High pregnancy rates in humpback whales (Megaptera novaeangliae) around the Western Antarctic Peninsula, evidence of a rapidly growing population

Antarctic humpback whales are recovering from near extirpation from commercial whaling. To understand the dynamics of this recovery and establish a baseline to monitor impacts of a rapidly changing environment, we investigated sex ratios and pregnancy rates of females within the Western Antarctic Peninsula (WAP) feeding population. DNA profiling of 577 tissue samples (2010–2016) identified 239 males and 268 females. Blubber progesterone levels indicated 63.5% of the females biopsied were pregnant. This proportion varied significantly across years, from 36% in 2010 to 86% in 2014. A comparison of samples collected in summer versus fall showed significant increases in the proportion of females present (50% to 59%) and pregnant (59% to 72%), consistent with demographic variation in migratory timing. We also found evidence of annual reproduction among females; 54.5% of females accompanied by a calf were pregnant. These high pregnancy rates are consistent with a population recovering from past exploitation, but appear inconsistent with recent estimates of WAP humpback population growth. Thus, our results will help to better understand population growth potential and set a current baseline from which to determine the impact of climate change and variability on fecundity and reproductive rates.


Introduction
As populations of baleen whales recover from past over-exploitation they are re-occupying ecological roles, particularly on their high latitude feeding grounds from which they may have been functionally absent for many decades. For example, in many areas of the Southern Ocean populations of humpback whales (Megaptera novaeangliae) are recolonizing feeding grounds in the Antarctic and sub-Antarctic [1,2]. In some of these areas, whales are encountering environmental conditions very different from those that existed prior to their exploitation. One of these areas, the Western Antarctic Peninsula (WAP), is experiencing some of the fastest rates of regional climate change on Earth. This region has experienced a rise in temperature of nearly 7°C since the 1950s, resulting in the collapse of ice shelves, the retreat of glaciers and the exposure of new terrestrial habitats [3][4][5]. A decline in seasonal sea ice has been observed along the WAP, resulting in an annual occurrence of sea ice (greater than 15% cover) that is, on average, 80 days shorter than it was four decades ago [5]. This warming is proceeding with a variety of impacts on the ecology of the system [5,6], including regime shifts among sympatric krill predators along the WAP, such as declines in the abundance of Adélie penguins and their replacement by sub-Antarctic gentoo penguins [5].
Like most mysticetes, humpback whales exhibit an annual migration from low-latitude breeding grounds to high-latitude feeding grounds where they exploit high levels of seasonal productivity [7]. The timing of this migration varies as a function of age, sex and reproductive state [8]. During the austral summer and fall, humpback whales feed throughout the Southern Ocean. For management purposes, the International Whaling Commission (IWC) divides this distribution into six feeding areas and includes the WAP within Area I [1,9,10]. The WAP is an important feeding ground for humpback whales (breeding stock G) that winter along the west coast of South and Central America (e.g. Ecuador, Colombia and Panama) and perhaps other populations as well [9,11]. The timing of migration, and therefore the duration of feeding during the austral summer and fall is critical to fuelling the migration of whales to their low latitude breeding grounds [7]. Energetic studies of blue and fin whales have shown that roughly 25% of their annual lipid reserves are used over the course of the annual migration [12]. These energetic demands are even greater for pregnant or lactating whales. Lockyer [12] suggested that lactation and pregnancy will consume roughly 19% of total energy stores in mature female fin whales.
Little is known about the exact timing of migratory movements and residency times of different age and sex classes of humpback whales along the WAP. Our lack of knowledge stems largely from the rapid depletion of humpbacks at the very beginning of Antarctic whaling during the first part of the last century [13], as well as the significant logistical challenges associated with studying whales in polar regions. During this period of whaling, humpback populations were a favoured target because they were relatively easy to catch, abundant in protected bays, and floated when killed [14]. The rapid depletion of these stocks has been well documented [13]. For example, humpbacks comprised the majority of the catch at South Georgia from 1904 to 1910, but by the 1913-14 season this species comprised less than 20% of the whales taken [15]. The same story of unconstrained over-exploitation was repeated in the South Shetlands and along the WAP [14]. As a result, there was little opportunity to study the ecology of humpback whales before they were reduced to remnant populations. An improved understanding of the demography and seasonal patterns of movements of humpback whales in this area would help place current research on their regional density and population dynamics into a broader ecological context.
Identification of the sex of individuals facilitates the study of population structure, behaviour, breeding patterns and social systems [16]. As both iconic animals and top predators, understanding such knowledge about humpback whales allows us to better understand how individuals interact and organize themselves within their environment, as well as provide insight as to the health, structure, and function of the marine ecosystem of which they are a part. Insights into the timing of migration for various age and sex classes are available from other populations of humpback whales (east and west Australia) in the Southern Ocean that were exploited later than those along the WAP [17]. In these areas, pregnant female humpback whales were the first to arrive on the feeding grounds, followed by immature animals, resting females and mature males, and lastly, lactating females and their young calves [7]. These observations were generated during the height of commercial whaling, and there have been no attempts to assess these patterns along the WAP until now.
Estimates of pregnancy are important for assessing density dependent effects predicted for populations as they recover and approach pre-exploitation or presumed carrying capacity [18]. Historically, assessment of the reproductive status of cetaceans was obtained from examination of carcasses taken in commercial hunts [19,20]. Since the protection of humpback whales by the IWC in 1966 this source of information has become unavailable for this species [19].
Since protection from hunting, reproductive rates have been estimated for some populations of humpback whales from crude birth rates (% of calves) [21] or by estimating calving rates derived from long-term sighting histories of individual mature females [22][23][24][25]. In this approach, estimates of reproductive rates were generated through repeated observations of adult females with and without calves. This approach provides an estimate of the number of recruits entering a population but, without correction, inherently underestimates true fecundity rate, because it does not account for fetal and perinatal mortality. Rates of fetal and perinatal mortality can vary and are potentially important indicators of population health [26].
More recently, biochemical techniques have been developed to detect pregnancy using non-lethal sampling and in some cases making it possible to estimate reproductive rates prior to parturition. For example, pregnancy status has been determined by assessing the concentrations of progesterone in the milk of lactating bottlenose dolphins [27], sex steroids secreted in the urine of killer whales [28], salivary steroids in Hawaiian monk seals [29], blood plasma, salivary, ocular, and vaginal secretions of false killer whales [30], and ultrasound evaluations of dolphins examined during capture-release health assessments [31]. However, such methods are not practical for use with large whales, which cannot be captured or handled. The measurement of progesterone in samples that can be collected in the field, such as faecal material [32], blow [33,34] and skin biopsies [19,20,35] offer a pragmatic alternative for these animals.
Progesterone, often referred to as the hormone of pregnancy, is a lipophilic circulatory steroid hormone produced by the corpus luteum, and is the primary regulator of oestrous cycling and pregnancy [36]. Progesterone's lipophilic properties make biopsy samples of skin and blubber a readily obtainable, non-lethal, analytical matrix for assigning pregnancy in free-swimming cetaceans. This work was pioneered in a large-scale study of short-beaked common dolphin (Delphinus delphis), northern rightwhale dolphin (Lissodelphis borealis) and Pacific white-sided dolphins (Lagenorhynchus obliquidens) killed incidentally in the California gill net fishery [20]. The pregnancy status of these female dolphins was determined by physical examination of the carcass, by noting number of corpora, corpus luteum size and/or length of the fetus (if present) [20]. The authors then validated blubber sample endocrine techniques by corroborating their progesterone values with the known reproductive state of the carcass; elevated progesterone concentrations were correlated with pregnant females.
The development of non-lethal tissue sampling techniques [37], and the methods to assess hormone levels from skin-blubber biopsy samples, now provides the capacity to assess demographic rates in wild cetacean populations [19,20,35]. The objective of our study was to assess variation in the sex ratios and pregnancy rates of humpback whales along the WAP, and in doing so, develop a baseline with which to assess the impact of climate change and variability, that affect both available habitat and prey.

Biopsy collection
We collected skin and blubber samples from humpback whales during the 2010 and 2013-16 austral summer and fall (January-June) field seasons along the WAP using standard biopsy techniques [37]. Samples were collected whenever whales were encountered during prey or visual surveys in an area within approximately 10 nautical miles of Palmer Station, Anvers Island. We also collected samples during the Palmer Long Term Ecological Research (LTER) research cruise in January and February of each year (figure 1). Samples were collected unselectively from all age/sex classes, except young of the  year calves, which were excluded. Calves of the year were assumed to have been born the previous breeding season as evident by their small size (less than half of the presumed mother's length) and close association with an adult, presumed to be the mother. Samples were collected using a crossbow with modified bolts and 40 mm tips (CetaDart) from a small vessel at distances of 10-30 m targeting the dorsal and lateral body surfaces near the dorsal fin [37]. Samples were stored frozen whole at −20°C until used for analysis.

DNA profiling
A standard DNA profile, including microsatellite genotypes and sex-specific markers, was used to identify individuals [38]. Total genomic DNA was extracted from the skin-blubber interface using a proteinase K digestion followed by a standard phenol-chloroform extraction [39]. The sex of each sampled whale was identified by amplification of sex-specific markers following the protocol of Gilson  et al. [40]. We compared results to controls for a known male and female using gel electrophoresis. Sex ratios, depicted as the ratio of males to females (M : F), were calculated for the entire dataset, within years, and across seasons. Each sample was also genotyped using 10 previously published microsatellite loci to determine individual identity and remove potential replicate samples (table 1) [41][42][43][44][45]. Alleles were sized and binned using the software program Genemapper v3.7 (Applied Biosystems). The total number of amplified loci for a given sample was considered as an added quality control threshold. Given the estimated probability of identity for these loci from previous studies [46], we considered that 7 loci were sufficient for individual identification. Samples with fewer than 7 microsatellite loci were re-analysed or excluded. The expected probability of identity (P ID ; the probability that two individuals drawn at random from a population will have the same genotype by chance) for each locus was calculated in GenAlEx v6.5 [47]. Cervus 3.0.7 [48] was used to compute the number of alleles (K), observed and expected heterozygosity (H O and H E ), deviation from Hardy-Weinberg equilibrium with a Bonferroni correction, and the probability of identity for all individual matches.

Hormone extraction and quantification
Hormone extraction followed standard methods [20,35]. In brief, a cross-sectional sub-sample (approx. 0.15 g) spanning from the epidermis-blubber interface to the most internal layer of the biopsy was subsectioned from female whales. These sub-samples were homogenized separately in 1400 µl of ethanol using an automated, multi-tube homogenizer (Bead Ruptor 12, Omni International). The resulting homogenates were then rinsed with a series of ethanol, 4 : 1 ethanol : acetone, and ethylether washes. Following each rinse, the supernatant was collected. Lastly, the progesterone was separated from the resulting lipid residue using a biphasic acetonitrile-hexane separation. The final hormone pellet was dried down and stored at −20°C until analysis. Progesterone concentrations were quantified using a progesterone enzyme immunoassay (EIA; Enzo Life Sciences, kit ADI-900-011). Prior to analysis, samples were re-suspended in 1 ml of phosphate buffered saline (pH 7.5) containing 1% bovine serum albumin and vortexed thoroughly. The progesterone EIA kit used in this study has 100% reactivity with progesterone and an assay detection limit between 15 and 500 pg ml −1 using the standard curve. Two additional standard dilutions were added to allow for a lower detection limit of the standard curve to 3.81 pg ml −1 . Samples were run blind and in duplicate. If a sample failed to fall within the detection limit of the assay curve, the samples were re-run at varying dilutions. The reported inter-assay coefficient of variation (COV) and intra-assay COV of the progesterone EIA were in the ranges 2.7-8.3% and 4.9-7.6% respectively. Progesterone concentrations are reported as nanograms of progesterone per gram of blubber (ng g -1 ).
We determined extraction efficiency with every progesterone extraction using samples from a dead, stranded animal of known pregnancy status by spiking subsamples of blubber with 150 ng of progesterone and including these with every extraction [20]. The percentage of progesterone recovered after the extraction was calculated and each sample concentration was adjusted to this efficiency prior to statistical analyses. An extraction efficiency greater than 60% was adequate and is based off the reported range of efficiencies that are seen using these methods [20]. If the efficiency of an extraction set was less than 60%, the sample extracts were discarded and the blubber samples were re-extracted and re-analysed.
A series of serially diluted pooled samples from five humpback whale individuals from the WAP were compared against the EIA kit progesterone standard to determine parallelism [49,50]. The series of serially diluted pooled samples and kit standards were run in duplicate and an analysis of covariance (ANCOVA) was used to determine if the slopes between the two lines were significantly different.

Pregnancy classification
Pregnancy was assigned using the methods developed by Pallin et al. [51]. Briefly, we used the progesterone concentrations from a collection of female humpback whales of known pregnancy status, based on annual calving observations, to develop a predictive model. We interpreted the pregnancy state of females sampled along the WAP based on the relationship of their progesterone concentration with the reference levels from known pregnant animals. The model returned the probability that each female was pregnant [51]. If the probability of being pregnant was greater than 99.9%, that female was assumed to be pregnant. If the probability of being pregnant was less than 0.1%, that female was classified as not-pregnant. If a female's probability of being pregnant was between those two bounds, she received an undetermined pregnancy assignment. Lastly, after all probabilities were assigned, we determined the probability of a sample being miss-classified.

Data preparation and statistical analyses
We used a two tailed exact binomial test [52] to test for deviations from a 1 : 1 sex ratio (parity) across the entire dataset, within a given year, and within each season. We separated the austral seasons of summer (January-March) and fall (March-June) using the autumnal equinox on 19 March. Additionally, to avoid re-sample bias in our analyses, we removed all within-year replicates from the inter-annual comparisons; we also removed within season replicates from the same year from the seasonal analyses. In both cases, the most recent sample was retained for the analyses.
We tested for differences in the sex ratios and pregnancy rates across all years and across both seasons by using a χ 2 test of independence. Additionally, we used a Tukey post hoc stepwise multiple comparison test to determine if there was a significant difference in pregnancy rates between any two individual years. For all statistical tests, we considered a p-value of less than 0.05 to be significant. All values are expressed as mean ± s.d., unless otherwise stated.

Individual identification and sex
We collected 583 biopsy samples from the waters around the WAP over the course of five field seasons from 2010 to 2016 (figure 1). On average, 9.8 loci were successfully genotyped per individual. The average P ID for any given combination of 7 loci ranged from 4.29 × 10 -11 to 4.83 × 10 -8 , consistent with previous studies. Two samples failed the initial genotype quality control and were re-analysed. Consequently, we considered samples with matching genotypes to be recaptures of the same individual. The DNA profiling was sufficient to identify and determine sex of 507 individual whales from these samples (table 2). In total, we sampled 239 individual males and 268 individual females over the course of the study. Details on annual sampling can be found in table 2. We resampled 54 individuals within the same year (table 3). Three of these individuals were sampled three times (two in 2014, one in 2016) and one individual was sampled four times in 2016. Additionally, we recaptured 11 individuals between years (table 3).

Annual and seasonal variation in sex ratios
Overall, we sampled more females than males (0.89 M : F), but this deviation from parity was not significant (p = 0.213, exact binomial test; table 2). Nor did the sex ratio differ significantly from unity in any given year (table 2)

Validation of humpback progesterone assays
Based on the concentrations observed from the series of spiked controls, our average extraction efficiency was 73.78% ± 0.09 (minimum 61.1%, maximum 95.6%). Additionally, our calculated inter-assay and intra-assay COV from a series of replicated samples was 6.37 and 7.75% respectively. The results from our assay parallelism test between the kit progesterone standard and humpback whale progesterone showed strong parallelism, indicating that the use of the progesterone assay to detect humpback whale progesterone is viable. This result was consistent with previous studies on humpback whales [53]. Specifically, the slopes of the two linear curves were not significantly different (p = 0.848; see electronic supplementary material, figure S1).

Pregnancy assignment
We measured progesterone concentrations in 264 samples obtained from 244 individual female humpback whales ( figure 1). A small number of samples were excluded from the analysis due to within year re-sampling or insufficient blubber for an extraction. Based on the relationship of their progesterone concentration with the reference levels from known pregnant animals, 89 individuals were assigned as not pregnant (p < 0.1% pregnant; blubber progesterone: mean = 2.10 ± 1.13 ng g -1 ;

Variation in sex ratios
The sex ratio of the sampled population was close to unity (0.89 M : F), and we observed no significant differences in sex ratios across years. Our analysis thus supports early observations, derived from catch data, that the sexes mix randomly on the feeding grounds in the Southern Ocean [7,17,54]. Chittleborough [17] reported that males comprised 52.8% of more than 18,000 humpback whales killed in Antarctic Area IV (south of Australia and New Zealand) between 1949 and 1962. Additionally, Matthews [54] reported that 45.4% of the 1057 whales killed at South Georgia were male. Similarly unbiased sex ratios have also been reported from feeding grounds of the North Atlantic [55] and North Pacific [45].
However, we observed a significant bias towards females as the feeding season progressed into the austral fall. This result corroborates prior observations of the temporal segregation of humpback migration in the Southern Hemisphere based on individual demographic status. Specifically, Dawbin [7] reported that during the northbound migration lactating females were the first whales observed to be migrating north to the breeding grounds near Cook Strait, New Zealand. These whales started their migration about one month earlier than the first northbound catches, and were followed by immature whales, mature males, resting females, and lastly, pregnant females [7], resulting in an increased proportion of females on the feeding grounds into the fall.

Variation in pregnancy rates
We observed a pregnancy rate of 63.5% across all years. Interestingly, we also documented evidence of an annual reproductive cycle in which postpartum ovulation and conception occurred in some females. The pregnancy rate of females varied significantly among the 5 years of sampling, ranging from 36 to 86%. This overall rate is higher than reports of the pregnancy rate of female humpbacks taken in Antarctic whaling areas IV and V (mean = 48%) from 1950 to 1956 [53,56]. Our estimate of pregnancy rates is also greater than expected from observed calving rates of 0.37-0.41 calves/year/mature female from Northern Hemisphere populations of humpback whales [57,58], which are also recovering from past exploitation [59][60][61], but these latter rates are not directly comparable to those presented here, because they include the unknown effects of fetal and perinatal morality.
Calving rates are generated by observing females accompanied by calves that have survived long enough to be observed. Absent sampling biases, the difference between this and pregnancy rates should reflect the rates of fetal loss and perinatal mortality that occur between the dates of pregnancy determination (in this case the biopsy date) and the dates of observations of mothers and calves. In addition, our pregnancy metric reflects the proportion of females that are pregnant, irrespective of maturity state. Typically, calving rates are reported as the fraction of mature females observed with a calf. These distinctions are extremely important when drawing comparisons between our work and other studies.
We observed the highest annual pregnancy rate (86%) in 2014. Similarly high rates of 72% and 82% were documented from catch data in Antarctic whaling areas IV and V from 1950 to 1956 and the Bellingshausen Sea, just west of the WAP, in February of 1957 [56,62]. However, these estimates may be positively biased, as lactating females accompanied by calves were protected by international regulation during this period [56]. Over the course of six years of whaling in Antarctic areas IV and V, pregnancy rates varied from 36% to 72%, similar to the variation we observed over the five years of our study [56].
Observed pregnancy rates along the WAP increased from summer to fall. We believe that pregnant females maximize time spent on the feeding grounds prior to migration. This interpretation is consistent with the bias in favour of females that we observed into the fall, as well as the demographic segregation of migration observed from northbound whales during the whaling era [7,17]. Specifically, there was an increase in the proportion of pregnant females caught in late July and August off Cook Strait, New Zealand [7]. Lockyer [12] suggested that baleen whales must increase their body mass up to 65% to maintain energetic costs for the next year, so pregnant females should stay as long as possible on the feeding grounds to increase their lipid reserves. Lockyer [63] demonstrated that this delay in migration allowed pregnant females to acquire an additional 10-15% of fat reserves than resting females.

Evidence of annual pregnancy
The frequency of post-partum pregnancy and annual calving in humpback whales is poorly understood. We sampled 44 mothers accompanied by calves over the course of our five field seasons, and more than half of these lactating females were pregnant. Chittleborough [64] concluded that female humpback whales are seasonally polyoestrous, with oestrus occurring on the breeding grounds from June to October [54]. There is a general lack of information on the termination of these cycles, but they are believed to occur at conception or when the southward migration begins. Most cycles are characterized by a single ovulation [17,65]. The total number of ovulations over a 2-3 year breeding cycle, however, still remains unclear [66]. Chittleborough [66] documented post-partum ovulation by examining the carcasses of female whales during the commercial whaling era. Specifically, of 19 lactating females he examined at Western Australian whaling stations between 1949 and 1955, eight (42.1%) were both pregnant and lactating [66]. This is similar to the proportion we observed (54.5%) among females sampled along the WAP, almost 60 years later.
There are few other records of annual reproduction in humpback whales in the literature derived from catch records, likely because the IWC prohibited the killing of lactating females accompanied by calves. However, several studies have documented annual reproduction in female humpback whales in the Northern Hemisphere [22,23,58,67,68]. For example, on the Hawaiian breeding grounds, Glockner-Ferrari & Ferrari [23] observed that four of 34 females experienced an annual reproductive cycle and one female was observed with different calves in four consecutive years. However, Glockner-Ferrari and Ferrari focused primarily on approaching mothers with a calf, leading to a likely ascertainment bias. Observations of annual calving are much lower on the feeding grounds. For example, Robbins [22] calculated that only 2% of females returned with calves in consecutive years in the Gulf of Maine.

Are observed rates of pregnancy 'too high'?
In 1966, the IWC prohibited the commercial takes of humpback whales; this protection has now been in place for over 50 years [69]. At a global level, the recovery of humpback whale populations under this protection has been very effective, as reflected by a recent decision to down-list the species to Least Concern by the International Union for Conservation of Nature (2016), although a few populations in the Northern Hemisphere remain in poor conservation status. It is reasonable to infer that the high abundance of krill along the WAP [70] has supported the recovery of this feeding aggregation of humpback whales. Unfortunately, it is not possible to reconstruct the demographic trajectory of these whales over the last half-century because there is no good historical baseline for this population and we have just recently been able to assess demographic parameters, such as pregnancy rates, in these whales.
We believe that the existence of annual reproduction in this population represents a response to favourable ecological conditions, as humpback whales along the WAP recover from past overexploitation. The frequent occurrence of an annual reproductive cycle, together with a high mean pregnancy rate of 63.5%, in female humpbacks suggests a population that is growing rapidly [23]. Indeed, the WAP feeding aggregation of humpbacks (and nearly all populations of this species) is increasing [60,71]. However, our pregnancy rates are much higher than expected for a population that is believed to be growing at a rate of only 3.4% per year [60]. Similarly high observations of annual reproduction and overall pregnancy rates were made among humpbacks sampled of the Kermadec Islands in Oceania [72]. The east Australian population of humpbacks is projected to be recovering at or near the estimated physiological limit of the species (approx. 10% per year) [73,74]. The differences between the high pregnancy rates we observed and the rather pedestrian rates of population growth estimated for breeding stock G could reflect one or more of the following: the estimate of population growth is biased low; pregnant females experience very high rates of perinatal loss, and or mortality of the calves on the breeding grounds; or there is significant spatial heterogeneity with respect to reproductive state on the foraging grounds.
The growth of a population depends not only on its reproductive rate, but on the survival and recruitment of calves into the population [75]. It is possible that the occurrence of an annual pregnancy cycle in this population results from the early mortality of a calf or late term fetus during the previous cycle [63,75]. There are no data on fetal loss in this species, but calf mortality is known to occur after birth but before arrival on the feeding grounds in some populations [76]. We believe that it is unlikely that significant fetal loss is occurring in humpback whales along the WAP, because the population is recovering from past depletion and because of a lack of other baleen whale populations having similarly recovered, it is likely that there is no competition for resources (e.g. food is not limiting) and whales can feed unrestricted. Resolution of this question will require more information on the reproductive histories of individual females and, particularly, on the fate of their dependent calves. Our future work, specifically genetic and photographic capture-recapture studies of mothers with calves, will focus on addressing this knowledge gap.
We sampled whales opportunistically, avoiding re-sampling an individual when possible. However, it is possible that the high rates of pregnancy we observed could be derived from some spatial artefact. Perhaps different demographic groups (e.g. lactating females, pregnant females) distribute themselves unevenly within the feeding area. We believe that this possibility is unlikely due to the fact that these animals range over very broad spatial scales during the feeding season [77]. However, future work is required to address this possibility.

Conclusion
Our research demonstrates that this feeding aggregation of humpbacks exhibits high pregnancy rates and a high proportion of females that are simultaneously pregnant and lactating. Both of these findings are consistent with a rapidly growing population, but are not supported by current estimates of population growth for the WAP population. Our work builds on the demographic and life-history data of Southern Hemisphere humpbacks derived from whales killed during the last century and published 50 years ago. Our intent here is to document the current demography of this population and establish a baseline with which to assess the impact of future climatic trends. As the extent of winter sea ice continues to decline, it is possible that we will witness a temporal shift and spatial expansion of the feeding grounds of these whales, at least in the short term, and prey availability will likely increase [78][79][80]. Long-term trends however, may be more problematic, due to the tight coupling of sea ice and the recruitment of krill in the WAP ecosystem [81,82]. We have already witnessed responses in the demography of other baleen whales to climate change [79,83]. We will continue to monitor this feeding aggregation of whales to document its recovery and response to future environmental changes. Authors' contributions. L.J.P. engaged in data collection, laboratory analyses, and interpretation of the results; A.S.F. assisted in project design, data availability and collection, fieldwork organization, interpretation of the results, and procured all necessary permits; D.W.J., A.J.R. and D.P.N. contributed to the initial conception of the project as well as data collection in the field, and interpretation of the results; C.S.B. and D.S. provided logistical support and a laboratory space for in laboratory genetic analyses as well as interpretation of the genetic results; N.M.K. provided training in the necessary laboratory extraction, analysation, and safety protocols needed for the hormone work and provided continued support in interpretation and analysation of hormone results; lastly J.R. provided a series of control samples from the Gulf of Maine that greatly aided in the interpretation of the results. All authors participated in drafting the manuscript and gave final approval for publication.