Foraging Tactics Of Humpback Whales Feeding Near Salmon Hatchery Release Sites In Southeast Alaska

Increases in the humpback whale (Megaptera novaeangliae) population have generated considerable interest in understanding the foraging habits of these large marine predators in the Gulf of Alaska. Globally, humpback whales are classified as generalist predators but are known to exhibit localized differences in diet. Intensified predation pressure is of particular concern to resource managers, who have observed whales feeding at juvenile hatchery salmon release sites in Southeast Alaska. We assessed the diets and behavioral tactics of humpback whales foraging near Hidden Falls Hatchery release sites (in Chatham Strait, 2016 to 2018) to better understand their predatory effects on juvenile hatchery-reared salmon. We used skin biopsies, prey sampling, and stable isotope analysis to estimate whales’ diet composition. Aerial footage and photographic sequences were used to assess the foraging tactics used on this prey source. We observed three individual whales repeatedly feeding on juvenile hatchery-reared salmon, and we were able to sample them multiple times over a period spanning shifts in diet. Overall, the diets of these whales were higher trophically than other humpback whales foraging in the area, even before feeding on juvenile hatchery salmon started. These hatchery-feeding whales may be generally more piscivorous than other whales, which focused on planktivorous prey. Our repeat sampling, in conjunction with scheduled introductions of a novel prey source, provided a semi-controlled feeding experiment that allowed for incorporation and turnover rate estimates from humpback whale tissue in a way that was not previously possible for large, free-ranging cetaceans. Finally, during the course of this study we discovered an undescribed feeding tactic employed by hatchery-associated whales. We observed the use of solo bubble-nets to initially corral prey, followed by calculated movements to establish a secondary boundary with the pectoral fins that further condensed prey and increased foraging efficiency. Our study provided the first empirical evidence for what we describe as “pectoral herding”. This work deepens our knowledge about humpback whale foraging ecology, how this innovative species is able to exploit newly available prey, and to what extent they feed on commercially valuable hatchery salmon.

Since 2008, humpback whales have been documented depredating hatchery-produced juvenile salmon, a novel prey, at points of their release in Southeast Alaska. The objectives of this dissertation are to determine the spatial distribution, seasonal distribution, and frequency of humpback whale foraging at release sites, determine whether whale presence is affecting the economic productivity of hatchery operations, and compare the bioenergetic benefits for whales feeding on juvenile salmon at hatchery release sites relative to typical prey. Five hatchery release sites were monitored over six years during the spring release season for whale presence/absence, numbers, and behaviors. Linear models were used to determine that for coho salmon, cohorts with frequent humpback whale presence had lower marine survival than cohorts with less or no humpback whale presence, but this was not seen for chum or Chinook salmon. Over six years, these sites lost an estimated 23% of revenue from coho salmon totaling almost a million dollars per year in addition to increased rearing costs to mitigate whale predation. A process model was developed to compare the net energy gain for humpback whales foraging on krill, herring and juvenile salmon. Whales were found to feed profitably on krill and chum salmon where they occurred in dense enough distributions and on herring when large coordinated groups impeded the escape of prey. Coho salmon typically distributed too diffusely for humpback whales to recuperate the full energetic costs of engulfment, indicating that behaviors such as bubble net feeding may be essential for increasing prey aggregation to an energetically profitable level, or humpback whales may be feeding to mitigate energetic losses. As intraspecific competition increases due to recovery and or changes to prey resources, generalist humpback whales may expand feeding to exploit new and less profitable prey resources.

Relative to other baleen whale populations, the humpback whales Megaptera novaeangliae inhabiting Southeast Alaska are noteworthy in that they form large, enduring groups when foraging on schools of Pacific herring Clupea harengus pallasi. These groups use a variety of unusual feeding tactics when attacking prey, including the production of loud feeding calls, the release of bubbles, and the waving of their large pectoral flippers. Past observations of these groups have been largely anecdotal so little is known about their social behavior or the function of their feeding tactics. In particular, it is unclear if these pods are based on cooperative interactions or if they represent groups of individuals who are competing for prey that occurs in dense aggregations. In addition, little is known about the underlying social structure of these pods and whether they are composed of kin. The objective of this study was to employ field and laboratory techniques to gain insight into the function of the unusual feeding tactics, and to describe the basic social structure of these pods. In order to test the prediction that bubbles, feeding calls, and flipper movements represent prey manipulation tactics, herring schools were brought into the laboratory and subjected to various simulated humpback whale feeding behaviors. I found that these simulated behaviors produced strong avoidance responses from the herring schools, suggesting that humpback whales use these tactics to manipulate the behavior of their prey. Seven years of field observations revealed that humpbacks with enduring social bonds (i.e., high coefficients of association) specialized on herring and represented a small proportion of the entire whale population. Calves born to these "core members" were never observed to return and feed with their mothers in subsequent years, suggesting that these groups are not composed of close kin. This was verified by analysis of haplotype variation in the mitochondrial genome and microsatellite variation in the nuclear genome. Individuals within these pods appear to invest in by-product benefits, with the enduring bonds between whales in Chatham Strait (cf. krill feeders in Frederick Sound) possibly representing combinations of individuals performing compatible tasks (Le., bubble blower, herder, vocalizer).

The North Pacific humpback whale (Megaptera novaeangliae) population has been increasing at an average annual rate of ~6% since the early 1990s. In northern Southeast Alaska alone, there are now more whales than estimated for the entire North Pacific several decades ago. An understanding of how this growing population is repopulating traditional foraging grounds will benefit from detailed investigations of their prey preferences and trends in whale abundance and distribution relative to those prey. This dissertation examines these issues from late May until early September 2008 in Frederick Sound and Stephens Passage, a Southeast Alaskan feeding area historically used by humpback whales. The foundation for the study is an analysis of the life histories and abundance patterns of euphausiids, the principal prey of humpbacks in the area, during late spring and summer. Four species, Thysanoessa raschii, T. longipes, T. spinifera, and Euphausia pacifica, were identified in plankton net samples collected at random locations throughout the study site (n = 49) and in locations where a strong scattering layer was observed on a 120 kHz echosounder (n = 48). Both sample types varied in euphausiid species composition. Abundance patterns of immature euphausiids coupled with observations of females carrying spermatophores indicated differences between species in spawning schedules. Thysanoessa spp. began spawning in early April with the spring phytoplankton bloom and continued until late June, whereas E. pacifica began spawning in early June and continued until late August. This protracted recruitment of immature euphausiids was geographically widespread throughout the summer in contrast to adults, which, although present all summer, were found primarily in slope and shallow (

Humpback whales (Megaptera novaeangliae, Borowski 1781) in the North Pacific migrate from mid- to high- latitude summer feeding grounds along the Pacific Rim, including areas off the coasts of the U.S., Canada, Russia and eastern Asia, to tropical breeding grounds each winter along Pacific coasts of Mexico and Central America as well as the offshore islands of Mexico, Hawaii, and Japan. Humpback whales in the North Pacific and elsewhere were reduced to very low numbers during a period of intense commercial exploitation that ended in 1965. As the population recovers in abundance, the range of cultural and genetic diversity that survived the exploitation-driven bottleneck is able to adapt, endure and evolve. My work uses genetic tools and photo identification data to investigate the population dynamics, mitochondrial (mt) DNA control region evolution and potential drivers of a specialized feeding behavior in a recovering subpopulation of humpback whales in the Glacier Bay and Icy Strait (GBIS) sub-region of the southeastern Alaska (SEAK) feeding ground. I first collated and reconciled available DNA profiles (mtDNA control region, 10 microsatellite loci and sex) from 556 individuals using tissue samples collected from 1987 to 2012. Photo identification records associated with 692 of 1,026 total genetic samples collected in SEAK (now archived within the SEAK DNA Register and Tissue Database) corresponded to extensive life-history information, extending back to the early 1970s, as archived within the SEAK Regional Database, curated by the National Park Service (NPS) and University of Alaska, Southeast (UAS). Changes in population structure in GBIS over 32 years (1973-2005) were investigated in order to determine whether the increase in local abundance was attributable to local fidelity and recruitment or immigration from outside of SEAK. Two temporal strata were defined: 'Founder' individuals identified between 1973-1985 (n = 74, n = 46 with DNA profiles) and 'Contemporary' individuals identified between 2004-2005 (n = 171, n = 114 with DNA profiles). There was no significant genetic differentiation between the strata, indicating that it is unlikely that the population increase within GBIS was due largely to immigration of whales from elsewhere in the North Pacific. However, two additional haplotypes were documented in the Contemporary stratum at low frequency, one of which was previously unreported in the North Pacific (haplotype A8, see below). This relative stability in haplotype frequencies over time argues for strong regional fidelity of the maternal lineages represented in GBIS between 1973 and 1985. After excluding the 42 Contemporary whales with no photo ID record of a mother or genotype available for maternity inference, at least 73.6% (n = 95) of the Contemporary stratum was either a returning Founder or a recruited descendant of a Founder female. Of all genetically confirmed females with genotypes in the Founder stratum, 96% (n = 24) were either represented in the Contemporary stratum, had at least one confirmed descendant in the Contemporary stratum, or both. This high proportion, in addition to the large proportion of the verifiable Contemporary stratum that were either returning Founders or a descendant of a Founder female, provides clear evidence for local fidelity and recruitment to GBIS. The discovery of the A8 haplotype, which differs by one base pair from a common haplotype referred to as A-, represents an increase in mtDNA diversity for the North Pacific humpback whale from 28 to 29 haplotypes. To investigate the origin of this new haplotype, we re-evaluated n = 1089 electropherograms of n = 710 individuals with A- haplotypes from both the SEAK DNA Register and Tissue Database and the ocean-wide program, SPLASH (Baker et al. 2013). From this review, we identified two individuals with the A8 haplotype (a cow and calf, both sampled in GBIS) and n = 20 individuals with clear heteroplasmy for haplotypes A-/A8. The majority of A-/A8 individuals (n = 15) were sampled in SEAK. Genotype exclusion and likelihood were used to identify one of the heteroplasmic females, #196 (first sighted in SEAK in 1982), as the likely mother of the A8 cow and grandmother of the A8 calf, establishing the inheritance and germ-line fixation of the new haplotype from the parental heteroplasmy. Based on life history records and estimates of pairwise relatedness from microsatellite genotypes, it appears likely that the A-/A8 and the A8 individuals are descendants from a common maternal ancestor one or more generations prior to the three generations documented here. Humpback whale sociality takes a distinct form in Icy Strait, where whales form large, coordinated groups with repeated membership across several decades. Twenty-one years of group association records (1985-2005, n = 2,204 groups) were used to investigate the hypothesis that kin selection influences membership in large, stable groups. Of the 2204 groups recorded, 113 consisted of 6 or more individuals; a size considered unexpectedly large assuming a Poisson distribution of group size with a mean of 1.7. A total of n = 71 individuals (n = 48 with DNA profiles) were encountered in a large group in at least one year, n = 38 individuals (n = 34 with DNA profiles) were encountered in a large group in at least two years, n = 29 individuals (n = 27 with DNA profiles) were encountered in a large group in at least three years, decreasing to n = 2 individuals (n = 2 with DNA profiles) that were encountered in a large group in at least 20 years. There were no significant differences in mtDNA frequencies between large group feeders and the Founder and Contemporary strata or when compared to whales never encountered in large groups in Icy Strait, indicating that group membership is not predominately passed through one maternal lineage. Sex ratios did not deviate significantly from 1:1 for those feeding in large groups over an increasing number of years, as would be expected if females were actively recruiting offspring into large groups. The average pairwise relatedness for large group feeders was not significantly greater than expected by chance and did not increase for those feeding in large groups over an increasing number of years. Of the 179 known offspring of females encountered in a large group, only 6% were also encountered in a large group in Icy Strait as an adult and only 2.2% in the same large group as their mother. These results indicate that kin selection is not the primary driver of membership in large, stable groups and pose an interesting dynamic in local habitat use: individuals are recruited to GBIS through local maternal fidelity but do not usually associate closely with direct maternal kin. The extensive collection of DNA profiles now archived with the individual-based data within the SEAK Regional Database allowed us to integrate genetics and photo ID to answer ecologically relevant questions regarding the whales in GBIS. Together, these results demonstrate that GBIS provide habitat for a distinct collection of individuals that exhibit strong fidelity and local recruitment, some of which engage in a highly specialized feeding behavior. Further, GBIS is a local feeding habitat for two individuals with a newly arising North Pacific mtDNA haplotype. These findings reveal local genotypic and cultural variation and highlight the importance of habitat protection for species with fine-scale habitat use and strong fidelity to local migratory destinations.

Group foraging is observed in many species as a means to increase the ability of members of the group to find and exploit patchy prey. Group foraging can be exhibited in a number of different contexts based on the relationships between the participants, including by-product mutualism. One variant of by-product mutualism is cooperation, in which individuals achieve a greater energetic gain by feeding together than they would alone. In cooperation, individuals adopt a role in the group, and in the most complex interactions there may be multiple roles, resulting in a division of labor that occasionally includes role specialization. Humpback whales (Megaptera novaeangliae) are one of the few baleen whale species that have been observed feeding in groups, utilizing behaviors that are hypothesized to be cooperative. One of these behaviors is group bubble-net feeding, which has been observed in the Northeastern Pacific, Northwestern Atlantic, and Southern Oceans. This study utilized multi-sensor archival tag data from 26 humpback whales from the southern Gulf of Maine, 4 from Southeast Alaska, and 1 from the Western Antarctic Peninsula to analyze individual bubble-net feeding behaviors and compare these across populations. Linear mixed effects models were used to determine if dive behaviors varied with group sizes to test the hypothesis that group size influences individual behavior. The results indicate that individuals in the southern Gulf of Maine, for which sufficient data were available, were consistent in their bubble-net feeding behaviors across group sizes, which suggests that individuals utilize set roles in group feeding events. There was evidence for a division of labor and role specialization among whales utilizing certain bubble-net feeding tactics in the southern Gulf of Maine. The three populations performed different variations of bubble-net feeding that are likely based on the speed and schooling patterns of the prey. The results are consistent with the hypothesis that bubble-net feeding is an example of by-product mutualism in these populations, though was not enough data to suggest that group bubble-net feeding in Southeast Alaska was a form of by-product mutualism. The prevalence of herding dives in feeding groups suggest that each individual takes on a role to herd the prey to the surface, and provide evidence against a producer-scrounger relationship in the southern Gulf of Maine, and potentially in the Western Antarctic Peninsula.

Efficient foraging strategies result in a predator spatially overlapping with its prey, foraging in the most profitable patches, and minimizing the time transiting between patches. Previous studies investigating baleen whale foraging strategies have generally focused on investigating spatial overlap with prey patches, patch profitability or movement within feeding grounds. The present study investigated the fine-scale strategies of movement between individual prey schools and larger prey patches in humpback whales (Megaptera novaeangliae) bubble-feeding on sand lance (Ammodytes spp.) in and around the Stellwagen Bank National Marine Sanctuary, USA. The goal was to investigate the presence of hierarchically nested spatial structures in both sand lance patches and whale bubble-feeding behavior, and to compare the scales and geometry of these patches between predator and prey behavior on each hierarchical scale. Furthermore, the profitability of sand lance schools in feeding areas was investigated. Using animal-borne tag technology to record underwater movement of whales in combination with surface observations of whale behavior, the locations of bubble-feeding events were identified. Concurrent hydroacoustic measurements of the prey distribution in the water column were used to identify the locations and energetic parameters of sand lance schools around tagged whales. First Passage Time analysis was used to determine the spatial scale of individual bubble-feeding events. Based on spatial proximity, feeding events and prey schools were grouped into larger feeding bouts and prey patches to investigate the presence of hierarchically nested scales. Up to three hierarchy levels were found in bubble-feeding behavior of nine whales tagged on six days between 2008 and 2012, and up to five hierarchy levels in the sand lance prey field around the tagged whales. There was a significant positive correlation between the lengths of whale bubble-feeding bouts and the lengths of sand lance patches over three common hierarchy levels. On each hierarchy level, the lengths of whale bubble-feeding bouts were significantly smaller than those of sand lance patches. Mean inter-feeding bout distances were significantly positively correlated with mean inter-prey patch distances over two hierarchy levels. Mean distances between feeding events were similar to the mean distances between prey schools. On larger hierarchy levels, mean inter-bout distances were greater than mean distances among prey patches. Mean school height and density tended to be greater in schools recorded inside than outside of feeding bout areas. The prey field structures found here were likely a result of the specific habitat requirements of sand lance. The results of this study suggest that the tagged whales were able to adapt their foraging movement to the structure of the prey field. By feeding on neighboring schools, whales could minimize the time spent between prey schools. On larger spatial scales, whales did not feed on neighboring prey patches. This could be a result of decreased abilities to find the nearest patch, or because, rather than restricting their foraging movement to neighboring patches, the whales were targeting specific patches. The foraging movement observed in this study led to spatial overlap of the tagged whales with sand lance schools that were characterized by properties rendering them more energetically profitable for bubble-feeding whales. While hierarchically structured foraging movement has been found in other marine predators, this is the first study that demonstrates this kind of foraging mechanism for baleen whales.