no influence of sex on the correlation between nitrogen stable isotopes and SST, it appears that females may primarily drive the weak positive relationship between nitrogen stable isotope values and SST, whereas males exhibit a more negative/variable trend. This may also contribute to the weak R² values associated with the correlation. The lack of statistical significance for the influence of sex may be due to the limited amount of data (9 individuals); thus, we are unable to draw conclusions. However, we speculate there is possible disparities in feeding strategies in mid-latitude regions, or differences in movement patterns between coastal waters off southern Australia and the Subtropical Convergence during winter.

Alternatively, since wind-driven upwelling brings cool, nutrient-rich waters in from other regions (e.g., the subantarctic (Lewis, 1981; Richardson et al., 2020)), the lower nitrogen and carbon values that do not overlap with potential prey items and represent the depleted section on baleen oscillations may reflect baseline stable isotope values further south. However, if this were the case, we would expect to observe lower stable isotope values during summer, due to the influx of higher-latitude waters (St John Glew et al., 2021), which was not the case. Further research is required to elucidate possible intersexual differences and understand potential drivers behind PRW movements between coastal waters off southern Australia and the Subtropical Convergence.

There was little statistical significance between PRW nitrogen stable isotope values and remote sensed chlorophyll-a concentrations. This may be due to the restrictions of remote sensed data, which rely on surface measurements of primary productivity for the eastern GAB region (van Ruth et al., 2010a; van Ruth et al., 2018). In the eastern GAB during summer upwelling periods, previous studies have found a relationship between the upwelled water mass and subsurface peaks in chlorophyll-a concentrations (Kämpf et al., 2004; van Ruth et al., 2010b; van Dongen-Vogels et al., 2011; van Dongen-Vogels et al., 2012; Kloser et al., 2017; Patten et al., 2018). Thus, remote sensed data that relies on surface measurements likely underestimate phytoplankton biomass by not accounting for productivity in sub-surface enriched waters (van Ruth et al., 2010a; van Ruth et al., 2018). Secondly, there was a shorter temporal coverage of chlorophyll a, which was only available from 2003 onwards, thus reducing our sample size for these models. Additional data, and a greater understanding of PRW movements will assist in interpreting the potential relationship between PRW nitrogen values and chlorophyll-a when lagged 2 months.

Based on the assumption that the distance between two consecutive nitrogen minimums represents an annual cycle, the average growth rate of PRW baleen was estimated to be 18.5 ± 4.8 cm/year. This was within the range of other species of baleen whales (see Supplementary Table 2 ). We also found PRWs exhibited a high degree of intraspecific variability in baleen growth rates (11 - 28 cm/year) alike other baleen whale species ( Supplementary Table 2 ). However, whilst there is clear variability in the length of each cycle (i.e., growth rate), it is unlikely baleen growth can almost triple in magnitude between years of adult PRWs, thus changes more likely reflects interannual variability in feeding behaviour that either elongates or reduces the yearly cycle. The assumption that baleen growth is linear may impact the attribution of time to baleen segments, as doing so relies on the baleen growing at a constant rate. However, using the species average baleen growth rate can assist to minimise the impact of such variability on temporal conclusions, although more research is necessary to assess the growth patterns of baleen.

Other factors besides diet and metabolism may impact the stable isotope ratios in PRW baleen. These include changes in reproduction (e.g., breeding/lactation, etc) as observed in other baleen whales (Clark et al., 2016; Trueman et al., 2019; Uchida et al., 2021), and changes in the isotopic baseline (McMahon et al., 2015). However, data limitations prevented us from analysing the impact of both these factors here. Additionally, as this study spans ~ 40 years and likely across periods of ecosystem-level changes, stable isotope ratios may be impacted by large-scale ecosystem shifts as observed by Mansouri et al. (2021). Given that PRWs rely on mid-latitude regions and Australian krill and copepod species, future work could investigate large-scale environmental shifts and how this impacts the diet and habitat use of PRWs, as well as their ecosystem sentinel potential for these regions.