-
Title
-
Eng
Feeding Ecology of “Southern Resident” Killer Whales (Orcinus orca): Benthic Habitat and Spatial Distribution
-
Date
-
2009
-
Creator
-
Eng
Lucas, Jeremy B
-
Subject
-
Eng
Environmental Studies
-
extracted text
-
Feeding Ecology of “Southern Resident” Killer Whales (Orcinus orca):
Benthic Habitat and Spatial Distribution
by
Jeremy Lucas
A Thesis: Essay of Distinction
Submitted in partial fulfillment
of the requirements for the degree
Master of Environmental Study
The Evergreen State College
June 2009
© 2009 by Jeremy Lucas. All rights reserved.
This Thesis for the Master of Environmental Study Degree
by
Jeremy Lucas
has been approved for
The Evergreen State College
by
Alison Styring
Member of the Faculty
Date
Abstract
Feeding Ecology of “Southern Resident” Killer Whales (Orcinus orca):
Benthic Habitat and Spatial Distribution
Jeremy Lucas
Though among the most geographically distributed mammals on earth, there are
separate, distinct populations of killer whales (Orcinus orca) occupying specific
geographic areas. In the coastal temperate northeast Pacific Ocean, one such
population, the so‐called “southern resident” killer whales (SRKWs), found in the inland
waters of Washington State and southern British Columbia, have been listed as
endangered by the governments of the United States and Canada. Possible reasons for
their population decline include contamination as a result of decades of pollution, a
decline in the numbers and density of their preferred prey, Pacific salmon
(Oncorhynchus spp.), and the impaired ability to find prey or perform other biologically
important functions due to excessive ambient noise in their environment. This study
examines the potential of classifying critical habitat for these killer whales based on
benthic topography and spatial distribution. From 2004‐2007, field data were collected
in the San Juan Islands and Puget Sound of Washington State to ascertain 1) if there are
any differences between where killer whales feed versus where they do not feed and 2)
if there are any spatial density patterns that describe where these animals are found. A
Kruskal‐Wallis and chi‐squared test revealed that there were no significant differences
between where killer whales were thought to feed versus where they were not.
Additionally, a chi‐squared test found that there were no differences in the geographic
location of feeding areas versus non‐foraging areas. A density analysis did reveal that all
areas of highest killer whale density were found in the Haro Strait, which is one of the
deepest areas within the archipelago’s marine habitat. This thesis’ findings support
previous studies that the Haro Strait is important habitat for SRKWs but does not
provide evidence that killer whales select for benthic characteristics when feeding.
Table of Contents
List of Figures…………………………………………………………………………………………………………….…….v
List of Tables…………………………………………………………………………………………………………………..vi
Acknowledgments………………………………………………………………………………………………………...vii
Introduction…………………………………………………………………………………………………………………… 1
Methods……………………………………………………………………………………………………………………….11
Field Methods……………………………………………………………………………………………………….. 11
Data Organization and Spatial Analysis………………………………………………………………….. 15
Follow Analysis……………………………………………………………………………………………………… 16
Habitat Analysis…………………………………………………………………………………………………….. 17
Results…………………………………………………………………………………………………………………………. 20
Follow Summaries…………………………………………………………………………………………………. 22
Follow Analysis……………………………………………………………………………………………………… 25
Habitat Analysis…………………………………………………………………………………………………….. 26
Discussion……………………………………………………………………………………………………………………. 31
Follow Type Comparison……………………………………………………………………………………….. 32
Habitat Analysis…………………………………………………………………………………………………….. 33
Conclusion…………………………………………………………………………………………………………………… 39
Literature Cited……………………………………………………………………………………………………………. 41
Appendix A: Follow Types within the San Juan Islands Study Area………………………………..49
Appendix B: Follow Types within the Puget Sound Study Area……………………………………. 50
Appendix C: Bathymetry Map of the San Juan Islands Study Area……………………………….. 51
Appendix D: Bathymetry Map with Killer Whale Density……………………………………………… 52
iv
List of Figures
Figure 1: The San Juan Islands Study Area…………………………………………………………………… 12
Figure 2: The Puget Sound Study Area…………………………………………………………………………. 13
Figure 3: The Grid Coverage of the San Juan Islands Study Area………………………………….. 18
Figure 4: The Grid Coverage of the Puget Sound Study Area……………………………………….. 19
Figure 5: Killer Whale Follow Tracklines in the San Juan Islands Study Area…………………. 21
Figure 6: Killer Whale Follow Tracklines in the Puget Sound Study Area………………………. 22
Figure 7: Histogram Showing the Duration of All Follows…………………………………………….. 23
Figure 8: Histogram Showing the Distribution of Seafloor Depths for Each Follow………. 24
Figure 9: Histogram Showing the Distribution of Seafloor Slopes for Each Follow……….. 24
Figure 10: Histogram Showing the Distance to Shore for Each Follow…………………………. 25
Figure 11: The Spatial Distribution of Follows within the San Juan Islands Study Area…. 27
Figure 12: The Spatial Distribution of Follows within the Puget Sound Study Area………. 28
Figure 13: The Density of Follows within the San Juan Islands Study Area…………………… 29
Figure 14: The Spatial Representation of Areas with High and Low Density Killer
Whale Follows as well as Areas where No Follows Occurred…………………………………. 30
v
List of Tables
Table 1: Days where Killer Whale Follows Occurred…………………………………………………….. 11
Table 2: Definitions of the Type of Behavioral Cues Associated with Foraging……………… 14
Table 3: The Number of Observations for Each Aspect Direction…………………………………. 25
Table 4: The Percentage of Follow Types per Density Areas………………………………………….31
vi
Acknowledgments
Several people provided insight that made the completion of this thesis
possible. First and foremost, I would like to thank Dr. Robin Baird from Cascadia
Research and Dr. Brad Hanson from NOAA for providing me with both the opportunities
to participate in the field data collection and to use these data for my thesis. Both Dr.
Baird and Dr. Hanson also provided valuable feedback in crafting this thesis. I would
also like to thank Dr. Alison Styring from The Evergreen State College for serving as my
primary thesis reader and aiding me in developing the concepts discussed in this
manuscript.
Greg Schorr from Cascadia Research and Candi Emmons from NOAA provided
instruction in the field covering a wide variety of concepts needed for successful field
research. In addition, Candi provided me with the database used for these analyses and
Greg provided equipment specifications.
Dr. Greg Stewart from the Evergreen State College provided a great deal of aid
in both GIS and statistical modeling, without which, accurate analyses may not have
been achieved. Lisa Schlender from Cascadia Research provided assistance and
motivation that were valuable to the completion of this project.
Dr. Tom Carlson, Dr. Peter Impara from the Evergreen State College, Dr. Rip
Heminway from The Evergreen State College, and Damon Holzer from NOAA all
provided GIS help and I am grateful for their support. Al Josephy from the Washington
State Department of Ecology and Dr. Carri LeRoy from The Evergreen State College
provided statistical help. I would also like to thank Ben Anderson from the Computer
Applications Laboratory at The Evergreen State College for help with formatting.
vii
Introduction
There are few animals as easily recognizable as killer whales (Orcinus orca).
Fewer animals seem to be a part of human culture. Until the mid 1970’s, killer whales
were thought to be a predator of humans (Ford et al 2000). Even now, some fishermen
may feel animosity for killer whales, and other marine mammal species, that compete
for dwindling fish populations.
Combined with their interesting (some might say “clown‐like”) coloration
pattern, their public display in aquariums such as Sea World and featured in popular
movies like the Free Willy series makes killer whales known to children and adults alike.
Killer whales are the largest member of the family Delphinidae, the ocean dolphins.
Males can grow up to 9.8m in length and weight 10,000kg and females can grow up to
8.5m long and weigh 7,500kg (Jefferson et al 2008). Their towering dorsal fins and
coloration make them relatively easy to spot and identify in the wild. As the ocean’s top
predator, killer whales serve an important ecological role in the marine ecosystem.
Globally, killer whales are second only to humans as the most geographically
distributed mammal on earth (Jefferson et al 2008). There are, however, areas of
higher‐than‐expected concentrations, such as the waters off of Washington State,
British Columbia, Alaska, Japan, Antarctica, Norway and Iceland (Ford et al 2000). Food
preferences and feeding strategies differ among populations, even in sympatric
populations (Ford and Ellis 1999; Ford et al 2000; Baird 2001). Off the Patagonian coast
of Argentina, killer whales frequently beach themselves to capture southern elephant
seals (Mirounga leonina) and southern sea lions (Otaria flavescens) and drag them into
deeper water (Lopez and Lopez 1985; Hoelzel 1991). Feeding strategies in Norway are
distinctly different as killer whales “herd” schools of herring (Clupea harengus) into
increasingly tight balls, use their flukes to swat at these herring balls, thereby stunning
and then consuming the affected individuals (Nottestad et al 2002).
In the coastal temperate northeast Pacific Ocean, there are three distinctly
different ecotypes of killer whales (Ford et al 2000). These populations have been
termed “residents”, “transients” and “offshore” killer whales. Though it is also
hypothesized that offshore killer whales are not a distinct ecotype but rather are
another population of the resident ecotype since their life histories are very similar
(Robin Baird pers. comm.).
Resident killer whales are primarily fish‐eaters. During predation studies
conducted on the resident killer whale populations in the Kenai Fjords and Prince
William Sound, Alaska, researchers found that 95% of the prey remains came from coho
salmon (Oncorhynchus kisutch) (Matkin et al 1999; Saulitis et al 2000). Further south,
northern resident killer whales (NRKWs), which occupy the waters of Johnstone Strait
and northern Vancouver Island, and southern resident killer whales (SRKWs), whose
1
core summer habitat is the waters surrounding the San Juan/Gulf Island chain of
Washington State and southern British Columbia, diet preferences are distinctly
different. Ford and Ellis (2006) found that 96% of southern and northern resident killer
whale prey was salmon, where 71.5% of all salmonid kills were Chinook (O.
tshawytscha), 22.7% were chum (O. keta) and a combination of coho, sockeye (O.
nerka), pink (O. gorbuscha) and steelhead (O. mykiss) represented the remaining 6% of
salmonid kills. It should be noted that there was a large bias towards NRKWs; 87.5% of
observations involved northern residents whereas only 12.5% involved southern
residents. Similar results have been found for southern resident killer whales in a study
that was independent of any northern resident individuals (Hanson et al in prep).
Even though transient killer whales are sympatric with both resident
populations of Washington State and British Columbia, they exhibit very different
feeding preferences. Unlike their fish‐eating relatives, transients feed primarily on
marine mammals and, on rare occasions, seabirds (Baird and Dill 1995; Ford and Ellis
1999; and Saulitis et al 2000). Despite overlapping ranges, evidence suggests that
transients are genetically isolated from residents—they do not interbreed (Hoelzel et al
1998; Barrett‐Lennard 2000). In fact, when in the vicinity of resident populations,
transients will change their route to avoid the fish‐eating animals (Baird and Dill 1995),
while residents, on the other hand, do not seem to alter their path.
Killer whale ecotype variations are not limited to the waters of Washington
State and British Columbia. Matkin et al (2007) found three similar ecotype populations
in the eastern Aleutian Islands of Alaska. Though not as intensively studied, there
appears to be a distinct difference in mammal‐eating and fish‐eating populations in
Antarctica as well (Berzin and Vladimirov 1982). In Antarctic waters, Type A and B are
primarily marine mammal eaters (Smith et al 1981; Berzin and Vladimirov 1983) while
Type C are fish‐eaters (Berzin and Vladimirov 1983). All three types of Antarctic killer
whales are morphologically different (Pitman and Ensor 2003).
There are numerous studies on resident populations that ascertain foraging
behavior. Current studies (Hoelzel 1993; Baird and Hanson 2004; and Ford and Ellis
2006) have found that resident killer whales possibly exhibit certain behavioral patterns
when foraging. A study on SRKWs revealed that a series of surface activity, including
rolls and turns, was common when foraging (Hoelzel 1993). Baird and Hanson (2004)
and Ford and Ellis (2006) found a suite of subtle behavioral cues that could indicate
hunting or successful capture of a fish. Ford and Ellis (2006) studied both northern and
southern resident communities and found that, when foraging, killer whales often swam
in zig‐zag patterns rather than straight lines. Directional and non‐directional chases,
long dives, and convergence of whales were also indicators of foraging behavior. Baird
and Hanson (2004) studied southern residents in the summer of 2004 and recorded
twenty‐seven behavioral cues, which included thirteen fast non‐directional surfacings
2
and eight moderate non‐directional surfacings. Ten of these twenty‐seven cues yielded
evidence of predation (fish scales and/or bits in the water).
Due to their cosmopolitan distribution, assessing killer whale conservation
status on a global scale rather than localized is very challenging. In 2008, their global
status was not determined because data were deficient (International Union for
Conservation of Nature and Natural Resources 2008). However, when tracking the
status of stocks, or meta‐populations, the task can be more manageable. Since 1976,
The Center for Whale Research has conducted a long‐term population assessment on
southern resident killer whales (Center for Whale Research 2008). At the time this
thesis was written, there were eighty‐seven individuals in the southern resident
community divided into three pods, J‐pod (n=25), K‐pod (n=20), and L‐pod (n=42).
Though this number is up by eleven individuals from when surveys first started, it is one
of the lowest population estimate numbers since 2002 and reflects a declining
population trend since the mid‐ and late 1990s when the population was as high as
ninety‐seven (Center for Whale Research 2008). As a result of these troubling trends,
the southern resident killer whale community was listed as endangered under the
Canadian Species at Risk Act (SARA) in 2002 and the United States Endangered Species
Act (ESA) in 2005. The challenges they face on the road to recovery are discussed
below.
Live Capture
The coastal temperate northeast Pacific population of killer whales was an
intensively used resource for the “live‐capture” industry for marine‐theme parks such as
Marineland, the Vancouver Aquarium, and the Seattle Aquarium (Center for Whale
Research 2008). The first animal was taken in 1961 and the “fishery” was closed in 1976
(Asper and Cornell 1977; Hoyt 1990; and Center for Whale Research 2008). As many as
303 individuals were captured from the waters of Washington, British Columbia and
California with fifty‐six kept (Asper and Cornell 1977). Many of the 303 animals
captured were repeated captures of the same individuals (Robin Baird pers. comm.).
There were individual mortalities associated with these capture attempts as well, which
were as high as ten (Asper and Cornell 1997) or eleven (Hoyt 1990). As many as sixteen
individuals were reported to have died within their first year of captivity (Hoyt 1990).
Community membership (i.e. a NRKW individual or a SRKW individual) was poorly
understood during the years of harvest. However, based on what is now known about
the range of each killer whale population, it can be safely assumed that a substantial
number of killer whales taken were from the southern resident community.
The decline of the SRKW population in the 1980s can most likely be attributed
to the intensive live‐capture industry of the previous two decades (Baird 2001).
However, given that the population rebounded in the mid‐ and late‐1990s it seems
3
unlikely that the history of the live‐capture industry continues to have an impact on the
southern resident killer whale community.
Salmon
As previously mentioned, salmon are the preferred prey of the southern
resident killer whales. Population numbers of all salmon species have declined in many
areas in Washington State and British Columbia (Augerot et al 2005). Understanding the
reasons for these declines is as complex as it is perilous and to provide an exhaustive
description is beyond the scope of this thesis. However, a few of the most widely‐
accepted reasons as to the current state of Pacific salmon (Oncorhynchus spp.), known
as the “4 H’s”; Harvest (overfishing), Hydro (damming of rivers), Hatcheries (artificial
propagation), and Habitat (loss of habitat) (Lichatowich et al 1999) are discussed below.
Of the “4 H’s”, Harvest is the most complex and dynamic. There are many rules
and regulations that allocate certain percentages of the quota to native and subsistence
fishermen and to recreational and commercial fishermen, time of year salmon can be
fished, who can fish, and who manages the fish (National Marine Fisheries Service
2007). Equally complex is understanding the toll overfishing takes on the salmon
population as a whole. While it is generally thought that overfishing is of lesser
importance than habitat loss (National Marine Fisheries Service 2007), it has also been
found that overharvesting is linked to population trends of Chinook salmon in both the
large scale area of the Pacific Northwest and the small scale area of Puget Sound
(Hoekstra et al 2007).
The intensive commercial salmon fishery began in 1866 (Lichatowich et al 1999)
and was so efficient that specific river runs were closed by 1915 (National Marine
Fisheries Service 2007). In British Columbia, total tonnes of all salmon species caught by
the commercial fishery declined by 62% from 1952 to 2004, declines were seen in all
species individually except chum salmon from the same time period (Irvine et al 2005).
In the Columbia River, total catch of all species declined drastically from 1866 to 1993
(Lichatowich 1999). Between 1976 and 2000, Chinook salmon catch in Washington
State declined by 84% (National Marine Fisheries Service 2007). In addition to the large
number of individuals being removed from the population, there are evolutionary
ramifications to intensively fished salmon populations. Although still under debate,
selection for salmon size as well as timing of fisheries could be causing a shift to smaller
salmon in general and a shift in timing for the migration back to native rivers to spawn
(Hard et al 2008).
The damming of rivers has also led to the decline of Pacific salmon. By the
1930s, dams designed to either generate hydroelectricity, divert water for irrigation, or
to create reservoirs for stored drinking water were being constructed throughout the
Pacific Northwest (Lichatowich et al 1999). Dams alter the characteristics of a natural
4
river system by decreasing flow velocity and changing water temperature, as well as
impact water quality by draining irrigated areas of water full of sediment and
agricultural contaminants (Waples et al 2007). Additionally, Waples et al (2007)
reported that river migration routes, especially for juvenile salmon migrating to
estuaries, have been severely impacted by dams.
The mitigation tactic of barging fish around these dams has been undertaken.
However, in a Columbia River case study (Keefer et al 2008), this process has been
shown to alter the return migration of both Chinook salmon and steelhead. Both
species are known for their homing ability (finding their way back to the stream in which
they were spawned) with some “not finding their way home” and returning to a
different river to spawn (straying). It is theorized that salmon collect information about
their home stream and their spawning grounds while swimming downstream (Quinn
2005), which is how they locate their native rivers. Keefer et al (2008) found that
juvenile Chinook salmon and steelhead that had been barged around dams were more
likely to stray as adults. The Columbia River is one of the most hydroelectrically‐
developed river systems in the world (Pacific States Marine Fisheries Commission 1997;
Augerot et al 2005) and many of its dams are impassable. Those that are passable are
treacherous as they are thought to kill 70‐96% of salmon juveniles swimming
downstream (Pacific States Marine fisheries Commission 1997). In Puget Sound, of the
“4 H’s”, dams may have the largest impact on salmon density (Hoekstra et al 2007).
Hatcheries present an interesting paradox as they were originally designed to
address the problem of declining salmon stocks (Lichatowich 1999). Hatchery‐reared
fish are those fish raised in human‐controlled environments (hatchery facilities or
hatcheries) with the purpose of increasing the number of salmon in the fishery.
Hatchery‐reared salmon are raised in very high densities with specialized diets and
conditions designed for increased survival of juvenile fish. The success of hatcheries has
been far from what was expected, however. The program as a whole has shown little, if
any, positive results (Lichatowich et al 1999) and over‐shadows the true issue of
declining wild salmon runs in the northwest (Larry Dominguez pers. comm.) potentially
increasing the declining rate of wild fish (Quinn 2005). Though interactions are complex,
there is evidence to suggest that hatchery fish may hinder the ability of wild fish
populations to recover (Levin et al 2001). Although some studies have shown evidence
to the contrary, it is generally shown that hatchery‐reared salmon are more aggressive
and larger than wild salmon within the same stream (Weber and Fausch 2003) creating
a population of salmon capable of out‐competing the native population for resources.
The sheer number of hatchery fish placed into a river may overwhelm native fish.
Broodstock used by a given hatchery may not be from that specific watershed thus
potentially altering of the gene pool (Weber and Fausch 2003; Goodman 2005).
Additionally, hatchery‐reared salmon suffer higher mortality rates than wild salmon
because hatchery fish do not have well‐developed predator‐avoidance behavior (Olla et
5
al 1994; 1998). There is also evidence to suggest that hatchery salmon may be less fit to
survive in the wild due to artificially controlled “favorable” conditions in hatcheries
(Araki et al 2008). Araki et al (2008) also summarized studies showing that hatchery
environments substantially increase the survival rate during the egg‐to‐smolt part of the
salmon life cycle, therefore unfavorable genes expressed by individuals that would
otherwise be filtered out via natural selection by the environment are now surviving
past the smolt stage, introducing these genes into the gene pool. The effectiveness of
hatcheries is still in question but there are a number of concerns about their impacts on
wild salmon.
The last of the “4 H’s”, and perhaps the most important in terms of its impact on
salmon, is loss of habitat. Salmon are anadromous, meaning they utilize a wide variety
of freshwater, estuarine, and oceanic habitat for different parts of their life cycle. In the
freshwater environment, there are numerous factors related to salmon survival, but two
of the most well‐known are gravel size and flow regimes. In general, the finer the
gravel, the lesser of value it serves as salmon spawning habitat. Fine gravel blocks the
exchange of good quality water into the salmon redds (“nests”) and can rob developing
salmon alevins in the gravel of water rich in oxygen (Quinn 2005). Development of land
within a given watershed can also have impacts on the distribution of salmon within
that watershed, causing salmon to shift to areas of less development and away from
areas of high development (Burnett et al 2007; Bilby and Mollot 2008).
As salmon begin to acclimate to the marine environment, they rely on estuarine
habitats to complete their physiological changes (known as smolting). These valuable
nearshore habitats have undergone tremendous anthropogenic changes. Estuarine
wetlands, which provide juvenile salmon areas for foraging and hiding from predators,
have been reduced to a fraction of their historic Puget Sound distribution (Tanner et al
2002; Puget Sound Action Team 2007). In the Fraser River delta of southern British
Columbia, less than 1% of valuable wetland habitat remains from historic times
(Environment Canada‐British Columbia Ministry of Environment, Lands and Parks 1992).
Altering shoreline habitat could also alter the behavior and the distribution of salmon
(Toft et al 2007).
It is unclear how southern resident killer whales will respond as food resources
continue to dwindle. Currently, there is no reason to suspect that they would switch to
another salmonid species. Chinook salmon are already the least abundant of salmon
species found in SRKW habitat (Quinn 2005) and although other salmon are readily
available, they are rarely taken. Puget Sound Chinook salmon populations are at only a
fraction of what their historical levels were (National Marine Fisheries Service 2007),
and the lack of their population restoration is undoubtedly hindering the recovery of
southern resident killer whales.
Pollution
6
Puget Sound has a long, rich history of human‐introduced contaminants into its
waters. The Puget Sound Action Team (2007) noted that degraded sediments, high
levels of polychlorinated biphenyls (PCBs), polybrominated diphenyl ethers (PBDEs),
polycyclic aromatic hydrocarbons (PAHs), and other persistent organic pollutants (POPs)
are impacting the ecosystem. Many of these toxins are capable of accumulating in the
tissue of living organisms and increase in levels and magnify in effects as they move up
the food chain (Puget Sound Action Team 2005), a process known as bioaccumulation.
Understanding the complexity of pollution inputs into the food web used by
southern resident killer whales is no easy task. The Puget Sound/Georgia Strait basin
surrounding the prime summer habitat of these whales is heavily populated by humans
and is projected to keep growing. In 2000, 7 million people lived in basin, and over 9
million are expected to reside in the area by 2020 (Fraser et al 2006). The increasing
population is going to call for more development of land. Nutrients, metals, toxic
chemicals and other contaminants are finding their way into the Sound via storm runoff
that is unable to penetrate and soak into the ground where development has occurred
(Washington State Department of Ecology 2009a). This has been cited as the number
one cause of pollution in Puget Sound (Puget Sound Partnership 2008). PCBs and
PBDEs, both suspected to have impacts on the health of the southern resident killer
whales, were developed primarily as electrical insulators and flame retardants,
respectively (Ross 2006). Despite having been banned in the United States and Canada
in 1977, PCBs are still present in the environment (Ross et al 2000; Missildine et al 2005;
Ross 2006; Krahn et al 2007; and Cullon et al 2009). In 2007, Washington State became
the first American state to ban PBDEs, however, unfortunately due to their persistent
nature similar to PCBs, they will most likely be found in the ecosystem for decades to
come. In fact, PBDEs are doubling every four years and by 2020 will surpass PCB levels
in many species, making it the most “important” contaminant in the habitat used by
southern resident killer whales (Puget Sound Action Team 2007).
In addition to causing cancer in animals, the Environmental Protection Agency
(EPA) reported that PCBs damage the immune, endocrine, reproductive, and nervous
systems (2008). PBDEs are not as well understood but animal studies show that PBDEs
impact brain development and impact an animal’s ability to learn (Washington State
Department of Health 2009). With chemical structure comparable to PCBs, long‐term
studies may reveal that these two compounds may have similar impacts on organisms.
Due to their industrial nature, these pollutants are found at much higher levels
in aquatic environments in close proximity to urban areas. Missildine et al (2005)
reported that Chinook salmon returning to hatcheries within the highly urbanized Puget
Sound had almost 2.5 times the PCB concentrations than Chinook salmon returning to
hatcheries along the sparsely populated Washington coast. Chinook salmon in the
relatively pristine Johnstone Strait of northern British Columbia had lower levels of PCBs
than Chinook found in the lower Fraser River (Vancouver, BC), Duwamish River (Seattle,
7
WA) and the Deschutes River (Olympia, WA) (Cullon et al 2009). PBDE concentrations in
English sole, herring, and harbor seals located within Puget Sound were much higher
than those species populations found within the Georgia Strait (Puget Sound Action
Team 2007).
The persistence of these chemicals is reflected in southern resident killer
whales. PCB and PBDE burdens in southern resident killer whales are three and five
times higher, respectively, than northern resident killer whales (Ross 2006). Unlike
southern residents, northern resident killer whale habitat is sparsely populated and has
been spared the habitat degradation experienced by the southern population. It should
be noted that transient killer whales have approximately 60% higher concentrations of
PCBs (but no real difference in PBDE concentrations) (Ross 2006). This probably reflects
the fact that transients feed on other marine mammals, which are higher up the food
chain, increasing the likelihood of higher PCB concentrations due to bioaccumulation.
PCB concentrations increase with age in males from both northern resident (Ross et al
2000) and southern resident killer whales (Krahn et al 2007). There is no evidence that
PCB concentration increases over the lifespan of females until post‐reproductive stage
(Ross et al 2000). Since PCBs and PBDEs are lipid‐soluble, these toxins are passed from
mother to offspring via milk, thus females offload a lot of their toxic burden to their
offspring. Therefore, it makes sense that J1, the oldest male in the southern resident
community would logically have the highest burden of PCB contaminants in the
population (Krahn et al 2007). PCB concentrations in sampled southern resident killer
whales are significantly higher than health effect thresholds (Ross et al 2000; Ross 2006;
and Krahn et al 2007). PBDEs were not age specific, which could reflect their continued
use and introduction into the environment long after the PCB ban. In Washington State,
a law was passed that calls for the gradual phasing out and banning of PBDEs
(Washington State Department of Ecology 2009b). Given that PBDEs are in many
products consumed in the northwest, the absence of a law banning PBDEs in British
Columbia, and their persistence in the aquatic environment, it is unclear what success
the Washington State ban will have in reducing PBDEs in the environment.
Despite no real difference in PBDE concentrations between southern residents
and transients, there does seem to be evidence that SRKW concentrations are higher
than northern residents (Rayne et al 2004). The pristine conditions of the northern
resident population habitat when compared to the southern resident’s is the most likely
explanatory variable for differences in contamination since they both feed at the same
level on the food chain. More troubling is as the human population increases in the
Puget Sound/Georgia Strait basin, forward planning and management for habitat
restoration will be challenging at best.
Effects from Boats
8
Southern resident killer whales’ core summer habitat is surrounded by a high
number and density of people. As a result, many large commercial shipping vessels
frequent the waters carrying goods to the millions of people living in the Puget
Sound/Georgia Strait region. Additionally, killer whales are charismatic megafauna,
therefore enthusiastic whale watchers congregate on commercially operated and
recreational private vessels, making southern residents easily accessible to a large
number of people.
Globally, whale watching has been a very profitable industry. From 1994 to
1998, the total number of whale watchers increased from 5.4 million to 9 million, with
an expenditure increase from $504 million to $1.049 million (Hoyt 2001). The whale
watch operations focused on southern resident killer whales has also experienced an
increase in the number of paying customers. From 1990 to 2000, the number of whale
watching vessels increased fivefold (Foote et al 2004). In 2005, there were a total of 74
commercially operated boats from various Canadian and American ports that carried
more than 250,000 people to view southern resident killer whales (Koski 2005).
The potential impacts of whale watching on killer whales are one of the more
recently studied phenomena. A study on northern resident killer whales in the
Johnstone Strait revealed that killer whales swam away from shore and out towards
open water when boats were present, but did not alter their speed or spacing (Jelinski
et al 2002). It was also observed that vessels violated both a motorized vessel‐restricted
area and recommended distance buffer from the whales. A study of southern resident
killer whales revealed that boat presence within 400 meters significantly decreased the
time the animals spent foraging and significantly increased the time the animals spent
traveling when compared to boat absence (Lusseau et al 2009). An increase in the
number of boats also led to a decrease in the amount of time between breaths, altered
dive durations, and increased swim speed (Williams et al 2009). Whales being cut off by
boats alter the direction of their migration path (Willams et al 2002).
Foote et al (2004) found that southern resident killer whales may adjust their
vocal calls in response to anthropogenic noise. From 2001 to 2003, they found that
killer whale call duration increased in the presence of boats. Although they did not find
a significant increase in call duration when exposed to vessel noise, Holt et al (2008) did
report that killer whales increase their vocal amplitude when exposed to higher levels of
ambient noise. The energy expended by killer whales in the presence of vessel noise
could also be higher than in the absence of boats (Williams et al 2006). Williams and
Ashe (2007) hypothesized that killer whale evasive tactics in the presence of boats
mimics evasive tactics used by other animals when being pursued by predators.
Considering all of this, the importance of whale‐watching should not be understated. In
addition to providing valuable economic opportunities, they also provide conservation
opportunities and give people a chance to see killer whales in their natural habitat.
9
The risk factors mentioned above are among the most common reasons as to why
the population of southern resident killer whales has declined and its lack of recovery.
These risk factors must be addressed if the recovery of the Puget Sound/Georgia Strait
population of this iconic species is to be achieved. The purpose of this thesis, however,
is to examine feeding habitat. Many cetacean habitat studies focus, at least in part, on
benthic characteristics, and/or spatial distribution in defining habitat characteristics for
a given population (for reviews of some of these studies please see the discussion
section). This thesis will use both benthic characteristics and spatial distribution to
determine differences between where killer whales were successful and where they
were not. From 2004‐2007 southern resident killer whales were studied in their core
summer habitat around the San Juan Islands, and to a lesser extent in Puget Sound
proper. Types of killer whale observed behaviors were numerous but were generally
defined as feeding (evidence of successful predation event), possibly feeding (no
evidence of successful predation event but behaviors that are thought to be signs of
foraging (Baird and Hanson 2004; and Ford and Ellis 2006)) and probably not feeding (no
evidence of successful predation event and no behavioral cues). The primary purpose of
this study was to:
•
Examine benthic habitat and spatial distribution of each of these killer whale
follow types and test for differences in benthic depth, benthic slope, benthic
aspect and distance to shore (habitat variables).
Additionally, areas of high killer whale densities were identified during this study. A
secondary question to this thesis became:
• Are there any benthic habitat characteristics guiding killer whale occurrence
densities?
Since it is logical to assume that predator foraging habitat is guided by prey distribution,
the results may provide as much insight into salmon habitat preference as much as it
does killer whale foraging habitat preference. As the Chinook salmon population
continues to decline, understanding where killer whales hunt for these fish is crucial if
SRKW conservation and restoration is to be achieved. This thesis will attempt to identify
critical killer whale feeding habitat.
Understanding how southern resident killer whales use habitat for feeding could
potentially have implications for their recovery. There are numerous studies (e.g.
Heimlich‐Boran 1988; Hoelzel 1993; and Hauser 2006) that identify the Haro Strait as
important habitat for these killer whales, but to‐date, no studies quantify feeding
habitat. Learning more about where this population feeds could, for example, identify
patches of marine protected areas that would guide management regimes. Quantifying
SRKW feeding habitat could create buffers of “no‐take” zones protecting Chinook and
other salmon species, reserving those resources in these newly‐created areas for killer
10
whales. In addition, boater restriction zones or areas of increased boater regulation
could be implemented.
Methods
Field Methods
Fieldwork was conducted during spring, summer and fall months during 2004‐
2007 (Table 1).
Table 1. Days Where Killer Whale Follows Occurred.
Year
2004
2005
Day
29‐Aug
30‐Aug
31‐Aug
1‐Sep
6‐Jun
7‐Jun
8‐Jun
9‐Jun
10‐Jun
12‐Jun
13‐Jun
6‐Jul
7‐Jul
8‐Jul
9‐Jul
11‐Jul
12‐Jul
9‐Aug
11‐Aug
12‐Aug
13‐Aug
15‐Aug
26‐Oct
30‐Oct
11
Year Day
2006
16‐May
18‐May
19‐May
21‐May
22‐May
23‐May
24‐May
13‐Jun
14‐Jun
15‐Jun
16‐Jun
17‐Jul
18‐Jul
19‐Jul
20‐Jul
19‐Sep
2007
7‐Jun
8‐Jun
10‐Jun
11‐Jun
12‐Jun
13‐Jun
14‐Jun
16‐Jun
17‐Jun
18‐Jun
19‐Jun
20‐Jun
6‐Sep
10‐Sep
11‐Sep
14‐Sep
15‐Sep
The primary sampling locations were the waters of the San Juan Islands in
Washington State (Fig. 1), with additional samples collected in Puget Sound (Fig. 2). The
research boat used was a 6.3m vessel with a custom pulpit designed for better visibility
and sample collection. Killer whale locations were generally known prior to the
research vessel leaving port as their movements within the inland waters of Washington
State are well observed and reported via citizen monitoring groups such as the Orca
Network or the commercial whale‐watch pager network which has been proven to be
very accurate in locating killer whales (Hauser 2006). On a few occasions, the general
location of the whales was unknown and the research vessel was launched in order to
find the animals.
Figure 1. The San Juan Islands Study Area.
12
Figure 2. The Puget Sound Study Area.
While on the water, a Garmin GPS (various models) was used to automatically
record the vessel location in five‐minute increments (referred to as timestamps
hereafter). General effort and survey information such as time of departure, weather
and sea state conditions was also recorded. Sea state conditions were continuously
monitored and updated.
Upon first observation of killer whales, the time of day, approximate number of
animals, probable pod membership (J, K, L, or some combination), general direction of
animal movement, and group envelope (geographic spread of group) was recorded.
This was termed an encounter. Encounters lasted for as long as the vessel was generally
close to the whales. Most days typically contained only one encounter and the
conclusion of an encounter generally meant the sampling day was complete. There
13
were a few days, however, that contained more than one encounter where the boat
would leave the whales and re‐establish contact later in the day. Once a group of
whales was spotted, the research vessel would stay in the general geographic vicinity of
that group until a smaller subset of the group became the target of more directed
follows.
A follow was defined as when a whale or group of whales was being tracked in
close proximity to the boat. To collect the data, the research vessel needed to follow
one or multiple animals and record information on movements and behavior. Follows
were classified as dedicated or opportunistic depending on how close the boat was and
footprint or side depending on if the vessel was directly behind the animal(s) or on
its/their side. When a follow started the exact time was recorded.
The number of animals and exact or probable identification of the whale were
recorded at the beginning of each follow. If the whale(s) demonstrated behavioral cues
(see Table 2 for a list) the type of cue and the exact time it occurred were recorded.
Additionally, if any samples were collected from the water, such as scales and/or bit
remains of fish (predation samples), fecal material from a whale, or other non‐
identifiable material, the time of these collections and location were also noted.
Table 2. Definitions of the Type of Behavioral Cues Associated with Foraging.
Behavioral Cue
Long Dive
Fish in Water
Directional Surfacing
(Moderate or Fast)
Directional Change
Non‐Directional Surfacing
Chase
Convergence
Change in Speed
Fish in Mouth
Milling
Brief Definition
A longer than average time spent below the surface of
the water
A fish spotted in close proximity to a whale
A surfacing that occurred with more velocity than
average surfacing but in the same direction as
previous surfacing
A sudden change in direction
A surfacing where the whale's direction changed by
more than 45 degrees from previous surfacing
A whale pursuing prey
A joining of two or more whales
A whale or group of whales speeding up or slowing
down
A whale spotted with a fish in its mouth
A multi‐directional slow movement
14
Follows ended for a number of reasons. If there was immigration or emigration
from the group being followed, that follow would end and a new follow would
immediately begin, reflecting a change in the group. A research permit was not
available for Canadian waters; therefore, if the whales entered Canada, the follow (and
the encounter) would end. Other reasons for ending a follow included a large number
of whale‐watching boats present, the animals getting too close to shore in areas with
high human habitation or the determination that a desirable amount of data had been
collected, and the vessel had the opportunity to follow a new whale or group of whales.
Data Organization and Spatial Analysis
The data collected from the field were transcribed into an Access database
(Microsoft Corporation, Redmond, Washington). To better work with the data,
spreadsheets were built using Microsoft Excel. All GPS timestamps and
latitude/longitude were imported into Excel spreadsheets and grouped by study day.
Encounters and follows were then added to the corresponding timestamp (it is
important to note here that since timestamps were in five‐minute intervals all data
added to the spreadsheet were added to the appropriate timestamp and not to its exact
time. For example if a follow began at 11:37 it is possible that the closest timestamp
was 11:35 and thus does not reflect actual start time of the follow). Number of
individuals followed, life history of the individual (individual name, matriline and pod
memberships, and gender), behavioral cues, if any, and collections of samples from the
water were also added to the spreadsheet in the same way.
Since the GPS recorded timestamps in five minute intervals starting from the
moment the boat left the dock, recording location even when follows were not taking
place, it was necessary to remove the timestamps that did not correspond to an actual
follow, leaving only those timestamps corresponding to follows. These timestamps
were imported as points into a GIS project using ArcView 9.2 (ESRI, Redlands, California)
and overlayed with a 90 meter resolution digital elevation model (DEM) (NOAA’s
National Geophysical Data Center, Boulder, CO,
http://www.ngdc.noaa.gov/mgg/coastal/coastal.html) that describes the depth of the
seafloor surrounding the San Juan Islands and Puget Sound. The spatial Analyst
extension was used to calculate both the slope of the seafloor (with the corrected z‐
factor value) and the aspect of the seafloor. Using Hawth’s Tools (Hawthorne Beyer,
www.spatialecology.com) the depth (meters), slope (degrees) and aspect (degrees) of
the seafloor were extracted for each point. A polygon shapefile representing
Washington State (Washington State Department of Ecology, Lacey, WA,
http://www.ecy.wa.gov/services/gis/data/data.htm) was then added to the map
project. Using the spatial join function, the distance to shore (meters) was calculated
for each point. All of the GIS‐derived variables were exported back into Microsoft Excel.
15
Benthic habitat points are highly correlated to one another, creating a problem
for analysis because data are not independent. To address this, waypoints were not
analyzed as raw data. Waypoints of corresponding follows had their depth, slope,
aspect, and distance to shore data (hereafter referred to as habitat data) averaged to
create one value for each habitat variable per follow. Though this does not remove all
dependency it does minimize it. Follows then had to be divided based on definitely
feeding, possibly foraging, and probably not feeding/foraging. Follows where predation
samples were collected were said to have had evidence of feeding and were called
“Feeding follows”. Follows where there were no predation samples collected but there
were behavioral cues generally associated with foraging were thought to be associated
with possibly foraging whales were given the designation “Potentially Foraging follows”.
Follows where there were no predation sample collections and no behavioral cues
observed to give any indication of feeding or foraging and were termed “Traveling
follows”. Because of their short duration, all follows less than five minutes (n=133) were
removed from the analysis and will not be represented in the data from this point on.
In addition, Hawth’s Tools were used to generate tracklines from the boat’s 5‐
minute waypoints to generate general movement patterns of the boat during killer
whale follows.
Follow Analysis
As mentioned above, to analyze the follows, waypoints of the same follows
were averaged to generate one value for each of their habitat variables. Aspect values,
which are directional data based on 0‐360 degrees in order to quantify the cardinal
directions, had to be converted to radians and have their sine and cosine values
calculated to get the true average direction. For an example of this, consider two
angles, 0 and 359 degrees, which are both quantitative representations of the direction
north. A simple average of the two numbers generates a value of 179.5 degrees, which
corresponds to an aspect of south. Since both observations correspond to north, the
direction south cannot be the average but when calculating the average by first
converting to radians, the value generated is 359.5, which is north.
Histograms for follow duration, benthic depth, benthic slope, and distance to
shore were created. For the follow duration histogram, the bin range was 0.25 decimal
hours. However, for each of the habitat variable histograms created, the bin range was
calculated using a formula designed by Scott (1979) because he presents evidence that
his formula provides the most accurate bin range for visually showing normality of data.
The next step was to evaluate differences in these variables between follows.
Depth, slope, and distance to shore data were first imported into the statistical package
R (http://www.r‐project.org) and tested for normality using the Shapiro‐Wilks test.
When all three datasets failed, they were placed back into Microsoft Excel and a non‐
16
parametric Kruskal‐Wallis test was used by leveraging the statistiXL package (Nedlands,
Western Australia, http://www.statistixl.com).
Aspect values, which are directional in nature, were analyzed differently.
Aspect values of each follow type were compared to a von Mises circular normal
distribution by using a modified chi‐square test (Greg Stewart pers. comm.).
Additionally, benthic depth, benthic slope, and distance to shore were pooled
for Feeding follows and Potentially Foraging follows and compared to Traveling follows
via Kruskal‐Wallis test.
Habitat Analysis
To examine spatial distribution of these follows within Washington State inland
waters, a grid system was created. Two grid coverages were created in GIS, one covered
the San Juan Island Study Area (Fig. 3) and the other covered the Puget Sound Study
Area (Fig. 4). Each cell was approximately 4.85km2. This cell size was determined to
match a previous grid study (Heimlich‐Boran 1988). Note that in the San Juan Islands,
the study area and some of the cells are irregularly shaped. This is because the research
was conducted in Canadian waters and the grid area was bounded by the international
border.
17
Figure 3. The Grid Coverage of the San Juan Islands Study Area.
18
Figure 4. The Grid Coverage of the Puget Sound Study Area.
Within the cells, each follow is represented by one point. All follows were
assigned to one cell, whether the follow stayed within that cell or not. If the follow
stayed entirely within one cell, the point representing the follow corresponds with the
first waypoint of the follow. If the follow spans two or more cells, the follow was
assigned to the cell where the most number of waypoints for that follow were, and
marked by the first waypoint within that cell. For follows that span two or more cells
but have equal number of waypoints in each cell, the follow was assigned to the cell
where the first waypoint was recorded.
The cell number for each follow was recorded. Follow duration for each follow
was also recorded and used to calculate the amount of time per follow type in each cell,
and the total amount of time spent in each cell. A chi‐square test was to assess whether
19
there were significant differences among follow type and follow type duration in each
cell.
To determine areas of higher killer whale follows, Hawth’s Tools were used to
calculate spatial density for the follows within the San Juan Islands Study Area. This was
not done in the Puget Sound Study Area because of the small number of follows in this
portion of the study area (n=29) when compared to the San Juan Islands study area
(n=443) and because there are only three clusters of points confined to the main Puget
Sound channel between Point No Point and the city of Tacoma.
The kernel density calculations were then classified into 3 areas, areas with a
high density of follows (0.04‐0.07 follows per square kilometer), areas with low density
of follows (0.002‐0.04 follows per square kilometer), and areas with no follows. These
areas were converted into a series of points and Hawth’s Tools was used to extract
depth, slope and aspect for each point. These data were used to find the average value
per variable for each area.
Results
From 2004‐2007 a total of approximately 170 hours of focal follows were
conducted. Figures 5 and 6 show the tracklines of all follows for the San Juan and Puget
Sound Study Areas respectively.
20
Figure 5. Killer Whale Follow Tracklines in the San Juan Islands Study Area. The lines indicate
follows where there were at least two waypoints whereas the dots display follows where only
one waypoint was recorded.
21
Figure 6. Killer whale Follow Tracklines in the Puget Sound Study Area. The lines indicate
follows where there were at least two waypoints whereas the dots display follows where only
one waypoint was recorded.
Follow Summaries
There were a total of 472 follows used in the analysis. Of those 472 follows,
16.74% were follows with confirmed successful predation events (Feeding Follows),
25.42% were follows with no confirmed predation events but probable foraging based
on foraging behavior (Potentially Foraging Follows), and 57.84% were follows with no
predation events and no foraging cues (Traveling Follows). The average follow lasted
twenty‐one minutes. Despite the fact that most follows were Traveling follows,
Traveling follows had the lowest average follow duration with eighteen minutes
whereas the average feeding and Potentially Foraging follows were twenty‐six minutes
and twenty‐five minutes, respectively. The differences in follow durations were
significant (H=22.75, df=2, p‐value=.00001). Figure 7 shows the histogram of all follow
22
durations, It should be noted that about 81% of all follows (n=381) fall within the first
two bins, indicating that follows with longer durations are uncommon.
Figure 7. Histogram Showing the Duration of All Follows.
The average depth for all follows was about 156 meters (standard deviation
approx 79). The average slope was 5 degrees (standard deviation approx. 8). The
average distance to shore was 2 kilometers (standard deviation approx. 3). For a review
of the distribution of these three variables see figures 8‐10. Since aspects are
directional values based on degrees of a circle, a histogram of a conventional
distribution does not apply. Table 3 shows the number of follows for each cardinal
direction, with their corresponding range of degrees. Of the nine possible directions, an
aspect of southwest is the overwhelmingly favored direction, representing 28.6% of all
aspect observations with an aspect of west in a distant second at 14.6%. A direction of
“flat”, which means there is no aspect because the slope at that point is zero, represents
the smallest percentage of observations at 2.5.
23
Figure 8. Histogram Showing the Distribution of Seafloor Depths for Each Follow.
Figure 9. Histogram Showing the Distribution of the Seafloor Slopes for Each Follow.
24
Figure 10. Histogram Showing the Distance to Shore for Each Follow.
Table 3. The Number of Observations for Each Aspect Direction.
Degrees
0‐22.5, 337.5‐
360
22.5‐67.5
67.5‐112.5
112.5‐157.5
157.5‐202.5
202.5‐247.5
247.5‐292.5
292.5‐337.5
Cardinal
Direction
North
Northeast
East
Southeast
South
Southwest
West
Northwest
‐1 Flat
Number of
Follows
53
32
21
37
59
135
69
54
12
All of the above summary data were conducted on data where the follows were
the unit of analysis. Each habitat variable for all waypoints within a follow were
averaged together to create one value per variable per follow.
Follow Analysis
To compare the depth, slope, and aspect of the seafloor between all three
follow types, a Kruskal‐Wallis test was conducted. The test did not produce significant
results for depth (H=1.57, df=2, p‐value=0.456), slope (H=0.598, df=2, p‐value=0.741) or
25
distance to shore (H=0.877, df=2, p‐value=0.645). A modified chi‐square test was
leveraged to compare the aspect values to a von‐Mises circular distribution. The results
show that each type of follow differs from a normal circular distribution (for Feeding;
χ2=37.1818, df=7, p‐value<0.05, for Potentially Foraing; χ2=29.5042, df=7, p‐value<0.05,
and for Traveling; χ2=91.81818, df=7, p‐value<0.05). Though these results do indicate
there is a potential difference in observed aspect values compared to a normal
distribution, there doesn’t appear to be any differences between follow types as
southwest makes up the majority of observations in each follow followed by west. This
is most likely explained by the fact that west and southwest facing aspects are the most
common in the study area (see below) and not by an actual selection for these habitat
criteria. In all cases east make up the least observed true direction for the seafloor for
each follow. A “flat” aspect is the least observed absolute value for all follow types but
was removed from the analysis because it is not a true direction on a circular
distribution.
For the analysis that compared the pooled Feeding and Potentially Foraging
follows to the Traveling follows, no significant differences existed for depth (H=0.143,
df=1, p‐value=0.705), slope (H=0.544, df=1, p‐value=0.461) or distance to shore
(H=0.784, df=1, p‐value=0.376).
Habitat Analysis
Figures 11 and 12 show the spatial distribution of the killer whale follows for the
San Juan and Puget Sound gridded study areas respectively. Each point represents one
follow.
A Chi‐Squared analysis was conducted to examine differences in follow type
duration per cell. Although there were two cells (one in each study area) that produced
significant results, these cells only had Traveling follows. None of the other cells that
had multiple follow types produced significant results and for the purposes of this study
it will be assumed that there are no significant differences in follow type durations over
the study areas as a whole. Individual Chi‐Square values will not be reported separately
in the text as there were too many cells examined and the vast majority was not
significant. In the San Juan Islands Study Area, the types of follows were spread
throughout the study area with the largest proportion of each follow type occurring in
the waters off the west and southwest coast of San Juan Island. In Puget Sound, on the
other hand, most of the Potentially Foraging and all of the Feeding follows were located
between north Seattle and the northern extent of Puget Sound whereas follows
between south Seattle and Tacoma were almost exclusively Traveling follows. Traveling
follows were also abundant in the northern area of Puget Sound.
26
The above process was also repeated for the number of each follow type per
cell. In this case, seven cells had significant differences. To the southwest of San Juan
Island, cells 17 (χ2=10.38, df=2, p‐value=0.005) and 27 (χ2=26.07, df=2, p‐
value=0.000002) had a significantly different number of follow types. Northwest of San
Juan Island, cells 3 (χ2=11.68, df=2, p‐value=0.002), 4 (χ2=14.8, df=2, p‐value=0.0006), 10
(χ2=19.76, df=2, p‐value=0.00005) and 12 (χ2=14, df=2, p‐value=0.0009) all had a
significantly different number of follows. Southwest of Lopez Island, the number of
follows per value type were significantly different in cell 49 (χ2=10.75, df=2, p‐
value=0.004). In all cases, Traveling follows were observed as the most frequent follow
type within these cells whereas Feeding follows were the least common in all but two
cells.
Figure 11. The Spatial Distribution of Follows within the San Juan Islands Study Area.
27
Figure 12. The Spatial Distribution of Follows within the Puget Sound Study Area.
28
Figure 13. The Density of Follows within the San Juan Islands Study Area (higher densities
shown in red).
Figure 13 shows the density of those follows in the San Juan Study Area. Note
that the highest density of follows (shown in red) are in areas southwest of San Juan
Island, and west of Henry and Stuart Islands, all located in the Haro Strait. These density
areas were redefined into GIS shapefiles (figure 14) and compared to one another. High
density areas, all located in the Haro Strait, averaged approximately 185 meters in depth
(stdev approx. 81), a slope of approximately 6 degrees (stdev approx. 11) and an
average aspect of west. Of the entire San Juan Island study area, high density areas
made up 1.5% of the area. Low density areas were more spread across the study area
though the majority of this area was also west or southwest of San Juan Island. Low
density areas made up 16.2% of the entire study area and had an average depth of
approximately 123 meters (stdev approx. 81), average slope of approximately 3 degrees
(stdev approx. 7) and an average aspect of southwest. The rest of the area, making of
82.3% of the entire study area, had an average depth of approximately 71 meters (stdev
aprrox. 55), average slope of approximately 2 degrees (stdev approx. 5) and an average
aspect of southwest.
29
Since the low density areas are spread across the study area, it was determined
to be valuable to compare the low density areas west and southwest of San Juan Island
with the rest of the low density areas scattered across the archipelago. The low follow
density areas west and southwest of San Juan Island had an average depth of
approximately 155 meters (stdev approx. 81) and average slope of approximately 4
degrees (stdev approx. 8) compared to an average depth of approximately 71 meters
(stdev approx. 48) and average slope of approximately 2 degrees (stdev approx. 4) for
the scattered areas throughout the rest of the archipelago. Aspect was deemed not
important since west and southwest facing slopes represent the majority of aspects
throughout the entire study area.
Figure 14. The Spatial Representation of Areas with High and Low Density Killer Whale Follows
as well as Areas where No Follows Occurred.
To examine if there are differences in the frequency of follow type, each follow
was tallied and assigned to a category based on the type of follow and in which spatial
density it occurred (see Table 4). The results indicate that the frequencies of follow
30
types do not differ between higher and lower densities, even though lower densities
cover more area and higher densities have more tightly clustered follows.
Table 4: The Percentage of Follow Types per Density Areas
Follow Type
Feeding
Potentially Foraging
Traveling
High Density
Low Density
Areas
Areas
16.57%
17.16%
29.71%
25.37%
53.71%
57.46%
Discussion
Although most follows (57.84%) were Traveling follows (follows with no
evidence of a successful predation event or foraging cues), they averaged only eighteen
minutes in duration. Feeding and Potentially Foraging follows represented 16.74% and
25.42% respectively yet were generally longer follows averaging twenty‐six and twenty‐
five minutes respectively. This may be explained by the priority that collecting prey
remains had on the project. Since it was most important to collect prey remains, follows
involving animals exhibiting traveling behavior were broken off in hopes of having better
luck finding prey remains from actively foraging animals. The average duration of all
follows was approximately twenty‐one minutes. Follows ended for a variety of reasons.
Southern resident killer whales attract a large number of commercial and recreational
whale‐watching vessels, with an average of 20 boats around groups of these whales
(Koski 2005). To prevent a negative impact on commercial operations, and in order to
continue to facilitate cooperative relationships with the industry, follows would
occasionally be terminated when a large number of whale‐watching boats were present
and had interest in the individual(s) that were the focus of the follow. Follows were also
ended if the followed individual(s) entered Canadian waters, traveled too close to shore
in areas with high human population density, or if daylight was waning. The primary
study objective was to collect any remains from a predation event, therefore, if the
followed individual(s) appeared inactive, and another individual or group of individuals
appeared to be exhibiting foraging cues we would switch the follow to potentially
foraging whales, thereby reducing the potential for harassment due to extended follow
duration. This could contribute to the longer duration of Feeding and Potentially
Foraging follows; if whales appeared to be foraging or if evidence of foraging existed,
follows may have been less likely to have been broken off, whereas non‐foraging whale
follows (Traveling) would be abandoned in favor of whales demonstrating potential
predatory behavior.
31
Follow Type Comparison
Habitat variables were compared to examine the differences, if any, between
the three follow types. Surprisingly, there were no significant differences in depth of
the seafloor, slope of the seafloor, aspect, or distance to shore for any of the follow
types. When the Feeding and Potentially Foraging follows were pooled to compare to
the Traveling follows, no significant differences were found in depth, slope or aspect.
Aspect was not tested in this pooled dataset because not only were the aspect values
and the proportions they represented in the dataset virtually identical between follows
but these values also reflect the available aspect in the study area. Thus aspect was
deemed unimportant. These results indicate that southern resident killer whales may
not pick certain areas, based on these criteria, to find salmon. It could, however,
potentially indicate that salmon do not select for habitat based on these criteria either,
logically assuming that the predator will selectively find food where its prey can be
found.
Two problems exist with testing differences in environmental variables between
follows in this fashion, both related to data dependency. Similar to an ANOVA, one of
the assumptions of a Kruskal‐Wallis test is that data are independent of one another.
The waypoints collected that ultimately made up a follow were not independent. With
bathymetry, each point is correlated (the depth at one point is related to surrounding
areas and thus is not truly “free” to be any depth). To minimize the impact of
bathymetric dependency, all waypoints for a given follow were averaged together to
create one value for benthic depth, benthic slope, benthic aspect, and distance to shore
for each follow instead of one value for every waypoint within a follow. The second
problem was that the follows themselves were spatially correlated. In many cases when
one follow ended, another would immediately begin, making the environmental
features of both follows correlated to one another. Again the combining of waypoints
to create one average for each of the four habitat variables for each follow was an
attempt to minimize this.
There may be other limiting factors to killer whales foraging ecology not
measured here. For New Zealand’s Hector’s dolphins (Cephalorhynchus hectori), water
clarity and sea surface temperature (SST) were significantly important for habitat
selection (Brager et al 2003). Physical and chemical data about the water were not
recorded in this study. However, changes in the water and its effects on killer whale
foraging would manifest itself in salmon distribution. Though it is clear that water
variables influence salmon presence/absence on a large scale and their vertical
migration (Quinn 2005) information about these factors’ influence on salmon
abundance at a fine scale are limited in coastal and estuarine waters. Hoelzel (1993)
found no correlations between bathymetry and fast non‐directional behaviors with
southern resident killer whales. Though the current study does not test correlations in
bathymetry and follow type, it does lend strength to Hoelzel’s findings that bathymetry
32
may not effect foraging. In addition to changes in the water, physical habitat
differences could explain feeding vs. non‐feeding follows. In areas of western Florida,
foraging bottlenose dolphins (Tursiops truncatus) selected for dredged channel and
spoil‐Island habitats in one area and natural channels in the other (Allen et al 2001).
There may be differences in the type of physical environmental features between these
follows that were not measured here that could have influenced the distribution of
salmon. This seems unlikely given that the greatest density of all three follow types
occur west of San Juan Island in a relatively small area but should not be ruled out until
properly investigated.
Another factor that could have impacted the results reported here is the issue
of spatial scale. With all spatial studies, the issue of spatial scale could impact the
results. The spatial scale chosen here was 90 meters, and this was chosen primarily
because it was the smallest spatial scale available. Follow waypoints were pooled to
create one value for each habitat variable per follow so increasing spatial scale was
deemed inappropriate given that resolution was being decreased by this process.
However changing the spatial scale would undoubtedly alter the habitat values for each
waypoint, which in turn could cause results that indicate significant differences between
follow types.
Habitat Analysis
A grid system containing cells roughly 4.85km2 was created to closely match a
previous study done by Heimlich‐Boran (1988), though his study used 4.6km2 cells. It
was determined that 4.85km2 were the dimensions to use because it could be easily
created in GIS and still closely match the dimensions of the comparative study. Both
time spent on each follow (duration per follow type) and number of observed follows
per cell were analyzed.
For follow durations, only two cells, one from each study area, yielded
significant results. Both of those cells, however, had only Traveling follows located
within them. Given this, it seems that spatial location did not influence follow duration
differences. In fact, the Haro Strait had the greatest number of all three follow types.
This information does seem to reveal that although spatial location does not influence
killer whale feeding, it does influence killer whale presence/absence.
Number of observed follows, on the other hand, revealed significant
differences. In the San Juan Islands Study Area, two cells southwest and four cells
northwest of San Juan Island had a significant different number of follows per follow
type. In one cell southwest of Lopez Island, significant differences in number of follows
per follow type were also observed. In all cells, Traveling follows were the most
numerous follow type. In most cases, Potentially Foraging follows were the second
most numerous while Feeding follows were the least observed follow type. Though
33
differences were not tested between these cells and the study area as a whole, nor was
the significance tested for the number of follow types for the entire study area, trends
for these cells seem to match the entire four year sample, where Traveling follows were
the most numerous and Feeding follows were the least numerous.
The sample size in Puget Sound is considered to be too small to draw any
conclusions about fine‐scale habitat use in that study area but what is known is that
Puget Sound plays an important role in southern resident killer whale ecology. Orca
Network (www.orcanetwork.org), which collects sighting data on many marine mammal
species including both resident and transient killer whales, gray whales (Eschrichtius
robustus), and minke whales (Balaenoptera acutorostrata), has logged 102 days where
resident killer whales were sighted within Puget Sound during 2004‐2007. During this
time period, there are very few days where resident killer whales were located south of
the Tacoma Narrows (n=5). The majority of the sightings (n=38) where around Vashon
Island. Other areas of higher number of sightings compiled by Orca Network were
Kingston (n=18) and Bainbridge Island (n=11). The majority of these sightings are from
recreational, non‐professional whale watchers therefore positive species and
pod/individual identifications have not been verified. It is possible that some of these
animals could be mammal‐eating transients. Another potential data bias could be found
in a larger number of sightings occurring around higher density human populated areas.
For instance, although Vashon Island itself is not heavily populated, it can be easily seen
and accessed by boaters from Tacoma to Seattle, the two largest cities in the Puget
Sound region. Bainbridge Island can be seen and accessed easily by boaters from
Seattle and can be easily accessed by Bremerton and Port Orchard. Three of the areas
with relatively few sightings, Cooper Point (n=2), Squaxin Island (n=2), and Fox Island
(n=1) are all south of the Tacoma Narrows where the population is less dense than the
Seattle‐Tacoma metropolitan area. Regardless of these potential sighting biases, Puget
Sound appears to be valuable habitat for the southern resident killer whales. More
research in the Puget Sound basin is warranted.
Comparisons with Heimlich‐Boran’s (1988) work were considered valuable to
examine trends in spatial‐habitat use. It should be noted that Heimlich‐Boran broke his
observed behavioral patterns into four categories (feeding, travel, rest, and socializing)
and sub‐divided those categories further down. To compare his data with data
presented here, some liberties must be taken in regards to the broader classifications
used in this study. For example, he differentiates between “resting” and “traveling”
whereas any follow that did not have any evidence of a predation event or foraging
behavioral cues was classified as a Traveling, regardless of other behavior observed, not
just whether they were resting or actively traveling.
Similar to the data here, Heimlich‐Boran’s study found a high density of killer
whale encounters west and southwest of San Juan Island. In addition, he had a high
number of encounters extending all the way into Canadian waters to the mouth of the
34
Fraser River. This study did not extend into Canadian waters therefore we cannot know
if that pattern would be seen here also. Although the results are not identical, similar
patterns do exist. Southwest of San Juan Island saw the longest Feeding follow
durations. In Hemlich‐Boran’s paper, these were areas defined as locations that saw
“increased feeding”. Traveling northwest up the coast of the Island, less time was spent
on Feeding follows and a greater proportion of time was spent on Traveling follows. In
the comparison paper, it was observed that, this area was generally used for either
traveling or resting. Though these results are not statistically significant, there does
seem to be similarities in the two datasets. The only major differences in the datasets
lie in the area around Point Roberts (figure 11, cells 24, 35, and 45; Heimlich‐Boran calls
this area central Georgia Strait). Hemlich‐Boran (1988) identified this as an area of
increased traveling behavior; however, this study found that Feeding follows were
observed a majority of the time (66.3%).
Hauser (2006) also found that the Haro Strait is heavily used by southern
resident killer whales during the summer months regardless of pod membership. She
also found that depth had a significant relationship with southern resident killer whales;
it was reported by Hauser that killer whales were found in greater density in deeper
cells. Slope and distance to shore were also found to be important. Likewise, in the
data presented here, killer whale density was also greatest in areas of deeper waters.
However, killer whale density in this study was also greatest in the Haro Strait (west and
southwest of San Juan Island), which is the deepest part of the waters around the San
Juan Island archipelago. In Hoelzel’s (1993) study, southern resident killer whales were
also found more often over deep areas. It is possible that factors besides depth, such as
its potential importance as a salmon migration route, could make the Haro Strait area
important to killer whales. There are six areas within the Haro Strait that have been
identified as “biological hotspots” (Bloch et al 2002), indicating that this area is
ecologically productive.
Although locations where killer whales were found in high density were in
deeper areas of the San Juan Islands, killer whales were also found, albeit in lower
densities, in shallower areas as well. All areas of high density follows were in Haro
Strait. Most of the lower density follows were also observed in Haro Strait, however,
there were other zones of lower density follows spread throughout the study region.
However, when comparing the lower density follow areas in the Haro Strait to those
scattered around the archipelago, the scattered areas were much shallower on average
and are comparable to the average depth of water with no killer whale follows. In
addition, prey remains have been collected in these shallower areas so thus importance
of these areas should not be ignored. In fact, data here show that although there are
areas where killer whales were found in higher densities (refer to Figure 14), the
frequency of follow types observed do not differ between high and low density areas.
On the surface, this indicates that areas where killer whales are less observed may be
35
just as important in terms of foraging ecology. Note that the density values reported
here reflect the spatial patterns of the follows that were conducted during the field
surveys and not killer whale occurrence. These patterns may not necessarily reflect the
density pattern of where the southern resident killer whale community can be found
(even though spatial patterns reported in other studies report similar patterns).
Hauser (2006) also found that there could be a potential preference for steeper
slopes by resident killer whales. As with the depth of the seafloor, there may be some
preference for killer whales that were observed in this study to also select for areas with
steeper slope. However, like depth, the Haro Strait has a steep slope gradient and thus
could bias the data presented in this study.
As with depth, the low density areas were divided to compare the average slope
of the low density areas within Haro Strait to the low density areas scattered around the
rest of the study area. Unlike depth, however, the slope of the scattered areas did not
match the slope of areas with no follows but rather was a mid‐point between the Haro
Strait data and the areas with no follows. Steeper slopes are more evenly distributed
than depth across the study area, therefore it is possible that this is merely the
byproduct of the spatial distribution of benthic slope rather than selection based on
preference.
Finally, Hauser (2006) examined distance to shore and discovered that southern
resident killer whales may prefer areas closer to shore. Although I examined distance to
shore as a variable for comparing follow types, distance to shore was not examined
when looking at presence/absence of killer whales within the study area. The research
vessel, in general, avoided populated nearshore areas in favor of areas further off land
or close to sparsely populated shorelines.
Aspect was also examined as a variable to determine killer whale density. A
southwest aspect is the dominant value throughout the entire study area. In high
density areas the dominant aspect was west. In low density areas and areas with no
follows, southwest was the dominant sea floor aspect. Aspect was deemed not
important in this study area as killer whale occurrence related to aspect does not differ
from the average aspect found in the study area.
As with the Chi‐Squared spatial analysis, the sample size within the Puget Sound
study area was too small to draw any conclusions about benthic habitat variables and
their influence on killer whale density. Therefore this analysis was not done in the Puget
Sound area as the small sample size would not produce any meaningful results. It would
be of value to examine densities of killer whale occurrence related to benthic habitat in
Puget Sound in the future.
Understanding predator‐prey interactions is important in ecological studies. As
spatial habitat studies become more refined, new methods and findings will emerge.
36
For example, movement patterns of top predators such as killer whales, will be
influenced, to some degree, by prey movements. In the Johnstone Strait of British
Columbia, Canada, data were collected that support this hypothesis. The timing of
northern resident killer whale occurrence and population increases coincide with
occurrences and increases in the salmon population (Nichol and Shackleton 1996).
Current research with southern resident killer whales reveals some interesting, and
surprising results. McClusky (2006) found that although southern resident population
fluctuations are correlated with population fluctuations of Chinook and chum salmon,
SRKW spatial density did not match the density of salmon. Heimlich‐Boran (1986) found
different results; upon examination of all of the inland waters of Washington State
(Puget Sound, Hood Canal, The San Juan Island archipelago and the Strait of Juan de
Fuca), he found that there was a positive correlation between salmon distribution
(based on catch data)and killer whale sightings in the study area as a whole. But more
importantly, and contrary to McClusky (2006), there were positive correlations in
salmon density and killer whale sightings in specific marine areas, such as the Strait of
Juan de Fuca, the San Juan Islands as a whole, and the two northern management
regions of Puget Sound stretching from Seattle to Point Wilson, WA.
A theme in ecology is that predators may maximize efficiency by matching their
distribution to that of their prey’s habitat and resources rather than actually following
or randomly searching for prey (Flaxman and Lou 2009). This is known as the optimal
foraging theory. As previously mentioned, Chinook salmon are the preferred prey item
of southern resident killer whales. Though the influence of bathymetry on Chinook
salmon is poorly understood, what is known is that Chinook salmon are found in deeper
parts of the water column when compared to the other salmon species (Quinn 2005).
Killer whales may be selecting for deeper areas in which to forage in response to the
water column partitioning observed by the Pacific salmonids.
Depth is one of the most widely used variables when examining cetacean
habitat. In the same study area as this study, it was discovered that harbor porpoise
(Phocoena phocoena) sightings occurred over depths greater than 100 meters (Raum‐
Suryan and Harvey 1998). In the Bay of Fundy, Canada, two species were examined. Fin
whales (Balaenoptera physalus) preferred waters less than 60 meters deep whereas
minke whales preferred waters deeper than 60 meters (Ingram et al 2007). Even in
other marine mammal species depth appeared to have some influence in distribution,
as is the case with a harbor seal (Phoca vitulina) population in Scotland (Tolit et al 1998).
This study found that harbor seals were generally found in depth ranges of 10‐50 meters
and would dive to the maximum depths within the foraging range.
A few studies revealed that depth was not important in determining
distribution. One study, in Cook Inlet, Alaska, determined that bathymetry was not
important in examining distribution of beluga whales (Delphinapterus leucas) (Goetz et
37
al 2007). However for most cetacean studies, depth was important either as a primary
explanatory variable or a secondary variable.
Slope is another widely examined and generally important variable in explaining
cetacean habitat use. The Bay of Fundy study by Ingram et al (2007) determined that
slope, specifically steeper slopes, was important to minke whales. In Scotland, grey
seals (Halichoerus grypus) seemed to show a preference for areas with a greater
variation in seabed slope (MacLeod et al 2007). Bottlenose dolphins along the west
coast of Ireland showed a preference towards estuaries with a greater benthic slope
(Ingram and Rogan 2002). The seabed gradient most likely has some affect on habitat
characteristics. It is known that benthic slope is a major driver in benthic complexity,
which in turn can create areas of biological “hotspots” (Ardron 2002). Habitat
complexity is already known to have a positive impact on species diversity, such as pool‐
riffle habitat in river ecosystems. Benthic slope could also affect nutrient upwellings,
which in turn could affect primary production.
Seabed aspect is not a commonly studied variable in marine mammal
distribution. In one study involving Blainville’s beaked whales (Mesoplodon densirostris)
around Great Abaco, The Bahamas, depth, slope, and aspect were examined (MacLeod
and Zuur 2005). Although these whales showed preference for depth and slope ranges,
aspect was discovered to be the most important variable describing distribution with a
preference of a northeast facing aspect.
Variables not examined here are also used to explain cetacean distribution. A
population of New Zealand dusky dolphins (Lagenorhynchus obscures) seemed to prefer
areas with currents exceeding 12km/hr (Markowitz et al 2004). Interestingly, a dusky
dolphin study conducted in Patagonia (Garaffo et al 2007) revealed that not only was
sea floor depth important to dusky dolphins but that preference changed over a
temporal scale from year to year. In many cases it is not one specific variable but a
combination of two or more that are important to cetacean use. In the Cook Inlet
beluga whale study, for example, proximity to mudflats and medium to high flow
accumulations (river basin discharge) were important in determining habitat features
(Goetz et al 2007). Chilean dolphins (Cephalorhynchus eutropia) in Yaldad Bay, southern
Chilie did not use the bay uniformly but selected for areas close to shore with depths of
5‐10 meters and in intermediate distances to rivers (Ribeiro et al 2007). In a study of
Hector’s dolphins (C. hectori), which are endemic to New Zealand and endangered
(Jefferson et al 2008), animals were frequently encountered in shallower, turbid waters
with an SST of greater than 14 degrees Celsius (Brager et al 2003).
It is clear from over 30 years of field studies that the waters west and southwest
of the San Juan Islands are important in killer whale ecology. This is one of the deepest
part of the marine habitat surrounding the archipelago, as well as one of the areas with
the highest slope. Its proximity to the Strait of Juan de Fuca, and salmon returning from
38
the open ocean to their spawning grounds in Washington State and British Columbia
may be an important factor for killer whale habitat preference. It is clear that the
protection this area will play a vital role in the recovery of the southern resident killer
whale population. The creation of the Orca Pass Stewardship Area as a marine
protected area (MPA) has been one proposed method of protecting important biological
hotspots (including important SRKW habitat) (Bloch et al 2002; Georgia Strait Alliance
2009). The degree to which MPAs provide protection is up to interpretation and debate
based on implementation
Though the designs of MPAs are developed with best‐available science and the
best of intentions, in many cases the desired ecological benefit is not the outcome. In
many cases, competing societal benefits can cause the size of the designated MPA to
decrease from the design phase to their actual size and lack of fishing regulation can
cause the productivity of these places to decrease (Le Quesne 2008). MPAs are more
than just areas of high productivity identified by fishery managers, however. They are
places that have ramifications on economics, politics, and tourism. As with any
potential MPA (Charles and Wilson 2008), the area valued by southern resident killer
whales will only be successfully protected if everyday citizens are actively involved in
every stage in the development.
Conclusion
This thesis has added to the growing knowledge of habitat use by the
endangered population of southern resident killer whales. A steady decline in the
population since the mid‐1990s, probably due to a decline in their favorite prey item,
Chinook salmon, as well as other risk factors, has created concern for the long‐term
viability of this population. This paper focused on benthic and spatial habitat related to
feeding and a secondary focus of benthic and spatial habitat related to density. The
general findings were that, at least in the focal follows “collected” for this study, there
are no habitat differences between three types of follows; follows where it was
confirmed they were feeding (due to collections of prey remains), follows where it was
thought that they were probably feeding (due to foraging cues displayed that have been
previously linked to foraging), and follows where it was thought that they may not be
foraging (no predation collections and no foraging cues). The lack of differences
between follow type may be due to an error in parameters tested here. Sea surface
temperate, turbidity, or proximity to marine benthic habitats (i.e. sea grass bed, river
mouths, etc) may be the explanatory variables in follow type. Also, as with all spatial
studies, the scale of data analysis could impact the results. Perhaps adjusting the spatial
scale of the digital elevation model could lead to significant results. The spatial
distribution of southern resident killer whale density matches what other studies have
found; the areas of highest densities were in a part of the San Juan archipelago known
39
as the Haro Strait. This area has already been identified as prime habitat for southern
resident killer whales. Restoration of the salmon population in addition to protecting
southern resident killer whale habitat will be needed if the goal is to restore the long‐
term viability of this species, especially in an area where population growth and
subsequent human development will only increase in the coming years.
40
Literature Cited
Allen, M. C., A. J. Read, J. Gaudet, and L. S. Sayigh. 2001. Fine‐scale habitat selection of
foraging bottlenose dolphins Tursiops truncates near Clearwater, Florida.
Marine Ecology Progress Series 222: pp 253‐264.
Araki, H., B. A. Berekikian, M. J. Ford, and M. S. Blouin. 2008. Fitness of hatchery‐reared
salmonids in the wild. Evolutionary Applications 1(2): pp 342‐355.
Ardron, J. 2002. A GIS recipe for determining benthic complexity: An indicator of species
richness. In: Breman, J. 2002. Marine Geography: GIS for the Oceans and Seas.
ESRI Press: Redlands, CA, USA. 204pps.
Asper, E. D. and L. H. Cornell. 1977. Live capture statistics for the killer whale (Orcinus
orca) 1961‐1976 in California, Washington and British Columbia. Aquatic
Mammals 5(1): pp21‐26.
Augerot, X., C. Steinback, N. Fobes. 2005. Atlas of Pacific Salmon: The First Map‐Based
Status Assessment of Salmon in the North Pacific. University of California Press:
Berkeley and Los Angeles, California, USA. 150pps.
Baird, R. W. 2001. Status of killer whale Orcinus orca in Canada. Canadian Field‐
Naturalist 115(4): pp 676‐701.
Baird, R. W. and L. M. Dill. 1995. Occurrence and behaviour of transient killer whales:
seasonal and pod‐specific variability, foraging behaviour, and prey handling.
Canadian Journal of Zoology 73: pp 1300‐1311.
Baird, R. W. and M. B. Hanson. 2004. Diet studies of “southern resident” killer whales:
prey sampling and behavioral cues of predation. Report prepared as partial
fulfillment of Contract No. AB133F‐03‐SE‐1070 to Cascadia Research Collective,
Olympia, WA from the National Oceanic and Atmospheric Administration,
Seattle, WA.
Barrett‐Lennard, L. G. 2000. Population structure and mating patterns of killer whales as
revealed by DNA analysis. PhD. Thesis, University of British Columbia.
Berzin, A. A. and V. L. Vladimirov. 1982. A new species of killer whale from the Antarctic
waters. Priroda 6: pp 31.
Berzin, A. A. and V. L. Vladimirov. 1983. A new species of killer whale (Cetacea,
Delphindae) from Antarctic waters. Zoologicheskii Zhurnal 62(2): pp 287‐295.
Bilby, R. E. and L. A. Mollot. 2008. Effects of changing land use patterns on the
distribution of coho salmon (Oncorhynchus kistutch) in the Puget Sound
region. Canadian Journal of Fisheries and Aquatic Science 65: pp 2138‐2148.
41
Brager, S., J. A. Harraway, and B. F. J. Manly. 2003. Habitat selection in a coastal dolphin
species (Cephalorhynchus hectori). Marine Biology 143: pp 233‐244.
Bloch, P., M. Sato, and J. White. 2002. The Orca Pass International Stewardship Area
process: Where marine science and policy intersect. In: Breman, J. 2002. Marine
Geography: GIS for the Oceans and Seas. ESRI Press: Redlands, CA, USA. 204pps.
Burnett, K. M., G. H. Reeves, D. J. Miller, S. Clarke, K. Vance‐Borland, and K. Christiansen.
2007. Distribution of salmon‐habitat potential relative to landscape
characteristics and implications for conservation. Ecological Applications 17(1):
pp 66‐80.
Center For Whale Research. 2008. (www.whaleresearch.com). Last accessed: Apr 9,
2009.
Charles, A. and L. Wilson. 2008. Human dimensions of marine protected areas. ICES
Journal of Marine Science 66(1): pp 6‐15.
Cullon, D. L., M. B. Yunker, C. Alleyne, N. J. Dangerfield, S. O’Neil, M. J. Whiticar and P. S.
Ross. 2009. Persistent organic pollutants in Chinook salmon (Oncorhynchus
tshawytscha): Implications for resident killer whales of British Columbia and
adjacent waters. Environmental Toxicology and Chemistry 28(1): pp 148‐161.
Environment Canada‐British Columbia Ministry of Environment, Lands and Parks. 1992.
The Lower Fraser Basin: A State of the Environment Synopsis. Vancouver, BC,
Canada.
Environmental Protection Agency. 2008. Health effects of PCBs.
(http://www.epa.gov/osw/hazard/tsd/pcbs/pubs/effects.htm). Date accessed
Apr 21, 2009.
Flaxman, S. M. and Y. Lou. 2009. Tracking prey or tracking the prey’s resource?
Mechanisms of movement and optimal habitat selection by predators. Journal
of Theoretical Biology 256: pp 187‐200.
Foote, A. D., R. W. Osborne, and A. R. Hoelzel. 2004. Whale‐call response to masking
boat noise. Nature 428: pp 910.
Ford, J. K. B. and G. M. Ellis. 1999. Transients: Mammal‐Hunting Killer Whales. University
Of British Columbia Press: Vancouver, British Columbia, Canada. 96pps.
Ford, J. K. B., G. M. Ellis, and K. C. Balcomb. 2000. Killer Whales (Second Edition).
University of British Columbia Press: Vancouver, British Columbia, Canada.
104pps.
Ford, J. K. B. and G. M. Ellis. 2006. Selective foraging by fish‐eating killer whales Orcinus
orca in British Columbia. Marine Ecology Progress Series 316: pp 185‐199.
42
Fraser, D. A., J. K. Gaydos, E. Karlsen, and M. S. Rylko. 2006. Collaborative science, policy
Development and program implementation in the transboundary Georgia
Basin/Puget Sound ecosystem. Environmental Monitoring and Assessment 113:
pp 49‐69.
Garaffo, G. V., S. L. Dans, S. N. Pedraza, E. A. Crespo and M. Degrati. 2007. Habitat use
by dusky dolphin in Patagonia: how predictable is their location? Marine Biology
152: pp 165‐177.
Georgia Strait Alliance. 2009. Orca Pass. (http://www.georgiastrait.org/?q=node/73).
Last accessed: April 8,2009.
Goetz, K. T., D. J. Rugh, A. J. Read, and R. C. Hobbs. 2007. Habitat use in a marine
ecosystem: beluga whales Delphinapterus leucas in Cook Inlet, Alaska. Marine
Ecology Progress Series 330: pp 247‐256.
Goodman, D. 2005. Selection equilibrium for hatchery and wild spawning fitness in
integrated breeding programs. Canadian Journal of Fisheries and Aquatic
Sciences 62: pp 374‐389.
Hard, J. J., M. R. Gross, M. Heino, R. Hilborn, R. G. Kope, R. Law, and J. D. Reynolds.
2008. Evolutionary consequences of fishing and their implications for salmon.
Evolutionary Applications 1(2): pp 1752‐4571.
Hanson, M. B., R. W. Baird, J. K. B. Ford, J. Hempelmann‐Halos, D. Van Doornik, J. R.
Candy, C. K. Emmons, G. S. Schorr, B. Gisborne, K. L. Ayres, S. K. Wasser, K. C.
Balcomb, K. Balcomb‐Bartok, J. G. Sneva and M. J. Ford. Species and stock
identification of prey selected by endangered “southern resident” killer whales
in their summer range. In prep.
Hauser, D. D. W. 2006. Summer space use of southern resident killer whales (Orcinus
orca) within Washington and British Columbia inshore waters. M.S. Thesis,
University of Washington.
Heimlich‐Boran, J. R. 1986. Fishery correlations with the occurrence of killer whales in
the greater Puget Sound. In: Kirkevold, B. C. and J. S. Lockard. 1986. Behavioral
Biology of Killer Whales. Alan R. Liss, Inc: New York, NY, USA. 457pps.
Heimlich‐Boran, J. R. 1988. Behavioral ecology of killer whales (Orcinus orca) in the
Pacific Northwest. Canadian Journal of Zoology 66: pp 565‐578.
Hoekstra, J. M., K. K. Bartz, M. H. Ruckelshaus, J. M. Moslemi, and T. K. Harms. 2007.
Quantitative threat analysis for management of an imperiled species: Chinook
salmon (Oncorhynchus twhawytscha). Ecological Applications 17(7): pp 2061‐
2073.
43
Hoelzel, A. R. 1991. Killer whale predation on marine mammals at Punta Norte,
Argentina; food sharing, provisioning and foraging strategy. Behavioral Ecology
and Sociobiology 29: pp 197‐204.
Hoelzel, A. R. 1993. Foraging behavior and social group dynamics in Puget Sound killer
whales. Animal Behaviour 45: pp 581‐591.
Holt, M. M., D. P. Noren, V. Veirs, C. K. Emmons, and S. Veirs. 2008. Speaking up: killer
whales (Orcinus orca) increase their call amplitude in response to vessel noise.
Journal of the Acoustical Society of America 125(1): pp 27‐32.
Hoyt, E. 1990. Orca: The Whale Called Killer. Firefly Books Ltd.: Altona, Manitoba,
Canada. 291pps.
Hoyt, E. 2001. Whale Watching 2001: Worldwide tourism numbers, expenditures, and
Expanding socioeconomic benefits.
(http://www.oceania.org.au/soundnet/features/hoytifaw.html). Date accessed:
Apr 21, 2009.
Ingram, S. N. and E. Rogan. 2002. Identifying critical areas and habitat preferences of
bottlenose dolphins Tursiops truncatus. Marine Ecology Progress Series 244:
pp 247‐255.
Ingram, S. N., L. Walshe, D. Johnston, and E. Rogan. 2007. Habitat partitioning and the
influence of benthic topography and oceanography on the distribution of fin and
minke whales in the Bay of Fundy, Canada. Journal of the Marine Biological
Association of the United Kingdom 87: pp 149‐156.
International Union for Conservation of Nature and Natural Resources. 2008. Orcinus
orca.
(http://www.iucnredlist.org/details/15421). Last accessed: January 11, 2009.
Irvine, J. R., L. Biagini, and M. Poon. 2005. An update on catch trends for Pacific salmon
in British Columbia Canada. (NPAFC Doc. 868). 14pps. Fisheries and Oceans
Canada, Science Branch, Pacific Biological Station, Nanaimo, B.C., Canada V9T
6N7.
Jefferson, T. A., M. A. Webber, and R. L. Pitman. 2008. Marine Mammals of the World.
Elsevier Publishing: London, UK, 573pps.
Jelinski, D. E., C. C. Krueger, and D. A. Duffus. 2002. Geostatistical analyses of
interactions between killer whales (Orcinus orca) and recreational whale‐
watching boats. Applied Geography 22: pp 393‐411.
Keefer, M. L., C. C. Caudill, C. A. Peery, and S. R. Lee. 2008. Transporting juvenile
salmonids around dams impairs adult migration. Ecological Applications 18(8):
pp 1888‐1900.
44
Krahn, M. M., M. B. Hanson, R. W. Baird, R. H. Boyer, D. G. Burrows, C. K. Emmons, J. K.
B. Ford, L. L. Jones, D. P. Noren, P. S. Ross, G. S. Schorr, and T. K. Collier. 2007.
Persistent organic pollutants and stable isotopes in biopsy samples (2004/2006)
from southern resident killer whales. Marine Pollution Bulletin 54(12):
pp 1903‐1911.
Koski, K. L. 2005. 2004‐2005 final program report. Contract No. AB133F‐04‐SE‐0835 from
the National Oceanic and Atmospheric Administration.
Le Quesne, W. J. F. 2008. Are flawed MPAs any good or just a new way of making old
mistakes? ICES Journal of Marine Science 66(1): pp 132‐136.
Levin, P. S., R. W. Zabel, and J. G. Williams. 2001. The road to extinction is pave with
Good intentions: negative association of fish hatcheries with threatened
salmon. Proceeding of the Royal Society 268: pp 1153‐1158.
Lichatowich, J. 1999. Salmon Without Rivers. Island Press: Washington, DC, USA. 317pps.
Lichatowich, J., L. Mobrand, and L. Lestelle. 1999. Depletion and extinction of Pacific
salmon (Oncorhynchus spp.): A different perspective. ICES Journal of Marine
Science 56: pp 467‐472.
Lopez, J. C. and D. Lopez. 1985. Killer whales (Orcinus orca) of Patagonia, and their
behavior of intentional stranding while hunting nearshore. Journal of
Mammalogy 66(1): pp 181‐183.
Lusseau, D., D. E. Bain, R. Williams, and J. C. Smith. 2009. Vessel traffic disrupts the
Foraging behavior of southern resident killer whales Orcinus orca. Endangered
Species Research 6: pp 211‐221.
MacLeod, C. D. and A. F. Zuur. 2005. Habitat utilization by Blainville’s beaked whales off
Great Abaco, northern Bahams, in relation to seabed topography. Marine
Biology 147: pp 1‐11.
MacLeod, C. D., C. R. Weir, C. Pierpoint, and E. J. Harland. 2007. The habitat preferences
of marine mammals west of Scotland (UK). Journal of the Marine Biological
Association of the United Kingdom 87: pp 157‐164.
Markowitz, T. M., A. D. Harlin, B. Wursig and C. J. Mcfadden. 2004. Dusky dolphin
foraging habitat: Overlap with aquaculture in New Zealand. Aquatic
Conservation: Marine and Freshwater Ecosystems 14: 133‐149.
Matkin, C., G. Ellis, E. Saulitis, L. Barrett‐Lennard, and D. Matkin. 1999. Killer Whales of
Southern Alaska. North Gulf Oceanic Society: Homer, Alaska, USA. 96pps.
Matkin, C. O., L. G. Barrett‐Lennard, H. Yurk, D. Ellifrit, and A. W. Trites. 2007. Ecotypic
variation and predatory behavior among killer whales (Orcinus orca) off the
eastern Aleutian Islands, Alaska. Fisheries Bulletin 105: pp 74‐87.
45
McClusky, S. M. 2006. Space use patterns and population trends of southern resident
Killer whales (Orcinus orca) in relation to distribution and abundance of Pacific
salmon (Oncorhynchus spp.) in the inland marine waters of Washington State
and British Columbia. M.S. Thesis, University of Washington.
Missildine, B. R., R. J. Peters, G. Chin‐Leo, and D. Houck. 2005. Polychlorinated biphenyl
Concentrations in adult Chinook salmon (Oncorhynchus tshawytscha) returning
to coastal and Puget Sound hatcheries of Washington State. Environmental
Science and Technology 39(18): pp 6944‐6951.
National Marine Fisheries Service. 2007. Puget Sound Recovery Plan.
(http://www.nwr.noaa.gov/Salmon‐Recovery‐Planning/Recovery‐
Domains/Puget‐Sound/PS‐Recovery‐Plan.cfm). Date accessed: Apr 13, 2009.
Nichol, L. M. and D. M. Shackleton. 1996. Seasonal movements and foraging behavior of
Northern resident killer whales (Orcinus orca) in relation to inshore distribution
of salmon (Oncorhynchus spp.) in British Columbia. Canadian Journal of Zoology
74: pp 983‐991.
Nottestad, L., A. Ferno, and B. E. Axelsen. 2002. Diggin in the deep: killer whales’
Advanced hunting tactic. Polar Biology 25: pp 939‐941.
Olla, B. L., M. W. Davis, and C. H. Ryer. 1994. Behavioural deficits in hatchery‐reared
fish: Potential effects on survival following release. Aquaculture and Fish
Management 25(Suppl. 1): pp 19‐34.
Olla, B. L., M. W. Davis, and C. H. Ryer. 1998. Understanding how the hatchery
environment represses or promotes the development of behavioral survival
skills. Bulletin of Marine Science 62: pp 531‐550.
Pacific States Marine Fisheries Commission. 1997. When salmon are dammed.
(http://www.psmfc.org/habitat/salmondam.html). Date accessed: Apr 12, 2009.
Pitman, R. L. and P. Ensor. 2003. Three forms of killer whales (Orcinus orca) in Antarctic
waters. Journal of Cetacean Research and Management 5(2): pp 131‐139.
Puget Sound Action Team. 2005. 2005‐2007 Puget Sound Conservation and Recovery
Plan. Puget Sound Action Team. Olympia, WA, USA. 48pps.
Puget Sound Action Team. 2007. Puget Sound Update: Ninth Report of the Puget Sound
Ambient Monitoring Program. Puget Sound Action Team. Olympia, WA, USA.
260pps.
Puget Sound Partnership. 2008. Looks can be deceiving.
(http://www.psp.wa.gov/pugetsound.php). Date accessed: Apr 21, 2009.
46
Quinn, T. P. 2005. The Behavior and Ecology of Pacific Salmon and Trout. University of
Washington Press: Seattle, WA, USA. 378pps.
Raum‐Suryan, K. L., and J. T. Harvey. 1998. Distribution and abundance of and habitat
use by harbor porpoise, Phocoena phocoena, off the northern San Juan Islands,
Washington. Fishery Bulletin 96(4): pp 808‐822.
Rayne, S., M. G. Ikonomou, P. S. Ross, G. M. Ellis, and L. G. Barrett‐Lennard. 2004.
PBDEs, PBBs, and PCNs in three communities of free‐ranging killer whales
(Orcinus orca) from the northeastern Pacific Ocean. Environmental Science and
Technology 38(16):
pp 4293‐4299.
Ribeiro, S., F. A. Viddi, J. L. Cordeiro, and T. R. O. Freitas. 2007. Fine‐scale habitat
selection ofChilean dolphins (Cephalorhynchus eutropia): interactions with
aquaculture activities in southern Chiloe Island, Chile. Journal of the Marine
Biological Association of the United Kingdom 87: pp 119‐128.
Ross, P. S. 2006. Fireproof killer whales (Orcinus orca): flame‐retardant chemicals and
the conservation imperative in the charismatic icon of British Columbia, Canada.
Canadian Journal of Fisheries and Aquatic Sciences 63(1): pp 224‐234.
Ross, P. S., G. M. Ellis, M. G. Ikonomou, L. G. Barrett‐Lennard, and R. F. Addison. 2000.
High PCB concentrations in free‐ranging Pacific killer whales, Orcinus orca: e
ffects of age, sex and dietary preference. Marine Pollution Bulletin 40:
pp 504‐515.
Saulitis, E., C. Matkin, L. Barrett‐Lennard, K. Heise, and G. Ellis. 2000. Foraging strategies
of sympatric killer whale (Orcinus orca) populations in Prince William Sound,
Alaska. Marine Mammal Science 16(1): pp 94‐109.
Scott, D. 1979. On optimal and data‐based histograms. Biometrika 66: pp 605‐610.
Smith, T. G., D. B. Siniff, R. Reichle and S. Stone. 1981. Co‐ordinated behaviour of killer
whales, Orcinus orca, hunting a crabeater seal, Lobodon carcinophagus.
Canadian Journal of Zoology 59: pp 1185‐1189.
Tanner, C. D., J. R. Cordell, J. Rubey, and L. M. Tear. 2002. Restoration of freshwater
intertidal habitat functions at Spencer Island, Everett, Washington. Restoration
Ecology 10(3): pp 564‐576.
Toft, J. D., J. R. Cordell, C. A. Simenstad and L. A. Stamatiou. 2007. Fish distribution,
abundance and behavior along city shoreline types in Puget Sound. North
American Journal of Fisheries Management 27: pp465‐480.
47
Tollit, D. J., A. D. Black, P. M. Thompson, A. Mackay, H. M. Corpe, B. Wilson, S. M. Van
Parijs, K. Grellier and S. Parlane. 1998. Variations in harbor seal Phoca vitulina
diet and dive‐depths in relation to foraging habitat. Journal of Zoology, London
244: pp 209‐222.
Waples, R. S., R. W. Zabel, M. D. Scheuerell, and B. L. Sanderson. 2007. Evolutionary
Responses by native species to major anthropogenic changes to their
ecosystems: Pacific salmon in the Columbia River hydropower system.
Molecular Ecology 17: pp 84‐96.
Washington State Department of Ecology. 2009a. Threats to Puget Sound.
(http://www.ecy.wa.gov/puget_sound/threats.html). Date accessed: Apr 21,
2009.
Washington State Department of Ecology. 2009b. Washington’s toxic flame retardant to
take effect. (http://www.ecy.wa.gov/news/2009news/2009‐017.html). Date
accessed: May 24, 2009.
Washington State Department of Health. 2009. Are PBDEs harmful to human health?
(http://www.doh.wa.gov/ehp/oehas/pbde/pbdehumanhealth.htm#How%20tox
ic%20are%20PBDEs). Date accessed: Apr 21, 2009.
Weber, E. D. and K. D. Fausch. 2003. Interactions between hatchery and wild salmonids
in streams: differences in biology and evidence for competition. Canadian
Journal of Fisheries and Aquatic Sciences 60: pp 1018‐1036.
Williams, R., D. E. Bain, J. K. B. Ford, and A. W. Trites. 2002. Behavioural responses of
male killer whales to a ‘leapfrogging’ vessel. Journal of Cetacean Research and
Management 4(3): pp 305‐310.
Williams, R., D. Lusseau, and P. S. Hammond. 2006. Estimating relative energetic costs of
human disturbance to killer whales (Orcinus orca). Biological Conservation
133(3): pp 301‐311.
Williams, R. and E. Ashe. 2007. Killer whale evasive tactics vary with boat number.
Journal of Zoology 272(4): pp 390‐397.
Williams, R., D. E. Bain, J. C. Smith, and D. Lusseau. 2009. Effects of vessels on behaviour
patterns of Individual southern resident killer whales Orcinus orca. Endangered
Species Research 6: pp 199‐209.
48
Appendix A
49
Appendix B
50
Appendix C
51
Appendix D
52