POPULATION STRUCTURE, RESIDENCY, AND INTER-ISLAND MOVEMENTS OF COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) OFF O‘AHU AND MAUI NUI

Item

Title

Eng POPULATION STRUCTURE, RESIDENCY, AND INTER-ISLAND MOVEMENTS OF COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS) OFF O‘AHU AND MAUI NUI

Date

Eng 2021

Creator

Eng Harnish, Annette

Identifier

Eng Thesis_MES_2021_Harnish

extracted text

POPULATION STRUCTURE, RESIDENCY, AND INTER-ISLAND MOVEMENTS OF
COMMON BOTTLENOSE DOLPHINS (TURSIOPS TRUNCATUS)
OFF O‘AHU AND MAUI NUI

by
Annette E. Harnish

A Thesis
Submitted in partial fulfillment
Of the requirements for the degree
Master of Environmental Studies
The Evergreen State College
June 2021

©2021 by Annette E. Harnish. All rights reserved.

This Thesis for the Master of Environmental Studies Degree
by
Annette E. Harnish

has been approved for
The Evergreen State College
by

_______________________________
John B. Kirkpatrick, Ph.D.
Member of Faculty

4 June 2021
_______________________________

Date

ABSTRACT
Population Structure, Residency, and Inter-Island Movements of
Common Bottlenose Dolphins (Tursiops truncatus) off O‘ahu and Maui Nui
Annette E. Harnish
Accurate descriptions of population structure are critical to inform effective management
of protected species. Here I present the results of a reassessment of the population structure and
residency of two common bottlenose dolphin (Tursiops truncatus) stocks from the main
Hawaiian Islands. Previous photo-identification and genetic studies have shown that bottlenose
dolphins in the main Hawaiian Islands live in four small (~100-200 individulas) demographically
independent and genetically differentiated island-associated populations designated as stocks
centered around Kaua‘i/Ni‘ihau, O‘ahu, Maui Nui (Maui, Lāna‘i, Kaho‘olawe, and Moloka‘i),
and Hawai‘i. However, photo-identification and satellite-tagging data has shown that some
individuals do occasionally move between island areas, especially between O‘ahu and Maui Nui.
These movements may have important consequences, as even a few dispersing individuals can
impact genetic diversity and allow for the transmission of culturally-mediated behaviors, both of
which could impact long-term population health..
I reassessed the population structures of the O‘ahu and Maui Nui stocks by analyzing
over two decades’ worth of photo-identification data representing 472 individuals, and satellitetag data from five individuals. While I found that social connections between the two
populations were minimal, there was geographic overlap in spatial use that crossed stock
boundaries. This was caused by a subset of individuals (n=14) from the O‘ahu population that
occasionally travel between island areas, using SW O‘ahu, SW Moloka‘i, and SW Lāna‘i.
Satellite-tag data from two suspected inter-island travelers reveals that these animals made
extensive use of Penguin Bank, indicating that this area may be of importance to inter-island
travelers. Inter-island travelers were sighted in both island areas at all times of the year, though
they were consistently sighted more frequently off O‘ahu than off Maui Nui. Further research
will be needed to identify the possible drivers of this behavior.

Table of Contents

List of Figures .............................................................................................................................. vii
List of Tables ................................................................................................................................ ix
Dedication ...................................................................................................................................... x
Acknowledgements ...................................................................................................................... xi
Chapter One: Literature Review: The Population Biology of Odontocetes and Global
Population Structures of Common Bottlenose Dolphins ........................................................... 1
Introduction ................................................................................................................................. 1
Population Biology of Odontocetes ............................................................................................ 1
Legislative Definitions and Applications of Population Structure for Managing Odontocete
Populations................................................................................................................................ 10
Common Bottlenose Dolphins – Globally Observed Population Structures ............................ 12
Conclusion ................................................................................................................................ 23
Chapter Two: Introduction: Hawaiian Resident Common Bottlenose Dolphins ................. 24
Ecotypes, Distribution, Population Structure, and Status ......................................................... 25
Anthropogenic Threats.............................................................................................................. 27
Conclusion ................................................................................................................................ 30
Chapter Three: Methods ............................................................................................................ 31
Primary Datasets ....................................................................................................................... 31

iv

Survey Effort, Quality Control, and Coverage ......................................................................... 35
Residency Assignments and Social Networks .......................................................................... 36
Subarea Stratification ................................................................................................................ 37
Inter-Island Movements ............................................................................................................ 40
Movements and Spatial Use...................................................................................................... 41
Interchange Indices ................................................................................................................... 42
Chapter Four: Results ................................................................................................................ 44
Survey Effort, Quality Control, and Coverage ......................................................................... 44
Residency Assignments and Social Networks .......................................................................... 55
Subarea-Stratification ............................................................................................................... 60
Inter-Island Movements ............................................................................................................ 68
Movements and Spatial Use...................................................................................................... 69
Interchange Indices ................................................................................................................... 82
Chapter Five: Discussion ............................................................................................................ 85
Introduction ............................................................................................................................... 85
Survey Coverage, Encounter Characteristics, and Depth Preferences ..................................... 87
Residency Assignments ............................................................................................................ 91
Inter-Island Movements ............................................................................................................ 94
Conclusion .............................................................................................................................. 100
References .................................................................................................................................. 102

v

Appendix .................................................................................................................................... 114

vi

List of Figures
Figure 1. Stock boundaries of the four resident island-associated stocks of common bottlenose
dolphins. ........................................................................................................................................ 26
Figure 2. Examples of potential anthropogenic stressors to Hawaiian insular bottlenose dolphins
that vary spatially, including fisheries interactions, tour boats, and military vessel activity. ...... 28
Figure 3. Examples of distinctiveness and photo-quality scores.. ............................................... 32
Figure 4. Example of a satellite-tag deployed on HITt0788 off Maui Nui in the limpet
configuration.. ............................................................................................................................... 34
Figure 5. Subarea locations (red text). ......................................................................................... 39
Figure 6. CRC effort tracklines (grey) off O‘ahu and Maui Nui from 2000-2018. ..................... 45
Figure 7. Barplots of encounters by year for each island area, including both CRC (black) and
non-CRC (grey) encounters. ......................................................................................................... 46
Figure 8. Encounter locations and subareas.. ............................................................................... 48
Figure 9. Histogram of bathymetric depths for CRC survey effort (top) and all encounters with
corresponding GPS locations, including CRC and non-CRC encounters (bottom).. ................... 49
Figure 10. Histograms of encounters by season for each island area.. ........................................ 52
Figure 11. Discovery curves of individuals by island area, restricted to distinctive or very
distinctive individuals with good or excellent quality photographs, with a reference 1:1 trendline
shown in red. ................................................................................................................................. 53
Figure 12. Social network with island areas of the individual nodes indicated by color, restricted
to distinctive or very distinctive individuals with good or excellent quality photographs. .......... 56

vii

Figure 13. Social network with revised residency assignments of the individual nodes indicated
by color, restricted to distinctive or very distinctive individuals with good or excellent quality
photographs. .................................................................................................................................. 58
Figure 14. Social network with subareas where individuals were encountered indicated by color
and shape, restricted to distinctive or very distinctive individuals with good or excellent quality
photographs.. ................................................................................................................................. 63
Figure 15. Map of subareas with the total number of animals identified within each subarea
following the subarea abbreviation in red, and the number of individuals identified within both
subareas adjacent to the red lines connecting subareas. ............................................................... 64
Figure 16. Discovery curves by subarea (indicated by color) for all encounters with
corresponding GPS location data, with a reference 1:1 trendline in black. .................................. 67
Figure 17. Histogram of mean inter-annual travel distances for all individuals (n=116) seen in
more than one year for all encounters where corresponding GPS locations were available. ....... 72
Figure 18. Douglas-filtered Argos tracklines of the five satellite-tagged inviduals.. .................. 74
Figure 19. Probability density representation of bottlenose dolphin home ranges by island area
where the animal was tagged, based on trimmed data from five satellite-tags deployed off O‘ahu
and Maui Nui. ............................................................................................................................... 77
Figure 20. Range of depths from photo-ID encounters for six Maui Nui long-term residents.. .. 79
Figure 21. Range of depths from satellite-tag data for five individuals. ..................................... 81

viii

List of Tables
Table 1. CRC and non-CRC survey effort and encounters .......................................................... 45
Table 2. Resighting rates with varying levels of restrictions on distinctiveness and photo quality
for all encounters........................................................................................................................... 50
Table 3. Revised residency assignment results by island area. .................................................... 57
Table 4. Total area (in km2) by depth (m) for each island area and subarea................................ 60
Table 5. Subareas versus revised residency classifications for all encounters where
corresponding GPS location data were available.......................................................................... 65
Table 6. Inter-island movements by season from photo-identification data ................................ 69
Table 7. Mean inter-annual travel distances for individuals seen in more than one year for all
encounters where corresponding GPS locations were available................................................... 71
Table 8. Summary of tag data. ..................................................................................................... 73
Table 9. Within-area resighting and interchange indices for individuals with designated subareas
based on GPS coordinates ............................................................................................................. 84

ix

Dedication

For my mother, Celeste, who first taught me the love of nature.

x

Acknowledgements
Words cannot adequately express the gratitude I have for my advisors, Dr. John
Kirkpatrick and Dr. Robin Baird. Your patient guidance through this project has brought it to
completion, and improved the quality of my work more times than I can count. Thank you both
for sharing your expertise, your wisdom, and your time with me. I treasure all of the lessons that
you have taught me, and I’ll carry them with me always.
Thank you also to Sabre Mahaffy, Michaela Kratofil, and Enrico Corsi for teaching me so
many of the skills that I applied throughout this process. I would be completely lost without you
guys! I also want to acknowledge the extensive work of Annie Gorgone, who started the
bottlenose dolphin catalog all the way back in 2003. Thank you as well to all of the countless
photo contributors and colleagues who have been a part of this work in a million ways, big and
small. Specifically, I would like to acknowledge the extensive photo contributions of the Pacific
Whale Foundation, Tori Cullins, Lynn Opritoiu, Paul Johnson, Chuck Babbitt, and Jim Ault.
Thank you to my friends and family for guiding me through this process. Whether it was
listening to a late night rant about histograms, bringing by a fresh chocolate chip cookie, or
encouraging me at my lowest points, I am eternally grateful. Dad - your love, support and
encouragement through this process has meant everything. Thank you for always believing in
me, and pushing me to pursue my dreams. My sisters Natalie and Karlyn – thank you for your
steadfast friendship over the years, and for always encouraging my passion for science! Last but
never least, Enrico – whales and dolphins have been a part of our story from the very beginning,
and I hope that never changes. Thanks for being my best friend and my partner in all things.
And finally, thank you to all of those who have worked over the course of generations to
conserve the beautiful animals that I am so fortunate to study today. I stand in the shadow of
your legacy, and I will do all that I can to honor it.
Cascadia Research Collective’s efforts to study Hawaiian common bottlenose dolphins
were funded by several groups, including the Pacific Islands Fisheries Science Center, the State
of Hawai‘i, the NOAA Bycatch Reduction Engineering Program, Dolphin Quest, the Marisla
Foundation, and Southwest Fisheries Science Center.

xi

Chapter One: Literature Review: The Population Biology of Odontocetes and
Global Population Structures of Common Bottlenose Dolphins
Introduction
As human populations and their consequent impact to the environment grow,
conservation efforts are becoming increasingly important to ensure the preservation of wildlife
species worldwide. Desirable conservation outcomes for wildlife are facilitated through effective
communication and feedback between scientists and managers, as well as broad scale support
from the public and policymakers. Scientists can identify and advocate for conservation priorities
and act as a source of knowledge for managers, who may then utilize scientists’ data and insights
to craft new policy solutions or management strategies whose outcomes can then be evaluated by
scientists as part of a continuous feedback loop. When considering conservation efforts at the
species-level, population information such as distribution, population structure, and abundance
are especially relevant, and these types of measures are assessed regularly for heavily managed
species, including several species of marine mammals. The literature review portion of this thesis
will explore and analyze how population structure is studied and evaluated among toothed
whales and dolphins (parvorder Odontoceti) and how this information is used by managers, with
particular emphasis on the common bottlenose dolphin (Tursiops truncatus).

Population Biology of Odontocetes
Toothed whales and dolphins include some of the most charismatic and highly intelligent
species on Earth. At the moment there are over 70 different identified species of odontocetes,
and continuing advances in genetics and taxonomy are likely to result in the classification of
additional species. Many species of odontocetes have wide distributions at a global scale, but

1

distribution is rarely continuous, and even when it is divisions within a species can quickly
become apparent with closer observation. These divisions create cohesive groups of individuals
within a species that are called populations, the exact definitions of which will be discussed at
length.

Factors that influence the division of populations
Several factors have been hypothesized to influence the division of odontocete species
into populations. Physical boundaries that separate populations in marine environments are not
as obvious as in terrestrial environments, but evidence indicates that land barriers, salinity, and
depth all limit dispersal ability of several marine species (Costello et al., 2017). While
odontocetes have the capacity for long-range movements, geographic boundaries do seem to
influence dispersal as has been shown for bottlenose dolphins in the Northeast Atlantic (Natoli et
al., 2005). Additionally, fine-scale population structure among odontocetes has been repeatedly
documented worldwide, proving that the capacity to disperse does not necessarily equate to
actual dispersion (e.g., Mirimin et al., 2011; Kiszka et al., 2012; Mirimin et al., 2011; Möller et
al., 2007; Urian et al., 2009).
Habitat preferences are suspected to limit dispersal between odontocete populations in
some circumstances. For example, populations of inshore and offshore ecotypes of common
bottlenose dolphins off Northwest Ireland have distributions that appear to be restricted by
distance from shore, with the inshore ecotype only sighted within the 3 km closest to the
shoreline, and the offshore ecotype only sighted at distances greater than 4 km from the shoreline
(Oudejans et al., 2015). Inshore vs. offshore habitat preferences among common bottlenose
dolphins are also reflected in dietary studies from United States East Coast, where the stomach

2

contents of stranded inshore animals were found to contain near-shore species of fish, while the
stomach contents of offshore animals had deep-water and pelagic species of fish and squid
(Mead & Potter, 1995). Habitat preferences between different bottlenose dolphin ecotypes have
been proposed as a driver of ecological specialization at an evolutionary scale, which will
continue to increase their behavioral, morphological, and genetic divergence (Louis et al., 2014a,
2014b).
Habitat preferences have also been identified within populations of the same species and
ecotype. For example, Indo-Pacific bottlenose dolphin (Tursiops aduncus) populations along the
southeast coast of Australia show significant genetic differentiation between coastal and bay
habitats, and a high degree of site fidelity in spite of their very close proximity (Möller et al.,
2007). Additionally, a genetic study of several U.S. East Coast bottlenose dolphin populations by
Richards et al. (2013) found significant genetic differentiation between inshore populations
occupying different habitat types along the coastline (Richards et al., 2013). The most startling
example of fine-scale habitat preference the authors identified in this study was in the Indian
River Lagoon of Florida, where one population of dolphins has a distribution restricted to the
northernmost regions of the lagoon even while a second, partially sympatric population is seen
throughout the entire range of the lagoon (Richards et al., 2013). The authors of both the
southeast Australia and the U.S. East Coast studies hypothesized that the genetic differences
between populations were the direct result of their specific habitat preferences (Möller et al.,
2007; Richards et al., 2013).
Behavior, in conjunction with habitat preference, may also play a role in the development
of population structures. Many odontocete populations have complex and specific social
structures and behavioral patterns that have come be viewed as cultures, and cannot be explained

3

by genetics alone (Rendell & Whitehead, 2001). Some aspects of culture are highly specialized
to particular habitats, and may affect the willingness of populations to disperse to new habitats.
For example, the Indo-Pacific bottlenose dolphins of Shark Bay exhibit habitat preferences that
coincide well with different foraging strategies and genetic haplotypes (Kopps et al., 2014). In
this case, social transmission of foraging strategies (usually from mother to calf) and a high
degree of site fidelity has increased habitat specialization among different groups (Kopps et al.,
2014). This in turn has increased philopatry, reduced mixing and led to genetic differentiation
between groups (Kopps et al., 2014). Additionally, complex social structures may limit dispersal
between populations. For example, Rosel et al. (2009) suggested that social relationships
between common bottlenose dolphins on the U.S. Atlantic coast might limit their dispersal to
new populations due to the relative cost of building new relationships.

Population definitions
In spite of its significance for both biological science and management, the precise
definition of a population has been hotly debated within science for decades. Two major
paradigms have emerged within this debate: the ecological paradigm, and the evolutionary
paradigm (Waples & Gaggiotti, 2006). The ecological paradigm states that a population is
defined by shared space and time and the opportunity to interact (demographic cohesion), while
in contrast the evolutionary paradigm states that populations are defined by proximity and the
opportunity to mate (reproductive cohesion) (Waples & Gaggiotti, 2006). The ecological
paradigm is especially relevant for exploring population structure among intelligent species with
a high degree of sociality, as has been frequently documented in odontocetes. Among these
species, population cohesion is largely determined by behavioral patterns, rather than proximity

4

or mating opportunity. A prime example of this is the existence of different sympatric ecotypes
of killer whales (Orcinus orca) in the nearshore waters of the eastern North Pacific. While both
fish-eating and mammal-eating ecotypes of killer whales frequently utilize the same territory,
their group sizes and foraging tactics differ in accordance with their prey (Baird et al., 1992).
Mating between ecotypes does not occur as a hypothesized consequence of these cultural
differences, and it has been proposed that the cultural differences between ecotypes will
eventually lead to their speciation (Baird et al., 1992; Riesch et al., 2012). In spite of their close
proximity and the opportunity to mate, these two ecotypes of killer whale clearly represent
different populations, and require different management strategies as such.
However, neither paradigm is ever “correct” to the exclusion of the other, and both
perspectives are useful to managers. The evolutionary paradigm’s emphasis on reproduction as a
cohesive factor makes it well-suited to examining questions about long-term evolution and
natural selection, and favors genetic approaches (Waples & Gaggiotti, 2006). In contrast, the
ecological paradigm is more adapted to exploring short-term conservation-oriented questions,
and can incorporate a larger range of methods, including studies of demographic connectivity
(Waples & Gaggiotti, 2006). Under the ecological paradigm, a population is best defined through
demographic independence (Waples & Gaggiotti, 2006; Martien et al., 2019). This occurs when
population trends such as abundance are more strongly influenced by internal birth and death
rates, rather than by emigration and immigration of individuals into/out of other populations.
Such populations are referred to as demographically independent populations (DIPs).
Quantitative measures for evaluating population structure also differ between paradigms.
Waples and Gaggiotti (2006) suggested that the most appropriate measure to define population
boundaries in the ecological paradigm is the migration rate (m), also known as a dispersal rate,

5

which is the percentage of individuals within a population that will disperse from their natal
population into a new population. The precise cutoff where demographic independence occurs is
highly dependent upon the species in question, but one estimate is that the cutoff occurs
somewhere around m = 10% (Hastings, 1993). In contrast, rather than describing a single
specific cutoff for demographic independence, Lowe and Allendorf (2010) argue that the type of
connectivity being evaluated must be taken into account when determining a cutoff point. For
example, while a low migration rate might be sufficient to theoretically prove ecological
connectivity, it does not necessarily demonstrate actual connectivity to a degree that is relevant
for conservation managers, unless the context in which migrants interact with the population is
further explored (Lowe & Allendorf, 2010).
One important caveat of migration rates is that they mean little in a genetic context; even
migration rates below 1% can reduce genetic differentiation within populations to the degree
where they are hard to distinguish using genetic methods and conversely, high migration rates
may not impact genetic differentiation at all if migrants fail to interbreed with a population
(Martien et al., 2019; Waples & Gaggiotti, 2006). For example, killer whales live in stable
matrilineal pods with only very rare dispersal of individuals between social groups, but
population-level genetic analyses indicate an appreciable degree of relatedness within the
population as a whole (Pilot et al., 2010). In this case, low dispersal between groups does not
result in low gene flow rates between groups as predicted, demonstrating that migration rates do
not necessarily equate to equivalent rates of gene flow. For this precise reason, Waples and
Gaggiotti (2006) suggested that analysis of genetic datasets should be redirected towards more
suitable approaches under the evolutionary paradigm, such as the number of migrants per
generation (Nm). Nm is typically determined using genetic data, and can be used in two ways to

6

determine if populations are demographically independent. First, Nm can be used to reject the
assumption of panmixia, in which all individuals in the populations are interbreeding, or it can be
proven that Nm falls below an arbitrarily set value for demographic independence (Waples &
Gaggiotti, 2006). The value of the different approaches has been contested though. Palsbøll et al.
(2007) caution that relying on a rejection of panmixia increases the risk of incorrectly
designating management units, especially when statistical power is low, while Lowe and
Allendorf (2010) again argue that any cutoff point for genetic connectivity must be considered in
light of the type of connectivity being assessed.
Metapopulation theory adds another layer of complexity to finding an operational
definition for a population. In metapopulation theory distinct localized populations of a species
may remain connected through periodic dispersal events, forming a larger metapopulation.
Classically, metapopulations are comprised only of isolated populations separated by unsuitable
habitat, without any large populations to act as a reservoir of immigrants (Levins, 1969).
Metapopulation theory is compatible with both the ecological and evolutionary paradigms, and
can have significant biological and conservation implications for the persistence of populations
in that it allows for even struggling localized populations to persist because of influxes of new
animals from surrounding populations (Hanski, 1991). However, metapopulation theory has little
relevance for delineating demographically independent populations for management purposes.
Quantitative methods for studying the dynamics of metapopulations usually involve assessing
both dispersal (i.e., m or Nm, as previously described) and connectivity (the ease of movement
between localized populations). Calabrese and Fagan (2004) separated measures of connectivity
into three flavors: 1) structural connectivity, which describes the physical spaces between
populations, 2) potential connectivity, which predicts connections between populations by

7

integrating dispersal information with spatial descriptions, and 3) actual connectivity, which is
derived directly from observation. Measures of structural connectivity can incorporate
information on the distance between suitable habitat spaces, the type of habitat, and the area of
different types of habitat, while potential connectivity measures usually incorporate knowledge
of behavioral responses to habitat type and dispersal ability, and actual connectivity measures
directly incorporate observational data of movements between populations (Calabrese & Fagan,
2004; Kindlmann & Burel, 2008).

Datasets that are useful in evaluating population structure
Several different types of data can be used to study population structure in odontocetes,
and have been broadly reviewed in Martien et al. (2019). The most reliable conclusions are
drawn from morphometric, genetic, and movement datasets, while moderately reliable
conclusions can be drawn from distributional data, contaminant ratios, habitat preferences, and
association data (Martien et al., 2019). However, these different lines of evidence are relevant at
different time scales and levels of divergence, and inherent tradeoffs also exist in data collection.
Morphometric and genetic datasets can be used to draw some of the most reliable longterm conclusions about population structure. Morphometric differences only emerge after a long
period of isolation between populations, while genetic differences can appear after a much
shorter period of isolation. For example, morphometric differences have been regularly used to
differentiate between stranded individuals of inshore and offshore ecotypes of common
bottlenose dolphins, which differ in their body size, bone structure, and hematology (Duffield et
al., 1983; Mead & Potter, 1995). However, morphometric differences between inshore

8

populations of common bottlenose dolphins are often subtle at best, and genetic methods are
more commonly applied to differentiate between these populations (e.g., Natoli et al., 2004).
Movement and distributional data can be used to draw reliable and moderately reliable
conclusions about population structure, but both may vary temporally, which presents a
challenge in interpreting these types of data. Movement data may vary both over time and
according to demographic variables. For example, adult male common bottlenose dolphins and
subadults of both sexes in Sarasota Bay, Florida, have been shown to explore larger areas of their
home ranges than the adult females within the same population (Wells et al., 1987). Additionally,
even if movement data reveals possible spatial overlaps between populations, the significance of
those overlaps cannot be interpreted without additional data to provide context (Martien et al.,
2019). For example, Baird et al. (2016) reports that a common bottlenose dolphin tagged off
Kaua‘i in October 2014 dispersed to O‘ahu nine days after being tagged, but a lack of further
resightings of the tagged animal makes it unclear whether this movement between islands
represents a permanent or temporary dispersal. Distributional hiatuses, when larger than home
ranges, can be especially helpful in defining stock boundaries, though careful consideration must
also be applied to where distributions change seasonally or over time (Martien et al., 2019). For
example, the U.S. East Coast migratory stocks of common bottlenose dolphins are most clearly
defined by their distributions during the summer months, when they have the smallest degree of
overlap with other stocks, whereas in the winter months a high degree of overlap between stocks
makes defining them by distribution very difficult (Hayes et al., 2018).
Habitat preferences and contaminant ratios can also be used to draw moderately reliable
conclusions about population structure. The use of habitat preferences as a line of evidence in
defining populations is informed by the idea that different populations will specialize in using

9

separate habitats, and contaminant ratios in individual animals may reflect this specialization as
well, provided that the habitat is contaminated. For example, common bottlenose dolphins in the
Indian River Lagoon in Florida are divided into two genetic clusters, with one restricted to the
northern portion of the lagoon while the other, partially sympatric cluster ranges primarily in the
southern portion of the lagoon, reflecting differences in habitat preference between these two
clusters (Richards et al., 2013). A 2015 study of blood samples from both clusters also showed
that these habitat preferences may be driving differences in contaminant ratios, as mercury
concentrations exist along a gradient that decreases from north to south (Schaefer et al., 2015).
Habitat preferences and contaminant ratios may also vary with demographic factors, and visiting
transient animals may further complicate attempts to define populations with these datasets
(Martien et al., 2019).
Association data have been repeatedly used to demonstrate the demographic
independence of populations (e.g., Baird et al., 2009; Mahaffy et al., 2015), but is not always
practical because of the need for long-term and intensive efforts to identify individual members
of a population (Martien et al., 2019). Additionally, association data is useful only for species
with a high degree of sociality, which may limit its applicability for certain species. Regardless
of the approach taken to explore population structure, the products of these efforts, and their
consequences all come down to the interpretation of results by policy makers.

Legislative Definitions and Applications of Population Structure for Managing Odontocete
Populations
Legislative definitions of populations may not necessarily equate to biological
distinctions, and may leave little room for nuanced explanations or the incorporation of scientific

10

uncertainty. However, these definitions form the basis for management decisions regarding
different species and have significant consequences for conservation efforts.
Within the U.S., odontocetes are primarily managed under the Marine Mammal
Protection Act (MMPA) of 1972, which utilizes the best available scientific information about
population structure to designate management units called stocks. The definition of a stock under
the MMPA has gradually evolved. The 1994 amendments to the MMPA first defined stocks as
“a group of marine mammals of the same species or smaller taxa in a common spatial
arrangement, that interbreed when mature”, a definition that falls under the evolutionary
paradigm because of its emphasis on interbreeding. However, in 2005 NMFS refined this
definition by stating that stocks should be DIPs (NMFS, 2005). This refinement shifted the stock
concept towards the ecological paradigm, and emphasizes the importance of dispersal over
genetic differentiation (Martien et al., 2019).
The National Marine Fisheries Service (NMFS) is responsible for designating odontocete
stocks based on the best available scientific evidence. Once designated, a new stock assessment
report is supposed to be produced every three years that incorporates the best available scientific
information to ensure that the abundance of designated stocks remains at the optimum
sustainable population level for ecological sustainability, though some vulnerable stocks are
assessed more frequently. However, it is also important to acknowledge that politics, both
internal and external to NMFS, can also play a role in determining the frequency with which
stocks are reassessed. The language of the MMPA makes clear that the conservation of stocks to
preserve ecosystem functionality is a priority for managers, but stock designation initially
favored large stock designations often corresponding with geopolitical boundaries, rather than
taking a precautionary approach that favored small stock designations (Martien et al., 2019). This

11

was problematic for achieving conservation goals, as the comparatively large allowable losses in
improperly designated large stocks may be devastating for small populations that are
disproportionately impacted. As a result, revisions have since broken many large stocks into
smaller stocks as scientific evidence shows that divisions are warranted (Martien et al., 2019).
This highlights the importance of accurate scientific information to guide management.
Population structure information is also utilized in the Endangered Species Act (ESA) of
1973, which is applied in addition to the MMPA for some odontocete species. In contrast to the
MMPA concept of stocks, the ESA utilizes management units called Distinct Population
Segments (DPS). This unit may not necessarily equate to a stock, and in practice DPSs are
frequently larger units because of different genetic divergence thresholds. One of the reasons for
this is that the MMPA prioritizes ecological function over genetic differentiation, while the ESA
prioritizes preserving genetic diversity to conserve small populations (Martien et al., 2019).
Regardless of the technical definition for management units however, both acts are dependent on
accurate assessments of population structure in order to be effective in achieving conservation
goals.

Common Bottlenose Dolphins – Globally Observed Population Structures
Common bottlenose dolphins are one of the most easily recognizable and well-studied
cetacean species in the world. They are globally distributed throughout tropical and temperate
waters, and divided into both inshore and offshore ecotypes that are occasionally sympatric. Site
fidelity is variable in this species, and various populations have been described as either
migratory or resident (e.g., Hayes et al., 2018). Inshore resident populations of common
bottlenose dolphins in particular have been identified as an ideal sentinel species for monitoring

12

changes in ecosystem health because of their long lifespans, high trophic level, and tendency to
store environmental pollutants in their blubber (Bossart, 2011; Wells et al., 2004).

Indo-Pacific bottlenose dolphins are the only other long-recognized species within the
same genus as common bottlenose dolphins, though their distribution is currently thought to be
limited to the Indian and Western Pacific oceans. While there is some evidence of hybridization
between the two species, genetic studies have shown that Indo-Pacific bottlenose dolphins are
more closely related to species within the Stenella genus than to common bottlenose dolphins,
indicating that perhaps this species has been misplaced taxonomically (Martien et al., 2011;
Möller et al., 2008). While many characteristics are shared between the two species, this review
will be limited to common bottlenose dolphins (hereafter just “bottlenose dolphins”) from this
point.
A defining characteristic of bottlenose dolphins is their well-developed cognitive
intelligence, which has allowed bottlenose dolphins to develop novel strategies for foraging that
are well-suited to their specific environments. For example, bottlenose dolphins in the Florida
Keys practice mud-plume feeding in shallow areas, where dolphins intentionally stir fine
sediments into a suspended plume, then lunge through the plume to capture fish inside (Lewis &
Schroeder, 2003). Further north in Florida, the bottlenose at Cedar Key practice fish herding,
where groups of dolphins work together to drive fish towards one another (Gazda et al., 2005).
Additionally, bottlenose dolphins in U.S. East Coast salt marshes have been observed stranding
themselves to drive fish onto mud banks for easier capture (Hoese, 1971; Rigley et al., 1981).
Each of these foraging strategies are uniquely suited to the habitat where they are practiced, and
constitute a form of ecological niche specialization within populations that may limit dispersal
and reinforce population structure (Hoelzel, 2009).

13

A recurring trend among bottlenose dolphin populations is the mixing of several different
residency categories within one geographical region, which can create confusion in delineating
populations. Bottlenose dolphins live within fission-fusion societies, where group sizes and their
individual compositions regularly fluctuate, though some long-term relationships between
individuals are maintained (Connor & Wells, 2000). An interesting consequence of this type of
social structure is the regular mixing of resident, transient, and migrant animals within the same
geographic area, which has been documented in several populations of bottlenose dolphins (e.g.,
Dinis et al., 2016; Estrade & Dulau, 2020; Silva et al., 2008; Speakman et al., 2010). While these
overlaps may increase the opportunities for gene flow, they do complicate the process of
delineating populations because they increase the chances of incorporating individuals that are
part of a separate population that is not actually resident to the area. To counter this, Martien et
al. (2019) recommends that associations be interpreted in the context of additional information.
Estrade and Dulau (2020) accomplished this by assigning residency classes to the individuals
that they studied, and by evaluating differences in observed movement patterns.
As previously mentioned, behavioral traits in conjunction with habitat preferences can be
a driver of population structure. The combined global distribution and widespread ecological
niche specialization of bottlenose dolphins has helped to establish an impressive array of
population structures within this species, several of which are reviewed below. These
classifications are broadly defined, but they do not necessarily capture the entire spectrum of
possible population structures.

14

Offshore Populations
Offshore bottlenose dolphins have been the subject of comparatively few studies, mostly
because of inherent difficulties in conducting deep-water surveys, and limited information is
available about their population structures and behaviors (Klatsky et al., 2007). Offshore
bottlenose populations have a high degree of genetic diversity based on samples taken from
stranded animals, which may be indicative of large, undifferentiated populations (Costa et al.,
2015; Hoelzel et al., 1998; Oudejans et al., 2015). There is also evidence that offshore bottlenose
may regularly travel long distances and show minimal site fidelity, covering up to 89 km/day in
extensive journeys that have been documented to reach 4,200 km (Wells et al., 1999). However,
only limited conclusions can be drawn from these two lines of evidence. The stranded animals
sampled in genetic studies may represent members of multiple offshore populations that have
drifted to strand at the same location (Hoelzel et al., 1998). Additionally, studies of movement
data for offshore bottlenose have had contradictory results: Wells et al. (1999) tagged two
animals that moved 89 km/day and 48 km/day, while Klatsky et al. (2007) tagged three dolphins
that travelled much smaller distances, averaging only 28 km/day. Klatsky et al. (2007) suggested
that their comparatively shorter average daily travel distances were the result of their tagged
animals lingering at specific sites during the time frame where movements were recorded,
possibly to take advantage of available prey. Additionally, differences in their movements may
have been driven by habitat differences in the areas where Wells and Klatsky tagged or released
dolphins; Klatsky deployed tags off the Bermuda pedestal, while Wells tagged dolphins that had
been previously stranded in Florida (Klatsky et al., 2007; Wells et al., 1999). Further studies of
offshore bottlenose dolphins will be needed to elucidate accurate information about their
population structures.

15

Coastal and Migratory Coastal Populations
Some inshore populations of bottlenose dolphins live in populations characterized by
long-distance movements and regular mixing. The best documented examples of this type of
population structure are found on the U.S. West and East Coasts. On the U.S. West Coast, the
Coastal California Stock of bottlenose dolphins has an extensive range that stretches from San
Francisco into northern Baja, Mexico (Carretta et al., 2019; Defran et al., 2015). Photo
identification studies have shown that individuals in this stock regularly travel long distances
along the shoreline, sometimes moving up to 95 km/day (Defran & Weller, 1999; Hwang et al.,
2014). In spite of its large range this stock remains genetically distinct from offshore
populations, and has more limited genetic diversity than offshore animals, indicating its
cohesiveness as a population (Lowther-Thieleking et al., 2014). Defran et al. (2015) suggested
that the long-distance movements within this population may be a consequence of limited prey
predictability, as animals are forced to travel farther in pursuit of foraging opportunities.
On the U.S. East Coast, two migratory stocks of bottlenose dolphins have been delineated
that perform seasonal migrations along the coastline (Hayes et al., 2018). The Western North
Atlantic Northern Migratory Coastal Stock spends the summer months between Assateague,
Virginia and Long Island, New York, then moves south between Cape Lookout, North Carolina,
and the North Carolina/Virginia border for the winter months (Garrison et al., 2017). The
Western North Atlantic Southern Migratory Coastal Stock, on the other hand, spends its summer
months to the north of Cape Lookout, North Carolina and Assateague, Virginia, then moves
south between Cape Lookout and northern Florida for the winter months (Garrison et al., 2017).
Between the two stocks, differences also exist in their abundance, and the number and type of
overlapping commercial fisheries, warranting their separate management (Hayes et al., 2018).

16

However, knowledge of the population structure of these two stocks is still evolving. Initially, all
Atlantic coast bottlenose dolphins were treated as members of a single depleted stock, but this
notion that was later disproven with genetic methods and stock designations were consequently
reassigned by NMFS (Rosel et al., 2009, 2011). Even with genetic methods the ease of
delineating these two migratory stocks changes seasonally due to spatiotemporal overlap though,
and understanding of the demographics and movement patterns of the Southern Migratory stock
remains especially limited (Hayes et al., 2018). Toth et al. (2012) has suggested that additional
divisions may even exist in the comparatively well-understood Northern Migratory Stock, based
on sighting locations, behavior, morphology, and the occurrence of a commensal barnacle.

Transient Archipelago-Associated Populations
Some archipelago-associated populations of inshore ecotype bottlenose dolphins also live
in transient populations with regular movement between islands and little genetic differentiation.
One example of this comes from the Canary Islands, Spain, where a photo-identification study
identified a small population of bottlenose dolphins (Tobeña et al., 2014). Over the course of ten
years, 313 individual dolphins were identified in the islands, 36 (10%) of which moved between
islands at least once, travelling distances between 30-130 km (Tobeña et al., 2014). Another
example comes from the Azores, Portugal, where out of 966 bottlenose dolphins identified in a
photo-identification study, 66 (7%) individuals moved between island areas (Silva et al., 2008).
While the average distance between sightings was only 25 km in the Azores study, movements
of up to 291 km were repeatedly detected (Silva et al., 2008). This suggests that the population is
not divided between island areas, as the distance between island groups ranges from 160-230 km,
well within the range of movement displayed by these animals (Silva et al., 2008). A third

17

example of this type of population structure comes from the Madeira Archipelago, Spain, where
out of 501 documented individuals, only 15 (4%) were resighted in more than one year, and
regular inter-island movements of approximately 50 km were likely based on calculated
movement probabilities (Dinis et al., 2016). Additionally, no genetic differentiation has been
detected within the Madeira Archipelago, indicating a lack of island-associated population
structuring (Quérouil et al., 2007). Interestingly, different portions of the Madeira archipelago
had varying probabilities of immigration and emigration, especially areas with shallow-water
habitats, where prey is likely more readily available (Dinis et al., 2016). This raises the question
of why bottlenose dolphins move between islands so readily.
Reasons suggested for inter-island movements include seeking out foraging or mating
opportunities (Hooker & Gerber, 2004). The waters surrounding both the Canary Islands and the
Azores are oligotrophic, which may contribute to limited prey availability that forces animals to
travel farther distances to forage (Baird et al., 2009; Silva et al., 2008; Tobeña et al., 2014). In
contrast, while in the Madeira archipelago there are shallow-water areas with assumed high prey
availability, the high rate of immigration and emigration in these areas suggests that other
factors, such as anthropogenic impacts or exposure to predators are preventing the establishment
of a resident population in this region (Dinis et al., 2016). Limited genetic differentiation within
the Azores population and within the North Atlantic region as a whole also suggests that
movement between islands increases mating opportunities, though this is more likely a
consequence rather than a driver of movements (Quérouil et al., 2007; Silva et al., 2008).

18

Resident Coastal Populations
Resident coastal populations of bottlenose dolphins have been repeatedly described in the
literature worldwide. These populations are characterized by a high degree of site fidelity and
genetic differentiation from neighboring populations, and frequently live within easily
distinguishable habitats like estuaries or lagoons. Fourteen distinct resident coastal populations
have been recognized along the U.S. East Coast alone, 11 of which are localized to estuarine
systems or embayments (Hayes et al., 2018). Globally, additional resident populations of
bottlenose have also been described in western Ireland (Mirimin et al., 2011; Nykänen et al.,
2018), Greece (Bearzi et al., 2008), the Black Sea (Gladilina et al., 2018), New Zealand
(Tezanos-Pinto et al., 2009), the Caribbean (Caballero et al., 2012), and southern Brazil (Costa et
al., 2015; Fruet et al., 2014), just to name a few.
Some of the most well-described populations of resident bottlenose dolphins are located
in the Gulf of Mexico, where 32 bay, sound, and estuarine stocks are currently recognized by
NMFS (Rosel et al., 2011). In particular, a 50 year examination of the Sarasota Bay, Florida
resident bottlenose has yielded incredibly detailed information that is fairly representative of
resident bottlenose communities described elsewhere. This population is small, with
approximately 160 individuals, and has been thoroughly documented over 40 years of study
(Wells, 2009). Both female and male Sarasota dolphins show site fidelity to the bay, though
males occasionally disappear for short periods, and have been seen in association with members
of neighboring populations (Wells et al., 1987; Wells & Scott, 1990). Immigration and
emigration are rare in spite of this however, and the dynamics of the population are thus
dependent on births and deaths (Wells & Scott, 1990). This degree of insularity has led to
significant genetic differentiation from other neighboring populations, and may be maintained

19

through specialized foraging behaviors (Sellas et al., 2005). Most of the population traits
described for the Sarasota Bay animals (small population size, high site fidelity, genetic
distinctiveness, and habitat-specific behaviors) are also represented in other documented coastal
resident populations, though variation does occur.

Resident Island-Associated Populations
Other than the main Hawaiian Islands stocks (reviewed in chapter two), resident islandassociated populations of bottlenose dolphins have only been described a few times. One of these
populations is found off the Little Bahama Bank in the Bahamas (Parsons et al., 2006). In
contrast to other island-associated populations discussed in this review however, Little Bahama
Bank is a single sand-bank system, meaning that the two major islands that comprise the bank
are not separated by deep-water channels (Parsons et al., 2006). This may increase the ability of
animals to move between areas in the islands because there is no break in suitable habitat
between areas (Parsons et al., 2006). However, photo-identification surveys of the population
revealed that dolphins showed site fidelity to surveyed sites on East and South Abaco, with only
infrequent movements between areas (Parsons et al., 2006). Additionally, genetic analyses
revealed that gene flow is limited between East Abaco, South Abaco, and a third study site at
White Sand Ridge, all of which are less than 200km apart (Parsons et al., 2006).
Recently, a resident island-associated population of bottlenose dolphins off Reunion
Island was described for the first time using photo-identification data collected over six years
(Estrade & Dulau, 2020). Dolphins were present at the island year-round, though only 1/3 of
animals were classified as residents, and immigration and emigration occurred regularly (Estrade
& Dulau, 2020). In spite of the documented presence of resident animals, there is still a sizable

20

proportion of individuals that are transient that associate with resident animals (Estrade & Dulau,
2020). An island-associated population of bottlenose dolphins similar to the Reunion population
has been described off São Tomé, in the Gulf of Guinea along the western coast of Africa
(Pereira et al., 2013). A photo-identification study of cetaceans off this island has revealed that
the bottlenose dolphin population there is very small, with only an estimated 37 individuals, and
regular resightings of roughly only 1/3 of the total number of identified animals (Pereira et al.,
2013). While the dataset used in the São Tomé study was limited in its scope, the regular
resightings of a few individuals do suggest some degree of residency (Pereira et al., 2013).
However, the high proportion of non-resident animals sighted off both Reunion and São Tomé
indicates that perhaps these populations as a whole should not be considered as truly resident to
the island, but may fall somewhere between the classifications of resident and transient island- or
archipelago-associated populations given in this review.
Additional resident island-associated populations of bottlenose dolphins may be
described in the future, but the present rarity of this particular type of population structure
indicates that the ecological conditions which favor it are generally uncommon.

Metapopulation Structures
Metapopulations are clasically defined as a series of connected but isolated habitat
patches surrounded by unsuitable habitat and no unlimited large population to provide a constant
influx of immigrants (Levins, 1969). This is compatible with many described resident bottlenose
dolphin populations, which show site fidelity to areas that are separated from one another by
habitat either unsuitable for foraging or where predation risk is increased. Additionally, most
resident bottlenose dolphin populations have no nearby large stock constantly supplying new

21

immigrants, though there is frequently some small degree of movement between populations.
This suggests, therefore, that metapopulation theory may be useful in evaluating the population
structures of this species, though to date it has only rarely been applied. Two metapopulation
structures of bottlenose dolphins have been identified by Gladilina et al. (2018): the wellconnected migratory Coastal California Stock (described earlier), and the slightly less wellconnected northwestern and central Mediterranean basin populations. The Mediterranean basin
populations fall somewhere between the previously explored classifications of migratory and
resident populations, and are characterized by regular long-distance movements of individuals
between stocks that generally show site-fidelity, resulting in genetic differentiation between the
western and eastern portions of the basin that becomes less apparent at a closer scale (Bearzi et
al., 1997; Bearzi et al., 2010; Gaspari et al., 2015; Gnone et al., 2011; Natoli et al., 2005).
Interestingly, Carnabuci et al. (2016) described the northwestern Mediterranean basin
populations as a metapopulation, but elsewhere the Coastal California Stock and Mediterranean
basin populations are generally not described as metapopulations in the literature.
Metapopulation structure has also been hinted at to describe the loosely connected resident
populations of Western Ireland and the United Kingdom, though this suggestion has drawn
criticism (Ingram & Rogan, 2003; Louis et al., 2014a; Nichols et al., 2007; Nykänen et al.,
2018). In general, metapopulation theory has not been regularly applied to examine population
structure and dynamics among bottlenose dolphins. This might be occurring because of
ambiguity over how to define metapopulations of bottlenose dolphins, the lack of relevance that
metapopulation theory has in delineating DIPs for management purposes, or because of
geographic limitations in study designs that restrict scientists to single localized populations and
prevent identification of connections between populations.

22

Conclusion
Common bottlenose dolphins exist in a variety of population structures worldwide that
are influenced by geography, oceanographic conditions, habitat preferences, behavior, and social
structure. While the precise definition of a population remains subject to debate, the most
relevant definition for this highly charismatic species is the DIP, which is theoretically intended
to be analogous to the MMPA’s definition of stocks. Accurate designation of stocks, performed
by NMFS, is of paramount importance to inform effective management, and is informed by the
best available scientific knowledge. This may include conclusions drawn from any number of
datasets, including morphometric, genetic, movement, and distributional data, as well as
contaminant ratios and habitat preferences (Martien et al., 2019).

23

Chapter Two: Introduction: Hawaiian Resident Common Bottlenose Dolphins

The Hawaiian Islands are one of the most isolated archipelagos in the world, located in
the middle of the Pacific Ocean and surrounded by a vast expanse of oligotrophic tropical waters.
Despite their isolation however, the islands and waters surrounding them have been repeatedly
colonized by wildlife of all varieties, and as a result are rich in biodiversity and home to many
endemic species (Eldredge & Evenhuis, 2003). One of the hypothesized factors driving the
marine diversity of the region is the comparatively high rate of productivity of the waters
surrounding the islands, known as the “island mass effect” (Doty & Oguri, 1956; Gove et al.,
2016). This effect is driven primarily by upwelling, but is also influenced by reef ecosystem
processes and natural and anthropogenic nutrient runoff (Gove et al., 2016).
One of the remarkable aspects of Hawai‘i’s biodiversity is the presence of 11 species of
odontocetes with resident populations that remain in the islands year-round, including dwarf
sperm whales (Kogia sima), Blainville’s beaked whales (Mesoplodon densirostris), Cuvier’s
beaked whales (Ziphius cavirostris), false killer whales (Pseudorca crassidens), pygmy killer
whales (Feresa attenuata), melon-headed whales (Peponocephala electra), short-finned pilot
whales (Globicephala macrorhynchus), rough-toothed dolphins (Steno bredanensis), pantropical
spotted dolphins (Stenella attenuata), spinner dolphins (Stenella longirostris), and bottlenose
dolphins (Baird et al., 2015). These populations were likely established from offshore
populations that “discovered” the islands and remained to take advantage of the increased
availability of prey that they found surrounding the islands compared to in offshore habitats
(Baird, 2016). The resident bottlenose dolphin populations, for example, show evidence of
genetic relatedness to offshore animals sampled in the area (Martien et al., 2011).

24

Ecotypes, Distribution, Population Structure, and Status
Both inshore and offshore ecotypes of bottlenose dolphins can be found in Hawaiian
waters, and are currently divided into one offshore pelagic stock and four inshore island area
associated stocks, centered around Kaua‘i/Ni‘ihau, O‘ahu, Maui Nui (Maui, Lāna‘i, Kaho‘olawe,
and Moloka‘i), and Hawai‘i (Figure 1; Carretta et al., 2019). Boundaries for the four islandassociated stocks are currently drawn at the 1,000 m depth contour, with the exception of the
O‘ahu and Maui Nui stocks, which are separated at the Ka‘iwi Channel between the islands at
approximately 500 m depth (Carretta et al., 2019). A 2009 abundance estimate based on photoidentification data placed the abundance of each of the four island-associated stocks in the low
hundreds, with the exception of O‘ahu, which was placed around 700 (Baird et al., 2009).
However, the relatively high abundance estimate for the O‘ahu stock in the 2009 estimate was
likely a consequence of data limitations, and a more recent abundance estimate has placed the
O‘ahu stock’s abundance in the low hundreds as well (Baird et al., 2009; Van Cise et al., 2021).

25

Figure 1. Stock boundaries of the four resident island-associated stocks of common bottlenose
dolphins. Stock boundaries are drawn at the 1,000 m bathymetric depth contour, with the
exception of the O‘ahu and Maui Nui stock boundary, which is drawn at approximately the 500
m bathymetric depth contour.

Hawaiian inshore bottlenose dolphins are distributed primarily throughout shallow
waters, with over 95% of sightings taking place at depths shallower than 1,000 m, and over 50%
of sightings taking place at depths shallower than 500 m (Baird et al., 2013; Baird, 2016).
Sightings have occurred year-round at each island area, with high resighting rates indicative of
resident populations (Baird et al., 2009; Baird, 2016). Additionally, genetic analyses have
revealed that the island-associated populations are genetically differentiated from one another
(Martien et al., 2011). Both satellite-tagging efforts and photo-identification have shown that
animals rarely leave their island areas, though occasional inter-island movements have been
documented through both methodologies, especially between Maui Nui and O‘ahu, which are the

26

two island areas that are closest in proximity (Baird, 2016). The significance of these inter-island
movements remains unknown at present, but one possible consequence is that it increases the
exposure of some individuals to spatially variable anthropogenic threats.
There is evidence that three of the four resident stocks are in decline, with significant
declines in the Maui Nui stock, and non-signficant declines in the O‘ahu and Kaua‘i/Ni‘ihau
stocks (Van Cise et al., 2021). Model-derived apparent survival rates are also lower than
expected in all four stocks, with the lowest apparent survival rate in the O‘ahu stock (0.84, se =
0.023; Van Cise et al., 2021). The factors driving these declines are uncertain at present, but
exposure to anthropogenic stressors likely plays a role.

Anthropogenic Threats
Today, the State of Hawai‘i, comprising the eight major islands at the southeastern end of
the archipelago, is home to over 1.4 million people, and boasts a booming economy largely
rooted in natural resources and tourism (State of Hawai‘i DBEDT, 2019). Dense human
inhabitation is known to have negative consequences for wildlife populations however, and
bottlenose dolphins in Hawai‘i are exposed to a variety of anthropogenic stressors, many of
which vary by island area (Figure 2).

27

Figure 2. Examples of potential anthropogenic stressors to Hawaiian insular bottlenose dolphins
that vary spatially, including fisheries interactions, tour boats, and military vessel activity. Photo
credits clockwise from top left: Deron Verbeck, Alicia Ward, Jessica Aschettino/CRC, and Tori
Cullins.
One example of a spatially-variable stressor is ambient noise from anthropogenic
activities. Shipping activity is most frequent in the Hawaiian Islands off the island of O‘ahu,
where Honolulu functions as a major port city, and there are considerable volumes of small boat
traffic off Maui Nui and Hawai‘i. Large commercial ships are especially significant noise
generators that can produce up to 195 dB of ambient noise, though even small boats can produce
up to 160 dB (Hildebrand, 2009). Increased ambient noise may interfere with the ability of
animals to communicate with one another by masking their vocalizations, as has been
documented in other species (e.g., Van Parijs & Corkeron, 2001). Increased boat traffic has also
been shown to induce changes in breathing among bottlenose dolphins, possibly as part of an
anti-predator response to boat noise (Hastie et al., 2003). Additional noise may come from
military activities, which are especially prevalent around Kaua‘i and Ni‘ihau. Military midfrequency sonar can produce sound levels on the order of 235 dB, and is regularly used on the

28

Pacific Missile Range Facility off Kaua‘i, where resident bottlenose dolphins may be repeatedly
exposed to sonar (Baird et al., 2014; Hildebrand, 2009). Exposure response studies in bottlenose
dolphins have shown that exposure to high levels of sound may induce behavioral changes,
which could have significant consequences for survival (Houser et al., 2013).
Bottlenose dolphins are also impacted by fisheries activity off the main Hawaiian Islands.
Their shallow distribution places them into regular contact with nearshore fisheries, and there is
evidence that animals are at least occasionally hooked or entangled, likely when trying to steal
bait or catch (Baird, 2016; Sims, 2013). Additionally, fishermen may sometimes attempt to
retaliate against animals that try to depredate lines, with potentially lethal consequences (Baird,
2016; Harnish et al., 2019). Another source of potential fisheries interactions is an offshore fish
farm on Hawai‘i Island’s Kona coastline, which has drawn repeated, long-term associations by
bottlenose dolphins (Harnish et al., 2021; Sims, 2013). Elsewhere, regular associations with fish
farms have caused changes in social structures and association patterns (Díaz López & Shirai,
2008). Similarly, there is evidence that the Kona farm may also be inducing behavioral changes,
including increased inter-species aggression and smaller group sizes (Harnish et al., 2021).
Additionally, long-term associations with the Kona farm hint at potential dependency, which
could have a devastating impact in the case that the farm is ever removed (Harnish et al., 2021).
Overall, the spatial variation in anthropogenic impacts to the resident bottlenose dolphins
emphasizes the need for accurate stock delineation to inform management efforts for these
populations, which is especially critical now given the evidence that some of these stocks may be
in decline.

29

Conclusion
One of the rarest types of population structure documented for bottlnose dolphins are
resident island-associated populations, four of which have been documented off the main
Hawaiian Islands and are designated as separate stocks for management. The long-term
residency of these stocks makes them vulnerable to island-specific anthropogenic threats, and a
recent abundance estimate indicates that three of these stocks may be in decline. Appropriate
management strategies must take into account any possible connections between stocks that
could allow for increased gene flow or the spread of socially transmitted behaviors. Long-term
photo-identification and satellite-tagging efforts off the islands of O‘ahu and Maui Nui have
shown that occasionally individual animals do move between areas, meaning that these
populations may not be demographically independent, though the significance of these interisland movements remains unknown at present (Cascadia Research Collective, unpublished
data).
The aim of this thesis will be to provide an updated assessment of the residency patterns
and possible connections between the O‘ahu and Maui Nui stocks. This information will help to
inform accurate management of these populations, which is especially critical now, given the
recent evidence of declines and lower-than-expected survival rates in both stocks.

30

Chapter Three: Methods

Primary Datasets

Photo-identification
Cascadia Research Collective (CRC) encounters with bottlenose dolphins took place as
part of a larger effort to study multiple species of Odontocetes through small boat surveys from
2000-2018. Survey effort varied both between island areas and over time, and was both
nonrandom and nonsystematic, though efforts were made to cover the greatest area and variety
of habitats possible, given limitations imposed by weather conditions. While surveying, GPS
locations of the survey vessel were recorded every five minutes. Whenever the survey vessel
detected a group of odontocetes, the group was approached until it was in range for species
identification, and the location, behavior, and estimated group size was recorded. Depending on
the species, photographs were taken of as many individuals as possible within the group,
including non-distinctive animals. Additional photographs of bottlenose dolphins were taken by
other researchers and community members from Hawaiʻi from 1996-2018, and contributed to
CRC. An encounter is considered an observation of a group of animals, and encounters ended
either when the group was lost, or the survey vessel left the group. A single day of survey effort
could include multiple encounters with the same group of animals. The circumstances of
encounters contributed by other researchers and community members varied widely, and
locations of encounters were not always recorded, though information about the island area
where encounters took place was always available.
All photographs of bottlenose dolphins were sorted within encounters by individual,
using the unique pattern of natural marking on each animal (notches on dorsal fin, pigmentation,

31

scarring, etc.). Then, following the protocols in Baird et al., (2009), individuals were compared
against the long-term CRC catalog of bottlenose dolphins to establish sighting histories. Each
individual was assigned a distinctiveness score between 1 and 4 (1 = not distinctive, including
unmarked fins; 2 = slightly distinctive, with 1-2 notches; 3 = distinctive, with 3-5 notches; 4 =
very distinctive, with 5 or more notches), and a best photo quality score between 1 and 4 (1 =
poor, 2 = fair, 3 = good, 4 = excellent) for every encounter (Figure 3). All unique individuals
were named using the following system: HITt####, where HI stands for Hawaiʻi, Tt for Tursiops
truncatus, and #### is a unique four digit number.

Figure 3. Examples of distinctiveness and photo-quality scores. Clockwise from top left:
distinctiveness 4, photo-quality 3, photo by Paul Johnson; distinctiveness 3, photo-quality 2,
photo by Tori Cullins; distinctiveness 1, photo-quality 4, photo by Tori Cullins; distinctiveness 2,
photo-quality 1, photo by Lynn Opritoiu.

32

Satellite-Tags
Satellite-tags were deployed on bottlenose dolphins by CRC staff on one occasion off
Oʻahu, and on four occasions off Maui Nui, in accordance with protocols described elsewhere
(Schorr et al., 2009). Tags were manufactured by Wildlife Computers, and included one MK10A tag (deployed in 2012 off Maui Nui, yielding location and dive data), one SPOT5 tag
(deployed in 2012 off Maui Nui, yielding location data only), and three SPOT6 tags (deployed
once in 2016 off Oʻahu and twice in 2017 off Maui Nui, yielding location data only). Briefly,
tags were deployed onto slow-moving adults via air rifle in the LIMPET configuration, using
two 4.7 cm titanium darts with backwards-facing petals that attach to the dorsal fin (Figure 4).
Tags were preprogrammed to maximize the number of transmissions during predicted Argos
satellite overpasses, extending battery life. Following deployment, tags transmitted location
information to Argos satellites when the tagged animal surfaced, following the preprogrammed
schedule. Tags were designed to remain on the animals and transmit for only a short time frame
(usually less than a month), and naturally worked their way out of the dorsal fin afterwards.
Photos of tagged individuals were compared against the catalog to determine sighting histories,
providing context for which to later interpret the tagging results.

33

Figure 4. Example of a satellite-tag deployed on HITt0788 off Maui Nui in the LIMPET
configuration, December 2012. Photo by CRC/Annie Douglas.

Satellite-tag processing was performed by CRC staff following established internal
protocols (e.g., Baird et al., 2021). Briefly, location data were processed by Argos with the
Kalman smoothing algorithm in the Wildlife Computers portal (Lopez et al., 2015), then with a
Douglas-Argos filter through Movebank (Kranstauber et al., 2011) to clean the location data of
any apparent errors based on a realistic measure of speed, with higher-quality locations (Argos
location-quality 2 or 3) exempt from filtering (Douglas et al., 2012). The user-defined settings of
the Douglas-Argos filter set the maximum sustainable rate of movement (MINRATE) at 20 km
per hour, the maximum distance between consecutive locations (MAXREDUN) at 3 km, and the
tolerance for turning (RATECOEF) at 25.

34

Biopsy Samples for Sex Determination
Skin biopsy samples were collected from 29 adult bottlenose dolphins on 32 occasions
between 2000-2010, 19 (representing 17 individuals) off Maui Nui, and 13 (representing 13
individuals) off Oʻahu. Briefly, a stainless steel dart (8 mm diameter and ~18 mm long) with
backwards-facing barbs, mounted onto an arrow with a float, was fired from a crossbow onto the
body of a passing animal when it surfaced, removing a small amount of skin and blubber. The
arrow and sample was retrieved, and frozen prior to subsampling and laboratory analysis.
Genetic analysis of the samples for sex determination was undertaken by Southwest Fisheries
Science Center.

Survey Effort, Quality Control, and Coverage
Survey effort by CRC and contributed encounters were evaluated to assess whether the
data represents a comprehensive sample of the O‘ahu and Maui Nui bottlenose dolphin
populations. Aspects of coverage that were examined include the spatial distribution of CRC and
non-CRC effort, the distribution of encounters across years, the distribution of encounters across
seasons, and the distribution of encounters across depths when GPS coordinates were available.
All depths were determined using the R package marmap v. 1.0.5 (Pante & Simon-Bouhet, 2013)
in conjunction with imported NOAA bathymetric data at a 1-minute resolution. Additionally,
discovery curves comparing the number of unique individuals against the number of total
identifications over time were constructed to evaluate the extent to which populations have been
comprehensively sampled.
To ensure that the data was robust and included no misidentifications or duplicate
identifications, and in accordance with previously established protocols (e.g., Baird et al., 2009),

35

the dataset was restricted to include only individuals with photo quality scores of 3 or 4, and
distinctiveness scores of 3 or 4.
The season in which encounters took place were defined by the month, with March-May
for Spring, June-August for Summer, September-November for Fall, and December-February for
Winter.

Demographics
Group size was considered only for CRC encounters where estimates of group size were
recorded in the field, in order to minimize any bias in photographic effort (i.e., partial coverage
of groups in contributed photographs). Group sizes for Oʻahu and Maui Nui encounters were
first tested for normality with the Shapiro-Wilk test, and then group sizes were compared using a
Mann-Whitney U test.
Sex was determined based on recorded calf presence, morphology (e.g., clear view of the
genital slit or penis), or on genetic analysis of biopsy samples undertaken at the Southwest
Fisheries Science Center. Sex distribution by island area was tested to examine whether it
differed significantly from random using Pearson’s Χ2 test.

Residency Assignments and Social Networks
Residency was initially assigned to each ID based on the island(s) and span of years that
each was encountered across. Individuals with a sighting history span greater than three years on
a single island were classified as long-term residents. Individuals with a sighting history span
greater than one year, but less than three years were classified as short-term residents, and

36

individuals with a sighting history spanning less than one year were classified as visitors.
Individuals that were seen off both Oʻahu and Maui Nui were classified as inter-island.
Residency assignments were then reassessed based on their social associations. A social
network containing all individuals with photo quality scores of 3 or 4 and distinctiveness scores
of 3 or 4 was built using a half-weight association index in SOCPROG 2.4 (Whitehead, 2008),
and visualized using Netdraw 2.158 (Borgatti, 2002) with spring embedding. Any visitors that
connected to the main components of the network were reassigned as associative residents for
the component that they most closely linked to (i.e., Oʻahu or Maui Nui), to account for the fact
that these individuals may be resident animals that were infrequently sampled, rather than true
visitors. Residency assignments for visitors that did not connect to the main components were
not revised.

Subarea Stratification
To explore how spatial use impacts residency, encounters with GPS locations were
divided into five different geographic subareas based on demographic and geographic separation,
three of which are located within the Oʻahu island area, and two of which are located in the Maui
Nui island area (Figure 5). These subareas are: O‘ahu North (ON), O‘ahu West (OW), O‘ahu
East (OE), Moloka‘i/Penguin Bank (MPB), and Maui Nui (MN, representing Maui, Lāna‘i , and
Kaho‘olawe). These subareas align with the subareas designated in Van Cise et al. (2021),
though the O‘ahu East subarea does represent a new addition. The O‘ahu North subarea
encompasses the northwest coast of O‘ahu, northward of Ka‘ena Point (~21.6° N, 158.3° W),
and West of Kahuku Point (~21.7° N, 158.0° W). The O‘ahu East subarea encompasses the
northeast coast of O‘ahu, eastward of Kahuku Point and southward to Makapu‘u Point (~21.3°

37

N, 157.7° W). The O‘ahu West subarea encompasses both the west and south coasts of O‘ahu,
south of Ka‘ena Point along the Wai‘anae coast, and southwest of Makapu‘u Point along the
south coast. The Moloka‘i/Penguin Bank subarea includes the waters surrounding Penguin Bank
(a large shallow water area to the southwest of the the island of Moloka‘i), and the waters
surrounding the island of Moloka‘i itself, extending to midway between Moloka‘i and Lāna‘i and
between Moloka‘i and Maui. The Maui Nui subarea is geographically the largest subarea,
encompassing the waters surrounding the islands of Maui, Lāna‘i, and Kaho‘olawe. Subareas
were further divided based on depth at the 500 m bathymetric contour in subareas where
deepwater encounters (i.e., encounters in water deeper than 500 m) took place, resulting in the
creation of three additional subareas: O‘ahu West Deep (OWD), O‘ahu East Deep (OED), and
Maui Nui Deep (MND).

38

Figure 5. Subarea locations (red text). Abbreviations are as follows: ON = O‘ahu North, OW =
O‘ahu West, OE = O‘ahu East, MPB = Moloka‘i/Penguin Bank, MN = Maui Nui (subarea).
Subarea divisions based on depth (O‘ahu West Deep, O‘ahu East Deep, and Maui Nui Deep) are
not shown. Depth contours (grey dashed lines) are shown for 500 m and 1,000 m bathymetric
depth.

To explore how characteristics of subareas might impact bottlenose dolphin distribution,
subareas were characterized by measuring the area within different depth ranges, and by
describing seasonal fluxes in chlorophyll-a concentrations. These particular variables were
chosen because Pittman et al. (2016) demonstrated that both are important predictors of
bottlenose dolphin distribution in the Hawaiian Islands, especially depth. The area within various
depth ranges was measured in 100 m bins using the R package marmap v. 1.0.5 and imported
NOAA bathymetric data at a 1-minute resolution. Chlorophyll-a concentrations were evaluated
using data from the National Aeronautics and Space Administration’s Aqua satellite, which

39

measures surface reflectance of blue and green light to indirectly gauge chlorophyll-a
concentrations (NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology
Processing Group, 2018). This data was mapped by season for three randomly selected years
spanning the study period with Oceanic Niño indices between -1.0 and 1.0 (2006, 2013 and
2018), using the R package oceanmap v. 0.1.1 (Bauer, 2020).
The subareas where individuals were encountered were mapped on the social network to
explore how spatial stratification impacts social relationships, and the distribution of residency
classifications by subarea was described. A rough estimate of “density” was calculated for each
subarea by dividing the total number of individuals encountered in a subarea by the total area.
However, as this measure does not account for variations in the presence or absence of
individuals over time or the amount of effort expanded in each subarea, it cannot be considered a
true measure of density. Discovery curves were also constructed for each subarea to examine
sampling coverage, and demographics were explored by comparing group sizes where CRC
encounters were available. Group sizes were first tested by subarea for normality using a
Shapiro-Wilk test, then tested for significant differences between subareas using a KruskalWallis ranked sums test in conjunction with Dunn’s test.

Inter-Island Movements
Inter-island movements were described at the individual level and summarized
seasonally. An inter-island movement was considered to be any movement that crossed the stock
boundary between island areas (Figure 1), either from Maui Nui to O‘ahu, or from O‘ahu to
Maui Nui. Inter-island individuals were also identified on both the residency class and subarea
social networks to explore social connections.

40

Movements and Spatial Use

Photo-Identification Data
To explore possible variation in movement patterns, mean inter-annual travel distances
were calculated for all individuals identified at any point within each island area, subarea and
residency class, following the methods from Van Cise et al. (2021). This value reflects the mean
distance between resightings from each year that an animal has been documented as a means of
assessing overall site fidelity. Briefly, one location was randomly selected using the sample_n()
function in R for each individual per year, and the distance between locations for each individual
was calculated using the Geosphere package (Hijmans, 2019). Then the mean was calculated for
all individuals sighted within the particular island area, subarea or residency class being
examined. Because this method samples individuals sighted over multiple years, all visitors and
associative residents were excluded from this analysis, as well as the O‘ahu East and O‘ahu East
Deep subareas.
Additionally, mean inter-annual travel distances were calculated for a simulated
randomly mixed population for both island areas as well as each island area separately, using the
encounter locations from the actual dataset while ignoring resightings, as in Van Cise et al.
(2021). Locations were randomly selected from the data for the same number of actual individual
sightings in each year, and the distances between locations were calculated using the Geosphere
package (Hijmans, 2019).

Satellite-Tag Data
Individual tracklines of Douglas-filtered datasets were first mapped to evaluate overall
spatial use. Previously conducted CRC distance analyses have shown that none of the animals

41

moved in concert while tags were deployed, so pseudoreplication of location data was not a
concern (CRC, unpublished data). To explore broad trends in spatial use, core home ranges were
constructed using probability-density distribution analysis, following the methods from Baird et
al. (2021). The first 24 hours of data from each tag deployment were excluded from the data to
reduce any bias in spatial use stemming from where the animals were tagged, and the data was
arbitrarily trimmed to include only every fourth record to reduce the effects of spatial
autocorrelation. Then kernel densities were constructed for individuals tagged off each island
area with AdehabitatHR (Calenge, 2006), and used to estimate 50, 95, and 99 percent home
ranges.

Individual Habitat Preferences
To explore whether there were individual habitat preferences, the distribution of depths
where each individual was encountered were determined for a small sample (n=6) of the
individuals with the greatest number of resightings with GPS coordinates. The distribution of
depths was also determined for all five satellite-tagged individuals, using the depths of all
Douglas-filtered locations between 0 and 1,000 m bathymetric depth. Significant differences in
depth preferences between individuals were tested among both the six individuals with the
greatest number of resightings and the five satellite-tagged individuals with a Kruskal-Wallis
ranked sums test, in conjunction with Dunn’s test.

Interchange Indices
Dispersal rates were calculated as the interchange index both between areas and subareas,
based on the methods of Urbán et al., 2000. Interchange indices increase when populations are

42

small, as well as when there is a high degree of movement between areas, and decrease when
populations or large or there is minimal movement between areas. First, to contextualize the
interchange indices, within-area resighting indices were calculated as:
RWA = (NR / (NA2)) x 1,000
where RWA is the within-area or within-subarea resighting index, NR is the number of animals
resighted over multiple years within the area or subarea, and NA is the total number of animals
seen in an area. Interchange indices were then calculated as:
RAB = (NAB / (NA x NB)) x 1,000
where RAB is the interchange index for areas A and B, NAB is the number of animals sighted in
both areas, NA is the total number of animals sighting in area A, and NB is the total number or
animals sighted in area B. Interchange indices that fall within the same magnitude as within-area
resighting rates for the corresponding areas indicate that movements between areas are just as
likely as movements within-areas, and are especially significant.

43

Chapter Four: Results
This study presents the results of a reassessment of the population structure and residency
of bottlenose dolphins off O‘ahu and Maui Nui, based on over two decades’ worth of photoidentification data, as well as satellite-tagging efforts. Several approaches were used to analyze
these datasets, including assigning residency classes based on sighting histories, social networks,
and spatial analysis.

Survey Effort, Quality Control, and Coverage

CRC Survey Effort
From 2000-2018, CRC conducted 588 hours and 9,626 km of survey effort off O‘ahu,
and 969 hours and 14.021 km of survey effort off Maui Nui (Table 1; Figure 6). Survey effort
was heavily biased towards the leeward sides of the islands, owing to restrictions caused by
weather conditions on the windward side, but covered a wide range of depths and potential
habitats. During this time, bottlenose dolphins were encountered on 18 occasions off O‘ahu and
on 65 occasions off Maui Nui. When restricted to the number of photographed encounters with at
least one individual of photo quality score ≥ 3 and distinctiveness score ≥ 3, these numbers drop
to 14 encounters off O‘ahu, and 63 encounters off Maui Nui (Table 1).
The number of CRC encounters with bottlenose dolphin varied across years on both
islands in accordance with research goals over time (Figure 7). The most CRC encounters took
place off Maui Nui in 2001, with 17 encounters, and off O‘ahu in 2003, with seven encounters.
CRC encounters occurred most frequently off both islands in the early 2000s.

44

Table 1. CRC and non-CRC survey effort and encounters. Two CRC encounters from the Maui
Nui island area lack corresponding GPS coordinates, but GPS coordinates are available for all
other CRC encounters. PQ = photo quality score, Dist = distinctiveness score.
Island Area

Total km of
CRC
Survey
Effort (on
effort)

Total Hours
of CRC
Survey
Effort

Total # of
Non-CRC
Encounters
With at
Least 1
Individual
PQ ≥ 3, Dist
≥3
234

Total # of
Non-CRC
Encounters
With
Corresponding
GPS Locations

588

Total # of
CRC
Bottlenose
Encounters
With at Least
1 Individual
PQ ≥ 3, Dist ≥
3
14

O‘ahu

9,626

Maui Nui

14,021

969

63

294

199

19

Figure 6. CRC effort tracklines (grey) off O‘ahu and Maui Nui from 2000-2018. Depth contours
(black dashed lines) are shown for the 500 m and 1,000 m bathymetric depth.

45

Figure 7. Barplots of encounters by year for each island area, including both CRC (black) and
non-CRC (grey) encounters. Top: O‘ahu. Bottom: Maui Nui.

46

Contributed Encounters
From 1996-2018, other researchers and community scientists have contributed photos
from over 700 encounters with bottlenose dolphins to CRC. When these are restricted to the
number of encounters with at least one individual with a photo quality score ≥ 3 and a
distinctiveness score ≥ 3, 234 encounters from O‘ahu and 294 encounters from Maui Nui were
retained (Table 1). Contributions off Maui Nui are in large part from the Pacific Whale
Foundation, while off O‘ahu they were largely from tour boat operators. Contributions off O‘ahu
especially have gradually increased over time (Figure 7b).

GPS Location Availability and Depth
GPS locations were available for all but two CRC encounters off Maui Nui (97% of all
CRC encounters), but were not available for all contributed encounters. Overall, of the total 528
contributed encounters, 218 (41%) had corresponding GPS locations. Of the 234 contributed
O‘ahu encounters, 19 (8%) had associated GPS location data, while of the 294 contributed Maui
Nui encounters, 199 (68%) had associated GPS location data (Table 1; Figure 8). Among the
encounters with associated GPS location data, depths for encounters off Maui Nui ranged from 1
m to 1,629 m, and off O‘ahu from 3 m to 872 m. However, encounters were heavily skewed in
favor of shallower encounters, with the vast majority (96%) of encounters taking place at depths
< 500 m (Figure 9). While the distribution of encounter depths generally resembled the
distribution of survey effort depth, encounters were more strongly skewed in favor of shallow
water (Figure 9).

47

Figure 8. Encounter locations and subareas. Encounter locations are represented by black dots,
with encounters where inter-island individuals were present shown as light blue diamonds.
Subareas are abbreviated in red, with ON for O‘ahu North, OW for O‘ahu West, OE for O‘ahu
East, MPB for Moloka‘i/Penguin Bank, and MN for Maui Nui (subarea). Subarea divisions
based on depth (O‘ahu West Deep, O‘ahu East Deep, and Maui Nui Deep) are not shown. Depth
contours (grey dashed lines) are shown for the 500 m and 1,000 m bathymetric depth.

48

Figure 9. Histogram of bathymetric depths for CRC survey effort (top) and all encounters with
corresponding GPS locations, including CRC and non-CRC encounters (bottom). One encounter
from the south side of Kaho‘olawe that took place in 1,629 m bathymetric depth has been
excluded from the bottom panel.

49

Photo-Identification Results and Quality Control
After encounters were processed by CRC staff, 3,278 total identifications representing
775 individuals from 748 encounters were part of the original dataset. However, resighting rates
varied substantially between different levels of distinctiveness and photo quality, with rates
improving as distinctiveness and photo quality increased (Table 2). Quality control measures
were therefore applied to restrict the dataset to only those individuals with photo quality scores
of 3 or 4, and distinctiveness scores of 3 or 4. This reduced the overall size of the dataset to
1,830 total identifications (56% of the original 3,278 total identifications) from 605 encounters
(81% of the original 748 encounters), representing 472 individuals (61% of the original 775
individuals).

Table 2. Resighting rates with varying levels of restrictions on distinctiveness and photo quality
for all encounters. Percentages indicate the proportion of the total number of individuals
identified within the same distinctiveness and photo-quality score range, rounded to the nearest
integer.
No Photo Quality Restrictions
Highest
#
# (%)
# (%)
Distinctiveness Individuals Seen
With Over
More
1 Year
Than
Resighting
Once
Span
110
37
15 (14%)
1 (Not
(34%)
Distinctive)
129
48
25 (19%)
2 (Slightly
(37%)
Distinctive)
136
102 (39%)
3 (Distinctive) 259
(53%)
277
151
128 (46%)
4 (Very
(55%)
Distinctive)
775
372
270 (35%)
Total
(48%)

Photo Quality ≥ 3 Only
#
# (%)
Individuals Seen
More
Than
Once
86
36
(42%)
110
42
(38%)
220
132
(60%)
252
144
(57%)
668
354
(53%)

# (%)
With Over
1 Year
Resighting
Span
15 (17%)
25 (23%)
102 (46%)
125 (50%)
267 (40%)

50

Coverage
Between CRC surveys and contributed photos, encounters off both O‘ahu and Maui Nui
have comprehensive coverage across years, though the overall number of encounters does
increase over time (Figure 7). The greatest number of encounters within a single year took place
off both O‘ahu and Maui Nui in 2018, with 57 encounters off Maui Nui, and 59 encounters off
O‘ahu. The mean number of encounters per year was 16 off Maui Nui (sd = ± 13.8, min = 0, max
= 57) , and 11 off O‘ahu (sd = ± 16.4, min = 0, max = 59).
The number of encounters varied seasonally off both O‘ahu and Maui Nui, with the
highest number occurring off Maui Nui during the Spring, and the lowest number occurring off
Maui Nui during the Fall (Figure 10). For O‘ahu the greatest number of encounters took place in
the summer, while the lowest number took place in the winter (Figure 10). For Maui Nui, the
number of encounters by season was Spring = 151 (42%), Summer = 69 (19%), Fall = 34 (10%),
Winter = 103 (29%), while for O‘ahu Spring = 76 (31%), Summer = 79 (32%), Fall = 46 (19%),
and Winter = 47 (19%).
A discovery curve displaying the number of unique individuals encountered over time
shows that coverage off Maui Nui is approaching an asymptote, indicating comprehensive
sampling (Figure 11). In contrast, the discovery curve for O‘ahu has mostly inflected, but
continues to rise and is higher overall, indicating a continued influx of new IDs in spite of fair
sampling effort (Figure 11).

51

Figure 10. Histograms of encounters by season for each island area, with seasons defined by the
month in which encounters took place (Spring = March-May, Summer = June-August, Fall =
September-November, Winter = December-February). Top: O‘ahu. Bottom: Maui Nui.

52

Figure 11. Discovery curves of individuals by island area, restricted to distinctive or very
distinctive individuals with good or excellent quality photographs, with a reference 1:1 dashed
trendline shown in red. Curves are constructed chronologically, with black dots showing the start
of each year. Top: O‘ahu (2002-2018). Bottom: Maui Nui (1996-2018).
53

Demographics
CRC group sizes of O‘ahu encounters tended to be larger than Maui Nui encounters. The
mean group size for O‘ahu encounters was 12.8 (sd = ± 10.4, min = 1, max = 40), and the mean
group size for Maui Nui encounters was 6.1 (sd = ± 4.4, min = 1, max = 18). Based on ShapiroWilk tests, the group sizes for O‘ahu were normally distributed (p = 0.061), but the group sizes
for Maui Nui encounters were not normally distributed (p < 0.001). Due to the mixed results
regarding normality, a Mann-Whitney U test was used to test whether the differences in group
sizes were statistically significant. The results of the Mann-Whitney showed that group sizes
were significantly different between island areas (p = 0.008).
Sex was determined for a total of 49 individuals, 25 of which were seen off O‘ahu, 21 of
which were seen off Maui Nui, and three of which were seen off both island areas (Appendix
Table A). In all three groups sexes skewed heavily towards females, with 20 females identified
off O‘ahu (80% out of the 25 sexed individuals total), 15 females identified off Maui Nui (71%
out of the 21 sexed individuals total), and three females identified off both island areas (100%
out of the three sexed individuals total). Calf presence and morphology contributed heavily
toward the high ratio of females vs. males for all three groups, but even when restricting to
individuals sexed through genetic sampling (n = 24), sexed individuals still included more
females than males for both O‘ahu and Maui Nui.
To examine whether the distribution of sexes differed significantly from random, a
Pearson’s Χ2 test was performed on the data. This test excluded the individuals seen off both
islands because of the small sample size, the fact that no males were documented off both
islands, and because these individuals could not be incorporated into the totals for both island

54

areas without counting them twice. The X2 statistic was 0.11015, with an associated p-value of
0.74, indicating that the variation in distribution of sexes did not differ from random.

Residency Assignments and Social Networks

Initial Residency Assignments
Island areas where individuals were encountered were mapped onto a social network of
all 472 individuals included in the study to evaluate initial impressions of connectedness between
island areas (Figure 12). The entire network includes 6,356 ties linking the 472 individual nodes,
and includes two easily identifiable main components that represent animals encountered largely
off O‘ahu, and animals encountered off Maui Nui. When restricted to the main components,
there are 5,966 ties connecting 380 nodes. Only two ties (representing < 0.1% of all ties in the
network) link the main O‘ahu and Maui Nui components. There are 17 peripheral clusters
including more than one connected individual, and 16 individuals that are unconnected to any
other individuals on the network. Of the 16 individuals unconnected to any other animals in the
network, eight were encountered by themselves, and the remaining eight were artifacts of quality
control measures (i.e., they linked to the main cluster when no restrictions were applied).
Initial residency assignments are fully reported in Appendix Table B for reference.
Briefly, of the 271 individuals encountered solely in the O‘ahu area, 59 are long-term residents,
22 are short-term residents, and 190 are visitors. Of the animals encountered solely in the Maui
Nui area, 66 are long-term residents, 26 are short-term residents, and 95 are visitors. Only 14
individuals were encountered off of both O‘ahu and Maui Nui, and therefore assigned as interisland individuals. Initial residency assignments were mapped onto the same social network
diagram as the island areas, revealing that most Maui Nui long and short-term residents cluster

55

together, and that most O‘ahu long and short-term residents also cluster together (Appendix
Figure A).

Figure 12. Social network with island areas of the individual nodes indicated by color, restricted
to distinctive or very distinctive individuals with good or excellent quality photographs. All
individuals with no included associations with other animals are shown in the upper left corner.
All tagged animals (n=4) are indicated by a square node shape. Red nodes are animals
encountered only off O‘ahu, blue nodes are animals encountered only off Maui Nui, and yellow
nodes are animals encountered in both island areas.

56

Revised Residency Assignments
Revised residency assignments are summarized in Table 3. Briefly, 152 of the 190 O‘ahu
visitors were reassigned as O‘ahu associative residents on the basis of their connection to the
main O‘ahu component of the network, along with the 16 of the 95 Maui Nui visitors that linked
most closely to the O‘ahu cluster through connections with inter-island individuals. These results
were also mapped onto the same social network diagram as the initial residency assignments
(Figure 13). All Maui Nui long-term residents were connected to one another in the social
network, along with all but two Maui Nui short-term residents. O‘ahu long-term residents were
not all connected to one another – they divided into three groups with 55 individuals, three
individuals, and one individual. These groups remained separate even with the addition of O‘ahu
short-term residents, which also resulted in the creation of another group with one individual.
When O‘ahu visitors were added to the social network, the two larger groups became connected
through a single individual, HITt1703 (seen twice off O‘ahu, both times in 2018).

Table 3. Revised residency assignment results by island area. Percentages indicate the
proportion of the total number of unique identified individuals from all island areas combined,
rounded to the nearest integer.
Island Area Total # (%) of
Individuals
O‘ahu
Maui Naui
InterIsland
All Island
Areas

# (%) ShortTerm
Residents
22 (5%)
26 (6%)
-

# (%)
Associative
Residents
168 (36%)
29 (6%)
-

# (%)
Visitors

287 (61%)
171 (36%)
14 (3%)

# (%) LongTerm
Residents
59 (13%)
66 (14%)
-

472

125 (26%)

48 (10%)

196 (42%)

88 (19%)

38 (8%)
50 (11%)
-

57

Figure 13. Social network with revised residency assignments of the individual nodes indicated
by color, restricted to distinctive or very distinctive individuals with good or excellent quality
photographs. All individuals with no included associations with other animals are shown in the
upper left corner. All tagged animals (n=4) are indicated by a square node shape. Red nodes are
O‘ahu long-term residents, orange nodes are O‘ahu short-term residents, purple nodes are O‘ahu
associative residents, pink are O‘ahu visitors, blue are Maui Nui long-term residents, light blue
are Maui Nui short-term residents, teal are Maui Nui associative residents, and green are Maui
Nui visitors.

58

Most individuals in the peripheral clusters were visitors, with two exceptions: HITt1145,
classified as an O‘ahu long-term resident, and HITt1169, classified as an O‘ahu short-term
resident. HITt1145 was first seen in 2008, then again in 2018, both times in the company of one
other individual. HITt1169 was first seen in 2016 by itself, then again in 2017 in the company of
three other individuals.
Interestingly, all inter-island individuals clustered most closely with the O‘ahu component, and
only represented a link between the two main components twice. These two links were identified
as HITt1095 (seen three times off Maui Nui in 2012, 2014, and 2017, and once off O‘ahu), and
HITt1152 (seen once off Maui Nui in 1997, and twice off O‘ahu in 2015 and 2016).
Additionally, several Maui Nui visitors clustered more closely with the O‘ahu component than
the Maui Nui component. When inter-island individuals were filtered out of the social network,
the two main components were completely separated, and all Maui Nui visitors that clustered
with the O‘ahu component became isolated from both main components.

Sex Determination
Consistent with the results for island areas, sex was biased towards females across all
residency classes (Appendix Table A). This trend was particularly apparent for sexed long-term
residents off O‘ahu, all nine of whom were female. This was also true for sexed inter-island
individuals, all three of whom were female. The sexed Maui Nui long-term residents had the
largest number of males (five males out of 16 sexed individuals total), and the highest
proportions of males identified were among the O‘ahu visitors (one male out of two sexed
individuals total), and the Maui Nui associative residents (one male out of two sexed individuals
total), though both of these groups had very limited sample sizes.

59

Subarea-Stratification
Spatial stratification of the dataset resulted in the creation of five distinct geographic
subareas: O‘ahu North (ON), O‘ahu West (OW), O‘ahu East (OW), Moloka‘i/Penguin Bank
(MPB), and Maui Nui (MN; representing Maui, Lāna‘i , and Kaho‘olawe; Figure 8). Three of
these subareas were further divided at the 500 m bathymetric depth (O‘ahu West (OW)/O‘ahu
West Deep (OWD), O‘ahu East (OE)/O‘ahu East Deep (OED), and Maui Nui (MN)/Maui Nui
Deep (MND)). This resulted in a total of eight subareas.

Subarea Characterization
The total area of each subarea ranged from 934 km2 (O‘ahu West) to 6,174 km2 (Maui
Nui), with all three of the O‘ahu subareas smaller than the Maui Nui subareas (Table 4). All
subareas had more shallow water (0-500 m bathymetric depth) than deep water (500-1,000 m
bathymetric depth), but the Maui Nui and Moloka‘i/Penguin Bank subareas in particular had
extensive shallow water habitat, with 4,030 km2 and 2,570 km2 of shallow water respectively.

Table 4. Total area (in km2) by depth (m) for each island area and subarea. Values have been
rounded to the nearest integer. For island areas, O = O‘ahu island area, MN = MN island area.
For subareas, OE = O‘ahu East, OW = O‘ahu West, ON = O‘ahu North, MN = Maui Nui
subarea, MPB = Moloka‘i/Penguin Bank.
Island SubArea area

O
MN
O
O
O
MN
MN

OE
OW
ON
MN
MPB

Amount of available habitat (km2) by depth (m) range
01000

0500

5001000

0100

100200

200300

300400

400500

500600

600700

700800

800900

9001000

3041

1683

1358

748

174

180

284

297

368

270

281

255

184

10103

6600

3503

2497

1283

1031

917

872

1049

1001

590

415

448

1145

612

533

296

58

74

100

84

126

132

152

71

52

962

559

403

262

32

42

94

129

155

90

52

74

32

934

512

422

190

84

64

90

84

87

48

77

110

100

6174

4030

2144

1267

801

707

652

603

541

603

419

253

328

3929

2570

1359

1230

482

324

265

269

508

398

171

162

120

60

The shape and relative exposure of shallow water within subareas varied substantially
(Figure 8). The O‘ahu West and O‘ahu West Deep subareas consist of a narrow band of habitat
along the west coast of O‘ahu that widens off the south coast, while the O‘ahu North subarea
consists of a broader expanse of shallow water. O‘ahu East consists of intermediate-sized bands
of shallow and deep water compared to O‘ahu West/O‘ahu West Deep and O‘ahu North.
Moloka‘i/Penguin Bank consists of an extensive shallow-water shelf (Penguin Bank) and
shallow area surrounding Moloka‘i, with a band of deep water encompassing Penguin Bank. The
Maui Nui subarea has a broad shallow-water area, with some nearshore deepwater habitat off the
west coast of Lāna‘i and the south shore of Kaho‘olawe. All of the O‘ahu subareas are highly
exposed to the open ocean, while a large portion of the Maui Nui subarea is enclosed by the
islands of Maui, Lāna‘i, and Kaho‘olawe. Penguin Bank is also highly exposed, but the waters to
the south of Moloka‘i are partially enclosed by Maui and Lāna‘i.
Chlorophyll-a concentrations varied substantially both between seasons and between
years, though all were consistently oligotrophic (Appendix Figure B). During the summer
months, chlorophyll-a concentrations rose markedly across all years examined, and generally
appeared to be lower during the fall and winter months. Within subareas, chlorophyll-a
concentration trends fluctuated within the same season between years, with no immediately
obvious recurring spatial patterns.

Social Networks, Residency, and Demographics by Subarea
Subareas where individuals were encountered were mapped onto the social network to
explore social consequences of spatial stratification. Spatial stratification between subareas
generally aligned with social relationships (Figure 14). O‘ahu North animals in particular

61

clustered together almost entirely separate from the main O‘ahu cluster, connected by one
individual, HITt1703, which was seen twice off O‘ahu in 2018, and only once with GPS
coordinates recorded. Individuals seen in both the O‘ahu West and O‘ahu West Deep subareas
were frequently intermixed with O‘ahu West individuals, though there was some peripheral
partitioning of a few O‘ahu West Deep individuals within the main O‘ahu cluster. Individuals
seen in the Moloka‘i/Penguin Bank subarea were generally located in peripheral clusters, though
a few were located on the periphery within the main O‘ahu cluster, and deep within the Maui Nui
cluster. Additionally, a couple of individuals with GPS locations in both the Moloka‘i/Penguin
Bank and Maui Nui subareas acted as cutpoints between the O‘ahu and Maui Nui main clusters.
These two individuals were identified as HITt0027 and HITt0070, both of whom were classified
as Maui Nui long-term residents.
Based on encounters with recorded GPS locations, two individuals were sighted in
subareas corresponding to two different island areas (Figure 15). One individual, HITt1104, was
sighted twice in the O‘ahu West and once in the Maui Nui Deep subarea. The other individual,
HITt0232 , was sighted once in the O‘ahu West Deep and once in the Maui Nui Deep subarea.
Neither individual was of known sex, and both individuals clustered within the main O‘ahu
cluster (Figure 14). Exchange between subareas within the same island area was repeatedly
documented, most frequently between the O‘ahu West and O‘ahu West Deep subareas, with nine
individuals documented in both subareas. Four individuals were documented in both the Maui
Nui and Moloka‘i/Penguin Bank subareas, two individuals were documented in both the Maui
Nui and Maui Nui Deep subareas, and one individual was documented in both the O‘ahu North
and O‘ahu West subareas.

62

Figure 14. Social network with subareas where individuals were encountered indicated by color
and shape, restricted to distinctive or very distinctive individuals with good or excellent quality
photographs. All individuals with no included associations with other animals are shown in the
upper left corner. Blue circles indicate the Maui Nui subarea, dark blue indicate both the Maui
Nui and Maui Nui Deep suabreas, blue triangles indicate the Maui Nui deep subarea, light blue
circles indicate the Moloka‘i/Penguin Bank subarea, green ciricles indicate both the Maui Nui
and Moloka‘i/Penguin Bank subareas, yellow circles indicate both the Maui Nui Deep and O‘ahu
West subareas, yellow triangles indicate both the Maui Nui Deep and O‘ahu West Deep
subareas, red triangles indicate the O‘ahu West Deep subarea, red circles indicate the O‘ahu
West subarea, orange circles indicate both the O‘ahu West and O‘ahu West Deep subareas, light
purple circles indicate both the O‘ahu North and O‘ahu West subareas, dark purple circles
indicate the O‘ahu North subarea, pink circles indicate the O‘ahu East subarea, pink triangles
indicate the O‘ahu East Deep subarea, and white circles indicate an individual lacking any GPS
coordinates to assign subareas.

63

Figure 15. Map of subareas with the total number of animals identified within each subarea
following the subarea abbreviation in red, and the number of individuals identified within both
subareas adjacent to the red lines connecting subareas. Depth contours (grey dashed lines) are
shown for the 500 m and 1,000 m bathymetric depth. ON = O‘ahu North, OW = O‘ahu West,
OWD = O‘ahu West Deep, OE = O‘ahu East, OED = O‘ahu East Deep, MPB =
Moloka‘i/Penguin Bank, MN = Maui Neui (subarea), MND = Maui Nui Deep.

The distribution of different residency classes between subareas varied substantially, and
is reported in full in Table 5. The greatest proportion of inter-island individuals (38%) was in the
Maui Nui Deep subarea, although the greatest number of inter-island individuals were
documented within the Maui Nui subarea. The greatest numbers of long-term residents were
documented within the Maui Nui and O‘ahu West subareas (66 and 32 individuals, respectively),
and the greatest number and proportion of short-term residents were documented within the
Maui Nui subarea (25 individuals, 17%). The greatest proportions of long-term residents were in
the O‘ahu East and O‘ahu West Deep subareas (50% each). Associative residents were most

64

frequently documented within the O‘ahu West subarea (41 individuals), and the greatest number
and proportion of visitors were documented within the Moloka‘i/Penguin Bank subarea (30
individuals, 67%). The greatest proportion of associative residents, however, was within the
O‘ahu North subarea (71%).

Table 5. Subareas versus revised residency classifications for all encounters where
corresponding GPS location data were available. Percentages indicate the proportion of the total
number of unique individuals identified within that particular subarea, rounded to the nearest
integer. Rough density is calculated as the total number of individiauls identified within a
subarea, divided by the total area (km2) for the appropriate depth range (0-500 m, 500-1,000 m,
or 0 – 1,000 m).
Subarea

#
Encounters

Total #
Identified
Individuals

Rough
Density

OE

1

2

0.003

#
Individuals
Resighted
Within the
Subarea > 1
yr
0

OED

2

11

0.021

0

OW

18

84

0.150

12

OWD

6

30

0.074

1

ON

6

24

0.026

3

MN

235

144

0.036

88

MND

7

8

0.004

0

MPB

18

45

0.011

2

# (%)
InterIsland

# (%)
LongTerm
Residents

# (%)
ShortTerm
Residents

# (%)
Associative
Residents

# (%)
Visitors

0
(0%)
0
(0%)
2
(2%)
1
(3%)
0
(0%)
7
(5%)
3
(38%)
2
(4%)

1 (50%)

0 (0%)

1 (50%)

0 (0%)

0 (0%)

1 (9%)

3 (27%)

7 (64%)

32 (38%)

9 (11%)

41 (49%)

0 (0%)

15 (50%)

1 (3%)

12 (40%)

1 (3%)

3 (13%)

3 (13%)

17 (71%)

1 (4%)

66 (45%)

25 (17%)

29 (20%)

1 (13%)

0 (0%)

2 (25%)

17
(12%)
2 (25%)

4 (9%)

1 (2%)

8 (18%)

30
(67%)

Rough calculations revealed that the greatest “density” of animals (0.150 animals/km2)
was within the O‘ahu West subarea, followed by the O‘ahu West Deep subarea (0.074
animals/km2), and then the Maui Nui subarea (0.036 animals/km2; Table 5). The lowest
“density” was within the O‘ahu East subarea, with 0.003 animals/km2. However, these
calculations should not be treated as true measures of density, as they fail to account for variation
65

in survey effort and spatial use, as well as variation in the number of animals over time.
Discovery curves constructed for each of the subareas revealed that sampling was not
comprehensive for the vast majority of subareas, with the exception of the Maui Nui and perhaps
the O‘ahu West subareas (Figure 16). This means that the densities of the remaining subareas are
likely biased low, due to a potentially large portion of the animals that use the area not having
been identified yet. However, differences in subarea “densities” may also reflect genuine
differences in habitat use.
Group sizes varied between subareas, with larger groups generally found in the O‘ahu
West subarea and O‘ahu West Deep subareas, and smaller groups found in the O‘ahu North,
Maui Nui, and Maui Nui Deep subareas (Appendix Table C). Group sizes were normally
distributed for only the Maui Nui and Maui Nui Deep subareas based on the results of ShapiroWilk tests, so to test whether group sizes differed significantly between groups, a Kruskal-Wallis
ranked sums test was applied. This test revealed significant differences in group size by subarea
(KW = 9.544, p = 0.049). Post-hoc pairwise comparison using Dunn’s test with the BenjaminiHochberg method revealed that the only significant difference in group sizes by subarea was
between the Maui Nui and O‘ahu West subareas at the 0.1 significance level (Dunn’s test padjusted = 0.074).

66

Figure 16. Discovery curves by subarea (indicated by color) for all encounters with
corresponding GPS location data, with a reference 1:1 dashed trendline in black. Curves are
constructed chronologically. Top: Full view of all discovery curves. Bottom: Zoomed-in view of
the bottom left corner of the graph. ON = O‘ahu North, OW = O‘ahu West, OWD = O‘ahu West
Deep, OE = O‘ahu East, OED = O‘ahu East Deep, MPB = Moloka‘i/Penguin Bank, MN = Maui
Neui (subarea), MND = Maui Nui Deep.
67

Inter-Island Movements
Fourteen individuals were documented moving between island areas based on photoidentification. Only three of the 14 were sexed, all of whom were identified as females based on
calf presence. HITt0518, one of the confirmed adult females, was seen the greatest number of
times, with 20 total sightings between 2006 and 2018, 19 of which took place off O‘ahu, and one
of which took place off Moloka‘i in 2007. Of the inter-island individuals, six were sighted off
O‘ahu in multiple years, and two were sighted off Maui Nui in multiple years. Inter-island
individuals were seen in association during eight encounters, of the 47 total encounters where at
least one inter-island individual was present. The mean number of sightings for inter-island
individuals was 4.4 (sd = ± 4.6), though the majority of resightings were often within the same
island area.
Most individuals were documented moving between island areas only once (e.g., moving
from O‘ahu to Maui Nui), but three individuals moved between island areas more than once
(e.g., moving from O‘ahu to Maui Nui and back to O‘ahu). The three individuals that moved
between island areas more than once were HITt1095 (sighted first off Maui Nui in December
2012, then off O‘ahu in June 2014, and then again off Maui Nui in November 2014), HITt1104
(first sighted twice off O‘ahu in June 2014 and February 2016, then off Maui Nui in March 2017,
and again off O‘ahu in November 2017) and HITt1439 (sighted first off O‘ahu in September
2012, then twice off Maui Nui in December 2012 and November 2014, and then again off O‘ahu
in June 2017). The two inter-island individuals documented moving between islands in
encounters with corresponding GPS locations were HITt0232 (sighted in the Maui Nui Deep
subarea in 2002, then in the O‘ahu West Deep subarea in 2003) and HITt1104 (sighted first in

68

the O‘ahu West subarea in 2016, then in the Maui Nui Deep subarea in March 2017, and then
again in the O‘ahu West subarea in November 2017).
There were no strong seasonal trends to encounters with inter-island individuals (Table
6). Encounters with inter-island individuals most frequently took place off O‘ahu, and were less
frequent off Maui Nui. The proportion of total encounters with inter-island individuals was
greatest in O‘ahu during the summer (18%), and at its lowest during the spring (12%), while off
Maui Nui it was greatest during the Fall (6%), and at its lowest during the summer (1%) and
winter (1%).

Table 6. Inter-island movements by season from photo-identification data. Percentages are of the
total number of encounters for that particular island area and seaon. II = Inter-Island.
Season

Spring
Summer
Fall
Winter

Total #
Encounters off
O‘ahu
76
79
46
47

# Encounters
with II off
O‘ahu (%)
9 (12%)
14 (18%)
7 (15%)
7 (15%)

Total #
Encounters off
Maui Nui
151
69
34
103

# of Encounters
with II off Maui
Nui (%)
6 (4%)
1 (1%)
2 (6%)
1 (1%)

Movements and Spatial Use

Photo-Identification Data
Mean inter-annual movement distances varied between island areas, subareas, and
residency classes, though the vast majority fell below 20 km (Table 7; Figure 17). The mean
overall inter-annual movement distance for all individuals was 16.5 km (sd = ± 16.8 km, rsd =
102%), but when broken down by island area was markedly lower for O‘ahu (10.3 km, sd = ±
8.5 km, rsd = 83%) than for Maui Nui (17.4 km, sd = ± 9.4 km, rsd = 54%). Inter-island
individuals had substantially larger inter-annual movement distances, with a mean of 65.0 km (sd

69

= ± 78.4 km, rsd = 121%), due the influence of the inter-island movements. When broken down
by subareas, the O‘ahu North, O‘ahu West, and Maui Nui subareas all had relatively small mean
inter-annual travel distances, at 8.0 km (sd = ± 1.5 km, rsd = 19%), 10.0 km (sd = ± 8.8 km, rsd
= 88%), and 16.9 km (sd = ± 9.2 km, rsd = 54%) respectively. In contrast, the O‘ahu West Deep
and Moloka‘i/Penguin Bank had mean inter-annual movement distances that exceeded the mean
for all individuals, at 25.4 km (sd = ± 49.1 km, rsd = 193%) and 26.3 km (sd = ± 12.5 km, rsd =
48%) respectively, and by far the largest mean inter-annual movement distance was for
individuals seen in the Maui Nui Deep subarea (86.9 km, sd = ± 74.8 km, rsd = 86%). When
broken down by residency class, O‘ahu long-term and short-term residents also had smaller
mean inter-annual movement distances when compared to Maui Nui long and short-term
residents, with the smallest distance for O‘ahu short-term residents (9.2 km, sd = ± 1.2 km, rsd =
13%), and the greatest distance for Maui Nui long-term residents (17.5 km, sd = ± 8.3 km, 47%).
Mean inter-annual movment distances for a simulated randomly mixing population were
also calculated to provide a point of reference for comparison between actual populations (Table
7). For the simulated population with random mixing between island areas, the mean interannual travel distance was 66.9 km (sd = ± 43.5 km, rsd = 65%). For the simulated populations
with random mixing within island areas, mean inter-annual travel distances were 19.5 km (sd = ±
13.3 km, rsd = 68%) for O‘ahu, and 22.0 km (sd = ± 9.6 km, rsd = 44%) for Maui Nui.

70

Table 7. Mean inter-annual travel distances for individuals seen in more than one year for all
encounters where corresponding GPS locations were available. For island area calculations, all
inter-island individuals have been excluded. Distances could not be calculated for the O‘ahu East
or O‘ahu East Deep subareas, or any associative residents or visitors.
Island Area, Subarea, or Residency
Class

# Individuals

Mean (SD) Inter-Annual
Travel Distance (km)

Simulated Random (All Encounters)
Simulated Random (O‘ahu Island

NA
NA

66.9 (± 43.5)
19.5 (± 13.3)

Simulated Random (Maui Nui Island
Area)

NA

22.0 (± 9.6)

All Individuals
O‘ahu (Island Area)
Maui Nui (Island Area)
Inter-island
O‘ahu North
O‘ahu West
O‘ahu West Deep
Moloka‘i/Penguin Bank
Maui Nui (subarea)
Maui Nui Deep
O‘ahu Long-Term Resident
O‘ahu Short-Term Resident
Maui Nui Long-Term Resident
Maui Nui Short-Term Resident

116
23
88
5
4
22
10
6
88
4
20
2
64
19

16.5 (± 16.8)
10.3 (± 8.5)
17.4 (± 9.4)
65.0 (± 78.4)
8.0 (± 1.5)
10.0 (± 8.8)
25.4 (± 49.1)
26.3 (± 12.5)
16.9 (± 9.2)
86.9 (± 74.8)
10.5 (± 9.0)
9.2 (± 1.2)
17.5 (± 8.3)
16.0 (± 10.6)

Area)

71

Figure 17. Histogram of mean inter-annual travel distances for all individuals (n=116) seen in
more than one year for all encounters where corresponding GPS locations were available.

Satellite-Tag Data
Satellite-tags were deployed off both island areas over the course of three different years,
and in a wide range of depths (Table 8). TtTag006 (HITt0788) and TtTag007 (HITt0794) were
both deployed in December 2012. HITt0788 (TtTag006) is a male dolphin classified as a Maui
Nui long-term resident, and has only been documented through photo-identification in the Maui
Nui subarea. On the social network, it is located within the main Maui Nui cluster, where it is
tied to 25 other individuals. This animal generally remained within the Maui Nui subarea over
the course of the 17.7 days where the tag was transmitting, though it did briefly cross up into the
Moloka‘i/Penguin Bank subarea when it moved to the south shore of Moloka‘i (Figure 18).
HITt0794 (TtTag007) is classified as a Maui Nui associative resident, and has only been

72

documented through photo-identification in the Maui Nui subarea as well. On the social network,
this individual is also located within the main Maui Nui cluster, where it shares ties with nine
other individuals. During the 9.0 days of signal contact, the animal generally remained within the
Maui Nui subarea, though it did venture out into the Maui Nui Deep subarea briefly (Figure 18).

Table 8. Summary of tag data. Sex was determined for only two individuals: TtTag006 was
identified as a male based on genetic analysis, and TtTag030 was identified as a female based on
calf presence.
Tag ID

Individual
ID

Date
Tagged

Tag
Deployment
Location

Bathymetric
Depth
Tagged (m)

Tag
Type

TtTag006

HITt0788
HITt0794

Southeast
Lānaʻi
West Lānaʻi

140

TtTag007
TtTag030

HITt0604

West O‘ahu

479

TtTag031

HITt1094
HITt1096

Offshore
West Lānaʻi
West Lānaʻi

708

TtTag032

13-Dec2012
19-Dec2012
17-Oct2016
07-Mar2017
17-Mar2017

MK
10-A
SPOT
5
SPOT
6
SPOT
6
SPOT
6

93

72

Duration
of
Contact
(Days)
17.7

# Kalman
Locations
(Unfiltered)

# Douglas
Filtered
Locations

345

337

9.0

131

130

13.0

236

230

16.3

343

332

13.0

253

246

73

Figure 18. Douglas-filtered Argos tracklines of the five satellite-tagged inviduals, one from
O‘ahu (top), and four from Maui Nui (bottom). In the top panel, TtTag030 (HITt0604, deployed
off west O‘ahu, 17-30 October 2016) is colored red. In the bottom panel, TtTag006 (HITt0788,
deloyed off southeast Lānaʻi, 13-31 December 2012) is colored red, TtTag007 (HITt0794,
deployed off west Lānaʻi, 19-28 December 2012) is colored orange, TtTag031 (HITt1094,
deployed offshore from west Lānaʻi, 7-24 March 2017) is colored light blue, and TtTag032
(HITt1096, deployed off west Lānaʻi, 17-30 March 2017) is colored dark blue. Depth contours
(black dashed lines) are shown for the 500 m and 1,000 m bathymetric depth.

74

TtTag030 (HITt0604) was deployed off O‘ahu in October 2016 onto a female O‘ahu
long-term resident documented through photo-identification in both the O‘ahu West and O‘ahu
West Deep subareas. This animal spent the duration of the 13.0 days of signal contact within the
O‘ahu West and O‘ahu subareas, with most locations in the O‘ahu West subarea (Figure 18).
The final two tags, TtTag031 (HITt1094) and TtTag032 (HITt1096), were both deployed
off Maui Nui in March 2017, onto two animals of unknown sex. HITt1094 (TtTag031) is
classified as a Mui Nui visitor, and has only ever been documented through photo-ID in the Maui
Nui Deep subarea. On the social network, it is disconnected from the main components, and
shares no ties with any individuals. Over the 16.3 days of signal contact, HITt1094 (TtTag031)
moved around extensively, crossing into the O‘ahu West, O‘ahu West Deep, Moloka‘i/Penguin
Bank, Maui Nui Deep, and Maui Nui subareas, with the majority of locations taking place on the
south end of Penguin Bank (Figure 18). HITt1096 (TtTag032) was not assigned a residency class
or included in the social network, on account of its distinctiveness falling below a score of Dist =
3. This individual has only been encountered once, on the day when it was tagged, when it was
encountered with three other individuals, HITt1095, HITt1097, and HITt1098. HITt1098 was
also not assigned a residency class or included in the social network on account of a low
distinctiveness score, but HITt1095 and HITt1097 were classified as an inter-island individual
and a Maui Nui short-term resident, respectively. On the social network, these two individuals
occupy locations on the periphery of the O‘ahu and Maui Nui clusters, and represent one of the
two links connecting the main clusters. Over the 13.0 days of signal contact, HITt1096 also
moved around extensively, moving between the Maui Nui and Maui Nui Deep subareas, as well
as the Moloka‘i/Penguin Bank and O‘ahu West Deep suabreas (Figure 18).

75

Probability densities were constructed for both island areas where tags were deployed
(Figure 19). The core area for the single animal tagged off O‘ahu was centered around the
southwest coast, incorporating parts of the O‘ahu West and O‘ahu West Deep subareas, while the
core area for the four animals tagged off Maui Nui were primarily centered around the island of
Lāna‘i, though a second, smaller core area existed on the south end of Penguin Bank. The 95 and
99% density polygons for Maui Nui-deployed tags extended to include the south shore of O‘ahu,
and overlapped with all three of the O‘ahu polygons.

76

Figure 19. Probability density representation of bottlenose dolphin home ranges by island area
where the animal was tagged, based on trimmed data from one satellite-tag deployed off O‘ahu
(top) and four off Maui Nui (bottom). Location data from the first 24 hours of each deployment
were discarded to reduce tagging area bias. Dark orange indicates the 50% polygon, orange
indicates the 95% polygon, and light orange represents the 99% polygon. Depth contours (grey
dashed lines) are shown for the 500 m and 1,000 m depth contours.
77

Individual Habitat Preferences
An examination of depth distributions for the six individuals with the greatest number of
sightings with corresponding GPS coordinates revealed a slight degree of individual preference
for different depth ranges (Figure 20). All six individuals (HITt0006, HITt0008, HITt0024,
HITt0044, HITt0056, and HITt0062) were Maui Nui long-term residents with sightings
exclusively in the Maui Nui subarea, and four of the six were identified as females based on
genetic sampling. Yet even within this mostly homogenous group, sightings of individuals were
distributed differently by depth. HITt0006, HITt0024, HITt0044, and HITt0056 exhibited a
strong preference for depths ≤ 100 m, with mean depths at 69 m (sd = ± 42 m), 101 m (sd = ± 93
m), 58 m (sd = ± 21 m), and 81 m (sd = ± 83 m) respectively. HITt0008 and HITt0062, on the
other hand, exhibited a preference for depths ≥ 100 m, with mean depths at 119 m (sd = ± 58 m)
and 147 m (sd = ± 50 m) respectively. Across the entire group, however, all sightings took place
in depths < 500 m, echoing the overall distribution of bottlenose sightings for all island areas
(Figure 9). A Kruskal-Wallis ranked sums test was performed to test whether encounter depths
differed by ID, revealing significant differences (KW = 23, p < 0.001). Post-hoc comparison of
IDs using Dunn’s test with the Benjamini-Hochberg method revealed significant differences in
encounter depths between HITt0006 and HITt0008 (Dunn’s test p-adjusted = 0.006), HITt0008
and HITt0044 (Dunn’s test p-adjusted = 0.007), HITt0008 and HITt0056 (Dunn’s test p-adjusted
= 0.012), HITt0006 and HITt0062 (Dunn’s test p-adjusted = 0.006), HITt0044 and HITt0062
(Dunn’s test p-adjusted = 0.010), and between HITt0056 and HITt0062 (Dunn’s test p-adjusted =
0.008).

78

Figure 20. Range of depths from photo-ID encounters for six Maui Nui long-term residents. Top
left: HITt0008. Top right: HITt0024. Middle left: HITt0062. Middle right: HITt0006. Bottom
left: HITt0044. Bottom right: HITt0056.

79

Tagged individuals also showed preferences for different depth ranges, though these
tended to be more extreme (Figure 21). TtTag006 (HITt0788) and TtTag007 (HITt0794) were
both tagged off Maui Nui in relatively shallow water ( < 150 m), and had the majority of their
Douglas-filtered Argos locations in depths ≤ 100 m, with mean depths of 75 m (sd = ± 74 m) and
78 m (sd = ± 151 m) respectively. However, a third individual tagged off Maui Nui in shallow
water at 72 m depth, TtTag032 (HITt1096), spent a significant portion of its time in waters
deeper than 100 m, and had a mean depth of 448 m (sd = ± 397 m). The remaining individual
tagged in relatively shallow water off O‘ahu in 479 m, TtTag030 (HITt0604), also displayed a
preference for waters ≤ 500 m, with a mean depth of 373 m (sd = ± 212 m) though it did spend
extensive time in deeper water. The one individual tagged in deep water off Maui Nui at 708 m,
TtTag031 (HITt1094), spent the majority of its time in water ≥ 500 m, with a mean depth of 809
m (sd = ± 545 m), though it did cross into shallow water areas as well. Another Kruskal-Wallis
ranked sums test was performed to determine whether the differences in tag location depths were
significant, revealing that they are (KW = 573.46, p < 0.001). Post-hoc pairwise comparison
using Dunn’s test with the Benjamini-Hochberg method revealed significant differences in tag
location depths between all pairs, with the exception of TtTag006 (HITt0788) and TtTag007
(HITt0794), and TtTag030 (HITt0604) and TtTag032 (HITt1096; Appendix Table D).

80

Figure 21. Range of depths from satellite-tag data for five individuals. Top left: TtTag006
(HITt0788, deployed off southeast Lāna‘i, 13-31 December 2012). Top right: TtTag007
(HITt0794, deployed off west Lāna‘i , 19-28 December 2012). Middle left: TtTag030
(HITt0604, deployed off west O‘ahu, 17-30 October 2016). Middle right: TtTag031 (HITt1094,
deployed offshore from west Lāna‘i, 7-24 March 2017). Bottom left: TtTag032 (HITt1096,
deployed off west Lāna‘i, 17-30 March 2017).

81

Interchange Indices
Within-area resighting indices and interchange indices for island areas were calculated
using the entire photo-identification dataset after quality control. Resighting indices were 1.07
for O‘ahu, and 2.33 for Maui Nui, indicating that a higher proportion of individuals are resighted
off Maui Nui than off O‘ahu, and reconfirming the earlier results of the discovery curves for
these areas (Figure 11). The interchange index between island areas was 0.28, falling an order of
magnitude below the Maui Nui and O‘ahu within-area resighting indices. This means that
movement between the island areas does occur, but not at levels that make it as likely as
movement within the island areas.
Within-area resighting indices by subarea for encounters with GPS coordinates ranged
from 0 to 5.21, with the lowest indices in the O‘ahu East (0.00), O‘ahu East Deep (0.00), and
Maui Nui Deep (0.00) subareas, and the highest index within the O‘ahu North subarea (5.21).
The second highest within-area resighting index was for the Maui Nui subarea (4.24), and values
for remaining subareas were 1.70 for O‘ahu West, 1.11 for O‘ahu West Deep, and 0.99 for
Moloka‘i/Penguin Bank. Sample sizes of resighted individuals in several subareas were very
small, including for the O‘ahu North, O‘ahu West Deep, O‘ahu East, O‘ahu East Deep,
Moloka‘i/Penguin Bank, and Maui Nui Deep Subareas, all of which had three or fewer
individuals resighted across multiple years. Total sample sizes of individuals within subareas
were comparatively small for all of the subareas as well, however. The greatest number of
within-area resighted individuals was within the Maui Nui subarea, with 88 out of 144 total
individuals having resightings across multiple years. Even with the generally small sample sizes,
however, movements between subareas were identified.

82

Interchange indices between subareas that were of the same or greater magnitude as both
within-area resighting indices for those subareas were found for four pairs of subareas: O‘ahu
West/O‘ahu West Deep, O‘ahu West/Maui Nui Deep, O‘ahu West Deep/Maui Nui Deep, and
Maui Nui/Maui Nui Deep (Table 9). This indicates that movements between these pairs of
subareas is as likely as movements within each of the subareas that compromise the pairs.
Interchanges indices between subareas that were not of the same or greater magnitude as both
within-area resighting indices for those subareas were calculated for two pairs of subareas:
O‘ahu West/O‘ahu North, and Maui Nui/Moloka‘i/Penguin Bank. Movements between these
pairs of subareas are therefore not as likely as movements within each of the individual subareas
that comprise the pairs, but do still occur. The interchange indices between all remaining pairs of
subareas were calculated to be zero, indicating that either no movement between those subareas
takes place, or that movement is infrequent enough that it was never documented.

83

Table 9. Within-area resighting and interchange indices for individuals with designated subareas
based on GPS coordinates. Resighting indices are based on the number of individuals seen across
multiple years, and located along the diagonal. Interchange indices that are of the same
magnitude as within-area resighting indices for both of the same subareas are shaded in grey. ON
= O‘ahu North, OW = O‘ahu West, OWD = O‘ahu West Deep, OE = O‘ahu East, OED = O‘ahu
East Deep, MPB = Moloka‘i/Penguin Bank, MN = Maui Neui (subarea), MND = Maui Nui
Deep.
Subarea (#
Within-Area
Resighted
Individuals, #
Total
Individuals)
ON (3, 24)
OW (12, 84)
OWD (1, 30)
OE (0, 2)
OED (0, 11)
MPB (2, 45)
MN (88, 144)
MND (0, 8)

ON

OW

OWD

OE

OED

MPB

MN

MND

5.21
0.49
0
0
0
0
0
0

1.70
3.57
0
0
0
0
1.49

1.11
0
0
0
0
4.17

0
0
0
0
0

0
0
0
0

0.99
0.62
0

4.24
1.74

0

84

Chapter Five: Discussion

Introduction
Accurate information about population structure is critical for effective management of
protected species. Bottlenose dolphins are found worldwide in tropical and sub-tropical waters in
a wide variety of population structures, including offshore, non-resident coastal, migratory,
transient archipelago-associated, resident coastal, and resident island-associated populations, as
well as possible metapopulation structures. While still subject to debate, the most relevant
definition for management purposes of a population for this species is the demographicallyindependent population (DIP), which is intended to be analogous to the NMFS-designated stocks
under direction of the MMPA.
Four small, demographically-independent and genetically differentiated populations of
resident island-associated bottlenose dolphins have been designated as stocks in the main
Hawaiian Islands, three of which showed evidence of decline in a recent abundance estimation
(Van Cise et al., 2021). These four stocks are centered around Kaua‘i/Ni‘ihau, O‘ahu, Maui Nui
(Maui, Lāna‘i, Kaho‘olawe, and Moloka‘i), and Hawai‘i, with statistically significant declines
occurring in the Maui Nui stock, and non-significant declines occurring in the Kaua‘i/Ni‘ihau
and O‘ahu stocks (Van Cise et al., 2021). These stocks are subject to unique anthropogenic
stressors that vary spatially, including shipping activity off O‘ahu, small boat traffic off Maui
Nui and Hawai‘i, military activity off Kaua‘i and Ni‘ihau, and fisheries activity off O‘ahu and
Hawai‘i. To ensure effective management, stock designations must therefore utilize accurate
spatial boundaries that encompass these spatially variable threats.
Long-term photo-identification and satellite-tagging efforts have revealed that some
resident dolphins may move between O‘ahu and Maui Nui (CRC, unpublished data). However,

85

the relative importance and impacts of these movements has remained unknown up to this point,
in spite of their potential for altering levels of genetic diversity, and transmitting culturallymediated behaviors. I used photo-identification and satellite-tag data on resident bottlenose
dolphins collected over the course of more than two decades to evaluate the population structure,
spatial use, and residency patterns of the O‘ahu and Maui Nui stocks using several approaches.
This sizable dataset, incorporating encounters from several areas, years, and seasons, allowed for
a robust evaluation. Dividing island areas into subareas allowed for greater resolution, and
helped to identify areas where minimal survey effort might bias conclusions.
Residency classifications revealed that while the number of long-term and short-term
residents in each island area were similar, there was a vast disparity in the number of visitors,
with a substantially larger number identified off O‘ahu. I also found that while social
connections between the two populations were minimal, there was geographic overlap in spatial
use. This was caused by a subset of individuals from the O‘ahu population that move between
the west and south coasts of O‘ahu, southwest Moloka‘i, Penguin Bank, and southwest Lāna‘i.
Generally, however, most individuals exhibited much more strict site fidelity and preference for
specific depth ranges, with mean inter-annual travel distances for the O‘ahu population at 10.3
km, and for the Maui Nui population at 17.4 km, compared to 65.0 km for animals that used both
island areas. There was no clear seasonal driver of inter-island movements, with inter-island
animals identified in both island areas at all times of the year. However, the drastically higher
number of visitors off O‘ahu in spite of a smaller total area of shallow-water habitat suggests that
perhaps this behavior is a response to limited resource availability caused by a high density of
bottlenose dolphins.

86

Survey Coverage, Encounter Characteristics, and Depth Preferences
Data spanned over two decades between CRC surveys and contributions from other
researchers and community scientists, allowing for a robust examination of residency and
movements. While CRC had considerable survey effort off Maui Nui in 2000-2001, the vast
majority of encounters (87%) were from groups other than CRC (Figure 7). This expansion of
the dataset allowed for a much more thorough evaluation of both residency and population
structure than would have been possible with CRC encounters alone, and improved the
likelihood of detecting inter-island movements. Additionally, contributed encounters increased
coverage in areas where CRC survey effort was not extensive, such as Penguin Bank and the
waters around Moloka‘i. However, both CRC effort and contributed encounters were heavily
biased towards the leeward sides of the islands due to weather conditions, likely excluding
bottlenose that primarily use the windward sides (Figure 6; Figure 8). Discovery curves for north
and east O‘ahu, two regions with minimal coverage, continue to rise at a steep incline, indicating
that either sampling has not been comprehensive or that there is a lower degree of site fidelity on
the windward sides (Figure 16). Additionally, the few encounters from the windward sides (six
from north O‘ahu, and three from east O‘ahu, as well as the three encounters from north of
Moloka‘i) were comprised of animals that were seen only in those areas, with the single
exception of an animal seen off both north and west O‘ahu. Furthermore, animals from these
areas cluster together, largely separate from the main components in the social network (Figure
14). This strongly suggests that additional groups of bottlenose dolphins are present on the
windward sides of the islands that are largely isolated from the leeward side groups, and may
therefore be demographically independent. Future survey efforts in these areas would be

87

beneficial in understanding whether windward side groups exist, and if so, what their
relationships to the leeward side groups are.
Dividing the island areas into subareas allowed for a more detailed analysis of population
structure and spatial use, and also helped to identify areas where encounter rates were low or
survey coverage was poor. Perhaps unsurprisingly given the small calculated mean inter-annual
travel distances (mean = 16.5 km for all resighted individuals), social stratification aligned well
with spatial stratification between subareas (Figure 14). Satellite-tag data also revealed spatial
stratification among tagged resident animals from the main island area clusters, with the two
Maui Nui residents generally remaining within shallow waters around Maui and Lānaʻi, and the
Oʻahu resident remaining in the waters around west and south Oʻahu (Figure 18). These five
tagged individuals help to demonstrate that Hawaiian resident bottlenose dolphins generally have
strict site fidelity, as previously reported by Baird et al. (2009), supported by Martien et al.
(2011), and later confirmed by Van Cise et al. (2021), something that is likely behaviorally
driven by distinct habitat preferences. Different subareas have unique degrees of open-ocean
exposure, total areas, and different expanses of nearshore habitat that could favor niche
specialization, as has been observed in other bottlenose dolphin populations worldwide (Hoelzel,
2009). Over a long time period, the limited mobility of Hawaiian bottlenose, in conjunction with
their social and spatial stratification, may continue to gradually increase the genetic and cultural
differentiation between groups. Already, the populations between island areas are genetically
differentiated (Martien et al., 2011), and future work to investigate how genetic haplotype ratios
differ between subareas would be informative.
Generally, encounter rates were highest in shallow subareas with extensive survey
coverage, including O‘ahu West, and Maui Nui, lowest in subareas with minimal survey

88

coverage, including O‘ahu North, O‘ahu East, and Moloka‘i/Penguin Bank. However, this trend
did not hold true for deep water areas, where in spite of extensive CRC survey effort, encounters
with bottlenose dolphins were less frequent than expected. This indicates that Hawaiian
bottlenose dolphins exhibit a general preference for shallow waters, especially waters shallower
than 200 m bathymetric depth, where the vast majority of encounters took place. These results
are well in line with previous work among other populations of inshore bottlenose dolphins
(Dinis et al., 2016; Silva et al., 2014). This also reaffirms conclusions in previous work on
Hawaiian bottlenose dolphins by Pittman et al. (2016), who demonstrated that depth is the single
most important predictor of bottlenose dolphin distribution in the Hawaiian Islands, with animals
exhibiting a strong preference for shallow water. This is likely driven by prey distribution, as
bottlenose are known to feed on nearshore and reef fish species (Baird, 2016).
A few encounters did take place in waters deeper than 500 m, raising the question of
whether the individuals in these encounters were from separate populations or social groups than
the shallow water animals. To explore this, subareas were further divided at the 500 m
bathymetric contour when encounters in both shallow and deep water were available. We found
that interchange between shallow and deep water subareas was generally high, with interchange
indices demonstrating that movements between O‘ahu West and O‘ahu West Deep, as well as
between Maui Nui and Maui Nui Deep were just as likely as movements within each subarea
(Table 9). Movements were not detected between the O‘ahu East and O‘ahu East Deep subareas,
but sample sizes in these subareas were small, limiting the ability to accurately assess
movements. This suggests that most bottlenose encounters in deep water (at least up to 1,000 m
bathymetric depth) likely do not represent different populations or social groups, but are
occasional excursions of the same resident populations that prefer shallow water.

89

Encounter characteristics and tag data revealed behavioral differences between animals in
different island areas. Significant differences in group size were detected between island areas,
with larger groups of animals encountered off O‘ahu compared to the groups encountered off
Maui Nui (Mann-Whitney U Test, p = 0.008). Additionally, satellite-tag data reveals possible
differences in depth preference. The two tagged animals that are positioned within the main
Maui Nui component of the social network remained exclusively in shallow water, never
venturing beyond 500 m depth, while in contrast the single animal tagged off O‘ahu and the
remaining two tagged animals from Maui Nui frequently ventured into deeper waters (Figure 18;
Figure 21). The two tags from Maui Nui that do not cluster with the main Maui Nui component
are more likely inter-island animals than members of the Maui Nui stock based on their
association and movement patterns, explaining the differences between the distributions of their
tag location depths and those of the other two animals tagged off Maui Nui. The similar depth
distribution between their tag locations and that of the single tagged O‘ahu animal lends further
support to the idea that they are inter-island animals, as all inter-island individuals are clustered
within the O‘ahu component of the social network and appear to represent a subset of this
population (Figure 13). However, tag data is inherently short-term, and caution must used in
interpreting these results. Crucially, the significant interchange between the Maui Nui and Maui
Nui Deep subareas revealed by photo-identification and interchange indices suggests that
members of the Maui Nui population do in fact make use of deep water, something which the tag
data failed to capture (Table 9).
Beyond population-level preferences, there is also evidence that preferences for different
depth ranges may exist at the individual level. An examination of the distribution of encounter
depths for the six Maui Nui long-term residents with the most encounters demonstrates that there

90

were significant differences in encounter depths between individuals (Figure 20; Kruskal-Wallis
ranked sums test, p < 0.001). This occurred in spite of the fact that these individuals are all
members of the same encountered only within the same subarea, and all are clustered together
within the Maui Nui main component of the social network (Figure 13). Significant differences
were primarily between individuals most frequently encountered between 100-200 m depth and
those most frequently encountered in ≤ 100 m depth. An analysis of the distribution of tag
location depths for the five satellite-tagged animals also revealed stark differences in tag location
depths between individuals (Figure 21). Almost all pairwise comparisons of tag location depths
between tagged individuals revealed significant differences, with only two exceptions – once
between two animals from the same island area, and once between two animals from different
island areas (Appendix Table D). This further highlights the need for caution in drawing
conclusions about population identity only from spatial use, and vice versa, as tremendous
variation can exist even within the same populations. Multiple lines of evidence in this case
produced slightly contradictory results regarding depth preferences, but in general the findings of
this study indicate that O‘ahu and inter-island animals tend to use deep water more regularly than
Maui Nui animals. This may be a behavioral adaptation indicative of ecological niche
specialization, though to some degree this may also reflect the greater total area of shallow water
habitat around Maui Nui compared to O‘ahu (Table 4).

Residency Assignments
Residency assignments were undertaken to explore potential variations between island
areas. There were similar numbers of long- and short-term residents for both island areas, though
there were almost six times as many associative residents in the O‘ahu island area compared to

91

the Maui Nui island area. The greatest numbers of both long- and short-term residents were
found in the O‘ahu West and Maui Nui subareas (Table 5). This is unsurprising given that these
subareas had both the greatest survey coverage and numbers of encounters, and improved survey
coverage increases the likelihood of repeated encounters with the same individuals, allowing for
a better assessment of how long individuals remain in a particular area. In contrast, the O‘ahu
East, O‘ahu East Deep, O‘ahu North subareas, all with poor coverage, have low numbers of
long- and short-term resident animals compared to the number of associative residents and
visitors. Interestingly, the Moloka‘i/Penguin Bank subarea had the greatest number of visitors
and one of the lowest resighting rates, in spite of having the second-largest number of encounters
within the subarea, and the third largest number of total identified individuals. This implies that
perhaps the animals using this subarea tend to be more transient, something reaffirmed by the
fact that the mean inter-annual travel distance of Moloka‘i/Penguin Bank animals is greater than
the mean inter-annual travel distances for all other subareas, with the exception of Maui Nui
Deep (Table 7).
Within the social network, two large main components representing the two main island
areas were easily distinguishable, along with several peripheral clusters (Figure 12). Of the 472
total animals included in the study, 380 (81%) individuals are part of the main components,
while 92 (19%) individuals are part of the peripheral clusters. Almost all long- and short-term
residents clustered within the main components, while the peripheral components were
comprised almost entirely of visitors (Figure 13). This is not unexpected, as their limited number
of resightings reduces the number of social relationships represented in the network. Peripheral
clusters may therefore be an artifact of inadequate sampling, but it is also possible that they
represent visiting dolphins from the pelagic stock, or from Kaua‘i or Hawai‘i Island, though no

92

such movements have been identified to date through photo-identification studies (Baird et al.,
2009; CRC, unpublished data). In terms of spatial use, most peripheral clusters were comprised
of individuals identified in the Moloka‘i/Penguin Bank subarea, though Moloka‘i/Penguin Bank
animals were also identified in both main components (Figure 14). Connections to the Maui Nui
component seem to be an artifact of the way that the subareas were drawn, however. HITt0075
and HITt0437 (the two Moloka‘i/Penguin Bank subarea animals deeply embedded within Maui
Nui component), and HITt0070 and HITt0027 (the two Moloka‘i/Penguin Bank animals
connecting the O‘ahu and Maui Nui components) were all encountered in the easternmost
portion of the Moloka‘i/Penguin Bank subarea, directly adjacent to the boundary of the Maui Nui
subarea. Revised subarea boundaries may counter this effect, and the further division of the Maui
Nui and Moloka‘i/Penguin Bank subareas should be considered in future work. The connections
to the O‘ahu component cannot be explained by subarea boundaries however, suggesting that
they are not artifacts of study design. This issue will be further discussed below.
The large number of associative residents within the O‘ahu island area is particularly
striking given how similar the relative numbers of other residency classes are for both island
areas (Table 3). All of these individuals are connected to the main O‘ahu component of the social
network, and of similar morphology to the other animals present around O‘ahu (CRC,
unpublished data), so they are likely not visiting offshore groups, but members of an inshore
population. Additionally, most associative residents were identified in either the O‘ahu West or
O‘ahu West Deep subareas (Table 5). These animals have all only been seen over time spans of
less than one year however, so they are either visitors from less frequently surveyed O‘ahu
subareas, or are not resident to the island at all. This raises the question of what draws these
individuals to the more heavily surveyed O‘ahu West subarea. Ecological conditions within the

93

O‘ahu West subarea are not markedly better than the ecological conditions elsewhere on O‘ahu
and Maui Nui, and in many ways are actually worse given the heavy volume of shipping traffic,
military activity, fishing activity, and the assorted environmental issues associated with dense
coastal human settlement. There is therefore no immediately identifiable ecological attractant to
explain such a large number of individuals passing through. An alternative explanation is that the
associative residents utilize certain areas of O‘ahu with high survey coverage (such as the west
coast of O‘ahu) as shallow-water travel corridors on their way to somewhere else with lower
survey coverage, and mingle along the way with the resident animals that use the area more
consistently. This would explain the limited resightings of these individuals and provide a
possible explanation for the larger group sizes encountered off O‘ahu. This still begs the
question, however, of where these animals are coming from and where they are going.
Additional tagging efforts dedicated to non-resident animals off O‘ahu may shed light on the
identity of these individuals.

Inter-Island Movements
Based on photo-identification evidence, very few individuals (14 out of 472 animals
included in this study, 3%) moved between island areas, all of which clustered within the main
O‘ahu component in the social network (Figure 12; Figure 13). In spite of over two decades
worth of data, and almost 400 identified animals with almost 6,000 links in the main
components, inter-island individuals only connected the two main components twice. Inter-island
travelers have associated with several individuals only seen once off Maui Nui, but without
additional resightings and social association data for these animals it is impossible to confirm
whether they are members of the Maui Nui stock, or travelers from the O‘ahu stock. Social

94

associations between the stocks, even with occasional inter-island movements occurring,
therefore appear infrequent at best. The fact that animals from separate stocks do not seem to
interact aligns well with the results of Martien et al. (2011), which found significant genetic
differentiation between the O‘ahu and Maui Nui stocks. Geographically, however, the areas used
by inter-island animals does overlap with the areas used by Maui Nui residents, indicating that
there is potential for interactions, and raising the question of why animals from different stocks
do not interact. A possible factor is the relative likelihood that separate groups of animals will
encounter one another. The Maui Nui stock is small (and declining), with only an estimated
abundance of 48-85 individuals (95% CI) in 2018 (Van Cise et al., 2021). The Maui Nui island
area has over 10,000 km2 of water between 0 and 1,000 m bathymetric depth, and over 6,000
km2 of water between 0 and 500 m of bathymetric depth, so this low abundance should
theoretically result in a very low density of bottlenose dolphins for the island area as a whole.
Combined with the fact that inter-island movements have only been captured a handful of times
through photo-ID, it seems unlikely that both an inter-island group and a Maui Nui group would
happen to be in the same place at the same time, allowing them the opportunity to interact. An
alternative explanation is that even in circumstances which spatially allow for interactions,
behavioral differences between groups may limit their ability or willingness to interact with one
another. Similar situations have been documented with different ecotypes of killer whales in the
nearshore waters of the temperate eastern North Pacific, where in spite of sympatry between
mammal-eating and fish-eating populations, the two ecotypes do not interact because of
behavioral differences (Baird et al., 1992). Similar to the different killer whale populations,
group sizes also significantly differ between the O‘ahu and Maui Nui stocks (Mann-Whitney U
Test, p = 0.008), suggesting that there is at least some degree of behavioral differentiation

95

between the two. Further research will be needed to explore whether there are additional
behavioral differences between the two stocks, such as different vocalization or communication
patterns or unique foraging specializations which might explain the lack of interaction between
them.
In spite of the lack of social interactions between stocks, there is clearly spatial overlap,
driven by the inter-island travelers. Compared to other island areas in the main Hawaiian
archipelago, connectivity between the O‘ahu and Maui Nui areas is quite good for bottlenose
dolphins. The channel between Hawai‘i and Maui Nui is almost 2,000 m deep and 28 km across,
and the channel between O‘ahu and Kaua‘i exceeds 3,000 m depth and is 116 km wide. While
distance is not necessarily an issue in that bottlenose dolphins are physically capable of travelling
long distances (e.g., Wells et al., 1999), most bottlenose in the Hawaiian islands have fairly
limited movements. The mean inter-annual travel distance of all bottlenose in the study was only
16.5 km, far smaller than the distances across the larger channels and the mean inter-annual
travel distance of a simulated population that randomly mixes across both island areas (Table 7).
Additionally, as previously discussed, bottlenose show a distinct preference for shallow water,
and especially water under 200 m bathymetric depth (Figure 9). This suggests that movements
across the larger channels are very unlikely, given that bottlenose dolphins are not likely to travel
that far or traverse into such deep waters. Only one movement of a tagged individual between
Kaua‘i and O‘ahu has been documented in the two decades that CRC has studied bottlenose in
the Hawaiian Islands (Baird, 2016), supporting this conclusion. In contrast, the Ka‘iwi channel
between O‘ahu and Maui Nui presents a much less significant barrier to movement. It reaches
only ~700 m depth, and is 42 km wide. While the distance across the channel does exceed the
mean inter-annual travel distance for all individuals, it is much smaller than the other channels.

96

Also, the much shallower depth does fall below the depths of the observed tag locations for three
out of the five satellite-tagged animals from Maui Nui and O‘ahu, and three encounters with GPS
locations that were included in this study. While bottlenose do not regularly move across the
larger channels, at least 14 individuals have been identified that have crossed the Ka‘iwi channel,
and the mean inter-annual travel distance for individuals confirmed to do this was 65.0 km, much
higher than the mean for all animals in the study.
As previously mentioned, all of the inter-island travelers clustered within the main O‘ahu
component in the social network, yet utilized both island areas. Encounters with GPS locations
where inter-island animals were identified were centered around southwest O‘ahu, southwest
Moloka‘i, and southwest Lāna‘I (Figure 8). Interchange indices based solely on photoidentification data with GPS locations also revealed that movements between the O‘ahu
West/O‘ahu West Deep subareas and the Maui Nui Deep subarea were just as likely as
movements within each of these areas (Table 9). Furthermore, based on associations and spatial
use, the two satellite-tags deployed on suspected inter-island animals (TtTag031 and TtTag032)
also made use of south O‘ahu, southwest Moloka‘i, west Lāna‘i, and south Lāna‘i. These tagged
animals also made extensive use of Penguin Bank, hinting that this area may also be of
importance to inter-island travelers (Figure 18). Combined, these data indicate that the interisland travelers seem to have a much larger range than the current O‘ahu stock boundaries,
extending from southwest O‘ahu, across Penguin Bank and southwest Moloka‘i to south Lāna‘i.
However, this range does not seem to extend into the more insular waters between western Maui
and eastern Lāna‘i, where the vast majority of encounters with Maui Nui stock members are
concentrated. Given the relative spatial use of the inter-island travelers versus the Maui Nui
residents, future research should incorporate revisions to the subarea boundaries (i.e., breaking

97

the Maui Nui subarea down into Maui Nui West and Maui Nui East subareas) to explore
movement patterns with a more refined lens.
The Moloka‘i/Penguin Bank subarea presents a few unique issues in regards to assessing
population structure. Animals identified in this subarea cluster within both the Maui Nui and
O‘ahu main components of the social network (Figure 14). The individuals clustered within the
Maui Nui component were all seen in the easternmost portion of the Moloka‘i/Penguin Bank
subarea, and therefore most likely an artifact of the way subarea boundaries were drawn, but
connections to the O‘ahu main component cannot be explained by subarea boundaries. Instead,
connections to the O‘ahu main component appear to be linked to inter-island movements. This
aligns well with the satellite-tag data of the two suspected inter-island travelers, which made
extensive use of the the Moloka‘i/Penguin Bank subarea. The area also has one of the greatest
proportions of visitors and lowest resighting rates, in spite of having 18 encounters, and the
discovery curve for this subarea is still sharply rising at almost a 1:1 ratio (Figure 16). This
suggests that the majority of animals within this subarea have not been identified, or that
residency to the area is very limited. This may be a result of inadequate sampling, but the sample
size is almost twice as large, or larger than the other regions with steep discovery curves. The
two regions with larger sample sizes (O‘ahu West and Maui Nui) had already shown some
indicators of leveling by the time their numbers of total identifications approached the number of
total identifications within the Moloka‘i/Penguin Bank subarea. This hints at either a larger
population within the Moloka‘i/Penguin Bank subarea, or that there is a smaller degree of
residency within the subarea. The large number of animals associated with peripheral clusters is
a further source of perplexity, as the majority of animals from this subarea don’t associate with
either the resident Maui Nui or O‘ahu animals. This could also be an effect of sample size, but

98

the relative proportion of animals in peripheral clusters vs in the main components exceeds that
of the other subareas with small sample sizes, making this theory less plausible. Instead, the most
likely hypothesis seems that the Moloka‘i/Penguin Bank subarea is frequently visited by
transient animals that do not remain in the area, though this cannot be proved without expanded
survey effort in the region.
While this study identified 14 inter-island travelers through photo-identification, and two
through satellite-tag data, it remains unclear what proportion of the O‘ahu stock moves between
island areas. When currently identified inter-island travelers are excluded, the mean inter-annual
travel distance of the O‘ahu stock is only 10.3 km, much smaller than the mean inter-annual
travel distance of a simulated population randomly mixing around O‘ahu at 19.5 km,suggesting
that at least while around O‘ahu animals do not tend to move about very much. Twelve
individuals from the O‘ahu stock have been resighted within the O‘ahu West subarea in multiple
years, sometimes dozens of times, so there is clearly an appreciable degree of site fidelity within
that subarea at least. However, it is possible that these animals also travel occasionally between
island areas, but that these movements are infrequent enough that they have not been
documented. Tour guides based on the western O‘ahu coast have said that they believe the
bottlenose dolphins in the region are not permanent residents, but regularly leave the area (R.W.
Baird, personal communication, March 5, 2021). Additionally, the O‘ahu West subarea is
comparatively small (962 km2 of water between 0 and 1,000 m bathymetric depth), especially
along the western coast where there is only a narrow band of shallow-water habitat (559 km2 of
water between 0 and 500 m bathymetric depth; Table 4). The O‘ahu West subarea also does not
seem exceptionally productive compared to other subareas, based on seasonal chlorophyll-a data
(Appendix Figure B). The limited availability of shallow-water habitat in the subarea may not

99

have enough resources to sustain the O‘ahu West animals, especially with the constant influx of
visiting animals that was previously discussed. It is possible, therefore, that these animals are
occasionally forced to travel greater distances to locate foraging opportunities, driving interisland movements. Similar circumstances have been identified among bottlenose dolphins in the
Azores, where the average distance between sightings in 25 km, but movements of up to 291 km
are repeatedly detected and hypothesized to be driven by limited prey availability (Silva et al.,
2008). The forces driving inter-island movements in the Hawaiian Islands remain unclear at
present though. Male sex-biased dispersal similar to that of the Sarasota resident bottlenose
dolphins (e.g. Wells et al., 1987) seems unlikely, as the three sex-identified inter-island animals
are females. There is also no obvious link between inter-island movements and seasonal
chlorophyll-a concentrations in the Hawaiian Islands, and no immediately discernable seasonal
trend to inter-island movements, with inter-island animals found in both island areas at all times
of the year. Continued research will be required to identify the ecological drivers of this
behavior.

Conclusion
Though somewhat constrained by uneven sampling across subareas, the broader patterns
revealed by this study point towards several interesting conclusions. I found that there are limited
social connections between the O‘ahu and Maui Nui island areas, which accounts for the
significant genetic differentiation between stocks revealed by Martien et al. (2011). However, in
spite of the limited social connections between stocks, there are important geographic overlaps
that cross current stock boundaries. These overlaps are caused by a subset of the O‘ahu
population that moves between island areas, using an expanded range that includes southwest

100

O‘ahu, Penguin Bank, southwest Moloka‘i, and the west and southwest coasts of Lāna‘i. This
suggests that current stock boundaries may be inadequate for the O‘ahu stock, as they fail to
account for exposure to any anthropogenic threats off Maui Nui. However, the proportion of the
O‘ahu stock that uses this extended range and the ecological cues that drive this behavior remain
unclear, and continued research will be required to address these factors. Continued research
should also focus on subareas with minimal coverage, such as O‘ahu North and O‘ahu East,
where social stratification in conjuction with spatial stratification suggests the possibility of
additional demographically-independent populations of bottlenose.
Finally, the use of multiple datasets yielded different conclusions in this study,
highlighting a general need to use multiple, long-term approaches in assessing population
structure. While genetic and photo-identification data initially suggested that animals do not
move across the Ka‘iwi channel, reassessment of long-term data and the inclusion of satellite-tag
data points to a different conclusion. Continued reassessment of the population structure of these
animals should be undertaken to test the accuracy of previous work, and ensure the effectiveness
of current management strategies to preserve these charismatic animals for future generations.

101

References
Baird, R. W., Abrams, P. A., & Dill, L. M. (1992). Possible indirect interactions between
transient and resident killer whales: implications for the evolution of foraging
specializations in the genus Orcinus. Oecologia, 89(1), 125-132.
Baird, R. W., Gorgone, A. M., McSweeney, D. J., Ligon, A. D., Deakos, M. H., Webster, D. L.,
Schorr, G. S., Martien, K. K., Salden, D. R., & Mahaffy, S. D. (2009). Population
structure of island-associated dolphins: Evidence from photo-identification of common
bottlenose dolphins (Tursiops truncatus) in the main Hawaiian Islands. Marine Mammal
Science, 25(2), 251-274. https://doi.org/10.1111/j.1748-7692.2008.00257.x
Baird, R. W., Webster, D. L., Aschettino, J. M., Schorr, G. S., & McSweeney, D. J. (2013).
Odontocete Cetaceans Around the Main Hawaiian Islands: Habitat Use and Relative
Abundance from Small-Boat Sighting Surveys. Aquatic Mammals, 39(3).
https://doi.org/10.1578/AM.39.3.2013.253
Baird, R. W., Martin, S. W., Webster, D. L., & Southall, B. L. (2014). Assessment of modeled
received sound pressure levels and movements of satellite-tagged odontocetes exposed to
mid-frequency active sonar at the Pacific Missile Range Facility: February 2011 through
February 2013. Prepared for U.S. Pacific Fleet.
https://www.cascadiaresearch.org/oldsite/Hawaii/Bairdetal2014_PMRFexposure.pd
f
Baird, R. W., Cholewiak, D., Webster, D. L., Schorr, G. S., Mahaffy, S. D., Curtice, C.,
Harrison, J., & Van Parijs, S. M. (2015). Biologically Important Areas for Cetaceans
Within US Waters-Hawai'i Region. Aquatic Mammals, 41(1), 54-64.
https://doi.org/10.1578/AM.41.1.2015.54
Baird, R. W., Cornforth, C. J., Jarvis, S.M., DiMarzio, N. A., Dolan, K., Henderson, E. E.,
Martin, S. W., Watwood, S. L., Mahaffy, S. D., Guenther, B. D., Lerma, J. K., Harnish,
A. E., & Kratofil, M. A. (2021). Odontocete studies on the Pacific Missile Range Facility
in February 2020: satellite-tagging, photo-identification, and passive acoustic
monitoring. Prepared for Commander, Pacific Fleet, Environmental Readiness Division,
Pearl Harbor, HI.
https://www.cascadiaresearch.org/files/publications/Bairdetal2021_Kauai.pdf
Baird, R.W. (2016). The Lives of Hawai‘i’s Dolphins and Whales: Natural History and
Conservation. University of Hawai‘i Press.
Bauer, R. K. (2020). Oceanmap: A Plotting Toolbox for 2D Oceanographic Data. R package v.
0.1.1. https://cran.microsoft.com/snapshot/2020-0408/web/packages/oceanmap/oceanmap.pdf
Bearzi, G., Notarbartolo-DI-Sciara, G., & Politi, E. (1997). Social ecology of bottlenose dolphins
in the Kvarnerić (Northern Adriatic Sea). Marine mammal science, 13(4), 650-668.
https://doi.org/10.1111/j.1748-7692.1997.tb00089.x

102

Bearzi, G., Agazzi, S., Bonizzoni, S., Costa, M., & Azzellino, A. (2008). Dolphins in a bottle:
abundance, residency patterns and conservation of bottlenose dolphins Tursiops truncatus
in the semi-closed eutrophic Amvrakikos Gulf, Greece. Aquatic Conservation: Marine
and Freshwater Ecosystems, 18(2), 130-146. https://doi.org/10.1002/aqc.843
Bearzi, G., Bonizzoni, S., & Gonzalvo, J. (2010). Mid-distance movements of common
bottlenose dolphins in the coastal waters of Greece. Journal of Ethology, 29(2), 369-374.
https://doi.org/10.1007/x10164-010-0245-x
Borgatti, S. P. (2002). NetDraw: graph visualization software (Version 2.158). Analytic
Technologies. https://sites.google.com/site/netdrawsoftware/home
Bossart, G. D. (2011). Marine Mammals as Sentinel Species for Oceans and Human Health.
Veterinary Pathology, 48(3), 676-690. https://doi.org/10.1177/0300985810388525
Caballero, S., Islas-Villanueva, V., Tezanos-Pinto, G., Duchene, S., Delgado-Estrella, A.,
Sanchez-Okrucky, R., & Mignucci-Giannoni, A. A. (2012). Phylogeography, genetic
diversity and population structure of common bottlenose dolphins in the Wider Caribbean
inferred from analyses of mitochondrial DNA control region sequences and microsatellite
loci: conservation and management implications. Animal Conservation, 15(1), 95-112.
https://doi.org/10.1111/j.1469-1795.2011.00493.x
Calabrese, J. M., & Fagan, W. F. (2004). A comparison-shopper's guide to connectivity metrics.
Frontiers in Ecology and the Environment, 2(10), 529-536.
https://doi.org/10.1890/1540-9295(2004)002[0529:ACGTCM]2/0/CO;2
Calenge, C. (2006). The package adehabitat for the R software: a tool for the analysis of space
and habitat use by animals. Ecological Modelling 197, 516–519.
https://doi.org/10.1016/j.ecolmodel.2006.03.017
Carretta, J. V., Forney, K. A., Oleson, E. M., Weller, D. W., Lang, A. R., Baker, J., Muto, M. M.,
Hanson, B., Orr, A. J., Huber, H., Lowry, M. S., Barlow, J., Moore, J. E., Lynch, D.,
Carswell, L., & Brownell Jr., R. L. (2019). U.S. Pacific Marine Mammal Stock
Assessments: 2018. U.S. Department of Commerce, NOAA.
https://repository.library.noaa.gov/view/noaa/20266
Carnabuci, M., Schiavon, G., Bellingeri, M., Fossa, F., Paoli, C., Vassallo, P., & Gnone, G.
(2016). Connectivity in the network macrostructure of Tursiops truncatus in the Pelagos
Sanctuary (NW Mediterranean Sea): does landscape matter?. Population ecology, 58(2),
249-264. https://doi.org/10.1007/s10144-016-0540-7
Connor, R. C., & Wells, R. S. (2000). The bottlenose dolphin: social relationships in a fission–
fusion society. In J. Mann, R.C. Connor, P.L. Tyack, & Whitehead, H. (Eds.), Cetacean
Societies: Field Studies of Dolphins and Whales (pp. 91–126). The University of Chicago
Press.
Costa, A. P. B., Fruet, P., Daura-Jorge, F. G., Simões-Lopes, P. C., Ott, P. H., Valiati, V. H., &
de Oliveira, L. R. (2015). Bottlenose dolphin communities from the southern Brazilian
103

coast: do they exchange genes or are they just neighbours?. Marine and Freshwater
Research, 66(12), 1201-1210. https://doi.org/10.1071/MF14007
Costello, M. J., Tsai, P., Wong, P. S., Cheung, A. K. L., Basher, Z., & Chaudhary, C. (2017).
Marine biogeographic realms and species endemicity. Nature Communications, 8(1), 110. https://doi.org/10.1038/s41467-017-01121-2
Defran, R. H., & Weller, D. W. (1999). Occurrence, distribution, site fidelity, and school size of
bottlenose dolphins (Tursiops truncatus) off San Diego, California. Marine Mammal
Science, 15(2), 366-380. https://doi.org/10.1111/j.1748-7692.1999.tb00807.x
Defran, R. H., Caldwell, M., Morteo, E., Lang, A. R., Rice, M. G., & Weller, D. W. (2015).
Possible stock structure of coastal bottlenose dolphins off Baja California and California
revealed by photo-identification research. Bulletin, Southern California Academy of
Sciences, 114(1), 1-11. https://doi.org/10.3160/0038-3872-114.1.1
Díaz López, B., & Bernal Shirai, J. A. (2008). Marine aquaculture and bottlenose dolphins’
(Tursiops truncatus) social structure. Behavioral Ecology and Sociobiology, 62(6), 887.
https://doi.org/10.1007/s00265-007-0512-1
Dinis, A., Alves, F., Nicolau, C., Ribeiro, C., Kaufmann, M., Cañadas, A., & Freitas, L. (2016).
Bottlenose dolphin Tursiops truncatus group dynamics, site fidelity, residency and
movement patterns in the Madeira Archipelago (North-East Atlantic). African Journal of
Marine Science, 38(2), 151-160. https://doi.org/10.2989/1814232X.2016.1167780
Doty, M. S., & Oguri, M. (1956). The island mass effect. ICES Journal of Marine Science, 22(1),
33-37.
Douglas, D. C., Weinzierl, R., Davidson, S. C., Kays, R., Wikelski, M., & G. Bohrer. (2012).
Moderating Argos location errors in animal tracking data. Methods in Ecology and
Evolution, 3(6), 999-1007. https://doi.org/10.1111/j.2041-210X.2012.00245.x
Duffield, D. A., Ridgway, S. H., & Cornell, L. H. (1983). Hematology distinguishes coastal and
offshore forms of dolphins (Tursiops). Canadian Journal of Zoology, 61(4), 930-933.
https://doi.org/10.1139/z83-123
Eldredge, L. G., & Evenhuis, N. L. (2003). Hawai‘i's biodiversity: a detailed assessment of the
numbers of species in the Hawaiian Islands. Bishop Museum Press.
Estrade, V., & Dulau, V. (2020). Abundance and site fidelity of bottlenose dolphins off a remote
oceanic island (Reunion Island, southwest Indian Ocean). Marine Mammal Science
36(3), 871-896. https://doi.org/10.1111/mms.12693
Fruet, P. F., Secchi, E. R., Daura-Jorge, F., Vermeulen, E., Flores, P. A. C., Simões-Lopes, P. C.,
Genoves, R. C., Laporta, P., Di Tullio, J. C., Freitas, T. R. O., Dalla Rosa, L., Valiati, V.
H., Beheregaray, L. B., & Möller, L. M. (2014). Remarkably low genetic diversity and
strong population structure in common bottlenose dolphins (Tursiops truncatus) from

104

coastal waters of the Southwestern Atlantic Ocean. Conservation Genetics, 15, 879-895.
https://doi.org/10.1007/s10592-014-0586-z
Garrison, L.P., A.A. Hohn and L.J. Hansen. (2017). Seasonal movements of Atlantic common
bottlenose dolphin stocks based on tag telemetry data. Southeast Fisheries Science
Center.
Gaspari, S., Holcer, D., Mackelworth, P., Fortuna, C., Frantzis, A., Genov, T., Vighi, M., Natali,
C., Rako, N., Banchi, E., Chelazzi, G., & Ciofi, C. (2015). Population genetic structure of
common bottlenose dolphins (Tursiops truncatus) in the Adriatic Sea and contiguous
regions: implications for international conservation. Aquatic Conservation: Marine and
Freshwater Ecosystems, 25(2), 212-222. https://doi.org/10.1002/aqc.2415
Gazda, S. K., Connor, R. C., Edgar, R. K., & Cox, F. (2005). A division of labour with role
specialization in group–hunting bottlenose dolphins (Tursiops truncatus) off Cedar Key,
Florida. Proceedings of the Royal Society B: Biological Sciences, 272(1559), 135-140.
https://doi.org/10.1098/rspb.2004.2937
Gladilina, E., Shpak, O., Serbin, V., Kryukova, A., Glazov, D., & Gol'din, P. (2018). Individual
movements between local coastal populations of bottlenose dolphins (Tursiops truncatus)
in the northern and eastern Black Sea. Journal of the Marine Biological Association of
the United Kingdom, 98(2), 223-229. https://doi.org/10.1017/S0025315416001296
Gnone, G., Bellingeri, M., Dhermain, F., Dupraz, F., Nuti, S., Bedocchi, D., Moulins, A., Rosso,
M., Alessi, J., McCrea, R. S., Azzellino, A., Airoldi, N. P., Portunato, N., Laran, S.,
David, L., Di Meglio, N., Bonelli, P., Montesi, G., Trucchi, R., Fossa, F., & Wurtz, M.
(2011). Distribution, abundance, and movements of the bottlenose dolphin (Tursiops
truncatus) in the Pelagos Sanctuary MPA (north‐ west Mediterranean Sea). Aquatic
Conservation: Marine and Freshwater Ecosystems, 21(4), 372-388.
https://doi.org/10.1002/aqc.1191
Gove, J. M., McManus, M. A., Neuheimer, A. B., Polovina, J. J., Drazen, J. C., Smith, C. R.,
Merrifield, M. A., Friedlander, A. M., Ehses, J. S., Young, C. W., Dillon, A. K., &
Williams, G. J. (2016). Near-island biological hotspots in barren ocean basins. Nature
communications, 7(1), 1-8. https://doi.org/10.1038/ncomms10581
Hanski, I. (1991). Single-species metapopulation dynamics: concepts, models and observations.
Biological Journal of the Linnean Society, 42(1-2), 17-38.
https://doi.org/10.1111/j.1095-8312.1991.tb00549.x
Harnish, A. E., Ault, J., Babbitt, C., Gulland, F. M., Johnson, P. C., Shaughnessy, N. L., Wood,
K. A., & Baird, R. W. (2019). Survival of a Common Bottlenose Dolphin (Tursiops
truncatus) Calf with a Presumptive Gunshot Wound to the Head. Aquatic Mammals,
45(5), 543-548. https://doi.org/10.1578/am.45.5.2019.543
Harnish, A. E., Baird, R. W., Corsi, E., Gorgone, A. M., Perrine, D., Ward, A., & Sepeta, E.
(2021). Common Bottlenose Dolphin Associations with a Fish Farm in Hawai‘i: LongTerm Associations and Impacts on Other Delphinids [Document PSRG-2021-09
105

presented to the Pacific Scientific Review Group]. Pacific Scientific Review Group
Annual Meeting, 16-19 March, 2021.
Hastie, G. D., Wilson, B., Tufft, L. H., & Thompson, P. M. (2003). Bottlenose Dolphins Increase
Breathing Synchrony in Response to Boat Traffic. Marine Mammal Science, 19(1), 7484. https://doi.org/10.1111/j.1748-7692.2003.tb01093.x
Hastings, A. (1993). Complex interactions between dispersal and dynamics: lessons from
coupled logistic equations. Ecology, 74(5), 1362-1372. https://doi.org/10.2307/1940066
Hayes, S. A., Josephson, E., Maze-Foley, K., & Rosel, P. E. (2018). US Atlantic and Gulf of
Mexico Marine Mammal Stock Assessments-2017: (second edition). U.S. Department of
Commerce, NOAA. https://repository.library.noaa.gov/view/noaa/22730
Hijmans, R. J. (2019). Geosphere: Spherical Trigonometry. R Package Version 1.5-10.
https://CRAN.R-project.org/package=geosphere
Hildebrand, J. A. (2009). Anthropogenic and natural sources of ambient noise in the ocean.
Marine Ecology Progress Series, 395, 5-20. https://doi.org/10.3354/meps08353
Hoelzel, A. R., Potter, C. W., & Best, P. B. (1998). Genetic differentiation between parapatric
‘nearshore’ and ‘offshore’ populations of the bottlenose dolphin. Proceedings of the
Royal Society B: Biological Sciences, 265(1402), 1177-1183.
https://doi.org/10.1098/rspb.1998.0416
Hoelzel, A. R. (2009). Evolution of population genetic structure in marine mammal species. In
Bertorelle, G., Bruford, M. W., Hauffe, H. C., Rizzoli, A., & Vernesi, C. (Eds.),
Population genetics for animal conservation (294-318). Cambridge University Press.
Hoese, H. D. (1971). Dolphin feeding out of water in a salt marsh. Journal of Mammalogy,
52(1), 222-223. https://doi.org/10.2307/1378455
Hooker, S. K., & Gerber, L. R. (2004). Marine reserves as a tool for ecosystem-based
management: the potential importance of megafauna. Bioscience, 54(1), 27-39.
https://doi.org/10.1641/0006-3568(2004)054[0027:MRAATF]2.0.CO;2
Houser, D. S., Martin, S. W., & Finneran, J. J. (2013). Exposure amplitude and repetition affect
bottlenose dolphin behavioral responses to simulated mid-frequency sonar signals.
Journal of Experimental Marine Biology and Ecology, 443, 123-133.
https://10.1016.j.jemb.2013.02.043
Hwang, A., Defran, R. H., Bearzi, M., Maldini, D., Saylan, C. A., Lang, A. R., Dudzik, K. J.,
Guzòn-Zatarain, O. R., Kelly, D. L., & Weller, D. W. (2014). Coastal range and
movements of common bottlenose dolphins off California and Baja California, Mexico.
Bulletin, Southern California Academy of Sciences, 113(1), 1-13.
https://doi.org/10.3160/0038-3872-113.1.1

106

Ingram, S., & Rogan, E. (2003). Bottlenose dolphins (Tursiops truncatus) in the Shannon
Estuary and selected areas of the west-coast of Ireland.
https://www.npws.ie/sites/default/files/publications/pdf/Ingram_&_Rogan_2003_Bot
tlenose_dolphins.pdf
Kindlmann, P., & Burel, F. (2008). Connectivity measures: a review. Landscape Ecology, 23(8),
879-890. https://doi.org/10.1007/s10980-008-9245-4
Kiszka, J., Simon-Bouhet, B., Gastebois, C., Pusineri, C., & Ridoux, V. (2012). Habitat
partitioning and fine scale population structure among insular bottlenose dolphins
(Tursiops aduncus) in a tropical lagoon. Journal of Experimental Marine Biology and
Ecology, 416-417, 176-184. https://doi.org/10.1016/j.jembe.2012.03.001
Klatsky, L. J., Wells, R. S., & Sweeney, J. C. (2007). Offshore bottlenose dolphins (Tursiops
truncatus): Movement and dive behavior near the Bermuda Pedestal. Journal of
Mammalogy, 88(1), 59-66. https://doi.org/10.1644/05-MAMM-A-365R1.1
Kopps, A. M., Ackermann, C. Y., Sherwin, W. B., Allen, S. J., Bejder, L., & Krützen, M. (2014).
Cultural transmission of tool use combined with habitat specializations leads to fine-scale
genetic structure in bottlenose dolphins. Proceedings of the Royal Society B: Biological
Sciences, 281(1782), 20133245. https://doi.org/10.1098/rspb.2013.3245
Kranstauber, B., Cameron, A., Weinzerl, R., Fountain, T., Tilak, S., Wikelski, M., & R. Kays.
(2011). The Movebank data model for animal tracking. Environmental Modelling &
Software, 26, 834-835. https://doi.org/10.1016/j.envsoft.2010.12.005
Levins, R. (1969). Some demographic and genetic consequences of environmental heterogeneity
for biological control. Bulletin of the Entomological Society of America, 15(3), 237-240.
https://doi.org./10.1093/besa/15.3.237
Lewis, J. S., & Schroeder, W. W. (2003). Mud plume feeding, a unique foraging behavior of the
bottlenose dolphin in the Florida Keys. Gulf of Mexico Science, 21(1), 92-97.
https://doi.org/10.18785/goms.2101.09
Lopez, R., Malardé, J., Danès, P., & P. Gaspar. (2015). Improving Argos Doppler location using
multiple-model smoothing. Animal Biotelemetry, 3, 32. https://doi.org/10.1186/s40317015-0073-4
Louis, M., Fontaine, M. C., Spitz, J., Schlund, E., Dabin, W., Deaville, R., Caurant, F., Cherel,
Y., Guinet, C., & Simon-Bouhet, B. (2014a). Ecological opportunities and specializations
shaped genetic divergence in a highly mobile marine top predator. Proceedings of the
Royal Society B: Biological Sciences, 281(1795), 20141558.
https://doi.org/10.1098/rspb.2014.1558
Louis, M., Viricel, A., Lucas, T., Peltier, H., Alfonsi, E., Berrow, S., Brownlow, A., Covelo, P.,
Dabin, W., Deaville, R., de Stephanis, R., Gally, F., Gauffier, P., Penrose, R., Silva, M.
A., Guinet, C., & Simon-Bouhet, B. (2014b). Habitat-driven population structure of

107

bottlenose dolphins, Tursiops truncatus, in the North-East Atlantic. Molecular Ecology,
23(4), 857-874. https://doi.org/10.1111/mec.12653
Lowe, W. H., & Allendorf, F. W. (2010). What can genetics tell us about population
connectivity?. Molecular ecology, 19(15), 3038-3051. https://doi.org/10.1111/j.1365294X.2010.04688.x
Lowther-Thieleking, J. L., Archer, F. I., Lang, A. R., & Weller, D. W. (2014). Genetic
differentiation among coastal and offshore common bottlenose dolphins, Tursiops
truncatus, in the eastern North Pacific Ocean. Marine Mammal Science, 31(1), 1-20.
https://doi.org/10.1111/mms.12135
Mahaffy, S. D., Baird, R. W., McSweeney, D. J., Webster, D. L., & Schorr, G. S. (2015). High
site fidelity, strong associations, and long-term bonds: Short-finned pilot whales off the
island of Hawai‘i. Marine Mammal Science, 31(4), 1427-1451.
https://doi.org/10.1111/mms.12234
Martien, K. K., Baird, R. W., Hedrick, N. M., Gorgone, A. M., Thieleking, J. L., McSweeney, D.
J., Robertson, K. M, & Webster, D. L. (2011). Population structure of island‐ associated
dolphins: Evidence from mitochondrial and microsatellite markers for common
bottlenose dolphins (Tursiops truncatus) around the main Hawaiian Islands. Marine
Mammal Science, 28(3), E208-E232. https://doi.org/10.1111/j.1748-7692.2011.00506.x
Martien, K. K., Lang, A. R., Taylor, B. L., Rosel, P. E., Simmons, S. E., Oleson, E. M., Boveng,
P. L., & Hanson, M. B. (2019). The DIP delineation handbook: a guide to using multiple
lines of evidence to delineate demographically independent populations of marine
mammals. U.S. Department of Commerce, NOAA.
https://repository.library.noaa.gov/view/noaa/22660
Mead, J.G., & Potter, C.W. (1995). Recognizing two populations of the bottlenose dolphin
(Tursiops truncatus) of the Atlantic Coast of North America – morphologic and ecologic
considerations. IBI Reports, 5, 31-44.
Mirimin, L., Miller, R., Dillane, E., Berrow, S. D., Ingram, S., Cross, T. F., & Rogan, E. (2011).
Fine-scale population genetic structuring of bottlenose dolphins in Irish coastal waters.
Animal Conservation, 14(4), 342-353. https://doi.org/10.1111/j.14691795.2010.00432.x
Möller, L. M., Wiszniewski, J., Allen, S. J., & Beheregaray, L. B. (2007). Habitat type promotes
rapid and extremely localised genetic differentiation in dolphins. Marine and Freshwater
Research, 58(7), 640-648. https://doi.org/10.1071/MF06218
Möller, L. M., Bilgmann, K., Charlton-Robb, K., & Beheregaray, L. (2008). Multi-gene evidence
for a new bottlenose dolphin species in southern Australia. Molecular phylogenetics and
evolution, 49(2), 674-681. https://doi.org/10.1016.j.ympev.2008.08.011
NASA Goddard Space Flight Center, Ocean Ecology Laboratory, Ocean Biology Processing
Group (2018). Moderate-resolution Imaging Spectroradiometer (MODIS) Aqua 2018
108

Seasonal Data; NASA OB.DAAC, Greenbelt, MD, USA.
https://doi.org/10.5067/AQUA/MODIS/L3M/CHL/2018
Natoli, A., Peddemors, V. M., & Hoelzel, A. R. (2004). Population structure and speciation in
the genus Tursiops based on microsatellite and mitochondrial DNA analyses. Journal of
evolutionary biology, 17(2), 363-375. https://doi.org/10.1046/j.1420-9101.2003.00672.x
Natoli, A., Birkun, A., Aguilar, A., Lopez, A., & Hoelzel, A. R. (2005). Habitat structure and the
dispersal of male and female bottlenose dolphins (Tursiops truncatus). Proceedings of the
Royal Society B: Biological Sciences, 272(1569), 1217-1226.
https://doi.org/10.1098/rspb.2005.3076
Nichols, C., Herman, J., Gaggiotti, O. E., Dobney, K. M., Parsons, K., & Hoelzel, A. R. (2007).
Genetic isolation of a now extinct population of bottlenose dolphins (Tursiops truncatus).
Proceedings of the Royal Society B: Biological Sciences, 274(1618), 1611-1616.
https://doi.org/10.1098/rspb.20007.0176
National Marine Fisheries Service. (2005). Revisions to the Guidelines for Assessing Marine
Mammal Stocks. Institution. https://www.nmfs.noaa.gov/pr/pdfs/sars/gamms2005.pdf
Nykänen, M., Dillane, E., Englund, A., Foote, A. D., Ingram, S. N., Louis, M., Mirimin, L.,
Oudejans, M., & Rogan, E. (2018). Quantifying dispersal between marine protected areas
by a highly mobile species, the bottlenose dolphin, Tursiops truncatus. Ecology and
evolution, 8(18), 9241-9258. https://doi.org/10.1002/ece3.4343
Oudejans, M.G., Visser, F., Englund, A., Rogan, E., & Ingram, S.N. (2015). Evidence for
distinct coastal and offshore communities of bottlenose dolphins in the North East
Atlantic. PLoS One, 10(4), e0122668. https://doi.org/10.1371/journal.pone.0122668
Palsbøll, P. J., Berube, M., & Allendorf, F. W. (2007). Identification of management units using
population genetic data. Trends in ecology & evolution, 22(1), 11-16.
https://doi.org/10.1016/j.tree.2006.09.003
Pante, E., & Simon-Bouhet, B. (2013). Marmap: A Package for Importing, Plotting, and
Analyzing Bathymetric and Topographic Data in R. PLoS One, 8(9), e73051.
https://doi.org/10.1371/journal.pone.0073051
Parsons, K. M., Durban, J. W., Claridge, D. E., Herzing, D. L., Balcomb, K. C., & Noble, L. R.
(2006). Population Genetic Structure of Coastal Bottlenose Dolphins (Tursiops truncatus)
in the Northern Bahamas. Marine Mammal Science, 22(2), 276-298.
https://doi.org/10.1111/j.1748-7692.2006.00019.x
Pereira, A., Martinho, F., Brito, C., & Carvalho, I. (2013). Bottlenose dolphin Tursiops truncatus
at São Tomé Island (São Tomé and Príncipe)—relative abundance, site fidelity and social
structure. African Journal of Marine Science, 35(4), 501-510.
https://doi.org/10.2989/1814232X.2013.850444

109

Pilot, M., Dahlheim, M. E., & Hoelzel, A. R. (2010). Social cohesion among kin, gene flow
without dispersal and the evolution of population genetic structure in the killer whale
(Orcinus orca). Journal of evolutionary biology, 23(1), 20-31.
https://doi.org/10.1111/j.1420-9101.2009.01887.x
Pittman, S. J., Winship, A. J., Poti, M., Kinlan, B. P., Leirness, J. B., Baird, R. W., Barlow, J.,
Becker, E. A., Forney, K. A., Hill, M. C., Miller, P. I., Mobley, J., & Oleson, E. M.
(2016). Ch. 6: Marine Mammals – Cetaceans. In Costa, B. M., & Kendall, M. S. (Eds.),
Marine Biogeographic Assessment of the Main Hawaiian Islands (pp. 227-265). Bureau
of Ocean Energy Management and National Oceanic and Atmospheric Administration.
https://www.cascadiaresearch.org/files/Projects/Hawaii/MarineMammalsMHI_NO
AATechMemo.pdf
Quérouil, S., Silva, M. A., Freitas, L., Prieto, R., Magalhães, S., Dinis, A., Alves, F., Matos, J.
A., Mendonça, D., Hammond, P.S., & Santos, R. S. (2007). High gene flow in oceanic
bottlenose dolphins (Tursiops truncatus) of the North Atlantic. Conservation Genetics,
8(6), 1405. https://doi.org/10.1007/s10592-007-9291-5
Rendell, L., & Whitehead, H. (2001). Culture in whales and dolphins. The Behavioral and brain
sciences, 24(2), 309-382.
Richards, V. P., Greig, T. W., Fair, P. A., McCulloch, S. D., Politz, C., Natoli, A., Driscoll, C.
A., Hoelzel, A. R., David, V., Bossart, G. D., & Lopez, J. V. (2013). Patterns of
population structure for inshore bottlenose dolphins along the eastern United States.
Journal of Heredity, 104(6), 765-778. https://doi.org/10.1093/jhered/est070
Riesch, R., Barrett-Lennard, L. G., Ellis, G. M., Ford, J. K. B., & Deecke, V. B. (2012). Cultural
traditions and the evolution of reproductive isolation: ecological speciation in killer
whales?. Biological Journal of the Linnean Society, 106(1), 1-17.
https://doi.org/10.1111/j.1095-8312.2012.01872.x
Rigley, L., Vandyke, V. G., Cram, P., & Rigley, I. (1981). Shallow water behavior of the
Atlantic bottlenose dolphin (Tursiops truncatus). Proceedings of the Pennsylvania
Academy of Science, 55(2), 157-159.
Rosel, P. E., Hansen, L., & Hohn, A. A. (2009). Restricted dispersal in a continuously distributed
marine species: common bottlenose dolphins Tursiops truncatus in coastal waters of the
western North Atlantic. Molecular Ecology, 18(24), 5030-5045.
https://doi.org/10.1111/j.1365-294X.2009.04413.x
Rosel, P. E., Mullin, K. D., Garrison, L., Schwacke, L., Adams, J., Balmer, B., Conn, P., Conroy,
M. J., Eguchi, T., Gorgone, A., Hohn, A., Mazzoil, M., Schwartz, C., Sinclair, C.,
Speakman, T., Urian, K., Vollmer, N., Wade, P., Wells, R., & Zolman, E. (2011). Photoidentification capture-mark-recapture techniques for estimating abundance of bay, sound
and estuary populations of Bottlenose dolphins along the US East Coast and Gulf of
Mexico, a workshop report. U.S. Department of Commerce, NOAA.
https://repository.library.noaa.gov/view/noaa/4014

110

Schaefer, A. M., Titcomb, E. M., Fair, P. A., Stavros, H. C. W., Mazzoil, M., Bossart, G. D., &
Reif, J. S. (2015). Mercury concentrations in Atlantic bottlenose dolphins (Tursiops
truncatus) inhabiting the Indian River Lagoon, Florida: Patterns of spatial and temporal
distribution. Marine Pollution Bulletin, 97(1-2), 544-547.
https://doi.org/10.1016/j.marpolbul.2015.05.007
Schorr, G. S., Baird, R. W., Hanson, M. B., Webster, D. L., McSweeney, D. J., & Andrews, R.
D. (2009). Movements of satellite-tagged Blainville’s beaked whales off the island of
Hawai ‘i. Endangered Species Research, 10, 203-213. https://doi.org/10.3354/esr00229
Sellas, A. B., Wells, R. S., & Rosel, P. E. (2005). Mitochondrial and nuclear DNA analyses
reveal fine scale geographic structure in bottlenose dolphins (Tursiops truncatus) in the
Gulf of Mexico. Conservation Genetics, 6(5), 715-728. https://doi.org/10.1007/s10592005-9031-7
Silva, M. A., Prieto, R., Magalhães, S., Seabra, M. I., Santos, R. S., & Hammond, P. S. (2008).
Ranging patterns of bottlenose dolphins living in oceanic waters: implications for
population structure. Marine Biology, 156, 179-192. https://doi.org/10.1007/s00227008-1075-z
Sims, N. A. (2013). Kona Blue Water Farms case study: permiting, operations, marketing,
environmental impacts, and impediments to expansion of global open ocean mariculture.
In Technical workshop proceedings: Expanding mariculture farther offshore: technical,
environmental, spatial, and governance challenges (pp. 263-296).
Speakman, T. R., Lane, S. M., Schwacke, L. H., Fair, P. A., & Zolman, E. S. (2010). Markrecapture estimates of seasonal abundance and survivorship for bottlenose dolphins
(Tursiops truncatus) near Charleston, South Carolina, USA. Journal of Cetacean
Research and Management, 11(2), 153-162.
State of Hawai‘i Department of Business, Economic Development, and Tourism. (2019).
Hawai‘i Facts and Figures.
https://files.hawaii.gov/dbedt/economic/library/facts/Facts_Figures_browsable.pdf
Tezanos-Pinto, G., Baker, C. S., Russell, K., Martien, K., Baird, R. W., Hutt, A., Stone, G.,
Mignucci-Giannoni, A. A., Caballero, S., Endo, T., Lavery, S., Oremus, M., Olavarría, &
Garrigue, C. (2009). A worldwide perspective on the population structure and genetic
diversity of bottlenose dolphins (Tursiops truncatus) in New Zealand. Journal of
Heredity, 100(1), 11-24. https://doi.org/10.1093/jhered/esn039
Tobeña, M., Escánez, A., Rodríguez, Y., López, C., Ritter, F., & Aguilar, N. (2014). Inter-island
movements of common bottlenose dolphins Tursiops truncatus among the Canary
Islands: online catalogues and implications for conservation and management. African
Journal of Marine Science, 36(1), 137-141.
https://doi.org/10.2989/1814232X.2013.873738
Toth, J. L., Hohn, A. A., Able, K. W., & Gorgone, A. M. (2012). Defining bottlenose dolphin
(Tursiops truncatus) stocks based on environmental, physical, and behavioral
111

characteristics. Marine Mammal Science, 28(3), 461-478.
https://doi.org/10.1111/j.1748-7692.2011.00497.x
Urbán, J., Jaramillo, A., Aguayo, A., Ladrόn de Guevara, P., Salinas, M., Alvarez, C., Medrano,
L., Jacobsen, J. K., Balcomb, K. C., Claridge, D. E., Calambokidis, J., Steiger, G. H.,
Straley, J. M., Von Ziegesar, O., Waite, J. M., Mizroch, S., Dahlheim, M. E., Darling, J.
D., & Baker, C. S. (2000). Migratory Destinations of Humpback Whales Wintering in the
Mexican Pacific. Journal of Cetacean Research and Management, 2(2):101-110
Urian, K. W., Hofmann, S., Wells, R. S., & Read, A. J. (2009). Fine-scale population structure of
bottlenose dolphins (Tursiops truncatus) in Tampa Bay, Florida. Marine Mammal
Science, 25(3), 619-638. https://doi.org/10.1111/j.1748-7692.2009.00284.x
Van Cise, A. M., Baird, R. W., Harnish, A. E., Currie, J. J., Stack, S. H., Cullins, T., & Gorgone,
A. M. (2021). Mark-Recapture Estimates Suggest Declines in Abundance of Common
Bottlenose Dolphin Stocks in the Main Hawaiian Islands. Endangered Species Research,
45, 37-53. https://doi.org/10.3354/esr01117
Van Parijs, S. M., & Corkeron, P. J. (2001). Boat traffic affects the acoustic behaviour of Pacific
humpback dolphins, Sousa chinensis. Journal of the Marine Biological Association of the
United Kingdom, 81(3), 533-538. https://doi.org/10.1017/S0025315401004180
Waples, R. S., & Gaggiotti, O. (2006). What is a population? An empirical evaluation of some
genetic methods for identifying the number of gene pools and their degree of
connectivity. Molecular ecology, 15(6), 1419-1439. https://doi.org/10.1111/j.1365294X.2006.02890.x
Wells, R. S., Scott, M. D., & Irvine, A. B. (1987). The social structure of free-ranging bottlenose
dolphins. In Genoways, H. H. (eds) Current mammalogy (pp. 247-305). Springer.
https://doi.org/10.1007/978-1-4757-9909-5_7
Wells, R. S., & Scott, M. D. (1990). Estimating bottlenose dolphin population parameters from
individual identification and capture-release techniques. Reports of the International
Whaling Commission, 12, 407-415.
Wells, R. S., Rhinehart, H. L., Cunningham, P., Whaley, J., Baran, M., Koberna, C., & Costa, D.
P. (1999). Long Distance Offshore Movements of Bottlenose Dolphins. Marine Mammal
Science, 15(4), 1098-1114. https://doi.org/10.1111/j.1748-7692.1999.tb00879.x
Wells, R. S., Rhinehart, H. L., Hansen, L. J., Sweeney, J. C., Townsend, F. I., Stone, R., Casper,
D. R., Scott, M. D., Hohn, A. A., & Rowles, T. K. (2004). Bottlenose dolphins as marine
ecosystem sentinels: developing a health monitoring system. EcoHealth, 1(3), 246-254.
https://doi.org/10.1007/s10393-004-0094-6
Wells, R. S. (2009). Learning From Nature: Bottlenose Dolphin Care and Husbandry. Zoo
Biology, 28(6), 635-651. https://doi.org/10.1002/zoo.20252

112

Whitehead, H. (2008). Analyzing animal societies: Quantitative methods for vertebrate social
analysis. University of Chicago Press.

113

Appendix

Figure A. Social network with initial residency assignments indicated by color, restricted to
distinctive or very distinctive individuals with good or excellent quality photographs. All
individuals with no included associations with other animals are shown in the upper left corner.
All tagged animals (n=4) are indicated by a square node shape. Red nodes are O‘ahu long-term
residents, orange nodes are O‘ahu short-term residents, pink are O‘ahu visitors, blue are Maui
Nui long-term residents, light blue are Maui Nui short-term residents, green are Maui Nui
visitors, and yellow are inter-island animals.

114

2006

2013

2018

Spring

Summer

Fall

Winter

Figure B. Chlorophyll-a levels in mg/m3 around Oʻahu and Maui Nui across different seasons
from 2006, 2013, and 2018 at a 4 km resolution, based on data from NASA’s MODIS-Aqua
satellite mission.

115

Table A. Sex distribution by island area and residency class for the 49 individuals where sex
could be determined, representing 10% of the 472 total individuals included in the study.
Residency
Class or
Island Area

#
Individuals
Sexed

# Males
(Confirmed
Through
Genetics)

# Presumed
Males
(Confirmed
Only
Through
Morphology)

# Females
(Confirmed
Through
Genetics)

O‘ahu
(Island
Area)
Maui Nui
(Island
Area)
Inter-Island
O‘ahu
Long-Term
Resident
O‘ahu
Short-Term
Resident
O‘ahu
Associative
Resident
O‘ahu
Visitor
Maui Nui
Long-Term
Resident
Maui Nui
Short-Term
Resident
Maui Nui
Associative
Resident
Maui Nui
Visitor

25

4

1

6

# Presumed
Females
(Confirmed
Only
Through Calf
Presence or
Morphology)
14

21

6

0

8

7

3
9

0
0

0
0

0
2

3
7

3

0

1

0

2

11

3

0

4

4

2

1

0

0

1

16

5

0

7

4

3

0

0

1

2

2

1

0

0

1

0

0

0

0

0

116

Table B. Initial residency assignment results by island area. Percentages indicate the proportion
of the total number of unique identified individuals from all island areas combined, rounded to
the nearest percentage.
Island Area
O‘ahu
Maui Naui
Inter-Island
All Island
Areas

Total # (%) of
Individuals
271 (57%)
187 (40%)
14 (3%)
472

# (%) LongTerm Residents
59 (13%)
66 (14%)
125 (26%)

# (%) ShortTerm Residents
22 (5%)
26 (6%)
48 (10%)

# (%) Visitors
190 (40%)
95 (20%)
285 (60%)

Table C. Subareas versus group size for subareas with CRC encounters. Group sizes are best
estimates from CRC encounters. CRC encounters were not available for the O‘ahu East, O‘ahu
East Deep, or Moloka‘i/Penguin Bank subareas.
Subarea
OW
OWD
ON
MN
MND

# CRC
Encounters
7
3
4
52
3

Mean Group
Size (SD)
17.3 (± 12.4)
11 (± 7.0)
6.3 (± 4.1)
6.3 (± 4.5)
4.3 (± 2.9)

Minimum
Group Size
2
6
1
1
1

Maximum
Group Size
40
19
10
18
6

Table D. Post-hoc pairwise comparisons of tag location depths between tags.
Comparison
TtTag006 – TtTag007
TtTag006 – TtTag030
TtTag006 – TtTag031
TtTag006 – TtTag032
TtTag007 – TtTag030
TtTag007 – TtTag031
TtTag007 – TtTag032
TtTag030 – TtTag031
TtTag030 – TtTag032
TtTag031 – TtTag032

Dunn’s Test P-Adjusted
8.055 x 10-1
2.058 x 10-29
1.687 x 10-99
2.549 x 10-40
1.955 x 10-19
5.553 x 10-58
7.039 x 10-26
4.842 x 10-15
1.076 x 10-1
6.164 x 10-10

117

Media

Thesis_MES_2021_Harnish.pdf