North Atlantic minke whale (Balaenoptera acutorostrata) feeding habits and migrations evaluated by stable isotope analysis of baleen

Abstract Isotopic analyses of the incrementally growing baleen in Mysticeti have been used to learn about their feeding and movement patterns. Using methods previously applied to Pacific minke whales, stable δ15N and δ13C isotope values were measured along the baleen plates of male and female minke whales from two locations in the Northeast Atlantic. The sample sizes used in this study are comparable to those previously used in the literature, and, although limited in size, the evidence suggests differences in isotopic signatures between whales caught at different locations. Both the δ15N and δ13C data suggest whales at the higher latitude site of Svalbard have a narrower diet than the whales from Lofoten/Vesterålen in Norway. Across all whales, the δ15N data indicate the whales primarily prey on fish for much of the year, only switching to zooplankton during the spring bloom. The δ13C data fail to confirm whether the whales migrate over long distances.


| INTRODUC TI ON
Minke whales are important constituents of the Northeast Atlantic ecosystem as predators and as a species harvested commercially in Norway and Iceland. Much emphasis has been placed on abundance estimation (Skaug et al., 2004) and studies of population genetic structure (Quintela et al., 2014). In terms of their trophic interactions and food-web roles, Nordøy and Blix (1992), Haug et al. (1996) and Windsland et al. (2007) have described minke whale stomach contents in summer, but stomach contents can only provide snapshots of what the whales eat in a specific area. As their behaviour is elusive and their size is small, consequently there is much that remains unknown regarding their behaviour, seasonal feeding habits, migrations, and winter distributions.
Large migrations and feeding activity across extensive areas are characteristics of a variety of mysticete whales. Given the ethical considerations of obtaining sufficiently large sample sizes to study whale feeding and movement habits, either non-destructive methods are employed, such as telemetry (Lee et al., 2017;Milmann et al., 2020) or biopsy darts that collect skin and blubber (Witteveen et al., 2009); or smaller samples are taken when opportunities allow, such as through strandings and museum samples (e.g. Best & Schell, 1996), licensed hunts (e.g. Hobson & Shell, 1998), or active research programmes (e.g. Mitani et al., 2006). Among the materials sampled, whale baleen is a keratinised structure that grows continuously. Application of stable isotopic analysis (SIA) to baleen plates can provide long records of whale feeding and movement habits. This method has been applied in a variety of whales, such as southern right whales (Eubalaena australis) (Best & Schell, 1996); bowhead whales (Balaena mysticetus) (Hobson & Shell, 1998;Schell et al., 1989); Pacific minke whales (Balaenoptera acutorostrata) (Mitani et al., 2006); and Antarctic minke whales (Balaenoptera bonaerensis) (Uchida et al., 2021). Findings from these studies have provided insights into migratory habits (Best & Schell, 1996;Schell et al., 1989) and seasonal feeding patterns and prey preferences (Mitani et al., 2006;Uchida et al., 2021).
The stable isotope analysis of Pacific minke whales provided insights into feeding shifts from krill to fish, but were unable to resolve latitudinal migrations (Mitani et al., 2006). In the case of the North Atlantic minke whale, while descriptions of their stomach contents in summer have been recorded (Haug et al., 1996;Nordøy & Blix, 1992;Windsland et al., 2007), a greater understanding of their year-round feeding habits is still needed. With regards to their movements, outside of the summer season little is known. It is assumed that they spend the summer at high latitudes and migrate to temperate or even tropical waters to breed and give birth in winter (Jonsgård, 1951;Risch et al., 2014;Vikingsson & Heide-Jørgensen, 2015), but others have questioned whether there may be an alternative to all individuals undertaking seasonal migrations to lower latitudes (Corkeron & Connor, 1999;Jonsgård, 1966;Kavanagh et al., 2018).
This study applies stable isotope analysis to understand feeding and movement patterns of minke whales caught along the coast of Northern Norway and Svalbard. The same methods used for Pacific (Mitani et al., 2006) and Antarctic (Uchida et al., 2021) minke whales are applied in the present study. Patterns in δ 15 N and δ 13 C are presented with an interpretation of what these reflect about the whale feeding habits. Also size measurements and estimated growth rates for the baleen plates are presented. In light of advancements to existing technology, the availability of new technologies and the increasing application of inter-disciplinary research, this study also takes the opportunity to collate existing literature on North Atlantic minke whale biology, behaviour and sightings to support an alternative hypothesis to the north-south migration assumed of these animals ( Figure 1).

| Baleen plate collection and sample preparation
Fresh baleen plates were collected, and the longest plate from each whale was stored at −20°C until analysis.
All samples were pre-treated using standard procedures for animal keratin samples (O'Connell & Hedges, 1999). Coarse materials were removed with a scalpel. The plates then received 30-min ultrasonic bath treatments in distilled water, a solution of 2:1 methanol:chloroform repeated twice, a solution of 1:2 methanol:chloroform repeated twice, and distilled water repeated three times. Samples were air dried overnight and then freeze dried for several hours.
The length of the baleen plates, excluding the bristles, ranged from 14.2 cm to 21.3 cm depending on the size of the whale (Table 1). Two females were just under 600 cm in length. If these (immature) females are excluded, the average length of baleen plates across whales greater than 700 cm in length is 18.2 cm. Sampling of baleen plates was done at 1-cm increments along the labial edge of the plate for the full length of the plate. Samples were standardized to weigh between 2 and 3 mg. Each sample was wrapped in a cleaned tin capsule (6 × 4 mm pressed tin capsules, standard weight, Elemental Microanalysis Ltd, UK) and stored in individual Eppendorf tubes labelled with the sample's identity.

| Fish and zooplankton sample preparation
Whole cod, herring, and krill obtained from the Barents Sea in 2014 were homogenized, and 500ml of their tissue homogenate was stored at −20°C in separate Nalgene bottles. For sample analyses, ~1 cm 3 of the sample was transferred into four 10 ml tubes, frozen at −80°C for 24 h, and freeze dried. After freeze drying, two sample replicates were weighed out at 1 mg, and two sample replicates were weighed out at 3 mg and wrapped in cleaned tin capsules (6 × 4 mm, pressed tin capsules, standard weight, Elemental Microanalysis Ltd, UK). The procedure of sample preparation was repeated a second F I G U R E 1 A minke whale Balaenoptera acutorostrata surfacing north of Svalbard (photo: Arnoldus Schytte Blix) TA B L E 1 Lofoten/Vesterålen (L/V) captured male (LV2,5,8,9,12) and female (LV1, 7, 10, 15) and Svalbard (S) captured female (S1, 4, 7, 9, 10, 11) minke whales, indicating whale ID; sex and whether pregnant; length of whale (cm); length of baleen plate (cm) (i.e. maximum length exclusive of free bristles; Jonsgård, 1951); length of baleen plate and bristles; distance measured between the time to obtain another complete set of four replicates for isotopic analysis.

| Isotopic analysis
Isotopic analysis was performed at the Godwin Laboratory for Paleoclimate Research at the University of Cambridge, UK, using an automated carbon and nitrogen analyser coupled in continuous flow to an isotope-ratio-monitoring mass spectrometer (cf-irm-ms) (Costech 4010 elemental analyser connected to a Thermo Delta V mass spectrometer). Results are reported using the delta scale in units of 'per mil' (‰) relative to internationally accepted standards, VPDB for carbon and AIR for nitrogen.

| Data treatment
Baleen growth was approximated as per methods of Mitani et al. (2006) for Pacific minke whales; the distance between isotope peaks was assumed to represent annual growth rate (Table 1) and 21 September to 20 December (Autumn).
Statistical analysis was done with R 3.6.1 (R Core Team, 2019) and figures were produced using the package 'lattice' (Sarkar, 2008).
Linear mixed-effect (LME) models were chosen as the most appropriate analysis given the limits in sample size. The 'lme' function from the 'nlme' package (Pinheiro et al., 2019) was used to examine the association between either of the δ 13 C or δ 15 N isotope measurements along the length of baleen and season, sex, and site. Models were set up with season, sex and site as fixed effects, and whale ID as a random effect. All isotope values of either δ 13 C or δ 15 N, measured along the baleen plates, were used for LME models. The association between sex and isotope signatures could not be verified from the Svalbard location given the lack of samples from male whales at this site. The 'corSpher' correlation term was included in LMEs to account for autocorrelation in data points. This 'corStruct' class was chosen by testing whether addition of the correlation term improved the model fit, assessed via Aikake Information Criterion, against the basic model without a correlation structure.
Isotope values (δ 13 C and δ 15 N), from other North Atlantic minke whales and from their fish and zooplankton prey, were obtained

F I G U R E 2
Map showing the areas (red line) within which the minke whales were harvested off Lofoten/Vesterålen and Svalbard F I G U R E 3 Examples of the variation in δ 15 N and δ 13 C isotope values along the length of baleen plates from one Svalbard female (S11). Blue dots are δ 15 N isotopes, and red triangles δ 13 C isotopes. Refer to the Appendix S3-S5 for isotope plots of the remaining individuals from the literature (Table S1 and associated references). For the δ 13 C and δ 15 N isotope values reported in each study as well as those from the present study, the mean, standard deviation, and sample size were obtained and a composite mean and standard deviation was calculated as per Burton (2016). In studies where δ 13 C and δ 15 N isotope values were presented in a figure, the isotope values were obtained from the figure using ImageJ 1.53e (Rasband, 1997(Rasband, -2018. Data supporting this publication are available at Eerkes-Medrano et al. (2021).

| RE SULTS
The average growth rate of baleen across all whales (>700 cm) is 11.0 ± 1.7 cm/year; broken down by sex, it is 12.8 ± 0.9 cm/year and 10.1 ± 1.2 cm/year for males and females, respectively. When Carbon isotopes also suggest differences between whales from the two locations, with an association between δ 13 C values and the site (β = −0.86, CI = −1.64 to −0.08, p<0.05). The spread of δ 13 C values in scatterplots also seem to differ by season (Figures 4 and S1) but this association between δ 13 C and season was not picked up in LME models. With regards to patterns between whale groups, L/V females had the largest variation in δ 13 C values and Svalbard females had a spread of δ 13 C that was consistently narrow.
With regards to the isotope values obtained from the literature for minke whales and their prey, the composite mean values for fish and minke whales cluster within a range of 12 and 13 ‰ δ 15 N while krill are found at 7.7 ‰ ( Figure 5).

| D ISCUSS I ON
The average rate of baleen growth from males in the present study was 12.8 ± 0.9 cm per year, in line with the growth rate of 12.9 ± 2.1 cm reported for male Pacific minke whales (Mitani et al., 2006) and the growth rate of ~11.9 cm reported for a single female North Atlantic minke whale (Riekenberg et al., 2020). The benefit in applying stable isotopic analysis to samples taken along the length of baleen plates is that a longer temporal record is obtained, as opposed to direct observations of feeding, samples from gut contents of dead animals, or faeces, which provide point samples of feeding activity.
With Pacific minke whales (Mitani et al., 2006) (Skaug et al., 1997). In the latter study, 18 individuals taken from the coast of Lofoten/ Vesterålen during July/August 1992 had 0-group fish, herring and pelagic fish forming the largest-by-biomass diet items. The δ 15 N values overlap sufficiently amongst these types of fish ( Figure 5) that the present study only distinguishes between periods when the North Atlantic minke whales eat fish or zooplankton, rather than attempting species identification of prey in the baleen records.
In line with observations for the Pacific minke whale (Mitani et al., 2006) and the single female North Atlantic minke whale (Riekenberg et al., 2020), the baleen plates for North Atlantic minke whales in the present study contain distinct δ 15 N depletions. For Lofoten/Vesterålen individuals, the timing of δ 15 N depletions occurs predominantly between May to July (Table 1), which can be interpreted to reflect annual shifts in the whale's foraging that coincide with the annual zooplankton productivity peak. Timing of this zooplankton productivity peak can be confirmed by observations from the Fugløya-Bjørnøya long-term zooplankton time series, located a couple of degrees north of Lofoten; here, the annual increase in zooplankton biomass starts in April and peaks in July ( Figure 6). Given the findings from Skaug et al. (1997) that minke whales taken from Lofoten/Vesterålen in 1992 contained mainly herring and pelagics in their stomachs, and, given that spring-spawning herring are known to track zooplankton dynamics in the Norwegian sea (Broms et al., 2012), it is likely the elevated periods in δ 15 N reflect a switch to forage fish species such as herring. While Svalbard individuals also exhibit δ 15 N depletions, the δ 15 N do not drop so much as from Lofoten/ Vesterålen whales. An overall tighter distribution of δ 15 N values for Svalbard whales suggests that Svalbard whales may have a more consistent diet than Lofoten/Vesterålen whales.
Baleen whales are thought to eat much less or not at all during the winter breeding season. In terrestrial vertebrates it is widely accepted that fasting periods lead to elevated δ 15 N in tissues (Hobson et al., 1993;Polischuck et al., 2001), but the evidence for baleen whales is inconclusive. Best and Schell (1996)  Trends in stable carbon isotopes have been used to study migrations in some baleen whales, such as southern right whales, whose baleen plates can be >200 cm in length and who conduct their seasonal migrations in the southern hemisphere, where strong gradients in plankton δ 13 C values can be detected. In these whales marked, δ 13 C oscillations (ranging up to 8 ‰) in baleen plates of four adults were suggested to reflect the whales' annual north-south migrations (Best & Schell, 1996). In bowhead whales, which also have long baleen plates (up to 400cm in adult whales), stable carbon isotopes have been used to track east-west migrations between winter areas in the western Bering Sea (>180°W) and summer areas in the eastern Beaufort Sea (136°W) (Schell et al., 1989). For these whales, δ 13 C oscillations ranged up to 4 ‰. In Pacific minke whales, the baleen plates are a fraction of the size (~20cm), for which Mitani et al. (2006) stated that mature whales from their investigations lacked the strong oscillations in δ 13 C that have been reported for southern right whales. They suggested the lack of marked δ 13 C oscillations in baleen plates of Pacific minke whales is because the gradient seen from lower latitudes to the poles in plankton δ 13 C values, is only 0.1 times in the northern hemisphere compared to the southern hemisphere. In addition to the potential use of δ 13 C to reflect latitudinal gradients in plankton, δ 13 C values are also reported to vary amongst prey items from coastal habitats versus offshore habitats, e.g. enriched δ 13 C values across trophic levels have been reported F I G U R E 5 Composite mean and standard deviation for δ 15 N and δ 13 C isotope values for minke whales and potential prey items (cod, haddock, herring, sandeel, and krill) from the present study and studies in the literature. All literature values are for studies conducted in the North Atlantic (see references in the Appendix S1). The sample sizes for isotope values represented from each type of animal, including those measured in the present study, are minke = 2, cod = 5, herring = 7, haddock = 5, sandeel = 4, krill = 5 from animals sampled closer to shore (Miller et al., 2008;Sydeman et al., 1997). Long-finned pilot whales (Globicephala melas) feeding on coastal prey reflect different δ 13 C values than those feeding in deeper habitats (Monteiro et al., 2015). In addition to the above, coastal upwelling (Miller et al., 2008) and season (Sadler et al., 2012) are other factors associated with changes in δ 13 C values.
In the present study of North Atlantic minke whales, the greatest range in δ 13 C occurred in spring and summer. The δ 13 C signatures were remarkably similar throughout the year with little variation for Svalbard females (Figures 4 and S1, S5). The Lofoten/Vesterålen whales show greater variation throughout the year, and particularly for the females, with a larger spread in δ 13 C values in spring (March-June) and summer (June-September). The Lofoten/Vesterålen males, on the other hand, show somewhat less variance than the females, except for in spring (March-June) (Figures 4 and S1, S3, S4).
Although the graphical examination appears to show visible seasonal patterns in δ 13 C values, these were not supported statistically and may require greater sample sizes to be resolved. There was a significant association between δ 13 C and site, which could reflect differences in foraging locations between whales of different sites.
For example, the Lofoten/Vesterålen animals may shift between onshelf and off-shelf foraging, whereas Svalbard animals may be more consistent in their foraging locations. The seasonal patterns of δ 13 C observed in plots (Figures 4 and S1-S5) could suggest migratory activity (as per Best & Schell, 1996), or that enriched δ 13 C values from spring and summer, are a result of different plankton species in the water column (as per Sadler et al., 2012). Indeed, other marine studies report the primary productivity peak occurring in summer as the main reason for δ 13 C enrichment (Sadler et al., 2012). However, without a significant association between δ 13 C and seasons, there is insufficient evidence to verify either speculation.
The stable isotopic signatures of whale baleen can be useful to reveal patterns in year-round foraging habits. However, even with a greater sample size, the baleen record of Northeast Atlantic minke whales may be unsuitable to provide accurate insights into their migrations. The field of dietary studies continues to evolve, with new techniques that provide greater accuracy in estimating trophic position and metabolic condition (e.g. amino acid nitrogen isotopic analysis, O'Connell, 2017; Riekenberg et al., 2020) and an increasing use of inter-disciplinary approaches that provide greater insights in dietary studies (e.g. biochemical and molecular approaches, Horswill et al., 2018). As this research is undertaken, it can benefit from questions and hypotheses formulated on the natural history of minke whales.
The paradigm of mysticete seasonal migrations between highlatitude summer feeding grounds and low-latitude wintertime breeding grounds is subject to an on-going debate that questions whether all individuals within a population migrate (Corkeron & Connor, 1999;Geijer et al., 2016;Kavanagh et al., 2018). Examining the case for the Northeast Atlantic minke whale, an implication that all whales undertake a wintertime migration and, in so doing, leave the abundant food resources in the Norwegian and Barents Seas, presents the question of why males and, particularly, immature animals of both sexes might migrate. Folkow and Blix (1992) show that young minke whales are thermoneutral in water close to freezing, which suggests that migration into warm water is to no energetic advantage. In Norwegian waters there is no direct information on to the newborn minke calf, which measures 2.74 m at birth (Jonsgård, 1951), but newborn narwhal (Monodon monoceros) and beluga whale (Delphinapterus leucas), which measure 1.6 m and 1.5 m, respectively, thrive and grow in ice-water. Killer whale (Orcinus orca) calves have been observed on the Norwegian coast at 63°N in January (Christensen, 1978).
For the sake of the argument, if minke whales are not migrating to the tropics en masse, why are they not caught or seen, in the Norwegian Sea in winter? This may be due to the darkness; shipbased activities, such as winter fisheries for cod (Gadus morhua) and herring (Clupea harengus) occurring near the coast; rough sea conditions; and most importantly, the fact that the minke whales in the northern hemisphere do not spout when they surface to breathe. All this make them extremely difficult to spot, even when on lookout under optimal (dead calm) sea-state conditions during summer.
Conversely, what observations are reported at lower latitudes?
There is a report of 200-250 minke whales sighted between 20 and 30°N and 30 and 40°W in May (Slijper et al., 1964). However, at those coordinates, neither southern nor northern hemisphere minke whales would be expected to be found in May when the whales should be feeding at higher latitudes. These sightings were made by laymen; thus it is possible the animals were not minke whales.  (Visser et al., 2011), whereas the larger baleen whales seem to congregate and linger in the area to take advantage of the krill resources (Silva et al., 2013), and the minke whales, which are not dependent on krill, are not abundant in the area (Visser et al., 2011).

Vikingsson and Heide-Jørgensen
In light of the evidence concerning minke whale biology and movements, an alternative hypothesis suggests that minke whales are not migrating to the tropics en masse. Males and immature animals may remain for most of the year in Norwegian waters. A similar view, suggesting that not all individuals migrate, has been presented by others (Corkeron & Connor, 1999;Jonsgård, 1966;Kavanagh et al., 2018), but has received little attention. One of these studies in particular is based on two wintertime sightings (once in November and once in March) of minke whale calves in deep waters off the shelf edge from Ireland at 53.684°N and 55.523°N. The individuals were approximately 2.5 m and 3.5 m and in company of an adult, so thought to be unweaned calves (Kavanagh et al., 2018). Potentially it is mainly the pregnant females ready to give birth that travel to lower latitudes, not of a requirement for warmer water, but far enough to avoid predation from killer whales as suggested by Corkeron and Connor (1999). These predators are abundant in the Norwegian Sea where they feed on herring and mackerel (Scomber scombrus) during summer and autumn (Nøttestad et al., 2014) and along the west coast of Norway in winter (Øien, 1988).
Age determination in northern minke whales is uncertain (Christensen, 1995), and therefore also the age at sexual maturity. Kato (1987) found that 50% of southern female minke whales became sexually mature at the age of 6 years, whereas 100% were mature at the age of 10 years. The same appears true for the whales of the northeast Atlantic (Olsen & Sunde, 2002). This would then mean that immature animals comprise a great number of animals that may have no need to be in the tropics in winter. It follows, that even if only these animals remain in the Norwegian Sea throughout the year, it will have a significant impact on the food webs of the area, which has hitherto not been accounted for. Further research applying advancements in methods, new technologies and interdisciplinary approaches might better support the alternative hypothesis to the seasonal migrations of minke whales by examining specifically whether the sexes and age classes of minke whales exhibit differences in movements.

ACK N OWLED G EM ENTS
This paper is dedicated to the memory of John Eerkes-Medrano.