Sterna paradisaea Sterna hirundo by

Sterna paradisaea Sterna hirundo by
Foraging area characteristics of Arctic Terns (Sterna paradisaea) and Common Terns
(Sterna hirundo) breeding on Machias Seal Island
by
Amie Lynn Black
BSc., University of New Brunswick, 2002
A Thesis Submitted in Partial Fulfilment of
the Requirements for the Degree of
Masters of Science
In the Graduate Academic Unit of Biology
Supervisor:
Antony Diamond, PhD, Biology, and Forestry and
Environmental Management
Examining Board:
Janice Lawrence, PhD, Biology
Charles Bourque, PhD, Forestry and
Environmental Management
This thesis is accepted.
Dean of Graduate Studies
The University of New Brunswick
November 2006
© Amie L. Black, 2006
1
Dedication
To my brother and sisters
2
Abstract
Arctic and Common Terns nesting sympatrically on Machias Seal Island (MSI) are
similar both in appearance and in the resources they use during the breeding season.
Because both species feed on similar prey types, and these prey are limited resources,
according to Gause’s Law some aspect of their biology must be different enough to allow
these two species to coexist. Extensive study of tern diet and foraging biology indicate
that Arctic and Common Terns occupy very similar ecological niches, leading us to
hypothesise that they may differ by feeding in separate areas. I conducted a radiotracking study to compare feeding locations of Arctic and Common Terns nesting on
MSI, to identify physical and biological characteristics of feeding areas to learn whether
they are foraging at random or in predictable areas, and to determine if there is a change
in foraging areas between the incubation, chick-rearing, and failed nest stages of the
breeding cycle. These questions were addressed by comparing tern foraging area
characteristics to available habitat in the searched area. Arctic and Common Terns did not
differ in the distance they travelled from MSI or the distance from the mainland they
foraged. Likewise, they foraged in areas where chlorophyll a concentration, sea-surface
temperature, and depth were not significantly different from the habitat where they were
not found. These results were consistent for all stages of the breeding cycle examined
(incubation, chick rearing, and failed nest). From this, I concluded that Arctic and
Common Terns nesting sympatrically on MSI were not reducing competition by foraging
in different areas. When I compared my results to the only comparable study, at a
different colony, I found that the distance both tern species travelled from the colony to
forage and ocean depth at foraging grounds were larger than previous estimates, which is
3
likely due to inter-colony differences such as breeding bird populations, colony location
with respect to the mainland, or prey abundance.
4
Preface
This thesis is written in traditional thesis format. The first chapter gives an
introduction to the objectives of the study, and chapter 2 provides background
information on the subject in the form of a literature review. Chapters 3 and 4 outline the
methods used and the results of the analyses. Chapter 5 discusses the findings of the
study, and provides insight into how my results compare to other studies.
The contents of this thesis will be modified and submitted to be published in a
peer-reviewed scientific journal. My supervisor, Dr. Antony Diamond, will be a coauthor of any manuscripts that are based on this work because he has been instrumental
to all aspects of the design, implementation, and synthesis of this work.
5
Acknowledgements
This project would not have been possible without the help and support of many
people. I would especially like to thank my supervisor, Dr. Tony Diamond for his ideas,
comments, and financial support.
Many thanks to my thesis committee, Drs. Steve Heard and John Chardine, who
offered statistical advice and helpful comments along the way, and my examining
committee.
From UNB, I would like to thank Graham Forbes, who provided radio-tracking
equipment; Jonathan Beaudoin of the Ocean Mapping Group for help with satellite
imagery; David Drolet for help with Matlab; Linda Allen, Marg Morton, Margaret
Blacquier, and Marni Turnbull in the Biology Office, and technicians, Roger Smith and
Mike Casey; and, of course all of the grad students who provided good fun and helpful
suggestions from the beginning.
Data from outside sources was an integral part of this thesis; I would sincerely
like to thank Andy Thomas and Ryan Weatherbee at the Satellite Oceanography Data
Centre at the University of Maine; Dave Greenberg at the Bedford Institute of
Oceanography; and the National Oceanic and Atmospheric Administration (NOAA). All
supplied free data and help processing them. Barry Turner, of the Department of
Atmospheric and Oceanic Sciences at McGill offered the use of his Linux computer, and
Falk Huettmann of the University of Alaska, Fairbanks, provided valuable guidance with
using ArcView.
Help in the field was provided by many people, but especially the always helpful
and entertaining Canadian Coastguard Lighthouse keepers (Paul, Ralph, Barry, Rick,
Gordon, and Russell); boat Captains Andy Patterson and Peter Wilcox, and First-mate
Durlan Ingersol, who shuttled me to and from MSI and brought fantastic treats; the expert
pilots, Klaus and Brad, of the Atlantic Charter Company; Marie-Paule, Ashley and Karel
who provided welcome field help; and last but definitely not least, the field crews of the
last few years: Kate (who introduced me to birds), Sarah D., Matt, Laura, Katie, Alex,
and Sarah F.
Comments and revisions were helpfully provided by Matt Charette, Kate Devlin,
Laura Minich, and Jesse Vermaire.
Finally, I would like to acknowledge my family, Greg, Sheila, Jamie, Angela, and
Jennifer for their love and commitment to helping me achieve my goals, and Jesse, for
reading many, many drafts, putting up with my summer absences, and providing me with
love and support; I cannot thank them enough.
This research was supported by the Atlantic Cooperative Wildlife Ecology
Research Network (ACWERN), Science Horizons, the New Brunswick Wildlife Trust
Fund, and the University of New Brunswick.
6
Table of Contents
Dedication…………………………………………………………………………...
ii
Abstract……………………………………………………………………………...
iii
Preface………………………………………………………………………………. v
Acknowledgements………………………………………………………………….
vi
Table of Contents……………………………………………………………………
vii
List of Tables………………………………………………………………………..
ix
List of Figures……………………………………………………………………….
x
Chapter 1. Introduction
1.1. General introduction…………………………………………………………… 1
1.2. Purpose of study………………………………………………………………... 3
Chapter 2. Literature Review
2.1. Study area……………………………………………………………………….
2.2. Gause’s Law, Optimal Foraging Theory, and tern time budgets……………….
2.3. Tern diet ………………………………………………………………………..
2.4. Foraging distance estimations and habitat characteristics……………………...
2.5. Previous Radio-tracking Studies………………………………………………..
Chapter 3. Methods
3.1. Productivity monitoring………………………………………………………...
3.2. 2004 data collection ……………………………………………………………
3.2.1. Aerial tracking…………………………………………………………..
3.2.2. Ground tracking……….…………………………………………………
3.2.3. Exclusion of 2004 data…………………………………………………..
3.3. 2005 data collection…………………………………………………………….
3.3.1. Aerial tracking…………………………………………………………...
3.2.2. Ground tracking…………………………………………………………
3.3.3. Environmental data……………………………………………………...
3.4. Analysis
3.4.1. Productivity analysis…………………………………………………….
3.4.2. Sex and body condition analysis…………………………………………
3.4.3. ArcView and ArcMap mapping…………………………………………
3.4.4. Randomization analysis…………………………………………………
4
5
9
12
14
16
16
18
19
20
20
22
24
24
26
26
27
28
Chapter 4. Results
4.1. Transmitter retention…………………………………………………………… 30
4.2. Transmitter effects on productivity…………………………………………….. 30
4.3. Sexing and body condition analysis..…………………………………………... 32
7
4.4. Foraging location mapping…………………………………………………….. 32
4.5. Habitat characteristic analysis………………………………………………….. 32
Chapter 5. Discussion
5.1. Transmitter attachment………………………………………………………….
5.2. Transmitter effects on productivity……………………………………………..
5.3. Sexing and body condition…..………………………………………………….
5.4. Foraging habitat………………………………………………………………...
5.5. Conclusion……………………………………………………………………...
36
36
37
37
44
References…………………………………………………………………………..
75
Vita
8
List of Tables
Table 1. Mean (SD) tern body measurements of adults trapped on Machias Seal
Island from 2003 to 2005..…………………………………………………………….
48
Table 2. Main prey items in tern chick diet on MSI from 1995 to 2005 (Bond et al.
2006). N indicates total number of prey identified. Prey items other than the 5
commonest prey are not included. Proportions of chick diet each prey item
comprises are reported as percentage by number of all identified prey………………. 49
Table 3. Number of prey types fed to chicks by Arctic and Common Terns on MSI
from 1995 to 2005. Numbers do not include unidentified prey items………………...
50
Table 4. Data layers obtained from external sources………………………………….
51
Table 5. Mean date (SD) of nest initiation for birds equipped with transmitters
(tracked) and without (untracked). Significantly different results within species
(ANOVA, p < 0.05) are denoted by a and b…………………………………………...
52
Table 6. Mean clutch size (SD) of birds equipped with transmitters (tracked) and
without (untracked). Mean clutch size is measured as the number eggs laid per nest.
No significantly different results within species (ANOVA, p < 0.05) were
detected.………………………………………………………………………………
53
Table 7. Mean hatching success (SD) of birds equipped with transmitters (tracked)
and without (untracked). Mean hatching success is measured by the number of eggs
that hatch per nest. Significantly different results within species (ANOVA, p < 0.05)
are denoted by a and b………………………………………………………………….
54
Table 8. Productivity (SD) of birds equipped with transmitters (tracked) and without
(untracked). Productivity is measured as the number of chicks fledged per nest.
Significantly different results within species (ANOVA, p < 0.05) are denoted by a
and b…………………………………………………………………………………… 55
Table 9. Mean (SD) body condition of birds with transmitters vs. no transmitters. No
differences were observed between tracked and untracked birds for either ARTE
(ANOVA, p= 0.142) or COTE (ANOVA, p=0.408)………………………………….
56
Table 10a-e. Results from randomization analysis comparing available habitat
location characteristics and tern feeding location characteristics at 3 nest stages and
all stages combined. Significant results (p < 0.05) are indicated in bold..……………
56
9
List of Figures
Figure 1. Map of the Gulf of Maine showing the location of Machias Seal Island…...
59
Figure 2. Figure 2. Histograms showing prey types fed to Common Tern chicks (a)
and Arctic Tern chicks (b) in 2004 and 2005. The 'other' category includes: pollock,
sticklebacks, lumpfish, insects and marine invertebrates……………………………... 60
Figure 3. Photos showing transmitter attachment to tail retrices in 2004 (a), and to
the leg band in 2005 (b)……………………………………………………………….
61
Figure 4. Map indicating locations of the lighthouse and Goofapuff, Foundation,
Gully, and Oceanspray plots…………………………………………………………..
62
Figure 5. Map outlining the approximately 40 km radius centred around MSI
searched for tern feeding areas in 2005. The perimeter of the north tip of Grand
Manan was also searched on one occasion…….……………………………………...
63
Figure 6. Map showing locations recorded along the flight path to represent
available tern foraging habitat………………………………………………………… 64
Figure 7. Map indicating mainland and coastal islands. Coastal islands were
excluded from the analysis of tern foraging distance from the mainland……………..
65
Figure 8. Map showing locations where Arctic Terns were found foraging during
each stage of the breeding cycle in 2005………………………..…………………….
66
Figure 9. Map showing locations where Common Terns were found foraging during
each stage of the breeding cycle in 2005……………………………..……………….
67
Figure 10. Box plots of the distance from the mainland (km) of Arctic Tern (ARTE,
n = 21) and Common Tern (COTE, n = 18) feeding areas, and available habitat
points (n = 2995). Dotted lines indicate mean, solid lines indicate median, boxes 2575% range and whiskers 10-90% range……..………………………………………..
68
Figure 11. Box plots of the distance from MSI (km) of Arctic Tern (ARTE, n = 21)
and Common Tern (COTE, n = 18) feeding areas, and available habitat points (n =
2995). Dotted lines indicate mean, solid lines indicate median, boxes 25-75% range
and whiskers 10-90% range…………………………………………………………..
69
Figure 12. Box plots of the depth (m) of Arctic Tern (ARTE, n = 21) and Common
Tern (COTE, n = 18) feeding areas, and available habitat points (n = 2995). Dotted
lines indicate mean, solid lines indicate median, boxes 25-75% range and whiskers
10-90% range………………………………………………………………………….
70
10
Figure 13. Mean depth (m) of Arctic Tern foraging areas when they no longer had a
nest (solid line) compared to the 95% confidence intervals (dashed line) of 10 000
iterations of available point depths (p = 0.0387)……………………………………...
71
Figure 14. Box plots of the chlorophyll concentration (mg/m 3) of Arctic Tern
(ARTE, n = 21) and Common Tern (COTE, n = 18) feeding areas, and available
habitat points (n = 2995). Dotted lines indicate mean, solid lines indicate median,
boxes 25-75% range and whiskers 10-90% range. ………………………………......
72
Figure 15. Mean chlorophyll concentration (mg/m 3) of all Common Tern foraging
areas (solid line) compared to the 95% confidence intervals (dashed line) of 10 000
iterations of available point chlorophyll concentration (p = 0.0030)…………………. 73
Figure 16. Mean chlorophyll concentration (mg/m 3) of Common Tern foraging areas
when they were incubating (solid line) compared to the 95% confidence intervals
(dashed line) of 10 000 iterations of available point chlorophyll concentration (p =
0.0019)………………………………………………………………………………...
74
Figure 17. Box plots of the sea surface temperature (0C) of Arctic Tern (ARTE, n =
21) and Common Tern (COTE, n = 18) feeding areas, and available habitat points (n
= 2995). Dotted lines indicate mean, solid lines indicate median, boxes 25-75%
range and whiskers 10-90% range...…………………………………………………..
75
11
Chapter 1. Introduction
1.1. General introduction
Seabirds live almost exclusively at sea, except during the breeding season when
they become tied to their nests, returning to the ocean to forage for both themselves and
their chicks (Shealer 2002; Garthe et al. 2003). Because they are top predators and are
connected closely to the marine environment, they are often used as indicators of marine
ecosystem health (Diamond and Devlin 2003; Weimerskirch et al. 2003a).
Seabird diet characteristics, such as composition and feeding rates, are often
topics in ecological research aimed at understanding, for example, both inter and intraannual diet variation, differences between sympatric species, and seabird energetics
(Lemmetyinen 1973; Erwin 1977; Diamond 1983; Ramos 2000; Hall et al. 2000; Nisbet
et al. 2002; Diamond and Devlin 2003; Adams et al. 2004). Despite this, there are still
major aspects of seabird biology that remain unstudied, yet are extremely important in
determining how a changing environment could affect populations. In particular, it is the
at-sea foraging ecology of seabirds where significant gaps in our knowledge exist (Hall et
al. 2000; Granadeiro et al. 2002; Shealer 2002), due in part to high costs, technological
challenges, and logistical constraints of conducting research at sea. In order to maintain
seabird populations it is important to know where these birds find food, and use this
information to ensure that necessary habitat is protected and available during the breeding
season.
Arctic Terns (Sterna paradisaea) are seabirds that breed inland and along the
coast in northern latitudes of North America, Greenland, Europe, and Asia. Along the
coast, they are found south to 41ºN, in Massachusetts (Hatch 2002). Common Terns
12
(Sterna hirundo) are also found throughout North America, Europe and Asia. In eastern
North America they are found south to the Gulf coast of Louisiana, (Nisbet 2002). These
species nest sympatrically south of 52ºN, in the southern portion of the Arctic Tern
breeding range, along the coast of eastern North America.
In the later part of the 19th century tern numbers were heavily depleted by hunters
collecting eggs and birds for food, as well as feathers for the millinery trade (Drury 1973;
US Fish and Wildlife Service 2004). While the end of these practices allowed the Gulf of
Maine tern population to recover, subsequent loss of nesting habitat to gulls and human
development, and increased gull predation, caused Gulf-wide numbers to fall (US Fish
and Wildlife Service 2004). In Maine, Arctic Terns are listed as a species of special
concern by the US Fish and Wildlife Service, and in the Atlantic Provinces they are
considered “sensitive” by the Canadian Wildlife Service (Hatch 2002). Common Terns
are listed as either endangered, threatened, or of special concern in US coastal states, but
are unlisted in all Atlantic Provinces except Nova Scotia, where they are considered
“sensitive” (Nisbet 2002). Since 1977 management plans in Maine and New Brunswick
have been implemented to protect vulnerable populations from further depletion (US Fish
and Wildlife Service 2004).
Arctic and Common Terns nesting on Machias Seal Island, N.B. (hereafter MSI,
44º30'N, 67º06'W), are an integral part of that seabird colony. They act as “umbrella”
species, chasing away predators, thereby helping to protect the other seabird species that
nest on the island. Much is known about their diet (Amey 1998; Diamond and Devlin
2003; Charette 2005) and daily activities (Paquet 2001) from past studies. Charette
(2005) concluded that Arctic and Common Terns nesting sympatrically feed at similar
13
trophic levels, and Paquet (2001) found that Arctic Terns spend much more time away
from their nests after chick-hatching than during incubation. Despite an abundance of
information on what terns breeding on MSI eat, we still do not know how far away or in
what direction they go to find their prey, or what foraging ground characteristics are
important to them.
1.2. Purpose of study
The objectives of this study are:
1. To determine if Arctic and Common Terns forage in different places.
Ecological theory (Gause’s Law) suggests that they likely forage in different geographic
areas, but since they feed at the same trophic level and on similar prey species, they may
also feed in similar habitat.
2. To identify and compare physical and biological characteristics of feeding areas
used by Arctic and Common Terns nesting on MSI, and determine whether they are
foraging at random or in predictable areas.
3. To determine if there is a change in foraging areas between the incubation and
chick-rearing nest stages of the breeding cycle of both tern species to reflect the time
budget switch documented in Arctic Terns by Paquet (2001).
14
Chapter 2. Literature Review
2.1. Study area
The Bay of Fundy and Gulf of Maine are productive areas of the ocean because of
a unique mix of physical and biological processes (Wahle 2000). The primary and
secondary productivity of this region are fuelled by the Eastern Maine Coastal Current,
which brings nutrient-rich water south from the Bay of Fundy into the Gulf of Maine
(Wahle 2000; Yoder et al. 2002; Fox et al. 2005). Additionally, the Gulf of Maine lies
just to the north of Georges Bank, one of the most productive ocean ecosystems in the
world (Yoder et al. 2002; Fox et al. 2005).The Gulf of Maine has its highest level of
phytoplankton chlorophyll (a measure of primary productivity) from April to October,
peaking in August (Yoder et al. 2002). Surface chlorophyll concentrations are relatively
stable from year to year, but show a considerable amount of within-season variability
(Yoder et al. 2001). Extensive vertical mixing in the waters adjacent to MSI also forces
nutrients from the bottom of the ocean to the surface, causing salinity and water
temperatures to be relatively uniform throughout the water column (Apollonio 1979;
Brooks and Townsend 1989).
Machias Seal Island is a small island located approximately 19 km southwest of
Grand Manan Island, New Brunswick, at the mouth of the Bay of Fundy, in the Gulf of
Maine (Figure 1). MSI is designated a Federal Migratory Bird Sanctuary and is managed
by the Canadian Wildlife Service. It is the summer breeding grounds for multiple seabird
species, including ca. 2,800 pairs of Atlantic Puffins (Fraturcula arctica), ca. 530 pairs of
Razorbills (Alca torda), Arctic and Common Terns (Sterna paradisaea and S. hirundo,
ca. 2,000 pairs and ca. 1000 pairs respectively), ca. 150 pairs of Leach’s Storm-Petrels
15
(Oceanodroma leucorhoa), and smaller numbers of Common Murres (Uria aalge)
(Bernard et. al. 1999; Diamond and Devlin 2003; Black et al. 2005). MSI is located near
the middle of the breeding range of Common Terns nesting in North America, and at the
southern edge of the breeding range of North American Arctic Terns (Nisbet 2002; Hatch
2002). The island consists mainly of granite bedrock with a vegetated central area.
Since 1995 Dr. Antony Diamond, of the University of New Brunswick (UNB)
and the Atlantic Cooperative Wildlife Ecology Research Network (ACWERN), and his
graduate students have been conducting research on the seabirds of Machias Seal Island.
The scope of this research has been wide; projects include investigating how seabird prey
can predict future commercial weir landings (Amey 1998), Arctic Tern time budgets
(Paquet 2001), Arctic Tern metapopulation dynamics (Devlin 2006), evaluating the
usefulness of seabirds as marine bioindicators (Diamond and Devlin 2003), Razorbill
habitat use (Grecian 2005), Atlantic Puffin metapopulation dynamics (Breton 2005;
Breton et al. 2006a, b), using stable isotopes to infer tern diet (Charette 2005), and longterm diet variability in 4 seabirds (L. Minich, in progress).
2.2. Gause’s Law, Optimal Foraging Theory, and tern time budgets
Gause’s Law
It is thought that seabird breeding is timed to coincide with peak food availability
(Perrins 1970; Safina and Burger 1985). Despite this, prey availability is considered a
limiting factor on breeding grounds, and seabirds in particular show notable declines in
breeding success corresponding to a reduction or collapse in prey stocks (Hunt 1972,
Safina et. al 1988, Newton 1998, Litzow et. al. 2002, Suryan et. al. 2002). Colony size
16
and spatial distribution with respect to other breeding colonies are other indicators that
prey may regulate seabird populations (Furness and Birkhead 1984, Ainley et. al. 2003).
This suggests that in a colonial setting there may be competition for limited food among
similar predatory species, such as Arctic and Common Terns (Ashmole 1968; Diamond
1978; Monteiro et al. 1996; Mori and Boyd 2004). On Machias Seal Island in 2004 and
2005 both Arctic and Common Terns experienced reduced chick growth rates and
extremely low productivity (Black et. al. 2005, Bond et. al. 2006). Although prey
availability was not measured, the poor performance in both years is likely due, at least in
part, to a preponderance of low quality food in chick diets, indicating that food was a
limited resource for both species.
Competitive exclusion is a basic principle in ecological studies (DeBach 1966;
Lack 1971). The theory was originally put forth by Grinnell in the early part of the 20 th
century (Grinnell 1904), but it was Gause who made the theory popular, and hence it is
usually referred to as Gause’s Law or Gause’s Principle (Hardin 1960; Lack 1971).
Gause (1934) describes competition between very similar species as unlikely, because
one species will tend to push out inferior competitors, and each species will eventually
occupy different niches. Like many scientific terms, this theory has evolved and can be
defined in several ways. Debach’s (1966) competitive displacement theory stated that if
two species share the exact same ecological niche, then they will not be able to co-exist
in the same habitat for long if resources are limited. Since then, scientists have continued
to explore inter-specific competition when resources are limited, and found that
coexistence is sometimes possible if resources are temporally or spatially variable
(Naeem 1988, Heard and Remer 1997).
17
Gause’s Law is important to consider when studying two congeners living
colonially. When discussing competition Gause (1934) specifically referred to 4 species
of sympatric-nesting terns in the Black Sea. He noted that each species occupied a
different feeding niche in order to coexist on a small island. Like the terns in Gause’s
discussion, Arctic and Common Terns are very similar in size and appearance. Both are
small seabirds; see Table 1 for a comparison of body measurements (Nisbet 2002; Hatch
2002). Their plumage is also strikingly alike; white body feathers underneath, with grey
back and wings and dark outer wing primaries, and a dark cap. Arctic Terns have slightly
redder bill and legs, whereas Common Tern bills and legs are more orange and the bill
has a black tip.
Despite their close physical resemblance, these species are biologically different,
especially on their wintering grounds. Arctic Terns spend the winter months feeding
along the pack ice of Antarctica (Hatch 2002), while Common Terns migrate to Central
and South America (Nisbet 2002). Considering these points, it is of interest to examine
the co-existence of Arctic and Common Terns when they are overlapping temporally and
spatially on the breeding grounds, because it is here that the issue of competition for prey
arises, not on their wintering grounds.
Ashmole (1968) described an example of competition between Brown Noddies
(Anous stolidus) and Sooty Terns (Sterna fuscata) that could also apply to Common and
Arctic Terns nesting together. Both species are of similar body and bill size, and are not
limited in nesting habitat where they are found together. They use similar fishing
methods to catch prey that are similar in both type and size. In describing where their
ecological niches differ, Ashmole concluded that Brown Noddies and Sooty Terns must
18
be separated by where they feed, and that it is this difference that allows them to continue
to co-exist on the same islands without one species driving the other out. In more recent
years this concept has been investigated by numerous seabird ecologists who found clear
foraging habitat segregation between sympatric seabirds (Abraham and Ankney 1984;
Safina 1990; Mori and Boyd 2004). As with the Brown Noddies and Sooty Terns in
Ashmole’s study, co-existence of Arctic and Common Terns on their breeding grounds
suggests superficially that they occupy the same niche, so possible differences- especially
in relation to feeding biology- are of broader theoretical interest.
Optimal Foraging Theory and time budgets
Optimal foraging theory (Stephens and Krebs 1986) suggests that animals should
try to maximise the energy gain from a food item per unit of time required to obtain it.
Time spent travelling to a food source is energetically expensive; during transit a bird is
unable to feed, and is therefore losing energy. To compensate for this, as transit time and
distance increase, the return in food must also increase (Stephens and Krebs 1986;
Wilson and Wilson 1988; Garthe et al. 2003). For example, Weimerskirch et al. (2003b)
found that Northern Gannets will travel small distances from their nest to find food for
their chicks, but will lose mass on these short foraging trips. When they travel further,
they make up for the loss of energy on short trips by gathering enough food to be able to
feed both themselves and their chicks. In building on optimal foraging theory, Stephens
and Krebs (1986) include models that allow for more realistic investigations of how
animals deal with constraints, such as central-place foraging, which will be described
further in section 2.4.
19
Time spent foraging changes from one part of the annual cycle to another in many
bird species (Goldstein 1988). Paquet (2001) found that Arctic Terns nesting on MSI
increase the time they spend away from the island from 38% during incubation to 70%
after chick hatching. The time spent away from the colony is presumed to be spent
foraging. It has also been shown that as the breeding season progresses, terns feed their
chicks larger fish, which translates into an increase in energy gain (Chapdelaine et al.
1985). With reference to optimal foraging theory, this observation indicates that Arctic
Terns could be foraging in different areas depending on the status of their nests, or they
could be selecting bigger fish at the same foraging sites as the season progresses. It
remains unclear which strategy is being employed.
2.3. Tern diet
Both tern species feed on a variety of juvenile schooling fish and invertebrates
(Burger and Gochfeld 1991; Hall et al. 2000; Nisbet 2002; Hatch 2002), mainly
euphausiid shrimp (Meganyctiphanes norvegica), Atlantic Herring (Clupea harengus),
Butterfish (Peprilus triacanthus), and three species indistinguishable in the field, White
Hake (Urophycis tennuis), Silver Hake (Merluccius bilinearis), and Four-beard Rockling
(Enchelyopus cimbrius) (Lemmetyinen 1973; Hall et al. 2000; Black et al. 2005). The
proportion of these prey species found in tern chick diet differs between Arctic and
Common Terns, and between years in both species (Burness et al. 1994; Black et al.
2005). This may be due to temporal and spatial variation in prey availability and foraging
habitat (Becker et al. 1993; Hall et al. 2000). Common Terns take slightly larger prey
than Arctic Terns (Lemmetyinen 1973; Black et al. 2005), and often feed on a larger
range of prey types (Hall et al. 2000).
20
The following is a general description of the main prey found in tern chick diets
on Machias Seal Island from 1995 to 2005 (Table 2).
Euphausiids:
Euphausiids have been an important part of tern diet since 2001, comprising up to
74.1% of Common Tern and up to 90.6% of Arctic Tern chick diet (Table 2). In 2004 and
2005 euphausiids were the most commonly delivered prey item for both species (Table 2,
Figure 2).
M. norvegica, also known as Northern Krill, are one of the most widespread
zooplankton species in the Atlantic Ocean (Saborowski et al. 2002), and is presumed to
dominate the euphausiid part of tern diets. The distribution of this species is regulated
predominantly by temperature, occurring in water temperatures of 2-15ºC (Saborowski et
al. 2002). Northern Krill are diurnal migrants in the water column; during the day they
are found in waters deeper than 70 m, and are usually found at the surface during the
night (Nicol 1984; Onsrud et al. 2004). Terns take advantage of this species when they
form sporadic and unpredictable surface swarms during the day because of ocean
upwelling, aquatic predators, or to forage (Nicol 1984).
Herring:
Historically Atlantic Herring were considered to be the preferred prey of Arctic
and Common Terns on MSI because both tern species fed their chicks more herring than
the 4 other main prey species (Table 2). Before 2001 herring made up from 40-85% of
tern chick diet and was the most abundant food item. Since then, the frequency of herring
has dropped to 0-39% of tern chick diet, and in 2005 it was the prey item taken least of
21
the 5 main Arctic Tern prey types, and the second least of the 5 main Common Tern prey
items (Table 2, Figure 2).
Juvenile Atlantic Herring are pelagic fish that inhabit the coastal waters of the
Gulf of Maine in May and June, but in July they begin to move further offshore into
deeper water for the winter months. In the Gulf of Maine juveniles are most often found
at depths of 15-135 m, where temperatures range from 6-9ºC (Reid et al. 1999). It should
be noted that these estimates are for juveniles matching the size class taken by terns
(approximately 3 - 6 cm, Bond et. al. 2006), but also for herring up to 25 cm long, which
is much larger than the size normally comprising tern diet. Habitat characteristics for only
very small herring juveniles were not found.
Hake and Four-beard Rockling:
Hake and Four-beard Rockling are indistinguishable in the field, so when
determining the proportion of tern diet that these species comprise they are referred to
collectively as hake. While hake is never the most important prey item in tern chick diet,
it is always present; since 1995 Common Tern chick diet has consisted of between 0.5
and 36.4% hake and Arctic Tern chick diet has been made up of between 1.9 and 27.2%
hake (Table 2). In 2005 hake was the third most abundant prey item delivered to chicks
of both tern species (Table 2, Figure 2).
Juvenile Silver Hake are abundant in coastal Gulf of Maine, and prefer bottom
temperatures of 6-9 ºC, and average water depths of 150-275 m (Morse et al. 1999).
Another species of hake found in large numbers in the Gulf of Maine is White
Hake. In trawl surveys in Massachusetts, juveniles caught inshore were usually found at
depths <75 m, and at bottom temperatures of 4-14 ºC. Offshore trawls found larger
22
juveniles at depths of <225 m, and at cooler bottom temperatures of 4-9 ºC. In a study of
near-shore hake in southwest Nova Scotia, their preferred habitat was described as warm
and turbid (Chang et al. 1999).
Information on juvenile Four-beard Rockling habitat is sparse. Adults of this
species are found in water depths ranging from 20-650 m, and no temperature preferences
are known (Atlantic Reference Centre 2006).
Butterfish
Butterfish makes up a varying portion of tern diet on MSI, although it is usually
represented in low numbers. From 1995 to 2005 between 0 and 33.3% of identified
Common Tern chick diet was Butterfish, with a large increase from the previous high of
7.5% in 1996, to 22.3% in 2004 and 33.3% in 2005 (Table 2, Figure 2). From 0 to 6.9%
of identified Arctic Tern chick diet was composed of Butterfish from 1995 to 2005 (Table
2, Figure 2). This species occupies the entire mid-Atlantic Shelf region during the
summer, including bays, estuaries, and open ocean up to 200 m beyond the shelf.
Juvenile Butterfish are pelagic and tend to be found at depths of a few meters to 180 m
where temperatures are generally 7-20ºC, which is warmer than the preferred
temperatures of the other major prey items (Cross et al. 1999).
2.4. Foraging distance estimations and habitat characteristics
Foraging distance
Using flight speed and time spent away from the nest, Pearson (1968) calculated
that Common Terns on the Farne Islands, U.K., travelled up to 21.9 km from the nest
while foraging. Becker et al. (1993) determined that Common Terns forage close to shore
23
(within 10 km), using coarse calculations based on radio-telemetry triangulation that
likely underestimated their true range due to the spatial limitations of this method; 12%
of the tracked birds travelled too far to be detected by triangulation. Nisbet (2002)
extended this feeding range to within 20 km of the breeding site, stating that although
they usually feed closer than this, they may also travel much further to forage. Nisbet’s
estimate was made using a combination of radio-telemetry data, trip duration estimates,
and boat surveys. The discrepancies between the range estimates of past studies may be
the result of either different distances to the feeding grounds at each breeding location, or
differences between sampling methods, or perhaps a combination of the two.
Arctic Tern foraging distances are poorly documented, but estimates vary from
less than 20 km (Hatch 2002) to a mean of 20.2 km (Pearson 1968). Like Pearson’s
Common Tern range estimates, his Arctic Tern range approximations were derived using
time spent away from the nest and estimated flight speed.
A prediction of central-place foraging models for seabirds that take a single prey
item per hunting episode is that the size of prey taken is related to travel time (Diamond
1983) and distance travelled to feeding grounds (Stephens and Krebs 1986). In some
studies Common Terns fed on slightly larger prey than Arctic Terns (Lemmetyinen 1973;
Black et al. 2005). Because the wing-loading (mass/wing area) of these 2 species does not
differ (Craik 1998) and they fly at similar speeds (Pearson 1968), comparing Arctic and
Common Tern foraging distances in the context of central-place foraging theory suggests
that Common Terns may travel further than Arctic Terns to find prey. Conversely,
Pearson (1968) and Hopkins and Wiley (1972) did not detect any difference in prey size
between Arctic and Common Terns. Conflicting results of prey size comparisons and
24
high overlap in foraging distance estimations indicate that perhaps the difference in prey
size taken by these two terns is too small or too ephemeral to drive a detectible difference
in foraging range, or that central-place foraging theory cannot be used to compare 2
species.
Habitat characteristics
Foraging studies have related tern feeding grounds to biological factors,
especially the presence of the predatory Bluefish (Pomatomus saltatrix) (Erwin 1978;
Safina and Burger 1988; Safina 1990; Ramos 2000; Nisbet 2002; Hatch 2002), and
environmental factors such as water depth and clarity (Abraham and Ankney 1984;
Safina 1990; Ramos 2000), tidal cycle (Becker et al. 1993; Hatch 2002), wind speed
(Dunn 1973; Safina 1990) and temperature (Safina 1990).
Diamond (1978, 1983) suggested that inshore feeders take a greater diversity of
prey types compared to pelagic predators (probably due to differences in environments).
Hall et al. (2000) hypothesised, based on prey diversity studies, that Common Terns
prefer to feed in inshore bays (even those birds nesting up to 15 kilometres off shore),
and Arctic Terns forage mainly over open ocean waters and along stony shores.
However, on MSI Common Terns do not always take a greater diversity of prey than
Arctic Terns (Table 3), so this may not be the case at this colony.
2.5. Previous Radio-tracking Studies
Radio-telemetry has become an important tool for ecologists trying to determine
individual animal activities through time and space. Radio-telemetry allows us to gather
information about animals when it would otherwise be difficult or impossible, such as
25
over dangerous or expansive terrain. This situation applies to island-nesting seabirds that
forage over tens to thousands of square kilometres of ocean. By attaching radio
transmitters to individual birds it becomes possible to collect data on where and when
these birds are foraging (Trivelpiece et al. 1986; Kenward 1987; Becker et al. 1993;
Monaghan et al. 1994).
Despite extensive studies of the breeding and foraging biology of both Arctic and
Common Terns, systematic investigation of their foraging grounds using radio tracking
has been limited, due probably to the high costs and logistical difficulties associated with
these types of studies. Tracking studies of Common Terns have been conducted in the
freshwater Great Lakes system (Burness et al. 1994), and in the Wadden Sea, Germany
(Becker et al. 1993). Studies of radio-tracked Arctic Terns are limited; however, a
Masters thesis describing Arctic, Common, and Roseate Tern foraging habits on Country
Island, Nova Scotia, using radio-telemetry has recently been compiled (Rock 2005). This
work is also the only known radio-tracking study on terns in the western Atlantic Ocean.
The purpose of my study is to describe the foraging ground characteristics of
Arctic and Common Terns breeding on MSI, and compare the habitat used between these
two species. An important aspect of the biology of any species is feeding ground
characteristics, an area where there is a considerable lack of information for these terns.
The results of this study will help determine which areas around MSI are important
feeding grounds, and facilitate conservation to ensure healthy populations of terns in the
Gulf of Maine.
26
Chapter 3. Methods
3.1. Productivity monitoring
Intrinsic to any wildlife radio-tracking study is accounting for changes in
behaviour potentially caused by the researcher. Because re-trapping terns during the
chick rearing period to directly measure changes in body condition has proved difficult,
measurements of nest and chick productivity were recorded to infer the impact of the
transmitters on adult tern behaviour.
We collected productivity data from nests in the radio-tracking study as well as tern
nests monitored for the long-term productivity study. The information collected included
nest initiation and hatch dates, and the number of eggs hatched and number of chicks
fledged per nest. Although nests with unknown fate were included in calculating average
nest initiation dates, hatch dates, and hatching success, only those nests with known fate
of all eggs or chicks were used in the final productivity calculations. This method follows
standard MSI protocol (Diamond et al. 2002).
3.2. 2004 data collection
2004 was an exploratory field season, and data collected were not used in the final
analyses. Trapping began on 12 June; 20 birds of each species (Arctic and Common
Terns) were captured approximately one week after clutch initiation. Early-nesting birds
were chosen to trap because older, more experienced birds lay earlier than younger birds
(Chapdelaine et al. 1985; de Forest and Gaston 1996; Arnold et al. 2004). This method
allowed variation due to age to be reduced. We used walk-in treadle traps (Weller 1957)
to trap adults on their nests. Decoy eggs were placed in the nest while trapping to avoid
27
breaking the real eggs, which were kept warm and dry until returned to the nest after the
trap was removed. Trapping methods are outlined in the Machias Seal Island Protocol
(Diamond et al. 2002). Pairs of terns were tagged in an attempt to compare foraging
areas between mates.
We recorded morphological measurements and banded the birds before attaching
a 3.4 g transmitter supplied by Lotek Wireless Inc (Newmarket, Ontario). Other tern
radio-tracking studies have attached the transmitter to the band on the leg (Morris and
Burness 1992), but this attachment method does not guarantee that the transmitter will be
shed before migration. Transmitters were attached to the 4 middle retrices and the
undertail coverts with cotton thread and epoxy to ensure that they would drop off during
the next molt (Figure 3a). Similar methods have been used on other species, such as
hawks and kittiwakes (Kenward 1987), as well as on Common Terns during a pilot study
discussed in Massey et al. (1988). It took us approximately 20 minutes to process each
bird using this method.
The birds used in this study were from areas monitored for productivity and
growth; two plots per species were used, with 1 plot containing both Arctic and Common
Terns, for a total of 3 plots (Figure 4). There were 10 tagged birds in each plot containing
only 1 species (Gully and Foundation plots), and 20 tagged birds in Oceanspray plot
containing both species (total = 40 tagged birds). These plots were monitored during
standard feeding watches and during tracking flights to provide information on prey type
and size delivered to chicks, and to link where the birds were found feeding to what they
fed their chicks.
28
Tracking usually began at least one week after the transmitters were attached to
allow the birds to acclimate to the equipment (Ministry of Environment 1998). One bird
had the transmitter attached for 2 days before the next tracking stint. This bird was not
located during the flight. Arctic and Common Tern frequencies were entered randomly in
a programmable receiver capable of polling through all frequencies at a specified rate.
During both aerial and land-based tracking stints the radio frequencies were searched for
9 second intervals before moving on to the next frequency.
3.2.1. Aerial tracking
Aerial radio-tracking was done using a fixed-wing twin-engine Seneca airplane
mounted with 2, 3-element Yagi antennas. One antenna was placed perpendicularly to the
plane on the footstep for the first 2 flights. Another antenna pointing directly forward was
installed in the nose of the plane for all subsequent flights. Tracking stints ranged from
1.75-3 hours between 29 June and 23 July, and were weather-dependent. During this time
6 flights were conducted, with birds being located on 5 of the 6 trips. This low encounter
rate is likely due to the loss of transmitters and the failure of nests early in the breeding
season.
Flights extended from Grand Manan Island to approximately 30 km around MSI.
The 30 km searched area (approximately 2827 km2) covered both Grand Manan and
Maine coastlines, as well as open ocean. The sea was explored in a spiral pattern centred
on MSI. The plane travelled at approximately 115 km/h, and flew at an altitude of 1000 ft
(approximately 300 m) to avoid disturbing foraging birds. Only a portion of the 2827 km 2
search area was covered in one flight. If a stint began around MSI one day, the next stint
29
would start around the outer perimeter of the search area (i.e. along the Maine and Grand
Manan coast). This usually allowed the entire area to be covered in 2 flights.
As each signal was detected the latitude and longitude were recorded and, if
visibility permitted, characteristics of the foraging area were noted. Due to foggy
conditions on most flights a foraging tern was visible on only one occasion. If it was
unclear where a signal was coming from a minimum distance from MSI was recorded.
This was the case for most of the points obtained during the first 2 flights when only 1
antenna was used. The ability to detect birds and determine their position improved after
the addition of the forward-facing antenna. Characteristics of the transmitter signal were
also noted, specifically whether the signal was steady or breaking up. In several cases the
signal was interrupted periodically, presumably as the bird hit the water while feeding
(versus steady and moving above water).
3.2.2. Ground tracking
The aerial flights were biased because if a signal was detected within
approximately 2 km of the island it was unclear if the bird was on the island or feeding
very close. To quantify how many birds were feeding close to the nesting site land-based
tracking was attempted using 2 methods.
Method 1
One person was stationed at the top of the lighthouse, and one person at the
northern area of MSI, near Goofapuff blind (Figure 4). The frequencies were scrolled
through simultaneously on both receivers, and when one was detected a compass bearing
was taken.
30
Method 2
One antenna was used to point directly at a foraging flock and all frequencies
were scanned to determine if a tagged bird was in the flock. If a bird was located a
bearing was recorded, the signal was detected from another location, and a second
bearing was recorded. The foraging location was then determined by triangulating these 2
bearings. These data were gathered opportunistically, and only twice due to limited
detection of flocks in foggy conditions. One bird was tracked during each of these stints.
3.2.3. Exclusion of 2004 data
The field work in 2004 mainly consisted of testing equipment, learning radio
tracking techniques, and determining what data collection methods were realistic. The
data collected in 2004 are of much lower quality than data collected in 2005 when
equipment set-up and methods were consistent; therefore 2004 was considered an
exploratory year and the data were not used in the final analysis.
3.3. 2005 data collection
Trapping began on 16 June; each nest was trapped one week after clutch initiation
or, if initiation date was unknown, one week after peak colony lay. The same trapping
methods used in 2004 were applied in 2005; see section 3.2. 2004 data collection for
details. As in 2004, we targeted earlier-nesting birds thought to be older and more
experienced (Chapdelaine et al. 1985; Arnold et al. 2004). Nineteen Arctic Tern adults
and 19 Common Tern adults were captured, measured, and fitted with a 1.3 g bandmounted transmitter supplied by Holohol Systems Ltd (Carp, Ontario; Figure 3b). See
31
Morris and Burness (1992) for a detailed description of the transmitter and mounting
technique. Because retention of feather-mounted tags in 2004 was poor (see section 4.1.
Transmitter retention), in 2005 we mounted the transmitter to the band to ensure that the
transmitter would remain attached to the bird for the duration of the study.
Since the nests used in this study in 2005 were not monitored for chick diet, they
were chosen based on stratified random sampling rather than only using nests located in
plots visible from blinds, which was the method used in 2004. The trapped nests were
chosen using data from the 2004 tern census (Diamond in Black et al. 2005), which
counted all tern nests in a 30m x 30m grid system over the entire island. The number of
terns nesting in each grid-square was used to determine how heavily each grid square was
weighted when selecting where to trap birds. The first 20 grid-squares picked in a
random sorting of weighted grid squares were surveyed as the terns began nesting in
2005. As each nest was found it was marked with a unique number and the date the nest
was initiated using a wooden stir-stick.
To determine if the birds fitted with transmitters were representative of the entire
population, I determined the sex of each bird, and also compared their body condition at
trapping to terns without transmitters. Arctic Terns were sexed according to the
discriminant function in Devlin et al. (2003), where BD = bill depth (mm) and HB = head
+ bill length (mm). All birds with D > 0.043 were classified as female, with 74% correct
classification.
D = BD(1.222) + HB(0.362) - 34.222
After ensuring there were no differences in head + bill length or mass between the birds
used to build the equation and the birds on MSI (t-test for unequal sample sizes, α = 0.05;
32
head + bill: t = 0.69, p = 0.49; mass: t = -1.50, p = 0.14), Common Terns were sexed
according to the discriminant functions in Nisbet et al. (unpublished MS). Birds with
head length > (77.9 + d) were classified as males, and those with head length < (77.9 – d)
were classified as females. All birds with values in the middle of this range were not
classified as either male or female. For an 85% accuracy rate d = 1.0 mm.
While the best measurement to define the body condition of terns has not been
determined, Chastel et al. (1995) examined Blue Petrel measurements to determine the
best indicator of condition. In their study, principal component analysis results showed
body mass/culmen length was the best measure (culmen PCI loading = 0.51), followed
closely by body mass/wing length (wing PCI loading = 0.49). Because culmen measures
were not taken during tern trapping, I used mass (g) over wing length 3 (mm3) as a
measure of tern body condition corrected for structural differences between birds.
3.3.1. Aerial tracking
In 2005, tracking began at least one week after transmitters had been attached,
and frequencies were entered randomly in a programmable receiver capable of scrolling
through all frequencies at a specified rate. During aerial tracking I searched each radio
frequency for 9 seconds before moving on to the next frequency.
The equipment set-up, flight altitude and speed, and data collection followed that
of 2004: a fixed-wing twin engine aircraft was mounted with 2 3-element Yagi antennas;
one antenna was placed perpendicularly to the plane on the footstep and another antenna
pointing directly forward was installed in the nose of the plane.
33
The radius of the area searched around MSI was increased from approximately 30
km in 2004 to approximately 40 km in 2005 (Figure 5). This increase was due to the
occurrence of terns at the outer edge of the searched area, indicating that the birds were
foraging further than anticipated. I used a handheld GPS unit to map the flight path of the
plane by recording its position at regular intervals (Figure 6). These points represent
available foraging habitat. The plane flew at approximately 115 km/h, and at a height of
1000 ft (approximately 300 m).
The data recorded on each flight consisted of the position (latitude and longitude)
and characteristic (steady or in-and-out) of the signal and, on 4 occasions when visibility
permitted, features of the foraging habitat observed from the plane. If the signal was
steady the bird was considered to be travelling, while a strong but in-and-out signal
indicated a bird was foraging. When the antenna of the transmitters was submerged as the
terns dipped into the water to catch prey the signal became undetectable, but was
immediately clear again when the terns returned to hovering over the water, giving the
transmitter signal a characteristic strong but intermittent quality. The tracking flights
began on 29 June and ended on 16 July. A total of 9 flights of 1.3 to 3 hours were
completed (total flight time = 22.3 hours) and terns were located on all flights.
A transmitter was retained to check the range of the transmitters. One check was
performed by placing the transmitter on the top of a fishing boat cabin, and tracking it
from the lighthouse while in communication with the boat captain. When the signal was
no longer audible the distance from the island measured by the GPS in the boat was
recorded. During each flight the transmitter was placed in an open area near the airstrip.
34
Each time the plane left and arrived from the airstrip I recorded the transmitter range
using the airplane GPS.
3.3.2. Ground tracking
The bias against detecting foraging terns close to the breeding colony was
addressed again in 2005 by attempting land-based tracking from Goofapuff and
Foundation blinds. This also allowed researchers to track birds on days when poor
weather conditions limited work in the colony to observations from blinds. One antenna
was mounted on a pole approximately 12 feet high. The pole was attached temporarily to
the side of the blind, beside a window, with a large hose clamp, allowing the pole to be
turned 360º. The bottom of the pole was anchored by inserting it into a hole in the top of
an oil bucket with rocks in the bottom to increase stability. The antenna was connected to
the receiver via coaxial cable leading into the blind through the window, allowing an
assistant to scan frequencies from inside the blind by turning the pole. Although this
method seemed to work fairly well, it was used on only 2 occasions, and did not yield
enough data to use in analyses.
3.3.3. Environmental data
A Geographic Information System (GIS) was used to map habitat data in relation
to MSI and the mainland, and then determine which data corresponded to tern feeding
areas. The data layers included in the GIS were water depth (m), surface chlorophyll
concentrations (mg/m3), and sea surface temperatures (SST, ºC). Other variables
examined were distance from the breeding colony (MSI) and distance from the mainland.
35
These environmental characteristics were chosen based on foraging habitat characteristics
perceived to be important to terns.
Unlike SST and chlorophyll concentrations, water depth (bathymetry) is relatively
constant over time. Because of the predictability of bathymetry, it is indicative of where
small schooling fish and invertebrates may be found (Maravelias 1999; Le Pape et al.
2003; Ressler et al. 2005), and may allow foraging seabirds to find prey consistently in
the same areas.
Chlorophyll from phytoplankton is used as an indicator of primary productivity in
marine systems (Feldman 2006). Near-surface chlorophyll concentrations have been
linked to seabird distribution (Ainley et al. 1998), as well as to the occurrence of
euphausiid shrimp patches (Ressler et al. 2005), so presumably affect the distribution of
plankton-feeding fish as well.
Because terns are surface feeders, taking prey from the top 50 cm of the ocean
(Nisbet 2002; Hatch 2002), SST may play an important role in where they forage. Sea
surface temperatures also indicate the occurrence of prey in the water column because
they signal major physical processes, especially upwelling (Burger 2006). Weimerskirch
et al. (1995) used albatross fitted with data loggers to validate SST measurements taken
by satellite imagery over large areas.
Foraging distance from shore indicates inshore versus offshore feeding.
When calculating the distance from shore where each bird was found, small islands along
the Maine and Grand Manan coast were not included. See Figure 7 for a map of areas
classified as mainland coast and islands excluded from distance analyses.
36
Data layers were gathered from multiple sources (Table 4) and were compiled
into ArcView 3.3 and ArcMap 9.1 GIS programs (Environmental Systems Research
Institute 2002, 2005). Coastal coverages were obtained from the Coastline Extractor, a
GEODAS database (NOAA/National Geophysical Data Center, 2006). Bathymetric data
were from the Bedford Institute of Oceanography (D. Greenberg, personal
communication, 2005). Sea-surface temperature (SST) and chlorophyll a concentration
([chl]) data coverages were produced by National Oceanic and Atmospheric
Administration (NOAA), and supplied by the Satellite Oceanography Data Center at the
University of Maine (A. Thomas, personal communication, 2005).
3.4. Analysis
3.4.1. Productivity analysis
Analysis of Variance (ANOVA) was used to determine if there were any
productivity differences between terns equipped with transmitters and those with no
transmitters. Only the nests of untracked birds that were laid within the same time period
as nests of tracked birds were used in the comparisons. The parameters included in the
analysis were nest initiation date (the date the first egg was laid in the nest), clutch size,
hatching success, and the number of chicks fledged per nest.
3.4.2. Sex and body condition analysis
The probability of obtaining the observed sex ratios of birds with and without
transmitters was tested using a Fisher’s exact test. This test is useful when comparing
categorical variables (male vs. female, transmitter vs. no transmitter) with small n’s.
37
Adult mass (g) and wing length (mm) measurements recorded during trapping
were used to calculate a ratio indicative of body condition. The mass:wing3 ratios of terns
with and without transmitters were analysed using Analysis of Variance (ANOVA).
3.4.3. ArcView and ArcMap mapping
Since [chl] and SST measurements were projected in UTM zone 19, using
NAD83 datum, all other data layers, which were unprojected in a geographic coordinate
system (decimal degrees), were transformed into this data format to ensure a proper
overlap of features in different data layers.
The bathymetry points were on a 0.33 km grid and both [chl] and SST were on 1
km grids. For each tern and available habitat point, corresponding mean weekly [chl] and
SST values were used in the analyses. Daily values were not used because many of the
daily satellite images were obscured by cloud and fog. The error on each location of a
tern was considered to be 2 km, which was consistently the maximum distance a
transmitter could be detected when testing transmitter reception from either the
lighthouse or airplane, as described earlier. Therefore, I used ArcMap to place 2 km
buffers around each bird location and available habitat points, and average depth, [chl],
and SST measurements from each buffer were used in my analysis. SST and [chl] were
tested for spatial autocorrelation using Moran’s I index.
ArcView was used to find the shortest distance between tern and available habitat
locations and the mainland, and between locations and MSI.
38
3.4.4. Randomization analysis
2005 was a particularly poor year for tern productivity; many of the tracked nests
failed, and adults were located most when they did not have an egg or chick in the nest.
Because of this, I was able to compare tern habitat use during egg, chick, and failed nest
stages. Failed nests included abandoned, broken, flooded, and depredated nests.
Few terns were located at each stage in the lifecycle (sample size varied from 4 to
9 locations per lifecycle stage), so assumptions of normality and variance were not
measurable, making the results of standard statistical tests unreliable (Manly 1991). One
way to deal with this is through randomization testing. Randomization techniques are
useful when assumptions about the underlying population necessary for most statistical
analyses are unknown (Manly 1991; Gonzalez and Manly 1998; Ninness et al. 2002a).
I used available habitat points sampled along the flight path to determine if terns
were found foraging at random by comparing these points with tern feeding points at
each stage in the lifecycle, and for all stages combined. For each species and each habitat
variable measured (SST, [chl], distance from the mainland, distance from shore, and
water depth), available habitat points were sampled using a randomization procedure.
This process chose n points randomly from the sample to determine the mean and 95%
confidence interval for 10 000 iterations, with n = n of the observed tern points. The
resulting p-values indicate the proportion of randomization iterations that had mean
values either greater than or less than the observed tern values. I used α = 0.05 throughout
to determine statistical significance.
Randomization analyses have been shown to control for type I error better than
traditional parametric and non-parametric statistical tests, especially with small sample
39
sizes (Gonzalez and Manly 1998; Ninness et al. 2002a; Ninness et al. 2002b). However,
relatively little literature discussing the calculation of statistical power, or the chance of
committing type II errors, exist (Peres-Neto and Olden 2001). Estimations of the power
of the tests used in this study would improve the confidence of my conclusions, but
power analyses are not feasible with my randomization methods because I did not
produce a standard test statistic (Peres-Neto and Olden 2001).
40
Chapter 4. Results
In the following results I have abbreviated tern names to conventional four letter
Bird Banding Laboratory species names; Arctic Terns are referred to as ARTE, and
Common Terns as COTE.
4.1. Transmitter retention
In 2004 transmitters were attached to the 4 middle tail feathers using cotton thread
and epoxy. This attachment site was unsuccessful; the terns either dropped their tail
feathers or there was too much stress on the feathers, causing the feathers to fall out. Of
the 40 transmitters deployed, 12 transmitters were found on the island no longer attached
to a tern, but an unknown number may have become detached and were not recovered.
Seven of the detached transmitters found around the island were re-attached to new terns.
In 2005, when transmitters were attached to the tern leg band, no transmitters
were found separated from the band. One year after the transmitters were put on the terns
3 Common Terns and 5 Arctic Terns were sighted with the transmitters still attached to
the band. One Arctic Tern still had both the transmitter and antenna attached to the band,
and, more seriously, one Common Tern appeared to be wounded as a result of the
transmitter attachment (Sarah Spencer, personal communication, 2006).
4.2. Transmitter effects on productivity
Nest initiation
41
ARTE selected to be equipped with transmitters in 2005 had a mean date of nest
initiation significantly earlier than the mean date for ARTE with no transmitters, but there
was no significant difference in 2004 (Table 5). COTE fitted with transmitters showed no
differences in mean lay date compared with birds without transmitters in either 2004 or
2005 (Table 5).
Clutch size
Mean clutch size of ARTE with transmitters was not significantly different from
birds without transmitters in either 2004 or 2005 (Table 6). Similarly, in both 2004 and
2005 there were no significant differences in clutch size between COTE with transmitters
and without transmitters (Table 6).
Hatching Success
In 2004 hatching success for ARTE with transmitters was significantly higher
than ARTE without transmitters, but there was no difference between COTE with and
without transmitters (Table 7). In 2005 no difference in hatching success was detected
between birds of either species with and without transmitters (Table 7).
Fledging success
ARTE with transmitters fledged significantly more young than those without
transmitters in 2004, but not 2005 (Table 8). COTE with transmitters did not fledge more
chicks than birds without transmitters in 2004 (Table 8). In 2005 the Common Tern
colony failed, and no monitored nests fledged any chicks.
42
4.3. Sexing and body condition analysis
Using discriminant functions I determined that 58% of Arctic Terns fitted with
transmitters were female, compared with 54% of Arctic Terns captured and not fitted
with a transmitter. The discriminant function used to sex Common Terns did not allow 5
of the 19 birds fitted with transmitters and 10 of the 27 birds without transmitters to be
sexed. Of those that were sexed, 43% of birds with transmitters were classified as female,
while 47% of birds without transmitters were classified as female. These proportions
were not significantly different (Fisher exact test, p = 1.00).
Body condition of birds of both species equipped with transmitters did not
significantly differ from those without transmitters (Table 9).
4.4. Foraging location mapping
In total, I located 30 ARTE and 23 COTE during tracking flights. Of these, 21
ARTE and 18 COTE were foraging when they were detected, so only these points were
mapped and used in my analysis. See Figure 8 for ARTE foraging locations and Figure 9
for COTE foraging locations.
4.5. Habitat characteristic analysis
Foraging locations of terns at three stages of the breeding cycle (egg incubation,
chick rearing, and failed nest) were compared to available habitat and examined for
changes in foraging ground characteristics. I also examined all 3 stages combined for
each species, and these are referred to as pooled points. The points gathered on each
43
flight to represent available foraging habitat are referred to as available points. Results
are summarized by each habitat variable examined.
Distance from the mainland
ARTE foraging grounds were located from 1 km to 46 km away from the
mainland; mean distance from the mainland for all locations was 20.9 km (± 9.3 km)
(Table 10a; Figure 10). COTE were found feeding as close as 0.3 km and as far as 37 km
from the mainland; average distance from the mainland for all locations was 16.0 km (±
12.7 km; Table 10a; Figure 10). The distance from the mainland for foraging locations
did not differ between each stage of the breeding cycle and the available points (Table
10a). No significant difference was detected for either ARTE or COTE when the average
foraging distance from the mainland of pooled points was compared to available points
(Table 10a).
Distance from MSI
ARTE were found foraging from 4 - 34 km from MSI (mean = 17.3 ± 8.8 km;
Figure 11), and COTE 9 - 30 km (mean = 19.8 ± 5.9 km; Table 10b; Figure 11). The
distance travelled from MSI to foraging areas during each period of the breeding cycle
did not differ for either ARTE or COTE during each breeding stage (Table 10b). No
significant difference was observed when I compared pooled points to the available
points (Table 10b; Figure 11).
Water Depth
On average, ARTE were found over water depths of 98.3 m (± 43.0 m; range = 19
to 200 m; Figure 12). COTE were found foraging in water depths ranging from 3 to 187
m, with an average water depth of 85.8 m (± 53.0 m; Figure 12). When the depth of the
44
feeding areas for each species at each nesting stage was compared to water depth at
available points, the only significant difference observed was for ARTE with failed nests
(Table 10c). During this stage ARTE were observed feeding in deeper water areas
(Figure 13). COTE showed no significant change in foraging area water depth between
breeding cycle stages. No significant foraging area depth preference was observed for
either species when all points were pooled (Table 10c).
Ocean Chlorophyll concentration
Average surface chlorophyll concentration in ARTE feeding areas was 1.5 mg/m 3
(± 0.4 mg/m3); concentrations ranged from 0.9 to 2.5 mg/m3 (Table 10d; Figure 14).
COTE were found feeding in areas with average chlorophyll concentration of 2.1 mg/m3
(± 0.7 mg/m3); values ranged from 1.3 mg/m3 to 3.3 mg/m3 (Table 10d; Figure 14).
ARTE showed no preference for areas of high or low chlorophyll concentration, either
during a particular breeding cycle stage or when all stages were combined (Table 10d).
COTE were found foraging in areas where chlorophyll concentrations were significantly
higher when all breeding stages were combined (Table 10d; Figure 15), as well as during
egg incubation (Table 10d; Figure 16) but not when they had chicks or no nest (Table
10d). Chlorophyll values across the study area were found to be spatially autocorrelated
(Moran’s I index = 0.14, p = 0.01).
Sea surface temperature
The average sea surface temperature of ARTE feeding grounds was 12.2 0C (± 4.9
0
C), with values ranging from 5.7 0C to 26.2 0C (Table 10e; Figure 17). COTE were
found foraging in surface temperatures from 1.9 0C to 28.4 0C, with an average of 11.6 0C
(± 6.1 0C; Table 10e; Figure 17). A comparison of the average sea surface temperature of
45
foraging grounds for each species at each stage of the breeding cycle, as well as for
pooled points, indicated that neither species were observed foraging in areas that were
significantly warmer or colder than the available points (Table 10e). SST values were
found to be spatially autocorrelated (Moran’s I index = 0.09, p = 0.01).
46
Chapter 5. Discussion
5.1. Transmitter attachment
Attaching transmitters to the retrices of terns in 2004 resulted in the detachment
of many transmitters before the end of the study, resulting in the loss of valuable data.
The stress of attaching the transmitters to the retrices may be only temporary, but if no
useful information is collected then this procedure is not practical and should not be
employed.
Alternatively, attaching the transmitter to the leg band does not carry the same
risk of transmitter loss during data collection. While this does not ensure complete
removal of the transmitter package after the end of the study period, it does ensure data
collection from the tagged birds. While some studies have found no indication that there
are any negative effects associated with carrying this type of transmitter (Morris and
Burness 1992; Rock 2005), in this study at least one of the 7 birds recaptured in 2006
with the transmitter still attached to the band showed signs of injury from the transmitter.
As a result, I recommend that future radio telemetry studies on terns weigh the
consequences of lost transmitters to the possibility of injuring the birds, and alternative
methods of transmitter attachment should be seriously considered.
5.2. Transmitter effects on productivity
For all breeding parameters measured (nest initiation, clutch size, hatching
success, and fledging success) Common Terns equipped with transmitters showed no
difference in performance when compared to terns with no transmitters. Arctic Terns
47
with transmitters demonstrated either no differences or increased productivity between
birds with and without transmitters. Rock (2005) also found no differences in
productivity, which leads me to conclude that attaching transmitters to terns appears to
have no harmful effect on breeding performance, at least of early-nesting terns.
5.3. Sexing and body condition
The male:female ratios of both Arctic and Common Terns with transmitters were
not significantly different from the sex ratio of birds without transmitters. Likewise, body
condition scores (mass:wing3 ratio) of both the Arctic and Common Terns used in this
study were not significantly different from the other terns trapped during 2005. This
indicates that the terns fitted with transmitters were not necessarily in better or worse
condition than the general population. It can therefore be assumed that the terns with
transmitters are representative of the tern population on MSI, both in the ratio of sexes
and body condition. These comparisons are important because the results obtained from
the tracked birds are therefore applicable to the local tern population.
5.4. Foraging habitat
The first two goals of this study was to compare and identify physical and
biological characteristics of feeding areas used by Arctic and Common Terns nesting on
MSI, and determine whether they segregate by feeding in different geographic areas and
whether they forage at random or in predictable areas. The second objective was to
determine if there is a change in foraging area characteristics between the incubation and
chick-rearing nest stages of the breeding cycle. Because of unusually high chick mortality
48
in 2005, a large number of the terns located on the feeding grounds had failed nests;
therefore foraging areas during this breeding stage were also examined.
Distance from the mainland
Regardless of the status of their nest, neither species showed a preference for
coastal or offshore areas. Diamond (1983) reported that seabirds that forage closer to
shore tend to feed on a greater diversity of prey, and from this Hall et al. (2000)
concluded that Arctic Terns probably feed more offshore than Common Terns. On MSI
the number of prey species taken by terns fluctuates from year to year, and Common
Terns do not always feed on a greater diversity of prey than Arctic Terns, as was the case
in 2005 (Table 3). This indicates that on MSI, prey diversity does not necessarily signify
inshore or offshore tern foraging areas.
Distance from MSI
Arctic Tern foraging ranges have not been studied extensively, but estimates of
how far from the colony they feed are in the range of 20 km or less (Pearson 1968; Hatch
2002). On Country Island, Rock (2005) found Arctic Terns foraging 8.1 ± 4.6 km from
the colony. My results indicate that these estimates are conservative; the mean distance
from MSI that Arctic Terns were found feeding was 17.3 ± 8.8 km, but individuals were
found foraging as far as 34 km away. The mean distances were consistent throughout
each of the breeding stages examined (egg: 15.9 ± 9.8 km; chick: 18.4 ± 7.9 km; failed:
18.2 ± 9.9 km; Table 10b), indicating that the larger foraging distances observed in my
study were not heavily influenced by one stage of the cycle.
Pearson (1968) estimated maximum Common Tern feeding range to be 21.9 km,
similar to the 20 km range reported by Nisbet (2002). The maximum distance I located
49
Common Terns from MSI was 30 km, much further away than previous estimates.
Similarly, the mean distance from MSI that Common Terns foraged in this study for each
stage in the breeding cycle and for all stages combined was nearly double that of the 9.4
± 4.7 km range described by Rock (2005).
It is interesting to note that my estimates of mean distance travelled for both
species may be conservative because some terns were observed feeding near the outer
limits of the searched area. The terns could be foraging further than 40 km radius from
MSI that was surveyed, so the reported ranges should be treated as minimum foraging
distances, and further research should expand the searched area to account for terns
feeding very far from the colony.
The disparity between earlier range estimates for both species and this study could
possibly be explained by where the colonies are situated with respect to the mainland,
colony size, prey abundance, or different methodologies for measuring foraging range.
The islands in the studies by Pearson (1968; Farne Islands, Northumberland), and Rock
(2005; Country Island, Nova Scotia) are located close to shore (approximately 1.6 to 7
km, respectively), whereas MSI is further offshore (approximately 13 km).
Additionally, Erwin (1978) found that birds nesting in larger colonies travelled
further when foraging compared to those nesting in smaller colonies. MSI has a larger
breeding population of both species of terns than Country Island, and although the tern
population of the Farne Islands is similar (Cramp et al. 1974) they are spread over a few
islands. This implies that terns from MSI may be travelling further to feed as a result of
their larger colony size.
50
Another possible explanation for the differences in foraging range estimates could
be variations in prey abundance. If prey abundance is low, foraging trip length of seabirds
has been shown to increase greatly (Monaghan et al. 1994; Suryan et al. 2000). In 2005
both Arctic and Common Terns on MSI had unusually low productivity, likely due to a
lack of food, so perhaps they were searching further than usual due to a scarcity of prey
near the colony.
When comparing the results of these studies, it is important to note that the
techniques used to measure foraging distances are widely variable. These differences are
especially pronounced between my work and Pearson’s (1968) study. While I used aerial
telemetry to determine where foraging grounds were, Pearson’s estimates were based on
feeding intervals observed at the nest. This method calculates a maximum distance by
assuming that the birds fly at high speed directly to foraging areas, immediately catch
prey, and return directly to the nest. Considering these marked differences, the
discrepancies in range estimates between studies may be entirely methodological.
Interestingly, my results show that Arctic and Common Terns feed at similar
distances from MSI. While Rock (2005) also found Arctic and Common Terns travelling
similar distances from the colony, she reported that they were travelling in different
directions (Arctic Terns were feeding offshore and Common Terns were feeding inshore),
which was not observed in this study (see discussion in distance from the mainland
section above).
Water depth
Curiously, Arctic Terns were observed feeding in areas of significantly deeper
water during the latter part of the breeding cycle, when they no longer had a nest (122.4 ±
51
63 m). The same pattern was not demonstrated by Arctic Terns during incubation, chick
rearing, or with pooled stages, or by Common Terns. Even so, during all stages of the
breeding cycle both Arctic and Common Terns were foraging over water depths
associated with their most abundant prey, mainly euphausiids and butterfish for Common
Terns and euphausiids and hake for Arctic Terns (Table 2). Why Arctic Terns were
foraging in deeper areas when their nesting attempts had failed is unclear. This may
simply be a chance finding. In all, 40 comparisons were made between tern foraging
points and available habitat, so the chance of a few tests showing false significant results
is likely high.
It is interesting to note that the mean depths where both species were found
foraging were much deeper than the relatively shallow foraging areas reported for terns
nesting on Country Island, Nova Scotia (Rock 2005). This is certainly due to differences
in the distance that terns from each colony travelled to foraging grounds; terns from MSI
fed further offshore away from land, while terns from Country Island remained relatively
close to shore and in the vicinity of other near-shore islands (Rock 2005).
Surface chlorophyll concentration
My results indicate that Arctic Terns do not choose to feed in areas of higher than
average chlorophyll, and foraging locations of Common Terns during the early part of the
breeding season when they are incubating are characterised by higher levels of
chlorophyll. While it appears as though Common Terns were feeding in areas of high
productivity when all stages were pooled, it is probable that this difference is driven by
the deviation from the mean observed during the incubation stage, since no difference
was found during the chick rearing and failed nest stages.
52
High temporal variability of chlorophyll has been observed on the scale of days to
weeks in the Gulf of Maine (Yoder et al. 2001), which is the same scale investigated in
this study. The fact that terns seemed to be indiscriminate with regards to chlorophyll
levels when foraging may indicate that terns or their prey do not respond to such a fine
scale of variation, and perhaps larger scales, such as the seasonal time periods at which
oceanographers generally examine both chlorophyll and sea surface temperature
variability (Thomas et al. 2002; Yoder et al. 2002; Perez et al. 2005), should be explored
for relationships to tern foraging grounds. Similarly, both chlorophyll and sea surface
temperature were spatially autocorrelated, and although too few tern foraging points were
collected to correct for this, spatial variability of chlorophyll and sea surface temperature
is usually examined at gulf or basin-wide scales (Platt and Sathyendranath 1988; Fox et
al. 2005) or at the ocean shelf or slope scale (Brooks and Townsend 1989; Thomas et al.
2002; Yoder et al. 2002), both of which are much larger than my approximately 5000
km2 study area. Perhaps a larger scale of investigation could mitigate the effects of
autocorrelation of habitat variables.
Why Common Terns foraged in areas with high chlorophyll only during
incubation is unclear. Perhaps they are less tied to the nest during incubation than when
they are provisioning chicks, and can afford to seek out especially productive areas,
despite the highly ephemeral and unpredictable nature of surface chlorophyll
concentrations within seasons (Yoder et al. 2001).
This hypothesis is based on the assumption that ocean chlorophyll a concentration
is indeed a strong predictor of primary production (Feldman 2006), and therefore
zooplankton and fish. The relationship between chlorophyll a and ocean productivity has
53
been shown to be regionally variable due to environmental variables such as SST and
wind (Hayward and Venrick 1982; Eppley et al. 1985), so a more in-depth analysis of
surface chlorophyll and productivity in the Gulf of Maine and Bay of Fundy area may be
required to make strong conclusions linking seabird foraging areas to chlorophyll a.
Another possible explanation lies in the small number of terns located; although
randomization tests are appropriate when analysing populations with unknown
distribution and variance, only 3 Common Terns incubating eggs were located. It is
possible that randomization analyses are overly sensitive to such a small sample size,
although discussions on this topic in published literature are sparse (Ninness et al.
2002a). Like the significant result found when examining water depth for ARTE, these
comparisons may be significant by chance due to the high number of comparisons made
between tern feeding areas and available habitat.
Sea surface temperature
No published studies describing sea surface temperatures at Arctic Tern foraging
areas were found, so I am unable to make a comparison with other colonies, although
they showed a similar range of surface temperatures as Common Terns. The mean sea
surface temperature of Common Terns foraging areas in my study (11.6 ± 6.1 0C) is
slightly cooler than temperatures observed in another foraging habitat study (16.9 ± 0.82
0
C) comparing Common and Roseate Terns (Safina 1990), likely because that study was
further south in the Atlantic Ocean, where water temperatures are certainly warmer than
around MSI.
Since Arctic and Common Terns did not use feeding areas preferentially with
regards to surface temperature, it appears as though this variable is not important in
54
determining where to feed (at least at the scale investigated; see comments above under
surface chlorophyll concentration), but could be indicative of what prey are present. In
2005 euphausiids were the most important prey item for both Arctic and Common Terns
nesting on MSI, followed by Butterfish (Common Terns) and Hake (Arctic Terns). These
3 prey types are found in a wide range of temperatures (euphausiids: 2 - 15 0C,
Butterfish: 7 - 20 0C, Hake: 4 - 14 0C), including temperatures where terns in my study
were found (Chang et al. 1999; Cross et al. 1999; Morse et al. 1999; Saborowski et al.
2002). Herring, which was not an important part of tern diet in 2005 but has been in past
years (Table 2), are found at a much narrower range of temperatures (Reid et al. 1999).
Tern foraging locations with respect to sea surface temperature are either diverse because
the range of temperatures where their prey are found is diverse, or they are unable to find
herring, formerly their preferred prey, in the narrow temperature range it is generally
found at, and are searching in different habitat for other prey items.
5.5. Conclusion
With respect to the first 2 objectives of this study, to compare the foraging
locations and habitat characteristics of Arctic and Common Terns breeding on MSI, it
was found that, for the most part, these two bird species were foraging in very similar
areas with similar habitat. The habitat partitioning of foraging terns documented in other
studies (Safina 1990; Rock 2005) was not observed in my study. The large overlap in the
foraging ground characteristics indicate that, at the scale and resolution of this study,
Arctic and Common Terns on MSI do not partition their foraging habitat, and therefore
these two species are not reducing competition by segregating spatially on foraging
55
grounds. One explanation for this behaviour may lie in research involving competition
between insects. Naeem (1988) found that many species can coexist on the same
resources if those resources were variable either spatially or temporally, and Heard and
Remer (1997) determined that competition between 2 co-existing species is easier if the
resource they both use is scarce. In light of these theories, Arctic and Common Terns
may be able to live sympatrically if their prey remain ephemeral. Further research into
multiple predator-prey dynamics, especially studies that quantify prey abundance and
compare years of high and low availability, could lead to interesting insights on interspecific tern competition.
Of the 40 comparisons made, only three showed significant deviations from the
average available habitat. From this, it appears as though terns on MSI forage at random,
sampling for prey in all conditions of the variables examined. While Shealer (2002) noted
that because prey are not distributed randomly, seabird foraging behaviour is not random,
this may be the case only when there is enough prey in predictable areas to sustain them.
In 2005 both Arctic and Common Terns nesting on MSI experienced near or total
breeding failure, apparently due, at least in part, to low food availability. When prey
abundance drops below a threshold, seabirds may turn to foraging at random to find
enough food for themselves and their chicks.
Earlier studies have concluded that seabirds respond to prey at scales of 2 to 10
km (Mehlum et al. 1999; Burger et al. 2004), yet weak associations between seabirds and
prey at scales of less than 5 km have also been noted (Shealer 2002). Recent research has
shown that seabird distribution and physical oceanic processes do not show associations
at fine scales (10’s of km), but are related at scales of hundreds to thousands of
56
kilometres (Pinaud and Weimerskirch 2005). This indicates that the scale of my study
(approximately 5000 km2) may be either too coarse to detect associations with prey or too
fine to detect differences in the physical parameters of tern foraging areas. Future studies
of seabird foraging ecology should focus on measuring the scale at which nesting birds
respond physical and biological processes in the Gulf of Maine.
Finally, although I chose to examine physical and biological factors thought to be
important to seabirds in general, many other elements of ocean environments have been
linked to seabird foraging areas, such as foraging flocks, windspeed, tides, and currents
(Dunn 1973; Irons 1998; Ainley et al. 1998; Spear and Ainley 1999; Shealer 2002; Yen et
al. 2004). While terns nesting on MSI in 2005 were not found to be foraging according to
the factors in my study, they could be responding to other variables that were not
examined.
My second objective was to examine whether terns nesting on MSI showed
differences in habitat characteristics between breeding cycle stages. Weimerskirch et al.
(1993) found that Wandering Albatross (Diomedea exulans) foraging distances and
strategies (long distance foraging vs. near-colony foraging) changed from the incubation
stage to the chick rearing stage. Paquet (2001) documented a large increase in the time
Arctic Terns spent foraging from incubation to chick rearing, suggesting that, like
Wandering Albatrosses, there could be differences in the areas where they were foraging
during each stage of the breeding cycle. The results from this study do not agree with
either of these earlier observations; in 2005 Arctic and Common Terns foraged in similar
areas, despite the status of their nest. Thus, it appears that terns spend more time on the
57
same foraging grounds as the breeding season progresses, instead of using different areas
during each stage.
This study indicates that in 2005 Arctic and Common Terns nesting on MSI
showed few foraging habitat preferences over different stages of the breeding cycle, with
respect to the variables and scale examined,. Further research should focus on examining
the relationships between feeding areas and prey abundance and the scale at which terns
respond to oceanic variables and prey.
58
Table 1. Mean (SD) tern body measurements of adults trapped on Machias Seal Island
from 2003 to 2005.
Measure
Arctic Tern
Common Tern
Body mass (g)
106 (11.9)
n = 216
121 (7.8)
n = 110
Wing length (mm)
267 (20.9)
n = 216
265 (8.8)
n = 110
Head + bill length
(mm)
71.1 (5.9)
n = 215
76.9 (7.2)
n = 110
59
Table 2. Main prey items in tern chick diet on MSI from 1995 to 2005 (Bond et al. 2006).
N indicates total number of prey identified. Prey items other than the 5 commonest prey
are not included. Proportions of chick diet each prey item comprises are reported as
percentage by number of all identified prey.
Prey
Common Tern
N
Herring
Hake
Euphausiid
Butterfish
Sandlance
Arctic Tern
N
Herring
Hake
Euphausiid
Butterfish
Sandlance
1995
1996
1997
1998
1999
2000
2001
2002
2003
2004
2005
146
84.9
8.2
0
0
0
212
77.8
0.5
11.8
7.5
0
120
89.2
0.8
0
0
6.7
110
53.6
36.4
0
3.6
1.8
368
50
13.9
2.4
0.8
11.4
179
82.7
8.9
1.1
0.6
0.6
722
28.3
14.8
2.4
0.6
44
444
19.1
5.4
74.1
0
0.7
281
39.6
7.9
44.6
5.0
0
112
8.0
8.0
29.5
22.3
1.8
12
8.3
8.3
41.7
33.3
0
97
78.4
16.5
0
2.1
0
175
64.0
4.0
22.3
6.9
0
209
68.9
10.0
10.0
1.0
10.0
169
58.6
27.2
4.7
0
6.5
208
44.7
23.1
0.5
4.8
23.1
241
47.7
12
35.7
1.7
1.7
588
7.1
19.2
0.5
1.0
71.1
891
7.0
1.9
90.6
0.1
0
414
12.8
15.7
65.9
2.1
1.1
470
2.3
3.0
60.4
6.0
0
62
0
24.2
46.8
6.0
3.2
60
Table 3. Number of prey types fed to chicks by Arctic and Common Terns on MSI from
1995 to 2005. Numbers do not include unidentified prey items.
Year
Arctic Tern
Common Tern
1995
4
10
1996
8
9
1997
5
6
1998
6
6
1999
8
9
2000
8
8
2001
8
11
2002
7
6
2003
10
8
2004
11
9
2005
8
5
Mean (SD)
7.6 (2.0)
7.9 (1.9)
61
Table 4. Data layers obtained from external sources.
Data
Type of data set
Data source
Bathymetry
Point coverages
Bedford Institute of Oceanography
Coastline
Point coverages
[Chlorophyll a]
Satellite images
Sea surface
temperature
Satellite images
National Atmospheric and Oceanic
Administration (NOAA)
Satellite Oceanography Data
Center, University of Maine;
NOAA
Satellite Oceanography Data
Center University of Maine,
NOAA
62
Table 5. Mean date (SD) of nest initiation for birds equipped with transmitters (tracked)
and without (untracked). Significantly different results within species (ANOVA, p <
0.05) are denoted by a and b.
Arctic Tern mean
nest initiation date
Year
2004
2005
Common Tern mean
nest initiation date
Untracked
Tracked
Untracked
Tracked
11 June
(6)
n=118
10 Junea
(1)
n=90
9 June
(8)
n=12
9 Juneb
(2)
n=17
11 June
(6)
n=41
12 June
(2)
n=22
10 June
(6)
n=14
11 June
(2)
n=12
63
Table 6. Mean clutch size (SD) of birds equipped with transmitters (tracked) and without
(untracked). Mean clutch size is measured as the number eggs laid per nest. No
significantly different results within species (ANOVA, p < 0.05) were detected.
Arctic Tern mean
clutch size
Year
2004
2005
Common Tern mean
clutch size
Untracked
Tracked
Untracked
Tracked
1.43
(0.50)
n=118
1.46
(0.50)
n=90
1.67
(0.49)
n=12
1.68
(0.49)
n=19
1.63
(0.66)
n=41
1.77
(0.53)
n=22
1.86
(0.66)
n=14
1.82
(0.39)
n=17
64
Table 7. Mean hatching success (SD) of birds equipped with transmitters (tracked) and
without (untracked). Mean hatching success is measured by the number of eggs that hatch
per nest. Significantly different results within species (ANOVA, p < 0.05) are denoted by
a
and b.
Arctic Tern mean
hatch success
Year
2004
2005
Untracked
0.82a
(0.80)
n=118
0.73
(0.79)
n=90
Tracked
1.33b
(0.49)
n=12
0.68
(0.75)
n=19
Common Tern mean
hatch success
Untracked
0.98
(0.72)
n=41
0.82
(0.85)
n=22
Tracked
1.07
(0.73)
n=14
1.24
(0.90)
n=17
65
Table 8. Productivity (SD) of birds equipped with transmitters (tracked) and without
(untracked). Productivity is measured as the number of chicks fledged per nest.
Significantly different results within species (ANOVA, p < 0.05) are denoted by a and b.
Arctic Tern mean
productivity
Year
Untracked
a
2004
2005
0.11
(0.32)
n=99
0.05
(0.22)
n=83
Tracked
b
0.38
(0.52)
n=8
0.13
(0.35)
n=15
Common Tern mean
productivity
Untracked
Tracked
0.03
(0.18)
n=31
0.00
(0.00)
n=22
0.08
(0.28)
n=13
0.00
(0.00)
n=16
66
Table 9. Mean (SD) body condition of birds with transmitters vs. no transmitters. No
differences were observed between tracked and untracked birds for either ARTE
(ANOVA, p= 0.177) or COTE (ANOVA, p=0.360).
Arctic Tern
Untracked
Average Body
Condition
(weight/wing3)
-6
5.54x10
(5.57x10-7)
n = 46
Common Tern
Tracked
-6
5.33x10
(5.36x10-7)
n = 19
Untracked
-6
6.27x10
(5.53x10-7)
n = 27
Tracked
6.63x10-6
(1.90x10-7)
n = 19
67
Table 10a-e. Results from randomization analysis comparing available habitat location
characteristics and tern feeding location characteristics at 3 nest stages and all stages
combined. Significant results (p < 0.05) are indicated in bold.
A. Distance from mainland (km)
Species
ARTE
COTE
Nest
status
Egg
Chick
Failed
All
Egg
Chick
Failed
All
N
9
7
5
21
4
6
8
18
Sample mean
(SD)
21.7 (10.8)
18.1 (8.4)
23.3 (8.4)
20.9 (9.3)
12.2 (11.5)
22.4 (9.0)
13.2 (15.0)
16.0 (12.7)
Randomization mean
(SD)
17.0 (8.2)
17.2 (9.2)
17.5 (10.8)
16.8 (5.3)
17.5 (12.0)
17.4 (9.8)
17.1 (8.5)
16.9 (5.8)
Sample mean
(SD)
15.9 (9.8)
18.4 (7.9)
18.2 (9.9)
17.3 (8.8)
22.9 (5.7)
17.4 (7.2)
20.0 (4.8)
19.8 (5.9)
Randomization mean
(SD)
18.9 (5.9)
18.8 (6.6)
18.8 (7.8)
18.8 (3.9)
19.0 (8.7)
18.9 (7.3)
18.8 (6.2)
18.7 (4.2)
Sample mean
(SD)
90.5 (25.5)
91.0 (44.7)
122.4 (63.0)
98.3 (43.0)
60.7 (40.4)
115.8 (46.7)
75.9 (57.6)
85.8 (53.0)
Randomization mean
(SD)
85.5 (27.7)
85.7 (31.4)
86.3 (36.7)
84.4 (18.0)
87.8 (42.3)
85.4 (34.1)
85.1 (29.3)
84.4 (19.4)
p
0.2365
0.7493
0.2381
0.1179
0.4982
0.2625
0.4393
0.8421
B. Distance from MSI (km)
Species
ARTE
COTE
Nest
status
Egg
Chick
Failed
All
Egg
Chick
Failed
All
C. Depth (m)
Species
Nest
status
ARTE
Egg
Chick
Failed
All
COTE
Egg
Chick
Failed
All
N
9
7
5
21
4
6
8
18
N
9
7
5
21
4
6
8
18
p
0.3524
0.9546
0.9297
0.4875
0.3634
0.7289
0.6769
0.5844
p
0.6486
0.6532
0.0387
0.1184
0.2698
0.0666
0.5859
0.8510
68
D. Chlorophyll concentration (mg/m3)
Species
Sample mean
(SD)
1.6 (0.5)
1.4 (0.2)
1.3 (0.3)
1.5 (0.4)
2.8 (0.9)
1.8 (0.4)
2.0 (0.7)
2.1 (0.7)
Randomization mean
(SD)
1.7 (0.4)
1.7 (0.5)
1.7 (0.6)
1.7 (0.3)
1.7 (0.7)
1.7 (0.5)
1.7 (0.5)
1.7 (0.5)
E. Sea surface temperature (0C)
Species
Nest
N
Sample mean
status
(SD)
ARTE
Egg
9
13.4 (6.8)
Chick
7
10.2 (1.6)
Failed
5
13.1 (4.0)
All
21
12.2 (4.9)
COTE
Egg
4
13.4 (6.8)
Chick
6
11.2 (1.9)
Failed
8
11.0 (8.0)
All
18
11.6 (6.1)
Randomization mean
(SD)
12.3 (2.8)
12.3 (3.2)
12.5 (3.8)
12.1 (1.8)
12.6 (4.2)
12.5 (3.5)
13.2 (3.8)
12.1 (2.0)
ARTE
COTE
Nest
status
Egg
Chick
Failed
All
Egg
Chick
Failed
All
N
9
6
4
19
3
6
5
14
p
0.8950
0.2853
0.2424
0.2013
0.0019
0.4268
0.1934
0.0030
p
0.3202
0.2696
0.5643
0.8045
0.5058
0.6411
0.5222
0.6163
69
100 km
Figure 1. Map of the Gulf of Maine showing the location of Machias Seal Island.
70
a.
50
COTE 2004
COTE 2005
% of Identified Prey Items
40
30
20
10
0
Herring
Hake
Euphausiid Butterfish Sandlance
Other
Prey Type
b.
70
ARTE 2004
ARTE 2005
% of Identified Prey Items
60
50
40
30
20
10
0
Herring
Hake
Euphausiid Butterfish Sandlance
Other
Prey Type
Figure 2. Histograms showing prey types fed to Common Tern chicks (a) and Arctic Tern
chicks (b) in 2004 and 2005. The 'other' category includes: pollock, sticklebacks,
lumpfish, insects and marine invertebrates.
71
a.
b.
Figure 3. Photos showing transmitter attachment to tail retrices in 2004 (a), and to the leg
band in 2005 (b).
72
N
Log s
t
Goofapuff
Vi ew 5
Lighthouse
Foundation
plots
Gully
Oceanspray
0
50
100
Mete rs
Figure 4. Map indicating locations of the lighthouse and Goofapuff, Foundation, Gully,
and Oceanspray plots.
73
Figure 5. Map outlining the approximately 40 km radius centred around MSI searched for
tern feeding areas in 2005. The perimeter of the north tip of Grand Manan was also
searched on one occasion.
74
Figure 6. Map showing locations recorded along the flight path to represent available tern
foraging habitat.
75
Figure 7. Map indicating mainland and coastal islands. Coastal islands were excluded
from the analysis of tern foraging distance from the mainland.
76
Figure 8. Map showing locations where Arctic Terns were found foraging during each
stage of the breeding cycle in 2005.
77
Figure 9. Map showing locations where Common Terns were found foraging during each
stage of the breeding cycle in 2005.
78
50
Distance from mainland (km)
40
30
20
10
0
ARTE
COTE
Available
Species
Figure 10. Box plots of the distance from the mainland (km) of Arctic Tern (ARTE, n =
21) and Common Tern (COTE, n = 18) feeding areas, and available habitat points (n =
2995). Dotted lines indicate mean, solid lines indicate median, boxes 25-75% range and
whiskers 10-90% range.
79
50
Distance from MSI (km)
40
30
20
10
0
ARTE
COTE
Available
Species
Figure 11. Box plots of the distance from MSI (km) of Arctic Tern (ARTE, n = 21) and
Common Tern (COTE, n = 18) feeding areas, and available habitat points (n = 2995).
Dotted lines indicate mean, solid lines indicate median, boxes 25-75% range and
whiskers 10-90% range.
80
0
Depth (m)
-50
-100
-150
-200
-250
ARTE
COTE
Available
Species
Figure 12. Box plots of the depth (m) of Arctic Tern (ARTE, n = 21) and Common Tern
(COTE, n = 18) feeding areas, and available habitat points (n = 2995). Dotted lines
indicate mean, solid lines indicate median, boxes 25-75% range and whiskers 10-90%
range.
81
Figure 13. Mean depth (m) of Arctic Tern foraging areas when they no longer had a nest
(solid line) compared to the 95% confidence intervals (dashed line) of 10 000 iterations
of available point depths (p = 0.0387).
82
6
5
Chlorophyll (mg/m3)
4
3
2
1
0
-1
-2
ARTE
COTE
Available
Species
Figure 14. Box plots of the chlorophyll concentration (mg/m 3) of Arctic Tern (ARTE, n =
21) and Common Tern (COTE, n = 18) feeding areas, and available habitat points (n =
2995). Dotted lines indicate mean, solid lines indicate median, boxes 25-75% range and
whiskers 10-90% range.
83
Figure 15. Mean chlorophyll concentration (mg/m 3) of all Common Tern foraging areas
(solid line) compared to the 95% confidence intervals (dashed line) of 10 000 iterations
of available point chlorophyll concentration (p = 0.0030).
84
Figure 16. Mean chlorophyll concentration (mg/m 3) of Common Tern foraging areas
when they were incubating (solid line) compared to the 95% confidence intervals (dashed
line) of 10 000 iterations of available point chlorophyll concentration (p = 0.0019).
85
35
Sea surface Temperature ( oC)
30
25
20
15
10
5
0
ARTE
COTE
Available
Species
Figure 17. Box plots of the sea surface temperature (0C) of Arctic Tern (ARTE, n = 21)
and Common Tern (COTE, n = 18) feeding areas, and available habitat points (n = 2995).
Dotted lines indicate mean, solid lines indicate median, boxes 25-75% range and
whiskers 10-90% range.
86
References
Abraham, D.M. and Ankney, C.D. 1984. Partitioning of foraging habitat by breeding
Sabine's Gulls and Arctic Terns. The Wilson Bulletin 96: 161-346.
Adams, J., Takekawa, J.Y., and Carter, H.R. 2004. Stable foraging areas and variable
chick diet in Cassin's auklets (Ptychoramphus aleuticus) off southern California.
Canadian Journal of Zoology 82: 1578-1595.
Ainley, D.G., Jacobs, S.S., Ribic, C.A., and Gaffney, I. 1998. Seabird distribution and
oceanic features of the Amundsen and southern Bellingshausen seas. Antarctic
Science 10: 111-123.
Ainley, D.G., Ford, G., Brown, E.D., Suryan, R.M., and Irons, D.B. 2003. Prey resources,
competition, and geographic structure of Kittiwake colonies in Prince William
Sound. Ecology 84(3): 709-723.
Amey, K.D.D. 1998. Seabirds as indicators of changes in availability and commercial
weir landings of herring. Unpublished M.Sc. thesis. University of New
Brunswick, New Brunswick.
Apollonio, S. 1979. The Gulf of Maine. Courier of Maine Books, Rockland, Maine. 59
pp.
Arnold, J.M., Hatch, J.J., and Nisbet, I.C.T. 2004. Seasonal declines in reproductive
success of the common tern Sterna hirundo: timing or parental quality? Journal of
Avian Biology 35: 33-45.
Ashmole, N.P. 1968. Body size, prey size, and ecological segregation in five sympatric
tropical terns (Aves: Laridae). Systematic Zoology 17: 292-304.
Atlantic Reference Centre. 2006. Fishes of Atlantic Canada: a Photographic
Compendium. Available from
http://collections.ic.gc.ca/compendium/enchelyopuscimbrius.html. Last accessed :
2 July 2006.
Becker, P.H., Dietrich, F., and Sudmann, S.R. 1993. Temporal and spatial pattern of
Common Tern (Sterna hirundo) foraging in the Wadden Sea. Oecologia 93: 389393.
Bernard, L.J., Devlin, C.M., and Diamond, A.W. 1999. Machias Seal Island 1995-1999
Progress Report. University of New Brunswick, Fredericton, N.B.
Black, A.L., Minich, L.I., and Diamond, A.W. 2005. Machias Seal Island 1995-2004
Progress Report. University of New Brunswick, Fredericton, N.B.
87
Bond, A.L., Black, A.L., McNutt, M.-P.F., and Diamond, A.W. 2006. Machias Seal
Island Progress Report 1995-2005. University of New Brunswick. Fredericton,
N.B.
Breton, A.R. 2005. Demographic parameter estimates and their biological implications
for a seabird metapopulation. Unpublished Ph.D. Thesis University of New
Brunswick, New Brunswick.
Breton, A.R., Diamond, A.W., and Kress, S.W. 2006a. Adult survival rates of Atlantic
Puffin (Fratercula arctica) at two colonies in the Gulf of Maine. The Auk 112:
773-782.
Breton, A.R., Diamond, A.W., and Kress, S.W. 2006b. Encounter, survival, and
movement probabilities from an Atlantic Puffin (Fraturcula arctica)
metapopulation. Ecological Monographs (in press).
Brooks, D.A. and Townsend, D.W. 1989. Variability of the coastal current and nutrient
pathways in the eastern Gulf of Maine. Journal of Marine Research 47: 303-321.
Burger, A.E. 2006. Effects of the Juan de Fuca eddy and upwelling on densities and
distributions of seabirds off southwest Vancouver Island, British Columbia.
Marine Ornithology 31: 113-122.
Burger, A.E., Hitchcock, C.L., and Davoren, G.K. 2004. Spatial aggregations of seabirds
and their prey on the continental shelf off SW Vancouver Island. Marine Ecology
Progress Series 283: 279-292.
Burger, J. and Gochfeld, M. 1991. The Common Tern: Its Breeding Biology and Social
Behavior. Columbia University Press, New York. 413 pp.
Burness, G.P., Morris, R.D., and Bruce, J.P. 1994. Seasonal and annual variation in brood
attendance, prey type delivered to chicks, and foraging patterns of male Common
Terns (Sterna hirundo). Canadian Journal of Zoology 72: 1243-1251.
Chang, S., Morse, W.W., and Berrien, P. L. 1999. Essential Fish Habitat Source
Document: White Hake, Urophycis tenuis, Life History and Habitat
Characteristics. NOAA Technical Memorandum. Woods Hole, Massachusetts.
Chapdelaine, G., Brousseau, P., Anderson, R., and Marsan, R. 1985. Breeding ecology of
Common and Arctic Terns in the Mingan Archipelago, Quebec. Colonial
Waterbirds 8: 166-177.
Charette, M. 2005. Variation in the Breeding and Wintering Ground Diets of Sympatric
Nesting Common and Arctic Terns. Unpublished M.Sc. thesis. University of New
Brunswick, New Brunswick.
88
Chastel, O., Weimerskirch, H., and Jouventin, P. 1995. Influence of body condition on
reproductive decision and reproductive success in the Blue Petrel. The Auk 112:
964-972.
Craik, J.C.A. 1998. Biometric differences between Common Terns Sterna hirundo and
Arctic Terns S. paradisaea. Ringing and Migration. 19(1): 75-78.
Cramp, S., Bourne, W.R.P., Saunders, D. 1974. The Seabirds of Britain and Ireland.
Collins and Sons, London. 287 pp.
Cross, J.N., Zetlin, C.A., Berrien, P.L., Johnson, D.L., and McBride, C. 1999. Essential
Fish Habitat Source Document: Butterfish, Peprilus triacanthus, Life History and
Habitat Characteristics. NOAA Technical Memorandum. Woods Hole,
Massachusetts.
DeBach, P. 1966. The competitive displacement and coexistence principles. Annual
Review of Entomology 11: 183-212.
de Forest, L.N. and Gaston, A.J. 1996. The effect of age on timing of breeding and
reproductive success in the Thick-billed Murre. Ecology 77: 1501-1511.
Devlin, C.M., Diamond, A.W., and Saunders, G.W. 2003. Sexing Arctic Terns (Sterna
paradisae) in the field and laboratory. Waterbirds 27(3):314 - 320.
Devlin, C.M. 2006. Birds crossing borders: A population study of Arctic Terns (Sterna
paradisaea). Unpublished Ph.D. thesis. University of New Brunswick, New
Brunswick.
Diamond, A.W. 1978. Feeding strategies and population size in tropical seabirds. The
American Naturalist 112:215-223.
Diamond, A.W. 1983. Feeding overlap in some tropical and temperate seabird
communities. Studies in Avian Biology 8: 24-46.
Diamond, A.W., Devlin, C.M., Amey, K.D. 2002. Machias Seal Island: guide to logistics
and seabird monitoring protocols, unpublished report, Atlantic Cooperative
Wildlife Ecology Research Network, University of New Brunswick, Fredericton,
N.B.
Diamond, A.W. and Devlin, C.M. 2003. Seabirds as indicators of changes in marine
ecosystems: ecological monitoring on Machias Seal Island. Environmental
Monitoring and Assessment 88: 153-175.
Drury, W.H. 1973. Population changes in New England seabirds. Bird-Banding 44: 267313.
89
Dunn, E.K. 1973. Changes in fishing ability of terns associated with windspeed and sea
surface conditions. Nature 244: 520-521.
Environmental Systems Research Institute. 2002. ArcView GIS. Redlands, California.
Environmental Systems Research Institute. 2005. ArcMap 9.1. Redlands, California.
Eppley, R.W., Stewart, E., Abbott, M.R., and Heyman, U. 1985. Estimating ocean
primary production from satellite chlorophyll. Introduction to regional differences
and statistics for the Southern California Bight. Journal of Plankton Research
7(1): 57-70.
Erwin, R.M. 1977. Foraging and breeding adaptations to different food regimes in three
seabirds: the Common Tern, Sterna hirundo, Royal Tern, Sterna maxima, and
Black Skimmer, Rynchops niger. Ecology 58: 389-397.
Erwin, R.M. 1978. Coloniality in terns: the role of social feeding. Condor 80: 211-215.
Feldman, G. C. 2006. SeaWiFS Project Information. Available at:
http://oceancolor.gsfc.nasa.gov/SeaWiFS/BACKGROUND/. Last accessed 1
February 2006.
Furness, R.W. and Birkhead, T.R. 1984. Seabird colony distributions suggest competition
for food supplies during the breeding season. Nature 311: 655-656
Fox, M.F., Kester, D.R., and Yoder, J.A. 2005. Spatial and temporal distributions of
surface temperature and chlorophyll in the Gulf of Maine during 1998 using
SeaWiFS and AVHRR imagery. Marine Chemistry 97: 104-123.
Garthe, S., Benvenuti, S., and Montevecchi, W.A. 2003. Temporal patterns of foraging
activities of Northern Gannets, Morus bassanus, in the northwest Atlantic Ocean.
Canadian Journal of Zoology 81: 453-461.
Gause, G.F. 1934. The Struggle for Existence. The Williams and Wilkins Company,
Baltimore. 163 pp.
Goldstein, D.L. 1988. Time and energy budgets of free-living birds. American Zoologist
28: 829-844.
Gonzalez, L. and Manly, B.F.J. 1998. Analysis of variance by randomization with small
data sets. Environmetrics 9: 53-65.
Granadeiro, J.P., Monteiro, L.R., Silva, M.C., and Furness, R.W. 2002. Diet of Common
Terns in the Azores, northeast Atlantic. Waterbirds 25: 149-155.
Grecian, V.D. 2005. The Effect of Physical and Biological Parameters on the Breeding
Success of Razorbills (Alca torda L.1758) on Machias Seal Island, NB, in 2000
90
and 2001. Unpublished M.Sc. thesis. University of New Brunswick, New
Brunswick.
Grinnell, J. 1904. The origin and distribution of the Chestnut-Backed Chickadee. The
Auk 21: 364-382.
Hall, C.S., Kress, S.W., and Griffin, C.R. 2000. Composition, Spatial and Temporal
Variation of Common and Arctic Tern Chick Diets in the Gulf of Maine.
Waterbirds 23: 430-439.
Hardin, G. 1960. The competitive exclusion principle. Science 131: 1292-1297.
Hatch, J.J. 2002. Arctic Tern (Sterna paradisaea). In The Birds of North America, No.
707 (A. Poole and F. Gill, eds.). The Birds of North America, Inc., Philadelphia,
PA.
Hayward, T.L. and Venrick, E.L. 1982. Relation between surface chlorophyll, integrated
chlorophyll and integrated primary production. Marine Biology 69(3): 247-252.
Heard, S.B. and Remer, L.C. 1997. Clutch-size behaviour and coexistence in ephemeralpatch competition models. The American Naturalist 150(6): 744-770.
Hopkins C.D. and Wiley, R.H. 1972. Food parasitism and competition in two terns. The
Auk 89: 583-594.
Hunt, G.L. Jr. 1972. Influence of food distribution and human disturbance on the
reproductive success of Herring Gulls. Ecology 53:1051-1061.
Irons, D.B. 1998. Foraging area fidelity of individual seabirds in relation to tidal cycles
and flock feeding. Ecology 79(2): 647-655.
Kenward, R. 1987. Wildlife Radio Tagging. Academic Press, London. 222 pp.
Lack, D. 1971. Ecological Isolation in Birds. Blackwell Scientific Publications, Oxford.
404 pp.
Lemmetyinen, R. 1973. Feeding ecology of Sterna paradisaea Pontopp. and Sterna
hirundo L. in the archipelago of southwestern Finland. Annals Zool. Fennici 10:
507-525.
Le Pape, O., Chauvet, F., Mahevas, S., Lazure, P., Guerault, D., and Desaunay, Y. 2003.
Quantitative description of habitat suitability for the juvenile common sole (Solea
solea, L.) in the Bay of Biscay (France) and the contribution of different habitats
to the adult population. Journal of Sea Research 50: 139-149.
Litzow, M.A., Piatt, J.F., Prichard, A.K., and Roby, D.D. 2002. Response of pigeon
guillemots to variable abundance of high-lipid and low-lipid prey. Oecologia 132:
286- 295.
91
Manly, B.F.J. 1991. Randomization and Monte Carlo Methods in Biology. Chapman and
Hall, New York. 281 pp.
Maravelias, C.D. 1999. Habitat selection and clustering of a pelagic fish: effects of
topography and bathymetry on species dynamics. Canadian Journal of Fisheries
and Aquatic Sciences 56: 437-450.
Massey, B.W., Keane, K., and Boardman, C. 1988. Adverse effects of radio transmitters
on the behavior of nesting Least Terns. The Condor 90: 945-947.
Mehlum, F., Hunt, G.L.Jr., Klusek, Z., and Decker, M.B. 1999. Scale-dependent
correlations between the abundance of Brunnich's Guillemots and their prey. The
Journal of Animal Ecology 68: 60-72.
Ministry of Environment, Lands and Parks Resources Inventory Branch. 1998. Wildlife
Radio-telemetry. Standards for Components of British Columbia's Biodiversity
No. 5. British Columbia.
Monaghan, P., Walton, P., Wanless, S., Uttley, J.D., and Burns, M.D. 1994. Effects of
prey abundance on the foraging behaviour, diving efficiency and time allocation
of breeding Guillemots Uria aalge. Ibis 136: 214-222.
Monteiro, L.R., Ramos, J.A., Furness, R.W., and Del Nevo, A.J. 1996. Movements,
morphology, breeding, molt, diet and feeding of seabirds in the Azores. Colonial
Waterbirds 19: 82-97.
Mori, Y. and Boyd, I.L. 2004. Segregation of foraging between two sympatric penguin
species: does rate maximisation make the difference? Marine Ecology Progress
Series 275: 241-249.
Morris, R.D. and Burness, G.P. 1992. A new procedure for transmitter attachment:
effects on brood attendance and chick feeding rates by male common terns. The
Condor 94: 239-243.
Morse, W.W., Johnson, D.L., Berrien, P.L., and Wilk, S.J. 1999. Essential Fish Habitat
Source Document: Silver Hake, Merluccius bilinearis, Life History and Habitat
Characteristics. NOAA Technical Memorandum. Woods Hole, Massachusetts.
Naeem, S. 1988. Resource heterogeneity fosters coexistence of a mite and midge in
pitcher plants. Ecological Monographs 58(3): 215-227.
Newton, I. Food-supply. Chapter 7 in Population limitation in birds. Academic Press,
New York. pp. 149-189.
92
Nicol, S. 1984. Population structure of daytime swarms of the euphausiid
Meganyctiphanes norvegica in the Bay of Fundy. Marine Ecology Progress Series
18: 241-251.
Ninness, C., Rumph, R., Vasquez III, E., Bradfield, A., and Ninness, S. 2002a.
Multivariate randomization tests for small-n behavioral research: a web-based
approach. Behavior and Social Issues 12: 64-74.
Ninness, C., Newton, R., Saxon, J., Rumph, R., Bradfield, A., Harrison, C., Vasquez III,
E., and Ninness, S. 2002b. Small group statistics: a Monte Carlo comparison of
parametric and randomization tests. Behavior and Social Issues 12: 53-63.
Nisbet, I.C.T. 2002. Common Tern (Sterna hirundo). In Birds of North America, No. 618
(A. Poole and F. Gill, eds.). The Birds of North America, Inc., Philadelphia, PA.
Nisbet, I.C.T., Montoya, J.P., Burger, J., and Hatch, J.J. 2002. Use of stable isotopes to
investigate individual differences in diets and mercury exposures among common
terns Sterna hirundo in breeding and wintering grounds. Marine Ecology Progress
Series 242: 267-274.
Nisbet, I.C.T., Bridge, E.S., Szczys, P., Heidinger, B.J. Sexual dimorphism, assortative
mating and female-female pairs in Common Terns. Unpublished manuscript.
NOAA/National Geophysical Data Center. 2006. GEODAS Grid Translator. Boulder,
Colorado. Available at:
http://www.ngdc.noaa.gov/mgg/gdas/gd_designagrid.html. Last accessed 14
March, 2006.
Onsrud, M.S.R., Kaartvedt, S., and Klevjer, T.A. 2004. Vertical distribution and feeding
patterns in fish foraging on the krill Meganyctiphanes norvegica. ICES Journal of
Marine Science 61: 1278-1290.
Paquet, J. 2001. Time-Budget Flexibility of Breeding Arctic Terns (Sterna paradisaea):
An Investigation of the Behavioral Buffer. Unpublished M.Sc. thesis. University
of New Brunswick, New Brunswick.
Pearson, T.H. 1968. The feeding biology of sea-bird species breeding on the Farne
Islands, Northumberland. Journal of Animal Ecology 37: 521-552.
Peres-Neto, P.R. and Olden, J.D. 2001. Assessing the robustness of randomization tests:
examples from behavioral studies. Animal Behavior 61: 79-86.
Perez, V., Fernandez, E., Maranon, E., Serret, P., and Garcia-Soto, C. 2005. Seasonal and
interannual variability of chlorophyll a and primary production in the Equatorial
Atlantic: in situ and remote sensing observations. Journal of Plankton Research
27(2): 189-197.
Perrins, C.M. 1970. The timing of birds’ breeding seasons. Ibis 112:242-255.
93
Pinaud, D. and Weimerskirch, H. 2005. Scale-dependent habitat use in a long-ranging
central place predator. Journal of Animal Ecology 74: 852-863.
Platt, T. and Sathyendranath, S. 1988. Ocean primary production: estimation by remote
sensing at local and regional scales. Science 241(4873): 1613-1620.
Ramos, J.A. 2000. Characteristics of foraging habitats and chick food provisioning by
tropical Roseate Terns. The Condor 102: 795-803.
Reid, R. N., Cargnelli, L. M., Griesbach, S. J., Packer, D. B., Johnson, D. L., Zetlin, C.
A., Morse, W. W., and Berrien, P. L. 1999. Essential Fish Habitat Source
Document: Atlantic Herring, Clupea harengus, Life History and Habitat
Characteristics. NOAA Technical Memorandum. Woods Hole, Massachusetts.
Ressler, P.H., Brodeur, R.D., Peterson, W.T., Pierce, S.D., Vance, P.M., Rostad, A., and
Barth, J.A. 2005. The spatial distribution of euphausiid aggregations in the
Northern california Current during August 2000. Deep-Sea Research II 52: 89108.
Rock, J.C. 2005. Foraging habitat and chick diets of Roseate Terns and co-nesting
Common and Arctic Terns. Unpublished M.Sc. thesis. Dalhousie University,
Nova Scotia.
Saborowski, R., Brohl, S., Tarling, G.A., and Buchholz, F. 2002. Metabolic properties of
Northern Krill, Meganyctiphanes norvegica, from different climatic zones. l.
Respiration and excretion. Marine Biology 140: 547-556.
Safina, C. 1990. Foraging habitat partitioning in Roseate and Common Terns. The Auk
107: 351-358.
Safina, C. and Burger, J. 1985. Common Tern foraging: seasonal trends in prey fish
densities and competition with Bluefish. Ecology 66: 1457-1463.
Safina, C. and Burger, J. 1988. Prey Dynamics and the Breeding Phenology of Common
Terns (Sterna hirundo). The Auk 105: 720-726.
Safina, C., Burger, J., Gochfeld, M., and Wagner, R.H. 1988. Evidence for prey
limitation of Common and Roseate Tern reproduction. The Condor 90: 852-859.
Shealer, D.A. 2002. Foraging Behavior and Food of Seabirds. Chapter 6 In Biology of
Marine Birds. (E.A. Schreiber and J. Burger, eds.). CRC Marine Biology Inc. pp.
137-177.
Spear, L.B. and Ainley, D.G. 1999. Migration routes of Sooty Shearwaters in the Pacific
Ocean. The Condor 101: 205-218.
Stephens, D.W. and Krebs, J.R. 1986. Foraging Theory. Princeton University Press,
Princeton, New Jersey. 247 pp.
94
Suryan, R.M., Irons, D.B., and Benson, J. 2000. Prey switching and variable foraging
strategies of Black-legged Kittiwakes and the effect on reproductive success. The
Condor 102: 374-384.
Suryan, R.M., Irons, D.B., Kaufman, M., Benson, J., Jodice, P.G.R., and Roby, D.D.
2002. Short-term fluctuations in forage fish availability and the effect on prey
selection and brood-rearing in the black-legged kittiwake Rissa tridactyla. Marine
Ecology Progress Series 236: 273-287.
Thomas, A., Byrne, D., and Weatherbee, R. 2002 Coastal sea surface temperature
variability from Landsat infrared data. Remote Sensing of Environment 81: 262272.
Trivelpiece, W.Z., Bengtson, J.L., Trivelpiece, S.G., and Volkman, N.J. 1986. Foraging
behavior of Gentoo and Chinstrap Penguins as determined by new radiotelemetry
techniques. The Auk 103: 777-781.
US Fish and Wildlife Service. 2004. Gulf of Main Seabird Working Group (GoMSWG)
2004 report. Available at:
http://www.fws.gov/northeast/mainecoastal/progsumm04.html. Last accessed 2
February 2006.
Wahle, R.A. 2000. Fisheries in a sea of change: ecology and oceanography of New
England's fishing grounds. Northeastern Naturalist 7: 317-328.
Weimerskirch, H., Salamolard, M., Sarrazin, F., and Jouventin, P. 1993. Foraging
strategy of Wandering Albatrossess through the breeding season: a study using
satellite telemetry. The Auk 110: 325-342.
Weimerskirch, H., Wilson, R.P., Guinet, C., and Koudil, M. 1995. Use of seabirds to
monitor sea-surface temperatures and to validate satellite remote-sensing
measurements in the Southern Ocean. Marine Ecology Progress Series 126: 299303.
Weimerskirch, H., Inchausti, P., Guinet, C., and Barbraud, C. 2003a. Trends in bird and
seal populations as indicators of a system shift in the Southern Ocean. Antarctic
Science 15: 249-256.
Weimerskirch, H., Ancel, A., Caloin, M., Zahariev, A., Spagairi, J., Kersten, M., and
Chastel, O. 2003b. Foraging efficiency and adjustment of energy expenditure in a
pelagic seabird provisioning its chick. Journal of Animal Ecology 72: 500-508.
Weller, M.W. 1957. An automatic nest-trap for waterfowl. Journal of Wildlife
Management 21: 456-458.
Wilson, R.P. and Wilson, M.T. 1988. Foraging behavior in four sympatric cormorants.
Journal of Animal Ecology 57: 943-955.
95
Yen, P.P.W., Huettmann, F., and Cooke, F. 2004. A large-scale model for the at-sea
distribution and abundance of Marbled Murrelets (Brachyramphus marmoratus)
during the breeding season in coastal Bristish Columbia, Canada. Ecological
Modelling 171: 395-413.
Yoder, J.A., O'Reilly, J.E., Barnard, A.H., Moore, T.S., and Ruhsam, C.M. 2001.
Variability in coastal zone color scanner (CZCS) chlorophyll imagery of ocean
margin waters off the US east coast. Continental Shelf Research 21: 1191-1218.
Yoder, J.A. Schollaert, S.R., and O’Reilly, J.E. 2002. Climatological phytoplankton
chlorophyll and sea surface temperature patterns in continental shelf and slope
waters off the northeast U.S. coast. Limnology and Oceanography 47: 672-682.
96
Vita
Candidate’s full name: Amie Lynn Black
Universities attended (with dates and degrees obtained):
University of New Brunswick 1998-2002 Bachelor of Science in Biology
Publications:
Charette, M.R., Black, A.L., Devlin, C.M., Diamond, A.W., Minich, L.I. 2004. Machias
Seal Island 1995 - 2003 Progress Report. University of New Brunswick, Fredericton,
N.B.
Black, A.L., Minich, L.I., and Diamond, A.W. 2005. Machias Seal Island 1995-2004
Progress Report. Fredericton, N.B., University of New Brunswick, Fredericton, N.B.
Bond, A.L., Black, A.L., McNutt, M.-P.F., and Diamond, A.W. 2006. Machias Seal
Island Progress Report 1995-2005. University of New Brunswick, Fredericton, N.B.
Conference Publications:
Black, A.L., and Diamond, A.W. 2004. Tracking terns on their feeding grounds: the trials
and tribulations. Gulf of Maine Seabird Working Group Fall Meeting, Hog Island, Maine,
U.S.A.
Black, A.L., and Diamond, A.W. 2004. Feeding Areas of Arctic Terns (Sterna
paradisaea) and Common Terns (Sterna hirundo) Breeding on Machias Seal Island, N.B.
6th Bay of Fundy Ecosystem Partnership Workshop, Cornwallis, N.S., Canada.
Black, A.L., and Diamond, A.W. 2004. Foraging areas of Arctic Terns (Sterna
paradisaea) and Common Terns (Sterna hirundo) breeding on Machias Seal Island, N.B.
9th Annual Atlantic Cooperative Wildlife Ecology Research Network Meeting,
Fredericton, N.B., Canada.
Black, A.L., and Diamond, A.W. 2005. Comparing Arctic and Common Tern foraging
ground characteristics during 3 stages of the breeding cycle. 10 th Annual Atlantic
Cooperative Wildlife Ecology Research Network Meeting, Kouchibouguac National
Park, N.B., Canada.
Black, A.L., and Diamond, A.W. 2005. Foraging habitat characteristics of Arctic and
Common Terns breeding on Machias Seal Island, NB, Canada. 29th Annual Waterbird
Society Meeting, Jekyll Island, Georgia, U.S.A.
97
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertising