Pagophila Eburnea) in the Canadian High Arctic

Spatial ecology, behaviour and critical winter habitat of the endangered

ivory gull (Pagophila eburnea) in the Canadian high Arctic

Nora Catherine Spencer

Thesis submitted in partial fulfillment of the requirements for

the Degree of Master of Science (Biology)

Acadia University

Fall Graduation 2014

The Thesis by Nora Catherine Spencer was defended successfully in an oral examination on 25 August 2014.

The examining committee for the thesis was:

Dr. Jun Yang, Chair

Rob Ronconi, External Reader

Trevor Avery, Internal Reader

Dr. Mark Mallory, Co-supervisor

Dr. Steve Mockford, Head

The thesis is accepted in its present form by the Division and Research and Graduate Studies as satisfying the thesis requirements for the degree Masters of Science (Biology) ……………………………………………

I, Nora Catherine Spencer, grant permission to the University Librarian at Acadia University to reproduce, loan or distribute copies of my thesis in microform, paper of electronic formats on a non-profit basis. I, however, retain the copyright in my thesis.

Author

Supervisor

Date

iii

Table of Contents

List of figure captions ...... viii List of tables captions ...... xiii Abstract ...... xvii List of Abbreviations ...... xviii Acknowledgements ...... xix

Chapter one: General introduction ...... 1 Canadian Arctic ...... 2 Sea ice and climate change ...... 2 Seabirds as environmental monitors and population trends ...... 3 The ivory gull ...... 5 Threats ...... 7 Satellite telemetry ...... 9 Objectives and predictions ...... 9 References ...... 11

Chapter two: Movement patterns and key habitats of the Canadian ivory gull through their annual cycle: the importance of sea ice ...... 16 Introduction ...... 17 Climate change and environmental monitors ...... 17 The ivory gull ...... 17 Satellite telemetry ...... 19 Objective and predictions ...... 19 Methods ...... 20 Study area ...... 20 Bird capture and tagging (2010) ...... 23 Processing of location data ...... 26 Annual cycle ...... 26 Duplicated data ...... 30 Calculation of distance and dates of travel ...... 30 Filtering unrealistic locations ...... 30 Adjusting for time periods larger than 24 h ...... 31 Analysis ...... 32 Assessing source of variation among transmitters ...... 32 Rate of travel...... 35

Distance traveled ...... 35 Sea ice ...... 35 Results ...... 36 Distribution ...... 36 Sea ice ...... 40 Seasonal spatial and temporal patterns ...... 42 Wintering ...... 42 Spring migration ...... 42 Breeding ...... 45 Non-breeding ...... 46 Fall migration...... 46 Rate of travel...... 50 Distance traveled ...... 54 Discussion ...... 58 Migratory movements and sea ice ...... 58 Breeding and non-breeding seabirds ...... 60 Rate of travel...... 62 Comparisons among other seabirds ...... 63 References ...... 66

Chapter three: Critical Winter Habitat of the Endangered Ivory Gull (Pagophila eburnea) in the Canadian Arctic ...... 72 Introduction ...... 73 Arctic sea ice...... 73 The ivory gull ...... 74 Protected areas ...... 75 Objectives and predictions ...... 76 Methods ...... 77 Study area ...... 77 Bird capture and tagging (2010) ...... 79 Processing of location data ...... 80 Classifying the winter period ...... 81 Duplicated data ...... 82 Filtering unrealistic locations ...... 82 Analysis ...... 83 Kernel density estimates ...... 83 Historical dataset ...... 85 Norwegian satellite data ...... 86 Results ...... 86

Variation of winter habitat across years ...... 86 Variation of winter habitat across years and months ...... 89 Individual variation of wintering habitat ...... 92 Winter habitat of all ivory gulls from Canadian satellite data ...... 92 Winter habitat of ivory gulls from PIROP dataset...... 92 Winter habitat of ivory gulls from Norwegian satellite data ...... 93 Overlap between ivory gulls from Canadian satellite transmitter data, PIROP data and Norwegian data ...... 97 Discussion ...... 100 Surviving winter in Davis Strait ...... 100 Area of international significance ...... 102 Future threats ...... 103 Critical habitat and marine protected areas ...... 104 References ...... 108

Chapter four: Future recovery and management of the ivory gull ...... 115 References ...... 123

Appendix A: Annual distances traveled by ivory gulls broken down by season ...... 125 Appendix B: Individual variation of habitat use among wintering ivory gulls ...... 128

List of figure captions

Chapter two

Figure 2.1: Bird capture and tagging of 12 ivory gulls took place on Seymour Island,

NU; a small low-lying island (3 km long), supporting the largest breeding colony for ivory gulls in Canada and a Migratory Bird Sanctuary in Canada...... 22

Figure 2.2: The distribution of the number of good quality transmissions (LC 0, 1, 2 and

3) used for analyses in relation to the time of day (early morning,0000-0559; morning,

0600-1159; afternoon, 1200-1759; and night, 1800-2359) and season (wintering, spring migration, breeding and fall migration) they arrived throughout the study period, July

2010 – July 2013...... 34

Figure 2.3: Annual distribution of the Canadian ivory gull. The 50% kernels represent distribution during breeding (red), post-breeding (orange), winter (light blue) and pre- breeding seasons (dark blue). General direction of fall migration is indicated by the arrows (south via Davis Strait and Foxe Basin in orange) from 'post-breeding' to 'winter' and direction of spring migration is indicated by the arrow (north via Davis Strait in blue) from 'pre-breeding' to 'breed...... 37

Figure 2.4: General direction traveled for nine bird-years south via Davis Strait during fall migration over the study period (July 2010- July 2013)...... 38

Figure 2.5: General direction traveled for three bird-years south via Foxe Basin Strait during fall migration over the study period (July 2010- July

2013)……………………….38

vii

Figure 2.6: General direction traveled for two bird-years south via Foxe Basin, cutting north across Baffin Island and south through Davis Strait during fall migration over the study period (July 2010- July

2013)………………………………………………………39

Figure 2.7: Median ( ̃) and 25th and 75th quartiles of proportion of observations of the 12 satellite-tagged ivory gulls over ice, land or water in a) winter (n= 15 bird-seasons, 403 days), b) spring migration (n=12, 62), c) breeding (n=21, 337) and d) fall migration

(n=19, 140). Whiskers represent values within 1.5 times the interquartile range. Total number of days with observations=942 over the study period, July 2010- July 2013...... 41

Figure 2.8: Six (n=12 bird-years) satellite-tagged ivory gulls during the study period

(July 2010- July 2013) migrated north in spring through Davis Strait, west through

Lancaster Sound and north to the breeding area...... 43

Figure 2.9: Breeding colonies used in 2011. One located on Seymour Island, that was abandoned (two ivory gulls) and two located on Grinnell Peninsula (one ivory gull at each colony)...... 45

Figure 2.10: Breeding colony used in 2013. One ivory gull was located on Cornwallis

Island, but transmissions cut out shortly after its arrival...... 45

Figure 2.11: Minimum Convex Polygon (MCP) representing the area used by non- breeding satellite-tagged ivory gulls (black outline; 768 199 km2; n= 4 bird-years) and breeding satellite-tagged ivory gulls (white outline; 34 343 km2; n= 2 birds) throughout the study period (July 2010- July 2013)...... 47

viii

Figure 2.12: Median ( ̃) and 25th and 75th quartiles of weekly rate of travel per month for all satellite-tagged ivory gulls throughout the study period, July 2010- July 2013.

Whiskers represent values within 1.5 times the interquartile range. Mean values for rate of travel are depicted by the black circles. The number above each month represents n bird-days that contributed to the analysis. December was excluded as there were too few records due to the blackout period for the transmitters...... 51

Figure 2.13: Median ( ̃) and 25th and 75th quartiles of hourly rate of travel per month for all satellite-tagged ivory gulls throughout the study period, July 2010- July 2013.

Chapter three

Figure 3.1: Bird capture and tagging of 12 ivory gulls took place on Seymour Island,

NU; a small low-lying island (3 km long), supporting the largest breeding colony for ivory gulls in Canada and a Migratory Bird Sanctuary in Canada...... 78

Figure 3.2: UD (95%) and core (50%) areas of wintering satellite-tagged ivory gulls pooled across each year of the study period a) 2010/2011 (n=8) b) 2011/2012 (n=4) and c) 2012/2013(n=2) in Davis Strait, NU...... 88

Figure 3.3: Monthly ice lines fading sequentially from dark blue to light blue, representing January, February, March and April, respectively in relation to the 50% core areas of pooled satellite-tagged ivory gulls across each year of the study period a)

2010/2011 (n=8); b) 2011/2012 (n=4); and c) 2012/2013(n=2) in Davis Strait, NU. The white area is comprised of cells with 625 km2 area representing the maximum sea ice extent of the respective year...... 90

Figure 3.4: UD (95%) area and core (50%) area of all Canadian satellite-tagged ivory gulls (n=8) throughout all years of the study (2010-2013), in Davis Strait, NU and

Labrador Sea, NL...... 94

Figure 3.5: UD (95%) area and core (50%) area of wintering ivory gulls sightings from ship-based surveys (PIROP database) over 41 years (1969-2010) in Davis Strait, NU and

Labrador Sea, NL...... 95

Figure 3.6: UD (95%) area and core (50%) area of all Norwegian satellite-tagged ivory gulls (n=9) throughout all years of the study (2010-2013), in Davis Strait, NU, Labrador

Sea, NL and east of Greenland...... 96

Figure 3.7: All observations from Canadian satellite-tagged ivory gulls over 3 years

(2010-2013; black outlined circles); Norwegian satellite-tagged ivory gulls over 3 years

(2010-2013; black circles); and PIROP at-sea observations of ivory gulls (1969-2010; white circles)...... 98

Figure 3.8: Overlapping core (50%) area of wintering Canadian and Norwegian satellite- tagged ivory gulls through 2010-2013 and wintering ivory gulls sighted from ship-based surveys (PIROP database) over 41 years (1969-2010) in Davis Strait, NU...... 99

Figure 3.9: Approximate historical hooded seal whelping locations (small circles) and bounds (large hollow circles) as described in sergeant (1974), Orr and Parson (1982),

Bowen et al., 1987 and Stenson et al. (1996) and the overlapping area of wintering

Canadian and Norwegian satellite-tagged ivory gulls and wintering gulls sighted from ship-based surveys (PIROP database) in Davis Strait, NU...... 107

Chapter four

Figure 4.1: The proposed protected area for the ivory gull selected by taking half the length of the widest part of the overlap to be the radius of a circle. The area is 90 885 km2……………………………………………………………………………………..121

Figure 4.2: Lines of latitude and longitude showing the orientation compared with the overlapping core (50%) area of wintering Canadian and Norwegian satellite-tagged ivory gulls and wintering ivory gull sightings from ship-based surveys (PIROP database) in

Davis Strait, NU. ……………………………………………………………………….122

Appendix B

Figure B1: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44523); a) over all years of the study (2010-2013); b) in winter 2010/2011; c) in winter 2011/2012; and d) in winter 2012/2013 in Davis Strait, NU……………………129

Figure B2: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44526); a) over all years of the study (2010-2013); b) in winter 2010/2011; c) in winter 2011/2012; and d) in winter 2012/2013 in Davis Strait, NU……………………130

Figure B3: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44526); a) over all years of the study (2010-2013); b) in winter 2010/2011; c) in winter 2011/2012; and d) in winter 2012/2013 in Davis Strait, NU……………………131

Figure B4: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44525); a) over all years the transmitter sent signals (2010-2013); b) in winter

2010/2011; and c) in winter 2011/2012 in Davis Strait, NU. ………………………… .132

Figure B5: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gull

(44530) over the period the transmitter was sending signals (2010-2011)………….…133

Figure B6: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gull

(44531) over the period the transmitter was sending signals (2010-2011)………….…133

xii

List of tables captions

Chapter two

Table 2.1: Description of each satellite transmitter (PTT number) and how the device was powered (battery or solar). The dates the transmitter ran is included as well as the total number of months for which data were collected and a count of the good quality location records (LC 0, 1, 2 and 3) that were available for analysis for the study period,

Juyl 2010- Jul y2013………………………………………………………………………..25

Table 2.2: Median ( ̃) start date, range of start dates and the median ( ̃) number of days and the range for each season of the annual cycle for the 12 satellite-tagged ivory gulls throughout the study period, July 2010- July

2013…………………………………………29

Table 2.3: The mean ( ̅) characteristics and standard deviation (SD) of each satellite- tagged ivory gull, including the mean distance traveled per day, mean duration traveled per day, the total number of transmissions and the number of days that the satellite transmitter sent signals throughout the study period, July 2010 – July

2013………………33

Table 2 4: Individual dates of arrival/departure for 12 satellite-tagged ivory gulls as well as migratory routes and breeding colonies used throughout respective annual cycles throughout the study period, July 2010- July 2013. ‘No Attempt’ indicates that no breeding attempt was made during the breeding season…………………………………………...48

xiii

Table 2.4 (cont’d): Individual dates of arrival/departure for 12 satellite-tagged ivory gulls as well as migratory routes and breeding colonies used throughout respective annual cycles throughout the study period, July 2010- July 2013. ‘No Attempt’ indicates that no breeding attempt was made during the breeding season…………………………………49

Table 2.5: Mean ( ̅) and standard deviation (SD) of distance flown per day (km/d) during breeding, wintering and migratory seasons of 12 satellite-tagged ivory gulls throughout the study period July 2010- July

2013………………………………………….55

Table 2.6: Total distance flown (km) for 12 satellite-tagged ivory gulls across breeding years (01 July – 30 June of 2010/2011, 2011/2012 and 2012/2013). Brackets indicate that a PTT did not transmit for the entire year. Mean ( ̅) and standard deviation (SD) was calculated for the distance traveled by the six ivory gulls that transmitted for the entire period of 2010/2011. N refers to the number of good quality locations received for each individual………………………………………………………………………………...56

Table 2.7: Comparison of seabirds’ rate of travel (km/h)……………………………….65

Chapter three

Table 3.1: The maximum distance (km) of 50% kernels (core area) west of the sea ice edge and the maximum distance (km) of core areas east of the ice edge in January,

February, March and April over the study period, 2010-2013. Two distances provided for one month indicate that there were two centres of activity…………………………………………91

xiv

Appendix A

Table A1: Sum of distances (km) traveled by 12 satellite-tagged ivory gulls each season throughout each calendar year of the study period (July 2010- July 2013) using the median start dates (Table 2). Bolded years indicate a full year of data was collected

(breeding season to breeding season). Brackets indicate the bird transmitted signals for only part of the season and were not included in calculating means and standard deviations (SD)...126

Table A2: Sum of distances (km) traveled by 12 satellite-tagged ivory gulls each season throughout each calendar year of the study period (July 2010- July 2013) using the median start dates (Table 2). Bolded years indicate a full year of data was collected

(breeding season to breeding season). Brackets indicate the bird transmitted signals for only part of the season and were not included in calculating means and standard deviations (SD)...127

Abstract

The ivory gull (Pagophila eburnea) is an endangered seabird that spends its entire year in the Canadian Arctic environment. In the past three decades, threats from various sources have contributed to an apparent >70% decline in the breeding population. As outlined in the Canadian ivory gull recovery strategy, information on annual movements is critical to obtain information of where and when they move as well as how they use these sites during breeding and non-breeding seasons. In 2010, satellite transmitters were attached to 12 ivory gulls on Seymour Island, NU providing up to four breeding seasons of tracking data. This is the first quantitative, multi-year tracking study conducted for the ivory gull in Canada, analyzing annual movements, distribution and critical habitat. Analysis of migratory behaviour revealed great individual variation of fall migratory route selection (with the surprising observation that the Greenland coast was not used), where travel was slow. In spring, gulls migrated quickly back to the breeding area using only Davis Strait. Data from Norwegian ivory gull satellite tracking as well as a Canadian at-sea seabird observation database were used to augment data from Canadian ivory gull satellite tracking data, with the goal of comparing winter distributions of ivory gulls from different populations and through time. Kernel density estimates (KDE) suggested that Davis Strait is critical winter habitat for Canadian and international populations. On average, Canadian ivory gulls preferred ≥50% sea ice concentration and were rarely found out over open water. Sea ice plays a large role in both migratory movements as well as determining winter habitat for this species. In winter, ivory gulls likely use breeding hooded seals (Crystophora cristata) as a main food source and by association follow and scavenge from polar bears (Ursus maritimus) in Davis Strait. I recommend that this area be considered for critical habitat and consequently protected for the ivory gull (both Canadian and international birds) for the monitoring and recovery of this endangered species.

xvi

List of Abbreviations

Abbreviation Description AO Arctic Oscillation BRF Brominated flame-retardants COSEWIC Committee on the Status of Endangered Wildlife in Canada EEZ Exclusive Economic Zone GIS Geographic Information System href Reference bandwith IUCN International Union for Conservation of Nature KDE Kernel Density Estimate KW Kruskal-Wallis LC Location Class MCP Minimum Convex Polygon MPA Marine Protected Area NAO North Atlantic Oscillation NSIDC National Snow and Ice Data Center OC Organochlorines PIROP Programme Intégré des Recherches sur les Oiseaux Pélagiques PTT Platform Terminal Transmitter SARA Species at Risk Act UD Utilization Distribution

xvii

Acknowledgements

First and foremost, I would like to thank Dr. Mark Mallory for providing me with opportunities that I could never have imagined happening two years ago. I would not have completed this project without your support and guidance. And of course, thank you for all the coffees and laughs. It’s no fun, if it’s no fun, right? Thank you also to my co-supervisor, Dr. Grant Gilchrist (National Wildlife Research Centre).

I’d like to acknowledge my funding source for the project: the Molson Foundation.

This project would not have been possible (literally) without Dr. Birgit Braune (National

Wildlife Research Centre), Dr. Karel Allard (Canadian Wildlife Service- Atlantic

Region), and Dr. Mark Mallory who flew in to Seymour Island four years ago to outfit some ivory gulls with satellite transmitters.

A big thank you to those who provided me with data for the project: Dr. Karel Allard, Dr.

Hallvard Ström (Norwegian Polar Institute, Norway) and Dr. Evan Richardson

(Environment Canada- Science and Technology Branch).

I would also like to thank Jason Duffy (National Wildlife Research Centre) for providing me with a space in his lab to hone my GIS skills and Jason Roberts (Duke University) for his help with ‘Marine Geospatial Ecology Tools’ for GIS.

Thank you to my family (Barbara, Philip and Chris) for your support and encouragement throughout my time here. Finally, to my partner and best friend, Gavin; thanks for always believing in me.

xviii

Chapter one

General introduction

Canadian Arctic

The Canadian Arctic is the region located above the 10°C isotherm, an inhospitable habitat for most species except those with specific adaptations to tolerate the thermal extremes (Thomas et al. 2008). It is known to be cold, dry and windy, where solar radiation is limited in its warmest season and distribution of land and open ocean contribute to the harsh climate (Thomas et al. 2008). In the past few decades monitoring data have shown that there has been an overall decline in Arctic sea ice, shifting to a younger and thinner type of ice due to pollution (and the consequent effects on temperatures) carried north by the atmosphere and sea currents (Lindsay and Zhang 2005;

Maslanik et al. 2007). This is concerning because many Arctic species are especially sensitive to climate change due to low fecundity, adaptations to cold temperatures, and having ecological connections to the cycle of sea ice cover.

Sea ice and climate change

Timing of sea ice formation and break up each year is an important cycle that influences the physical connection between islands in the Arctic Archipelago, affects the formation and deformation of foraging areas for wildlife and has a significant influence on the timing and abundance of marine food resources (e.g. Mallory et al. 2010). The majority of the Canadian Arctic freezes during the winter months; however wind and water currents keep polynyas and shore leads (a gap of water between pack ice and the shore) of varying sizes open in March and April (Stirling 1997). Polynyas are areas of open water within the pack ice that allow seabirds, marine mammals, fish and invertebrates to forage throughout the breeding season (Thomas et al. 2008). They allow increased light penetration around and under sea ice, which supports the development of

epontic algae growth on the undersurface of the ice, and consequently these “islands of open water in a sea of ice” become biodiversity hotspots where organisms from many different trophic levels come to feed (Bradstreet 1982). Although climate models agree that there is a general warming trend, region-specific trends caused by cyclical patterns created by the North Atlantic Oscillation (NAO) and the Arctic Oscillation (AO) introduce inter-annual climate variability in the North Atlantic and Arctic (Jenouvrier et al. 2005; Irons et al. 2008). An increasing ice extent could cause some polynyas and leads to close during the winter periods, decreasing primary productivity and causing more marine mammals to become trapped in the ice (Heide-Jørgensen and Laidre 2004).

However, decreasing ice cover and thickness can also have detrimental effects on many arctic species including hooded seals (Crystophora cristata) and harp seals (Phoca groenlandicus), which rely on the sea ice surface to haul out and raise their pups

(Johnston et al. 2005). With the combination of climate change and inter annual and decadal variability, large fluctuations in climate are expected to intensify and could have long term negative impacts on the sensitive Arctic ecosystem (Salinger 2005; Thomas et al. 2008; Gilg et al. 2012).

Seabirds as environmental monitors and population trends

Seabirds make particularly good environmental monitors to determine effects of climate change in higher latitudes because of their colonial nature and the relative ease with which they can be tracked (Furness and Camphuysen, 1997). Population trends and changes in diet are just two of the many parameters that can be monitored and which provide an indicator of ecosystem health (Gaston et al. 2009a). Population trends within the past few decades have varied for the approximately four million breeding pairs of

seabirds in the eastern Canadian Arctic (Gaston et al. 2012). For example northern fulmars (Fulmarus glacialis), and glaucous gulls (Larus hyperboreus) are declining, while numbers are increasing for species such as the thick-billed murre (Uria lomvia) and black-legged kittiwake (Rissa tridactyla) (Gaston et al. 2012). These population changes may, in part, be dictated by long-term changes in food availability, which in turn is affected by climate change (Irons et al., 2008). The overall warming trend in ocean temperatures may affect the timing of prey availability and/or may cause changes in prey distribution, forcing seabirds to adjust their migratory movements and initiation of breeding and may negatively affect reproductive success (Fiedler 2003; Gaston et al.

2009b; Mallory et al. 2010).

Seabirds require open water to forage, which may only occur in the form of polynyas or shore leads throughout a majority of the year in the environmental extremes of the Canadian Arctic (Stirling 1997). The North Water Polynya is the largest and best known polynya in the Canadian Arctic (Stirling 1980). Located in Baffin Bay and spreading to Lancaster Sound, this area is known to host large concentrations of marine animals in May and June as they migrate back to their breeding areas (Stirling 1980;

McLaren 1982; Thomson 1982; Heide-Jørgensen and Laidre 2004). With the overall warming trend creating earlier ice break up, opportunities to return to breeding areas earlier to feed and reproduce may occur. Seabirds can adapt to the changing environment; however there is likely a cost to their reproductive success, foraging, growth, future population trends and distributions (Gaston et al. 2005a; Irons et al. 2008;

Gaston et al. 2009a; Mallory et al. 2009, 2010). Prey distribution in Hudson Bay has recently shifted to smaller, warmer water fish species, capelin (Mallotus villosus) and

sandlance (Ammodytes spp.; Gaston et al. 2005b; Mallory et al. 2010). This has forced the thick-billed murre to lengthen foraging periods to make up for smaller sized prey and consequently chick growth has suffered (Gaston et al. 2005a; Mallory et al. 2010).

Temperate seabirds may also adjust to changes in their prey distribution, potentially creating more competition with very specialized Arctic species of which are accustomed to few competitors (Mallory et al. 2010; Gilg et al. 2012). To compensate, Arctic seabirds such as the little auk (Alle alle) that rely on the sea ice edge to forage will be pushed farther north by both competition and retreating sea ice (Fort et al. 2013).

Unfortunately, declining sea ice extent is also beneficial for human activities like mining operations and offshore oil developments, and these activities will only accelerate the negative impacts on seabird migration, feeding and reproduction (Stirling 1997; Gaston et al. 2012). As a sentinel of changing environmental conditions in the Arctic, few seabirds would make better biomonitors than the year-round Arctic inhabitant, the ivory gull

(Pagophila eburnea), one of the most poorly known seabirds in the world (Gilchrist et al.

2008).

The ivory gull

The ivory gull is a medium-sized, colonial gull commonly associated with sea ice

(Gilchrist and Mallory 2005; Mallory et al. 2008). Its breeding range spans across the

Canadian Arctic in the North Atlantic and Arctic oceans (Robertson et al. 2007). Despite the large range, there are only approximately 49 nesting locations known on Ellesmere,

Devon, Seymour, and Cornwallis islands, and the Brodeur Peninsula of Baffin Island

(COSEWIC 2006; Robertson et al. 2007). These nesting areas are comprised of small colonies primarily in three areas: (1) on steep granite cliffs, known as nunataks, up to

1400m in elevation; (2) inland on gravel limestone plateaus; and (3) isolated, low lying islands (Robertson et al. 2007; Mallory et al. 2012). Ivory gulls exhibit breeding site fidelity, where they may produce 1-3 eggs in a clutch during the breeding season

(Mallory et al. 2008). Their colonies are often located close to polynyas.

Outside of Canada, breeding populations can be found on Svalbard, Greenland and Russian Arctic Islands in similar habitats to the Canadian population (COSEWIC

2006). A recent satellite tracking study showed that ivory gulls from European populations winter primarily in Davis Strait and Labrador Sea; however, southeast

Greenland, Bering, Chukchi, Okhotsk, and Barents seas are also known to host smaller numbers of wintering birds (Royston 2007; Gilchrist et al. 2008; Gilg et al. 2010). There are 35 known nesting colonies in north and southeast Greenland, estimated to hold at least 1800 individuals, but the southern population is thought to be declining (Gilg et al.

2009). Population trends of northern Greenland, Russia and Svalbard are difficult to assess, due to lack of information available. Nonetheless, internationally, the ivory gull has been listed as “Near Threatened” by the World Conservation Union (IUCN) since

2005 (Miljeteig et al. 2012). Nationally, the ivory gull has been legally protected in

Greenland since 1977, in Svalbard since 1978 and is registered as a Category 3 (Rare) species in the Red Data Book of the Russian Federation (Gilchrist et al. 2008).

As the Canadian high Arctic is covered in sea ice and snow for the majority of the year, ivory gulls must rely on polynyas for foraging (Bradstreet 1982). Ivory gulls are opportunistic feeders, consuming mainly juvenile Arctic cod (Boreogadus saida) and a number of invertebrate species (Mallory et al. 2008; Karnovsky et al. 2009). They are also known to feed on polar bear (Ursus maritimus) kills and garbage in human

settlements (Mallory et al. 2008). On the other hand, during the winter months, ivory gulls appear to be associated with hooded seal whelping patches on ice edges north of

Labrador and in Davis Strait (Orr and Parsons 1982). To date, very little is known about the ecology of the ivory gull due to its remote locations year-round, and consequent limited accessibility to scientists (COSEWIC 2006; Gilg et al. 2010).

Breeding range and approximate numbers (< 2000 breeding pairs) of ivory gulls have been estimated recently in the high Arctic, but little is known about the behaviour and migration patterns of this species (Gilchrist and Mallory 2005; Gilg et al. 2010). As well, a 70% decline in population numbers since the 1980s has led to the uplisting of ivory gulls as endangered in Canada (COSEWIC 2006; Gilg et al. 2009). The

COSEWIC report and a more recent Recovery Strategy (Environment Canada 2014) identified several potential threats to ivory gulls. Most of the recent research on the ivory gull has focused on contaminants (Braune et al. 2006, 2007; Miljeteig et al. 2009).

Threats

As a pelagic seabird that feeds at a high trophic level, the ivory gull accumulates pollutants from dietary items which have assimilated contaminants principally derived from anthropogenic sources (Funess and Camphuysen 1997). Contaminants such as mercury, organochlorines (OCs) and brominated flame retardants (BFRs) are passed to eggs through the females, reducing fitness, causing developmental problems and decreasing the likelihood of survival of embryos and hatchability (COSEWIC 2006;

Braune et al. 2007; as reviewed in Miljeteig et al. 2012). These contaminants are becoming problematic in Canada, where ivory gull eggs are reported to have the highest mercury levels of any seabird in the Arctic which may have long term lethal and sub

lethal effects (Braune et al. 2001, 2007). An international study has found ivory gull eggshells from Russia and Svalbard to be 17% thinner than in 1930 (Miljeteig et al.

2012). However, concentration of OCs and BFRs in ivory gull eggs in Canada between

1976 and 2004 were below lethal levels for the bird (Braune et al. 2007). The influence of other possible threats to the ivory gull (harvest in Greenland, climate change, industrial activities, disturbance in key migration and wintering areas) have been identified but not examined quantitatively.

Many of the threats faced by ivory gulls may occur during migration or during winter; times when Canadian ivory gulls are extremely difficult to observe because they move away from their colony. During this period, the exposure of ivory gulls to various threats must be understood to formulate effective management recommendations for species’ conservation. For example, a significant threat to the ivory gull is illegal shooting in Greenland, which has historically occurred during periods of migration when

Canadian birds fly along the coast of west Greenland (Stenhouse et al. 2004;

Environment Canada 2014). The typical timing of arrival and departure of birds that move along that gauntlet is not known and this timing may change as climate change occurs. Birds also may winter in areas of increased shipping activity that could result in increased disturbance and enhanced exposure to oil pollution (Environment Canada

2014). Overall, the ivory gull wintering range or core wintering area is largely speculative (Orr and Parsons 1982). Determining the timing and duration of ivory gull exposure to these threats presents a considerable challenge in the remote and harsh conditions of the Canadian Arctic because the birds spend much of their time out of sight.

Satellite telemetry is an excellent tool to acquire information on movements and habitat

use of a remote species and is a valuable and an increasingly popular approach to addressing remote monitoring challenges (e.g. Sittler et al. 2011; Adams et al. 2012; Gilg et al. 2013; Nuijten et al. 2014).

Satellite telemetry

Satellite telemetry is a monitoring technique generally used on migratory animals to determine their movements throughout a year (Lindberg and Walker 2007). This technique provides scientists with a tool to understand population movements remotely

(Falk and Møller 1995), and recent Arctic examples include the northern fulmar

(Fulmarus glacialis), the long-tailed duck (Clangula hyemalis), and the thick-billed murre (Mallory 2006; Mallory et al. 2006; Gaston et al. 2011). While traditional bird banding provides information on bird movements from capture and recapture locations, satellite telemetry bridges the gap produced by this technique, providing researchers with instantaneous data on movement patterns and behavioural information otherwise unknown between these capture-recapture locations (Falk and Møller 1995).

In brief, birds are fitted with small satellite transmitters that periodically transmit locations to satellites that are then transmitted to the scientists (CLS 2014). Program R

(R Core Team 2012) and Geographic Information Systems (ArcGIS 10.1; Environmental

Systems Research Institute, Redlands, CA) can be employed to manipulate, analyze and visualize these satellite data to characterize the migratory movements of birds (e.g. Hillis et al. 1998; Merkel et al. 2006; Seney and Landry 2008; Peterson et al. 2012). Both applications can be used in concert to create filters discarding unrealistic satellite locations and create kernel density estimates (KDE) to identify home ranges and distinguish movement patterns of individuals.

Objectives and predictions

Very little is known about habitat requirements and year-round habitat use by the ivory gull. Consequently, the ivory gull Recovery Strategy recommended that a satellite tracking study be undertaken to record positions of ivory gulls throughout its yearly migration cycle. For the current study, satellite transmitters were attached to ivory gulls to track their movements, and subsequently I used GIS to identify and describe the annual movements and habitat use of the ivory gull in the Canadian Arctic. Secondly, GIS was used to define critical wintering habitat for the Canadian population; a key element required by the Recovery Strategy for the species (Environment Canada 2014) but has yet to be defined. This study provides the first continuous, multiyear monitoring of movement and habitat used by the same individuals from any seabird colony in Arctic

Canada (Gilchrist and Mallory 2005; Mallory et al. 2008; Mallory et al. 2012).

The breeding locations and suspected wintering locations for Canadian ivory gulls have been identified previously (Orr and Parsons 1982; Mallory et al. 2008). Because ivory gulls appear to have strong preferences for ice habitats (Mallory et al. 2008), my overall hypothesis was that sea ice distribution and formation would influence the route(s) flown by ivory gulls from their breeding colony in the central Canadian high

Arctic, to their suspected wintering area in Davis Strait and back to their breeding colony the following year. Consequently, I predicted that ivory gulls would primarily use marine habitats during migration and avoid terrestrial routes. Secondly, I predicted that the timing of formation of sea ice would act as a cue for ivory gulls to initiate fall migration.

Finally, suitable wintering habitats need to provide a reliable source of food; therefore I expected that ivory gulls would aggregate near hooded seal whelping patches, which are

thought to be a critical element of wintering habitat for the species (Orr and Parsons

1982; Mallory et al. 2008).

References

Adams J, MacLeod C, Suryan RM, Hyrenbach KD, Harvey JT (2012) Summer-time use of west coast US National Marine Sanctuaries by migrating sooty shearwaters (Puffinus griseus). Biological Conservation 156:105–116.

Bradstreet MSW (1982) Occurrence, habitat use, and behavior of seabirds, marine mammals, and Arctic cod at the Pond Inlet ice edge. Arctic 35:28–40.

Braune BM, Donaldson GM, Hobson K (2001) Contaminant residues in seabird eggs from the Canadian Arctic. Part I. Temporal trends 1975-1998. Environmental Pollution 114:39–54.

Braune BM, Mallory ML, Gilchrist GH (2006) Elevated mercury levels in a declining population of ivory gulls in the Canadian Arctic. Marine Pollution Bulletin 52:969– 987.

Braune BM, Mallory ML, Gilchrist GH, Letcher RJ, Drouillard KG (2007) Levels and trends of organochlorines and brominated flame retardants in ivory gull eggs in the Canadian Arctic, 1976 to 2004. Science of the Total Environment 378:403–417.

CLS (2014) Argos user’s manual. CLS Argos, Toulouse, France. 68 p.

COSEWIC (2006) Assessment and update status report ivory gull Pagophila eburnea. 42 + vi p.

Environment Canada (2014) Recovery strategy for the ivory gull (Pagophila eburnea) in Canada. Species at Risk Act Recovery Strategy Series. 26 + iv p.

Falk K, Møller S (1995) Satellite tracking of high-arctic northern fulmars. Polar Biology 15:495–502.

Fiedler W (2003) Recent changes in migratory behaviour of birds: a compilation of field observations and ringing data. In: Berthold P, Gwinner E, Sonnenschein E, editors. Avian Migration. Springer Berlin Heidelberg, pp 21–38.

Fort J, Moe B, Strøm H, Grémillet D, Welcker J, Schultner J, Jerstad K, Johansen KL, Phillips RA, Mosbech A (2013) Multicolony tracking reveals potential threats to little auks wintering in the North Atlantic from marine pollution and shrinking sea ice cover. Diversity and Distributions 19:1322–1332.

Furness RW, Camphuysen KCJ (1997) Seabirds as monitors of the marine environment. J Marine Science 54:726–737.

Gaston AJ, Bertram DF, Boyne AW, Chardine JW, Davoren G, Diamond A, Hedd A, Montevecchi WA, Hipfner JM, Lemon MJF, Mallory ML, Rail J-F, Robertson GJ (2009a) Changes in Canadian seabird populations and ecology since 1970 in relation to changes in oceanography and food webs. Environmental Review 17:267–286.

Gaston AJ, Descamps S, Gilchrist HG (2009b) Reproduction and survival of glaucous gulls breeding in an Arctic seabird colony. Journal of Field Ornithology 80:135– 145.

Gaston AJ, Gilchrist GH, Hipfner MJ (2005b) Climate change, ice conditions and reproduction in an Arctic nesting marine bird: Brunnich’s guillemot (Uria lomvia L.). Journal of Animal Ecology 74:832–841.

Gaston AJ, Gilchrist H, Mallory M (2005a) Variation in ice conditions has strong effects on the breeding of marine birds at Prince Leopold Island, Nunavut. Ecography 3:331–344.

Gaston AJ, Mallory ML, Gilchrist G (2012) Populations and trends of Canadian Arctic seabirds. Polar Biology 35:1221–1232.

Gaston AJ, Smith P, Tranquilla LM, Montevecchi WA, Fifield D, Gilchrist G, Hedd A, Mallory ML, Robertson GJ, Phillips RA (2011) Movements and wintering areas of breeding age Thick-billed Murre Uria lomvia from two colonies in Nunavut, Canada. Marine Biology 158:1929–1941.

Gilchrist G, Mallory ML (2005) Declines in abundance and distribution of the ivory gull (Pagophila eburnea) in Arctic Canada. Biological Conservation 121:303–309.

Gilchrist G, Strøm H, Gavrilo MV, Mosbech A (2008) International ivory gull conservation strategy and action plan. CAFF International Secretariat, Circumpolar Seabird Group (CBird), CAFF Technical Report No. 18. 20 + i p.

Gilg O, Boertmann D, Merkel F, Aebischer A, Sabard B (2009) Status of the endangered ivory gull, Pagophila eburnea, in Greenland. Polar Biology 32:1275–1286.

Gilg O, Kovacs KM, Aars J, Fort J, Gauthier G, Grémillet D, Ims RA, Meltofte H, Moreau J, Post E, Schmidt NM, Yannic G, Bollache L (2012) Climate change and the ecology and evolution of Arctic vertebrates. Annals of the New York Academy of Science 1249:166–90.

Gilg O, Moe B, Hanssen SA, Schmidt NM, Sittler B, Hansen J, Reneerkens J, Sabard B, Chastel O, Moreau J, Phillips RA, Oudman T, Biersma EM, Fenstad AA, Lang J, Bollache L (2013) Trans-equatorial migration routes, staging sites and wintering areas of a high Arctic avaian predator: the long-tailed skua (Stercorarius longicaudus). PLoS ONE 8:1–10.

Gilg O, Strøm H, Aebischer A, Gavrilo MV, Volkov AE, Miljeteig C, Sabard B (2010) Post-breeding movements of northeast Atlantic ivory gull Pagophila eburnea populations. Journal of Avian Biology 41:532–542.

Heide-Jørgensen MP, Laidre KL (2004) Declining extent of open-water refugia for top predators in Baffin Bay and adjacent waters. Ambio 33:488–95.

Hillis TL, Mallory FF, Dalton WJ, Smiegielski AJ (1998) Preliminary analysis of habitat utilization by woodland caribou in northwestern Ontario using satellite telemetry. Rangifer 18:195–202.

Irons DB, Anker-Nilssen T, Gaston AJ, Byrd GV, Falk K, Gilchrist G, Hario M, Hjernquist M, Krasnov Y, Mosbech A, Olsen B, Petersen A, Reid JB, Robertson GJ, Strøm H, Wohl KD (2008) Fluctuations in circumpolar seabird populations linked to climate oscillations. Global Change Biology 14:1455–1463.

Jenouvrier S, Barbraud C, Cazelles B, Weimerskirch H (2005) Modelling population dynamics of seabirds: importance of the effects of climate fluctuations on breeding proportions. Oikos 108:511–522.

Johnston D, Friedlaender A, Torres L, Lavigne D (2005) Variation in sea ice cover on the east coast of Canada from 1969 to 2002: climate variability and implications for harp and hooded seals. Climate Research 29:209–222.

Karnovsky N, Hobson K, Brown ZW, Hunt GJ (2009) Distribution and diet of ivory gulls (Pagophila eburnea) in the North Water Polynya. Arctic 62:65–74.

Lindberg MS, Walker J (2007) Satellite telemetry in avian research and management: sample size considerations. Journal of Wildlife Management 71:1002–1009.

Lindsay R, Zhang J (2005) The thinning of Arctic sea ice, 1988-2003: have we passed a tipping point? Journal of Climate 18:4879–4894.

Mallory ML (2006) The northern fulmar (Fulmarus glacialis) in Arctic Canada: ecology, threats, and what it tells us about marine environmental conditions. Environmental Reviews 14:187–216.

Mallory M, Akearok J, Gaston A (2009) Status of high Arctic black-legged kittiwake (Rissa tridactyla) colonies in Barrow Strait, Nunavut, Canada. Arctic 62:96–101.

Mallory ML, Akearok J, North NR, Weseloh DV, Lair S (2006) Movements of long- tailed ducks wintering on lake Ontario to breeding areas in Nunavut, Canada. Wilson Journal of Ornithology 118:494–501.

Mallory ML, Allard KA, Braune BM, Gilchrist G, Thomas VG (2012) New longevity record for ivory gulls (Pagophila eburnea) and evidence of natal philopatry. Arctic 65:98–101.

Mallory ML, Gaston AJ, Gilchrist G, Robertson GJ, Braune BM (2010) Effects of climate change, altered sea-ice distribution and seasonal phenology on marine birds. In: Ferguson SH, Loseto LL, Mallory ML, editors. A Little Less Arctic: Top Predators in the World’s Largest Northern Inland Sea, Hudson Bay. Springer, Dordrecht, pp 179– 195.

Mallory ML, Stenhouse I, Gilchrist H, Robertson G, Haney C, MacDonald S (2008) Ivory gull (Pagolphila eburnea). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/175. Accessed 2014 Apr 25.

Maslanik J, Fowler C, Stroeve J, Drobot S, Zwally J, Yi D, Emery W (2007) A younger, thinner Arctic ice cover: increased potential for rapid, extensive sea-ice loss. Geophysical Research Letters 34:1–5.

McLaren PL (1982) Spring migration and habitat use by seabirds in Eastern Lancaster Sound and Western Baffin Bay. Arctic 35:88–111.

Merkel F, Mosbech A, Sonne C, Flagstad A, Falk K, Jamieson SE (2006) Local movements, home ranges and body condition of common eiders Somateria mollissima wintering in Southwest Greenland. ARDEA 94:639–650.

Miljeteig C, Gabrielsen GW, Strøm H, Gavrilo MV, Lie E, Jenssen BM (2012) Eggshell thinning and decreased concentrations of vitamin E are associated with contaminants in eggs of ivory gulls. Science of the Total Environment 431:92–9.

Miljeteig C, Strøm H, Gavrilo MV, Volkov AE, Jenssen BM, Gabrielsen GW (2009) High levels of contaminants in ivory gull Pagophila eburnea eggs from the Russian and Norwegian Arctic. Environmental Science & Technology 43:5521–5528.

Nuijten RJ, Kölzsch A, van Gils JA, Hoye BJ, Oosterbeek K, de Vries PP, Klaassen M, Nolet BA (2014) The exception to the rule: retreating ice front makes Bewick’s swans Cygnus columbianus bewickii migrate slower in spring than in autumn. Journal of Avian Biology 45:113–122.

Orr C, Parsons J (1982) Ivory gulls, Pagophila eburnea, and ice edges in Davis Strait and the Labrador Sea. Canadian Field-Naturalist 96:323–328.

R Core Team (2012) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna Austria. Available from: http://www.R-project.org/.

Robertson GJ, Gilchrist G, Mallory ML (2007) Colony dynamics and persistence of ivory gull breeding in Canada. Avian Conservation Ecology 2:8.

Royston S (2007) Genetic structure, diversity and evolutionary history of the ivory gull (Pagophila eburnea) and Rross’s gull (Phodostethia rosea). M.Sc. Thesis, Memorial Univeristy of Newfoundland. Available: http://www.collectionscanada.gc.ca/obj/thesescanada/vol2/002/MR42056.PDF. Accessed 18 September 2013.

Salinger MJ (2005) Climate variability and change: past, present and future – an overview. Climate Change 70:9–29.

Seney E, Landry A (2008) Movements of Kemp’s Ridley sea turtles nesting on the upper Texas coast: implications for management. Endangered Species Research 4:73–84.

Sittler B, Aebischer A, Gilg O (2011) Post-breeding migration of four long-tailed skuas (Stercorarius longicaudus) from north and east Greenland to west Africa. Journal of Ornithology 152:375–381.

Stenhouse IJ, Robertson GJ, Gilchrist G (2004) Recoveries and survival rate of ivory gulls banded in Nunavut, Canada, 1971-1999. Waterbirds 27:486–492.

Stirling I (1980) The biological importance of polynyas in the Canadian Arctic. Arctic 33:303–315.

Stirling I (1997) The importance of polynyas, ice edges, and leads to marine mammals and birds. Journal of Marine Systems 10:9–21.

Thomas D, Fogg G, Convey P, Fritsen C, Gili J-M, Gradinger R, Laybourn-Parry J, Reid K, Walton D (2008) The Biology of Polar Regions. 394 p.

Thomson D (1982) Marine benthos in the eastern Canadian high Arctic: multivariate analyses of standing crop and community structure. Arctic 35:61–74.

Chapter two

Movement patterns and key habitats of the Canadian ivory gull through their annual cycle: the importance of sea ice1

1 Submitted to PLoS ONE 06 July 2014 17

Introduction

Climate change and environmental monitors

The Arctic is changing more rapidly in response to global warming than any other area of the world. With climate trends and models showing that global temperatures are increasing and anticipated reductions in sea ice cover, there is greater interest in resource exploitation in the Arctic (Irons et al. 2008; Gaston et al. 2012). Warming air temperatures and thinning sea ice results in annual ice taking place of multi-year ice

(Lindsay and Zhang 2005; Gilg et al. 2012). In fact, in 1987, 57% of sea ice in the central Arctic basin was ≥ 5 years old; however in 2007 that number dropped to 7%

(Maslanik et al. 2007). Continued assessment and monitoring of this sensitive environment and how warming trends are influencing Arctic flora and fauna is pertinent

(Gilg et al. 2012). Arctic seabirds have proven to be an appropriate and effective group of organisms with which to monitor variation in the Arctic environment, because different species feed at different trophic levels in Arctic food webs and consequently their reproductive success and habitat use reflects conditions of the food webs on which they depend (Boyd et al. 2006; Irons et al. 2008).

The ivory gull

Of the Arctic seabird species that breed in Canada, the ivory gull (Pagophila eburnea) should be an excellent bioindicator of the effects of global warming. However, to date we know relatively little about the species’ breeding biology or its habitat needs in

Canada (Mallory et al. 2008). It is an endangered species listed under the Species at Risk

Act (SARA) in Canada, spending its entire year in the circumpolar Arctic (COSEWIC

2006). It is a small, all-white gull with black legs and dark green bill with yellow tip

(Mallory et al. 2008). Populations are also found in Svalbard, Greenland and Russia

(Gilg et al. 2009). Current numbers are estimated to be < 1000 breeding pairs in Canada with a dramatic (>70%) decline in the population since the 1980s (Gilchrist and Mallory

2005).

The ivory gull breeds in remote sites in the Canadian high Arctic, where 49 known breeding colonies are found on nunataks (mountain peaks surrounded by glaciers), low-lying islands and inland plateaus of Ellesmere, Devon, Seymour, Cornwallis and northwestern Baffin Island (Brodeur Peninsula; COSEWIC 2006; Mallory et al. 2012b).

During the non-breeding season, wintering areas for Canadian ivory gulls have been identified as Davis Strait and Labrador Sea (Orr and Parsons 1982), although information on these locations is very limited. In the winter, the ivory gull is most likely found over pack ice of 70-90% ice concentration and along ice edges, but is rarely observed over open water > 5 km from ice (Mallory et al. 2008). This close association with sea ice has been documented often in ivory gull studies (Orr and Parsons 1982; Renaud and McLaren

1982; Thomas et al. 2008; Karnovsky et al. 2009). Other than some recent colony surveys and analyses of contaminant levels, most research on the ivory gull in Canada is limited to some descriptive work from Seymour Island in the 1970s (Mallory et al. 2008) and even that is outdated. Very little is known of the migratory behaviour, movements and key habitat needs of the ivory gull other than at their breeding colonies (Gilchrist and

Mallory 2005; COSEWIC 2006).

At a time when Arctic environments are changing, species that appear to be closely tied to sea ice might be expected to be deleteriously affected. Polar bears, Ursus maritimus, are iconic in this regard (Stirling and Parkinson 2006), but ivory gulls are no

less important as a single species under similar habitat threats. Clearly, further research is required to better manage this endangered species and to identify possible actions to promote its recovery and protection (Environment Canada 2014). However, given the remote location and limited accessibility of the ivory gull, classic seabird observation studies (e.g. Mallory et al. 2009, 2012b) are difficult to conduct effectively. Instead, satellite telemetry can be employed to define year-round distribution, migratory movements and behaviour of the species with little effort (as reviewed in Burger and

Shaffer 2008).

Satellite telemetry

Satellite telemetry has become a widely used technique in the seabird world, to provide information on bird behaviour at times when birds cannot practically be observed from land or sea (Falk and Møller 1995; Furness et al. 2006; Burger and Shaffer 2008;

Mallory et al. 2008; Hatch et al. 2010; Gilg et al. 2010). Improving technology is continuously making satellite transmitters smaller and lighter, opening up the possibilities for the use of this technology on increasingly smaller study subjects (Burger and Shaffer

2008). The satellite transmitters, or Platform Terminal Transmitters (PTTs), transmit messages to nearby satellites on a pre-defined schedule and these messages are collected by the Argos satellite processing centres and sent to the user of the program (CLS 2014).

By linking some basic knowledge of the activities of the birds at different times of year

(e.g. breeding, foraging trips for breeding birds, migration movements), these locations and the time spent to move between them may then be analyzed and interpreted to reveal previously unknown information on annual movements, thus lend insights into habitat use trends and ultimately, inform habitat requirements.

Objectives and predictions

The breeding locations and suspected wintering locations for Canadian ivory gulls have been identified previously (Orr and Parsons 1982; Robertson et al. 2007; Mallory et al. 2008). However, information on migration timing and movements was largely based on speculation. The exception is that we believed many Canadian birds migrated along the Greenland coast in spring and fall, because birds banded in Canada were subsequently shot in Greenland (Stenhouse et al. 2004). Because ivory gulls appear to have strong preferences for sea ice habitats (Mallory et al. 2008), my overall hypothesis was that sea ice distribution and timing of ice formation are the primary factors influencing their migration routes and timing from their breeding colony in the central Canadian high

Arctic to their suspected wintering area in Davis Strait and back to their breeding colony the following year. Based on this hypothesis, I made three following predictions: 1) ivory gulls would use primarily marine habitats and travel routes during migration, not terrestrial routes; 2) gulls would migrate over ice, spending a disproportionate amount of time over this habitat and would not migrate in advance of its formation; and 3) wintering habitats over ice would be closely linked to food supplied and therefore birds would spend the majority of their time close to the floe edge, or in the vicinity of known food supplies such as hooded seal (Crystophora cristata) whelping patches. In fact, I expected that the vicinity of these whelping patches forms critical wintering habitat for the species

(Orr and Parsons 1982; Mallory et al. 2008).

Methods

Study area

Seymour Island (78.80° N, 101.27° W), located in Nunavut, is a Migratory Bird

Sanctuary established in 1975 (Fig. 1). This low lying rocky island is approximately 3km long and an example of a nesting location that is subject to predation by Arctic fox

(Vulpes lagopus) and polar bears (Mallory et al. 2012a). Seymour Island has supported over 300 ivory gulls in the past; however, despite population declines of approximately

2.7% between 1974 and 2006, it remains the largest nesting colony for the Canadian population of ivory gulls (Robertson et al. 2007; Latour et al. 2008). This colony is located near the Penny Strait polynya, thought to be an important foraging area for the ivory gulls during the breeding season (Mallory and Gilchrist 2003; Mallory et al. 2008).

This study site was selected as it has provided consistent reports of annual ivory gull presence and is one of the few accessible locations to capture individuals easily

(Robertson et al. 2007).

The geographic study range extended from the bird collection area, Seymour

Island, to Davis Strait and Labrador Sea, based on previous banding work and at-sea surveys (Orr and Parsons 1982; Stenhouse et al. 2004).

Figure 2.1: Bird capture and tagging of 12 ivory gulls took place on Seymour Island,

NU; a small low-lying island (3 km long), supporting the largest breeding colony for ivory gulls in Canada and a Migratory Bird Sanctuary in Canada.

Bird capture and tagging (2010)

Twelve ivory gulls were captured using a modified version of a bownet trap

(Salyer 1962) from a single colony on Seymour Island on 29 and 30 June 2010. Field work was conducted by Mark Mallory, Birgit Braune and Karel Allard. Five individuals were tagged with 20g battery powered PTTs made by North Star Technologies (King

Georges, Virginia). The remaining seven individuals were tagged with 15g solar powered PTTs (a customized PTT-100 12g model in a larger case to fit a larger solar chip) by Microwave Telemetry, Inc. (Columbia, Maryland). Individuals were caught during incubation to ensure they were actively breeding in Canada and a leg loop harness design was used to attach the transmitters, leaving flight muscles and major fat deposits unencumbered (Mallory and Gilbert 2008). The transmitter plus the harness represented approximately 3% of ivory gull body mass; the preferred load to minimize deleterious effects on individuals (Phillips et al. 2003). Ivory gulls were released and observed for approximately 15 min (unless they flew out of sight) to ensure birds were accepting the harnesses. All transmitter attachments on birds appeared to be successful.

The PTTs were compatible with the Argos satellite positioning system (CLS

2014). The duty cycle of battery-powered PTTs was programmed to send signals within an 8 h period and shut off for 72 h. Solar powered PTTs had 10 hr on and 48 hr off with customized modifications to the voltage output done by the manufacturer to accommodate the low incident light conditions of the Arctic fall and winter. Each message received from Argos was given an accuracy of the location estimate if four or more messages were sent to the satellite: location class (LC) 0 ≥ 1500 m; LC 1 ≤ 1500 m;

LC 2 ≤ 500 m; LC 3 ≤ 250 m. Location classes A and B did not have accuracies (i.e. only

two or three messages were received) calculated and LC Z described a failed accuracy calculation. A description of how the each PTT was powered, start and end dates of transmission and number of useable locations is in Table 2.1.

Messages began transmitting on 1 July 2010. As of July 2013, two of the PTTs continued to provide locations for this project, while eight others lasted between four and

12 months and two others lasted until August and September 2012. A black out period occurred each year for approximately 8-10 weeks between November and January where the solar powered PTTs were not able to transmit data because of a lack of sunlight and consequently, insufficient recharging and power to send signals. Nonetheless, 59 439 locations were available for analyses. There was a trade-off between the low number of individual birds tracked and the long periods of tracking (repeat observations) of the same individual birds.

July 2010- July 2013.

PTT Battery/ Dates running # months Useable locations solar power transmitting (LC 0,1,2,3) 44509 Battery 1 Jul, 2010 – 23, Jan, 2011 7 550 44516 Battery 1 Jul, 2010 – 18, Nov, 2010 5 467 44517 Battery 1 Jul, 2010 – 18, Feb, 2011 8 762 44519 Battery 1 Jul, 2010 – 03, Mar, 2011 7 653 44522 Battery 1 Jul, 2010 – 26, Nov, 2010 5 476 44523 Solar 1 Jul, 2010 - current date 48 + 10276 44524 Solar 1 Jul, 2010 – 26, Sep, 2012 38 7958 44525 Solar 1 Jul, 2010 – 20, Aug, 2012 38 7385 44526 Solar 1 Jul, 2010 – 30, May, 2013 49 13415 44529 Solar 1 Jul, 2010 – 22, Nov, 2010 5 1876 44530 Solar 1 Jul, 2010 – 11, Jul, 2011 12 3366 44531 Solar 1 Jul, 2010 – 20, May, 2011 11 3093

Processing of location data

The statistical programs R (R Core Team 2012), Geospatial Modelling

Environment (Spatial Ecology, 2012) and ArcMap 10.1 (ArcMap; Environmental

Systems Research Institute, Redlands, CA) were used for the analysis.

With a large dataset, I chose to include only the highest quality location data for analyses. Thus, I excluded positions with location classes “A”, “B”, “Z”, those that were blank as well as any with erroneous dates (before 1 July 2010). One of the sensors on the transmitters detected bird movement and indicated that birds were still active if the sensor value changed each time a message was received. If the motion sensor numbers remained the same, it meant the harness had been discarded or the bird had died; any data collected after the point at which motion sensor values stopped changing were excluded from further analyses.

Annual cycle

To properly assess how ivory gulls behaved during different times of the year, I estimated periods that described the breeding and wintering seasons as well as fall and spring migrations. However, individual rates of travel and distances flown varied within and across years, and therefore defining a single date to begin and end a migration would introduce unnecessary bias. Arrival to the breeding and wintering areas were defined as the dates that a bird had obviously slowed down and was no longer making large directional movements (for the breeding season this was most often when they arrived in

Parry Channel, above 74.5°N).

Defining the beginning of both fall and spring migration for this species was subjective as ivory gulls have many short stopovers during their migration, and in some cases even traveled back in the direction from which they came. However, I defined the start of migration periods as the date when a bird began to fly long distances in an obvious linear pattern, generally away from the breeding or wintering area. Using the range of arrival and departure dates for individual ivory gulls, a median date was given to the beginning of each season (winter, spring migration, breeding and fall migration) to standardize analyses across birds and years, as in Gilg et al. (2013) (Table 2.2).

For a more detailed analysis of the breeding season, individuals were divided into those that bred and those that did not (as determined from birds returning and remaining at the breeding colony). The breeding season was also broken into three stages: 1) the dates between arrival at the breeding area and arrival at the colony; 2) the dates spent at the colony; and 3) the date the individual left the colony to the beginning of fall migration. Arrival at the breeding colony was defined as the date when an ivory gull was first seen over land and continued to return to the colony. The date of colony departure was defined as the first day an ivory gull flew away from the colony and did not return.

Those birds that did not make a breeding attempt or abandoned their nesting attempt (yet had returned to the vicinity of the colony) in a given year, were described as non- breeding birds. Bateson and Plowright (1959), described 17 July as the threshold date after which most copulations did not succeed. Therefore, this was taken as the date when a bird not already attending the colony and was categorized as a “non-breeder”. This date was not the date they began breeding because the birds did not breed every year of the

study. For these non-breeding birds, I grouped their subsequent behavioural period into one category: the date from arrival at the breeding area to the beginning of fall migration.

A second challenge included the blackout period of 24 h darkness between

November and January when the tags transmitted few signals because the transmitters were solar powered. Only two birds arrived at their wintering location before the blackout in late November (44517, 2010; 44523, 2012). One bird arrived at the wintering location in mid-January after the tags began transmitting again. The remaining birds had a data gap of one to three months, arriving at the wintering location at some point within the blackout period. Therefore the median date of arrival at the wintering location was taken to minimize bias.

Table 2.2: Median ( ̃) start date, range of start dates and the median ( ̃) number of days

and the range for each season of the annual cycle for the 12 satellite-tagged ivory gulls

throughout the study period, July 2010- July 2013.

Season ̃ start date Range ̃ number of days (range) Arrival to wintering area 19 Dec 20 Nov – 17 Jan 154 (129-171) Start of pre-breeding migration 15 May 02 May – 28 May 18 (8-28) Arrival to breeding area 05 Jun 23 May – 19 Jun 118 (89-127) Start of post-breeding migration 26 Sep 05 Sep – 18 Oct 74 (50-121)

Duplicated data

Duplicated observations sharing location, date and time (making up approximately 10% of the overall data) needed to be deleted from the dataset. Duplicate observations either had differing location classes or latitude and longitude. For those records, I retained the observation with a more accurate location class, or if these were the same, I retained the record with the smaller error radius (i.e. I always kept the most accurate record). No cases arose where all variables were the same.

Calculation of distance and rates of travel

Distance was calculated as the orthodromic (great circle route) distance and was calculated in kilometres (km) between consecutive locations for each individual (as in

Gilg et al. 2013). Rate of travel was then calculated from the distance and the time difference between two consecutive locations in km/h. Rate of travel was also calculated on a daily and weekly time scale.

Filtering unrealistic locations

The final step before analysis of the data was to create a filter to extract any implausible rates of travel (~flight speeds) for the birds. There were no reports of rates of travel for ivory gulls, only that they are reported to fly faster than black-legged kittiwakes

(Rissa tridactyla; Mallory et al. 2008). Oldén and Peterz (1985) recorded black-legged kittiwake ground speeds of 96km/h, so I used 100km/h as the threshold rate of travel for ivory gulls. This was probably a liberal estimate, as birds never appeared to be flying this fast even when evading a helicopter landing near the colony (M. Mallory, pers. observ.); however, a lower threshold may have removed satellite messages that were real bird

movements, as wind-assisted flights could have resulted in true, high rates of travel.

Wind data were not incorporated in this study.

Distances and times were re-calculated for the rates of travel that were not rejected by the filter. These recalculations still showed a few records with rates greater than 100 km/h so the filter was repeated until there were no longer any speeds greater than 100 km/h.

Adjusting for time periods larger than 24 h

Blackout periods were typically one to three months in length. Using the last point before the blackout period and the first point after the blackout period led to the calculation of unrealistic time differences and distances (if a linear pattern were assumed). Thus, I zeroed the beginning of each year for these variables, eliminating the inflation of future calculations. Similarly, since the difference in time between two consecutive points was often greater than 24 h, the number of days (24 h period) between consecutive points was calculated. If the number of days was greater than one, the distance was divided by the number of days (then divided by 24 h to get km/h) to provide a more realistic representation of how far an ivory gull could travel within 24 h. Rates of travel were adjusted accordingly for analysis. As this method underestimated how far an ivory gull flew when large gaps in time were produced between transmissions, I did not use these calculations for summarizing total distances flown by individuals.

Analysis

Assessing sources of bias and variation among transmitters

To assess sources of variation or bias among satellite transmitters, I calculated the total distance traveled per day, per bird and per total number of transmissions. This tested differences in distances flown between individuals and whether the distribution of the time of day when messages were received varied across seasons (potentially biasing rates of travel in one season compared to another). The mean distance traveled per day varied across individuals (K-W test, χ2=264, p < 0.001; Table 2.3). The satellite transmitters’ duty cycles did not allow continuous transmission across 24 h and poor transmission conditions (i.e. weather) also shortened transmission periods. Duty cycles may have experienced frequency drift over time, meaning that there may have been a difference over time between the actual location and the location received (Nicholls and Robertson

2007). Therefore, the distance traveled per day and total number of transmissions represented an index (minimum) of the distance an ivory gull can travel in a day.

Consequently, I expected that distance traveled would be positively correlated with the number of transmissions sent per day, but this was not the case (r12 = -0.09, p = 0.78).

Secondly, transmissions were not biased by time of day (K-W test χ2= 3, p = 0.39) and the distribution of the time of day when messages were received were relatively similar across seasons (Fig. 2.2). Therefore, variation of the rates of travel between individuals must be explained across seasons, years or by behavioural differences between individuals. Comparisons between seasons, years and individuals were made using indices of rates of travel on different time scales.

Table 2.3: The mean ( ̅) characteristics and standard error (SD) of each satellite-tagged ivory gull, including the mean distance

traveled per day, mean duration traveled per day, the total number of transmissions and the number of days that the satellite

transmitter sent signals throughout the study period, July 2010 – July 2013.

PTT 44509 44516 44517 44519 45522 44523 44524 44525 44526 44529 44530 44531 ̅ km/d ± SD 111 ± 127 ± 179 ± 145 ± 82 ± 92 ± 97 ± 107 ± 163 ± 144 ± 120 ± 117 ± 122 90 217 202 65 105 119 114 131 115 131 108 ̅ transmissions/d 5.90 6.41 5.35 4.49 5.03 4.75 4.76 5.37 4.98 4.10 5.02 5.54 Transmissions 286 229 384 359 218 5543 4161 3695 4823 429 1712 1533 # Days 31 24 48 53 28 785 602 504 811 105 262 203

Figure 2.2: The distribution of the number of good quality transmissions (LC 0, 1, 2 and

3) used for analyses in relation to the time of day (early morning,0000-0559; morning,

0600-1159; afternoon, 1200-1759; and night, 1800-2359) and season (wintering, spring migration, breeding and fall migration) they arrived throughout the study period, July

2010 – July 2013.

Rate of travel

Mean rates of travel (km/h and km/wk) were calculated per month using all records and all good quality transmissions (Gilg et al. 2010). Because most data distributions did not approximate normality (as assessed using Q-Q plots), I used a non- parametric Kruskal-Wallis test in program R to assess whether the rates of travel of ivory gulls differed by season. A pairwise Wilcoxon test was then used to assess in which seasons the rates of travel varied significantly using the Holm (1979) method to adjust the p-value for greater than three comparisons.

Distance traveled

The total distance traveled per bird in a year (1 July – 30 June) was calculated summing all locations of distance traveled by season. This is a minimum estimate as distances flown during blackout periods could not be calculated.

Sea ice

To analyze sea ice concentration of dates ranging 01 July 2010 to 30 June 2013 in

ArcGIS, the ‘Interpolate Time Series of Rasters at Points’ tool was used from the open source extension Marine Geospatial Ecology Tools (Roberts et al. 2010). This correlated daily sea ice charts (NSIDC) with all ivory gull locations in relation to date. Values of

254 and 253 indicated land, while values of 0 indicated no sea ice was present (i.e. open water) and all values in between represented sea ice concentration (Maslanik and Stroeve

1999). I used 3 discreet habitat categories; ice (of any concentration), water and land.

Proportions of time spent over each habitat type by each bird per year and season were calculated. To assess how proportions of birds over each habitat type changed across

seasons, a Kruskal-Wallis test and Dunn’s Multiple Comparisons test was applied in

InStat (GraphPad Software, LaJolla, CA, USA, 1997). All numbers were reported ± SD.

Results

Distribution

Ivory gulls breeding at the Seymour Island colony used areas north of Parry

Channel to forage during breeding, non-breeding and pre/post-breeding (Fig. 2.3). Prior to migration, gulls moved into Parry Channel, Lancaster Sound (part of Parry Channel) and in one case up to the North Water Polynya. Foraging and migration locations varied among years, but key geographic regions that they used included Davis Strait, Foxe

Basin, and (in one case) Hudson Bay (Figs. 2.4-2.6). Ivory gulls spent the winters in

Davis Strait and the Labrador Sea. Ivory gulls rarely traveled over open water that was devoid of ice, but tended to remain along the ice edge in the winter. Only three transmissions were recorded within territorial waters (22 km) of Greenland, two being from the same individual (44526) on consecutive days in the spring and the other observation was during the fall (44523). The total area, defined by a maximum convex polygon, used by individuals during the study was 6 175 895 km2.

Figure 2.3: Annual distribution of the Canadian ivory gull. The 50% kernels (methods are discussed in Chapter 3) represent distribution during breeding (red), post-breeding (orange), winter (light blue) and pre-breeding seasons (dark blue). General direction of fall migration is indicated by the arrows (south via Davis Strait and Foxe Basin in orange) from 'post-breeding' to 'winter' and direction of spring migration is indicated by the arrow (north via Davis Strait in blue) from 'pre-breeding' to 'breed. Three years and 12 birds are represented.

Figure 2.4: General direction traveled for nine bird-years south Figure 2.5: General direction traveled for three bird-years via Davis Strait during fall migration over the study period south via Foxe Basin Strait during fall migration over the study

(July 2010- July 2013). period (July 2010- July 2013).

Sea ice

Across all years and individuals, ivory gulls had three habitat types over which they could be recorded: sea ice, land, or open water. Time spent over ice, land and water differed among seasons (K-W test, χ2 = 24.4 p < 0.0001; χ2 = 14.2 p = 0.003; χ2 = 28.6 p< 0.0001, respectively; Fig. 2.7). Ivory gulls spent less time over sea ice in the breeding season than during spring migration (Dunn’s Multiple Comparisons test, p < 0.01), whereas fall migration showed that fewer individuals spent time over ice compared with spring migration (p< 0.001) and winter (p< 0.01). Time spent over land in winter was significantly less than during breeding (p< 0.05) and fall migration (p< 0.01). Finally, time spent over water during the breeding season was greater than during spring migration (p< 0.001) and more time was spent over water compared to winter (p< 0.05) and spring migration (p< 0.001).

Using data from five ivory gulls that provided at least one full year of information, the average ivory gull flew 31 118 ± 9035 km in a year (n= 10 bird-years).

This total corresponded to a mean rate of travel of 10 ± 11.4 km/h and an average daily movement of 160 ± 82 km. Individually, however, gull behaviour varied among season, individual and year.

(n=19, 140). Whiskers represent values within 1.5 times the interquartile range. Total number of days with observations=942 over the study period, July 2010- July 2013.

Seasonal spatial and temporal patterns

Wintering

Ivory gulls stayed in Davis Strait and the Labrador Sea for a median duration of

154 d (129-171 d) during the winter. Median date of arrival was 19 December (range 20

November – 17 January). One ivory gull (44509) remained in Barrow Strait (part of

Parry Channel; 76.05ºN -105.01ºW) during the winter of 2010, until the transmitter failed on 23 January 2011. The activity sensor indicated that this bird remained alive until that day.

Spring migration

Spring migration occurred over 18 d (8-28 d). The earliest date of departure was 2

May and the latest departure was 29 May. Gulls flew relatively directly to breeding colonies and with few stops during the spring migration. Throughout all years tracked during spring migration, ivory gulls only traveled north though Davis Strait and Baffin

Bay, then flew west though Lancaster Sound and subsequently passed Cornwallis Island to get to their breeding area (Fig. 2.8).

Figure 2.8: Six (n= 12 bird-years) satellite-tagged ivory gulls migrated north in spring through Davis Strait, west through Lancaster Sound and north to the breeding area during the study period (July 2010- July 2013)

Breeding

All ivory gulls abandoned breeding after being tagged in 2010. Before birds headed to the breeding colony in the following years they foraged in the vicinity of the colony for a median of 10 d (7-26 d). Two individuals appeared to have bred successfully in 2011, 44523 and 44525, as indicated by the period of time spent at the breeding colony

(52 d and 65 d, respectively). Each individual bred at a separate colony; neither of which had been previously discovered, located on Grinnell Peninsula of Devon Island (Fig. 2.9).

However, 44523 visited Seymour Island between 23-27 June. Gulls 44524 and 44530 both arrived on Seymour Island in 2011 but the transmitter for 44530 stopped transmitting 42 d later, and 44524 abandoned the island after spending 10 d there (Table

2.4). Of four gulls still transmitting to this time, none bred in 2012, but 44525 landed at a known colony on eastern Devon Island before the satellite tag stopped transmitting on the day it arrived at the colony (the activity sensor also began indicating that there was no movement later in the day that it arrived). This suggests that the bird died after arriving at this site. Gull 44526 stopped on eastern Cornwallis in 2013 at a previously unknown breeding site, just north of some known colonies; however, the activity sensor indicated that the transmitter was no longer moving one day after arriving at the colony, meaning the gull had either died or the transmitter had been detached (Fig. 2.10).

Figure 2.9: Breeding colonies used in 2011. One located on Figure 2.10: Breeding colony used in 2013. One ivory gull

Seymour Island, that was abandoned (two ivory gulls) and two was located on Cornwallis Island, but transmissions cut out located on Grinnell Peninsula (one ivory gull at each colony). shortly after its arrival.

Non-breeding

Ivory gulls that did not breed foraged as far north as the Arctic Ocean (82º N, Fig.

2.11), extending north to Ellesmere Island, west to Victoria Island and Parry Channel until fall migration. The area used by breeding and non-breeding ivory gulls originally tagged at Seymour Island encompassed approximately 768 199 km2 (as determined by a

Minimum Convex Polygon). Non-breeding birds remained in the breeding area between

23 May and 17 October (median 79 d, range 12-127 d).

Fall migration

Ivory gulls left the breeding area at between 5 September and 18 October (median date 26 September). The median length of the fall migration was74 d (range 50-121 d).

Migratory movements were punctuated by regular stops, presumably to forage, rather than long, continuous flights observed in spring migration. Ivory gulls flew one of three routes during fall migration: 1) east though Lancaster Sound and then south though

Baffin Bay and Davis Strait (Fig. 2.4); 2) south from Lancaster Sound though Prince

Regent Inlet and Foxe Basin (Fig. 2.5); and 3) a combination of moving south through

Prince Regent Inlet, north to Baffin Bay and south through Davis Strait (Fig. 2.6).

Individuals exhibited variation in fall migration: one gull used the Baffin Bay / Davis

Strait route three years in a row, while three other gulls used different routes in successive years (Table 2.4).

Table 2.4: Individual dates of arrival/departure for 12 satellite-tagged ivory gulls as well as migratory routes and breeding colonies

used throughout respective annual cycles throughout the study period, July 2010- July 2013. ‘No Attempt’ indicates that no breeding

attempt was made during the breeding season.

Bird Year Arrival to Depart for Migration Arrival Arrival to Colony Depart Depart for Migration wintering spring route to breeding name from fall route area migration breeding colony (island) breeding migration area colony 44509 2010 ------02 Jul 05 Sep - 44516 2010 ------02 Jul 23 Sep - 44517 2010 ------02 Jul 09 Sep Davis Strait 2011 20 Nov ------44519 2010 ------03 Jul 06 Sep Davis Strait 2011 04 Dec ------44522 2010 ------07 Jul 22 Sep - 44523 2010 - - - - - 03 Jul 25 Sep Davis Strait 2011 09 Dec 28 May Davis Strait 19 Jun 29 Jun Devon 20 Aug 09 Oct Davis Strait 2012 22 Dec 23 May Davis Strait 09 Jun No Attempt - - 14 Oct Davis Strait 2013 03 Dec 27 May Davis Strait 11 Jun No Attempt - - - - 44524 2010 ------04 Jul 21 Sep Foxe Basin 2011 11 Dec 23 May Davis Strait 02 Jun 30 Jun Seymour 09 Jul 18 Oct Davis Strait 2012 13 Dec 22 May Davis Strait 09 Jun - - - - -

Table 2.4 (cont’d): Individual dates of arrival/departure for 12 satellite-tagged ivory gulls as well as migratory routes and breeding

colonies used throughout respective annual cycles throughout the study period, July 2010- July 2013. ‘No Attempt’ indicates that no

breeding attempt was made during the breeding season.

Bird Year Arrival to Depart for Migration Arrival to Arrival to Colony name Depart from Depart for Migration wintering spring route breeding breeding (island) breeding fall route area migration area colony colony migration 44525 2010 ------16 Jul 18 Sep Davis Strait 2011 17 Jan 26 May Davis Strait 08 Jun 20 Jun Devon 24 Aug 05 Sep Foxe Basin 2012 03 Dec 22 May Davis Strait 10 Jun 10 Jun Devon - - - 44526 2010 ------05 Jul 02 Oct Foxe Basin 2011 14 Dec 02 May Davis Strait 30 May No Attempt - - 26 Sep Foxe/Davis 2012 09 Dec 04 May Davis Strait 23 May No Attempt - - 17 Sep Foxe/Davis 2013 09 Dec 02 May Davis Strait 27 May 28 May Cornwallis - - - 44529 2010 - - - - - 10 Jul 18 Sep - 44530 2010 ------05 Jul 04 Oct Davis Strait 2011 11 Dec 16 May Davis Strait 24 May 31 May Seymour - - - 44531 2010 ------31 Jul 18 Oct Davis Strait 2011 14 Dec 09 May Davis Strait - - Devon - - -

Rate of travel

Mean weekly rate of travel for an ivory gull was 512.2 ± 393.8 km/wk (range 0-

2023 km/wk). Between January (n=728 bird-days) and February (n=1647) ivory gulls increase their weekly movements from 346.0 ± 369.4 km/wk to 661.4 ± 350.8 km/wk.

Their rate peaked for the wintering period in March (n=2230) at 717.4 ± 451.9 km/wk.

During spring migration, the distance peaked again at 703.2 ± 449.2 km/wk in May

(n=2280) and then declined though the breeding season into August (n=3689; 411.7 ±

245.9 km/wk). This distance slowly increased during fall migration until November

(n=205; 649.2 ± 483.6 km/wk). December was excluded as there were too few records due to the blackout period for the transmitters. Mean rates of weekly travel can be seen in Fig. 2.12. Weekly rates of travel differed among seasons (K-W test, χ2=26.5, p<0.001), with rates of travel for the winter being significantly higher than rates during the non-breeding season (Pairwise Wilcoxon test; p<0.001) and fall migration (p<0.009).

Figure 2.12: Median ( ̃) and 25th and 75th quartiles of weekly rate of travel per month for all satellite-tagged ivory gulls throughout the study period, July 2010- July 2013.

Mean hourly rate of travel for ivory gulls across all months was 10.0 ± 11.3 km/h

(range 0 - 99 km/h). Peak in hourly rates of travel appeared during November (n= 205 bird-days; 12.2 ± 11.5 km/h) when all birds had initiated migration (Fig. 2.13). Between

January (n= 728) and April (n= 2112), ivory gulls flew at average speeds between 10.4 ±

12.2 km/h to 11.6 ± 12.4 km/h and decreased to the lowest point during the breeding season in July (n= 3910; 8.7 ± 10.6 km/h). Hourly rate of travel increased again after

July when weekly travel rates decreased, as individuals likely conserved energy while resuming molt before migration (traveling at higher hourly rates but for shorter periods of time). Ivory gulls flew at significantly different hourly rates of travel across seasons (K-

W test, χ2 = 231.5, p <0.0001). Mean rates of travel during the breeding and non- breeding seasons were significantly lower as compared to the wintering season and spring and fall migrations (Pairwise Wilcoxon; all p<0.001). Mean rates of travel of non- breeding individuals were significantly higher (n=6958; 9.4 ± 10.9 km/h; p <0.001) from those birds that were breeding (n= 5025; 8.7 ± 10.7 km/h).

On a daily scale, the average distance flown varied by season (K-W test, χ2 = 42.5, p <0.0001); Table 2.5). Results from the pairwise Wilcoxon test showed that the daily distances traveled during the breeding season were lower than wintering, spring or fall

(all p< 0.0001). Within the breeding season, breeding birds flew greater distances per day (n= 5025; 109.1 ± 120.7 km/d) than non-breeding birds (n=6958; 93.2 ± 77.3 km/d; p

< 0.0001).

Figure 2.13: Median ( ̃) and 25th and 75th quartiles of hourly rate of travel per month for all satellite-tagged ivory gulls throughout the study period, July 2010- July 2013.

Distance traveled

While annually ivory gulls flew approximately 31 118 ± 9034.9 km (n=10 bird- years), one ivory gull traveled for a minimum of 132 213 km over three years (Table 2.6).

Using total distances traveled by birds that provided a full season’s worth of data, seasons were found to vary significantly (K-W63= 31.5, p<0.0001), with distance flown during winter significantly greater than during any other season (Dunn’s Multiple Comparisons tests, all p≤0.01). All other inter-seasonal comparisons were not significantly different

(all p>0.05). Although the direct distance between their wintering and breeding area was

2300 km, ivory gulls traveled a mean of 3674.1 ± 2205.6 km (n=14 bird-seasons) during fall migration, stopping often and/or changing their route. By contrast, individuals used a direct route with few stops during spring migration, traveling a mean of 2961.7 ± 994.5 km (n=11 bird-seasons; i.e., 20% shorter flight in spring). Birds that abandoned nests or did not breed during the breeding season traveled a mean of 5002.6 ± 2499.4 km (n=17 bird-seasons) between spring and fall migration. Individual distances are available in

Appendix A.

Table 2.5: Mean ( ̅) and standard deviation (SD) of distance flown per day (km/d) during breeding, wintering and migratory periods of 12 satellite-tagged ivory gulls throughout the study period July 2010- July 2013.

Season Wintering Spring Breeding Fall Migration Migration ̅ ± SD 126 ± 123.0 152 ± 163.2 100 ± 97.3 157 ± 180.7 N 7889 1442 11994 2090

Table 2.6: Total distance flown (km) for 12 satellite-tagged ivory gulls across breeding years (01 July – 30 June of 2010/2011, 2011/2012 and 2012/2013). Brackets indicate that a PTT did not transmit for the entire year. Mean ( ̅ ) and standard deviation (SD) was calculated for the distance traveled by the six ivory gulls that transmitted for the entire period of 2010/2011. N refers to the number of good quality locations received for each individual.

Bird 2010/2011 2011/2012 2012/2013 Total N 44509 (3428) - - 3428 286 44516 (3055) - - 3055 229 44517 (9152) - - 9153 284 44519 (7712) - - 7712 359 44522 (2294) - - 2294 218 44529 (15074) - - 15074 429 44531 23858 - - 23858 1533 44523 22634 19606 29954 72194 5543 44524 26819 25518 (5848) 58185 4161 44525 28185 25921 - 54106 3695 44526 45234 46612 40367 132213 4823 44530 31454 - - 31454 1712 ̅ ± SD 29697 ± 8234.4 - - - -

Discussion

As expected, the breeding distribution was consistent with previous banding and observational work (Thomas and MacDonald 1987; Gilchrist and Mallory 2005).

However, the wintering distribution has been speculative, historically, based on casual observations (to describe wintering areas and movements in winter) and a single winter survey (Orr and Parsons 1982). These results show that Davis Strait and the Labrador Sea are the two main winter areas that ivory gulls from Seymour Island return to every year and these transmitted observations are consistent with earlier reports (Orr and Parsons

1982; Renaud and McLaren 1982). Gilg et al. (2010) showed that ivory gulls from

Russia, Greenland and Svalbard also use the northwest Atlantic to winter, suggesting that this area is of international significance to the ivory gull (this will be analysed in detail in

Chapter 3). Hypotheses explaining the importance of Davis Strait and Labrador Sea to the ivory gull for the winter include the presence of hooded seal whelping patches and the location of the sea ice edge. Sea ice edges are not only important for the ivory gull during winter, but also during migratory periods (Table 2.4).

Migratory movements and sea ice

Results showing the distribution of the ivory gull during spring and fall have markedly changed our perception of ivory gull migration, including their heavy reliance on sea ice formation and recession. It has long been believed that ivory gulls migrate principally along the Greenland coast (Stenhouse et al. 2004). However, my study found that individuals showed great plasticity in their migration routes and timing, which can probably be attributed to their affinity for sea ice and their apparent avoidance of completely open water (COSEWIC 2006; Mallory et al. 2008; Gilg et al. 2010). It is

possible that the ivory gulls sampled in my study have never flown this route, or my results may have documented a shift in their migratory movements due to changing ice conditions or a response to years of harvest in Greenland (Stenhouse et al. 2004; Gilg et al. 2012).

Sea ice formation begins around October (when the majority of ivory gulls initiated fall migration) and extends slowly south through Davis Strait until it has reached its full extent in March (Maslanik and Stroeve 1999). The formation of sea ice in the fall may be a contributing factor to the variety of migratory routes that ivory gulls take, selecting those that offer the best foraging opportunities. Gehrold et al. (2014) showed that migrating European gadwalls (Ana strepera) exhibit individual variation in migratory movements, suggesting that it is an advantageous strategy to exploit feeding grounds.

Secondly, sea ice extent may explain, in part, why fall migration was four times longer and rates of travel approximately 20% slower than spring migration. It is possible that individuals made many stopovers in the fall (to forage), moving quickly between the stops and rarely extending east of the ice edge. In contrast, spring migration was very short and, in comparison with fall migration, rates of travel very high. Ivory gulls moved west of the receding sea ice in May of each year of the study, back to the breeding area where ice was still heavily concentrated. This association with sea ice formation and recession during migratory periods has been noted in another Arctic species, the

Bewick’s swan (Cygnus columbianus bewickii), and is probably correlated with food availability (Nuijten et al. 2014). The ivory gull’s migratory behaviour may correspond with sea ice in a similar manner.

How are ivory gulls able to survive or even exploit extensive ice cover during migration? First, the combination of wind and water currents keep polynyas and shore leads of varying sizes open during the winter, while this effect also causes recurring polynyas and shore leads to open in March and April (Stirling 1997). Thus, gulls likely seek out these sites as they move, possibly also queuing in on polar bears hunting near these sites (Bateson and Plowright 1959). Secondly, primary productivity and associated foraging “hotspots” are intricately tied to the links between landfast ice, ice edges, epontic algae growth, snow cover on the ice, and solar irradiance (Bradstreet and Cross

1982; Stirling 1997; Mundy et al. 2009). These factors collectively lead to patched of prey richness and foraging opportunities in an apparent endless sea of ice, and ivory gulls may exploit these locations (like many other Arctic marine birds; McLaren 1982). In the case of ivory gulls, these productive areas, such as Lancaster Sound, may dictate harp seal (Pusa hispida) abundance (Johnston et al. 2005), which in turn influences polar bear foraging opportunities (Stirling and Parkinson 2006), that provide scavenging opportunities for migrating gulls (McLaren 1982). In fact, Inuit local ecological knowledge has indicated that ivory gulls nesting on northwestern Baffin Island stopped annually during late migration to scavenge on Inuit-killed marine mammal carcasses left on the sea ice near Arctic Bay, Nunavut (Mallory et al. 2003). Alternatively, many Arctic nesting migrants use residual body stores from the wintering sites or stopovers to survive until the ice has broken up (Klaassen 2003) and although this may not be their sole strategy, it may have an effect on their reproductive success (Sénéchal et al. 2011).

Breeding and non-breeding seabirds

For a long-lived seabird, exhibiting breeding site fidelity may depend on age, costs associated with changing nest sites, reproductive success in the previous year, probability of adult mortality and the individual’s knowledge of other colonies (Switzer

1993; Blackmer et al. 2004; Naves et al. 2006). Despite the lack of reproductive success in 2010 and the disturbance caused by tagging the ivory gulls, two individuals returned to the Seymour Island colony in 2011. Conover and Miller (1979) suggested that when there are few suitable habitats located in the vicinity for nesting, ring-billed gulls (Larus delawarensis) should remain at their current nesting site, regardless of previous predation history. Neighbours at the Seymour Island colony in 2010 may have been successful; a cue to the disturbed individuals to return the following year (Switzer 1993). However, in

2011, Seymour Island was probably disturbed by predators (which happens frequently e.g. Mallory et al. 2008). One individual abandoned after 11 days at the colony and the other did not continue transmitting. The two individuals that switched nest sites in 2011 probably chose higher quality habitats for breeding, located on Grinnell Peninsula of northwest Devon Island; previously undiscovered nesting sites. A 2011 survey found these colonies supported 69 birds, representing >8% of the estimated Canadian population in an area previously unknown as an ivory gull nesting location (Robertson et al. 2007; M. Mallory, pers. comm.). This finding suggests that current population numbers may be underestimated and that there may be other undiscovered colonies.

Ivory gulls, similar to ring-billed gulls, likely arrive at the breeding area in the spring to predict the current situation at the breeding colony and may use previous outcomes as well as experience to dictate whether they will breed or not (Conover and Miller 1979).

In my study however, the majority of ivory gulls did not breed in 2011 or any of the subsequent years.

It is common for ivory gulls to breed intermittently; however, reasons for this intermittent behaviour are suggested to vary from physical and anthropogenic effects to colony disruption by predators (Gilchrist and Mallory 2005; COSEWIC 2006; Mallory et al. 2012). Attachment of satellite transmitters and handling were suggested to be the cause of abandonment and lack of reproductive success in the following year for northern fulmars (Fulmarus glacialis) and Leach’s storm petrel (Oceanodroma leurochroa), both known to exhibit high site fidelity (Blackmer et al. 2004; Mallory and Gilbert 2008).

Other studies have found no effects on reproductive success, however and the method of leg-loop harness attachment to be effective when fit properly, allowing for changes in body mass (Phillips et al. 2003; Naef-Daenzer 2007; Mallory and Gilbert 2008). From the current data, three of five ivory gulls did not breed in 2011, although all but one individual at least made an attempt to breed that year. The exception did not attempt again during the period of study, potentially because of some negative affect of the transmitter. As mentioned above, the colonies could have been disrupted by mammalian predators.

Rate of travel

My study found that ivory gulls traveled an average of 10 km/hr at any time of the year. This rate is similar to those of European populations of the ivory gull (between ~1-

11 km/hr; Gilg et al. 2010), but there were some differences between studies in the calculated distances traveled. Discrepancies were likely due to the differences in calculations and underestimates of rates of travel from polar nights (Gilg et al. 2010).

Rates found in the literature may depend on the type of calculation used, or how and when the data were collected. Instantaneous speeds from a radar gun taken on board a ship in spring may be quite different from calculating a speed that includes wind vectors from geolocator data. For example, Gilg et al. (2010) reported an average distance traveled for July to December for Greenland ivory gulls of > 50 000 km based on the ivory gulls’ rate of travel from that period. This is a much larger estimate than the ones calculated for yearly travel of the Canadian population of ivory gulls (Table 2.6) but likely differs because the calculations were based on extrapolations for individual rates of travel during a smaller time period.

The study period of Gilg et al. (2010) covered July to December and therefore I could not compare rates of travel at the overwintering sites. However, ivory gulls in my study moved more than expected, averaging 10 – 12 km/h and exhibited the highest weekly rates of travel during the winter (717 ± 56 km/wk). Although there are few studies that have examined winter energetics, the little auk (Alle alle) and Brunnich’s guillemot (Uria lomvia) have a large increase in energetic requirements after December, and these requirements remain high throughout the winter in the northwest Atlantic (Fort et al. 2009). Ivory gulls likely may have similar energy requirements to compensate for the environmental challenges (i.e. storms, increased thermoregulatory needs) occurring in winter (Fort et al. 2009). Although winter, spring and fall migratory hourly rates of travel did not vary significantly from each other, ivory gulls had the highest rates of movement by week in the wintering period. Thus, ivory gulls did not lead a sedentary lifestyle in winter; rather, they appeared to be searching almost constantly for foraging opportunities, perhaps to offset energetic needs.

Comparisons among other seabirds

The ivory gull is unique in the seabird world being the only species, other than the black guillemot (Cepphus grylle), to remain north of 70º N in the winter (Gaston 2004).

In comparison with other high Arctic nesting seabirds, they migrate short distances and do not appear to have migratory staging areas, unless they were used during the blackout periods of the study (González-Solís et al. 2011; Stenhouse et al. 2012). In contrast,

Arctic breeding gulls that are long distance migrants, such as the black-legged kittiwake

(Rissa tridactyla) and Sabine’s gull (Xema sabini), use staging areas pre- and post- breeding (Day et al. 2001; Hatch et al. 2009). Long distance migrants should be less selective for favourable weather, because the costs of slowing migration progress and rates of travel and may explain why the black-legged kittiwake and Sabine’s gull travel approximately 50 km/hr more than the ivory gull during migration (Åkesson and

Hedenström 2007; Klaassen et al. 2011; Table 2.7). Or it may be simply that staging areas for molting may be unnecessary for ivory gulls as the waters remain open in the breeding area until October, allowing continued foraging opportunities as molting resumes (Howell 2001). Gilg et al. (2010) noted that some ivory gulls of European populations also remained close to their breeding colony until the end of September. The availability of prey near the breeding colony well into the fall may explain the varied dates of departure of the Canadian ivory gull.

The ivory gull’s relationship with sea ice is apparent throughout the fall migratory period, probably altering their movements according to sea ice formation as shown in this study. This close association with dense pack ice (on average 50% concentration) and ice edges was clear in all seasons and years for all individuals; a well-documented

relationship that is not known in other seabirds (Orr and Parsons 1982; Renaud and

McLaren 1982; Falk and Møller 1995), but is seen in its analog in the Antarctic, the snow petrel (Pagofroma nivea; Ainley et al. 1984). Unlike other gulls which generally use land or coasts for migration (Gaston 2004), ivory gulls tended to avoid land. In one exceptional case, I noted that one gull covered almost 500 km within 6 h across Baffin

Island to access the next ice covered destination. This affinity for moving over ice may be part of a strategy to avoid land-based predators (similar to their breeding strategy;

Mallory et al. 2008) or to maximize their chances of observing polar bears and scavenging kill.

Table 2.7: Comparison of seabirds’ rate of travel (km/h).

Species Rate of travel Source (km/h) Long-tailed skua (Stercorarius longicaudus) 14 Sittler et al. 2011 Brünnich's guillemot (Uria lomvia) 72-74 Elliott and Gaston 2005 Northern fulmars (Fulmarus glacialis) 33-38 Elliot and Gaston, 2005 ~ 47 Mallory et al. 2012a Ross's gulls (Rhodostethia rosea) 44- 53 Hedenström, 1998 Atlantic puffin (Fratercula arctica) 69 Pennycuick 1987 77-82 Lowther et al. 2002 Common murre (Uria aalge) 58 Pennycuick, 1987 Razorbill (Alca torda) 54 Pennycuick, 1987 Great skua (Catharacta skua) 48 Pennycuick, 1987 Parasitic Jaeger (Stercorarius parasiticus) 69 Pennycuick, 1987 Black-legged kittiwake (Rissa tridactlya) 47 Pennycuick, 1987 ~60 Hatch et al. 2009 Great-black backed gull (Larus marinus) 45 Pennycuick, 1987 Herring gull (Larus argentatus) 41 Pennycuick, 1987 Northern gannet (Morus bassanus) 54 Pennycuick, 1987 Common Shag (Phalacrocorax aristotelis) 55 Pennycuick, 1987 Ivory gull (Pagophila eburnea) 1-11 Gilg et al., 2010 Red Phalarope (Phalaropus fulicarius) 32-50 Tracey et al. 2002 Black guillemot (Cepphus grylle) 58 Butler and Buckley 2002 Sabines’s gull (Xema sabini) 50 Day et al., 2001 Thick-billed murre (Uria lomvia) 65-75 Gaston and Hipfner 2000

References

Ainley D, O’Connor E, Boekelheide R (1984) The marine ecology of birds in the Ross Sea, Antarctica. Ornithological Monographs 32:1-97 + iii-x.

Åkesson S, Hedenström A (2007) How migrants get there: migratory performance and orientation. Bioscience 57:123–134.

Bateson P, Plowright R (1959) The breeding biology of the Ivory Gull in Spitsbergen. British Birds 52:105–114.

Blackmer AL, Ackerman JT, Nevitt GA (2004) Effects of investigator disturbance on hatching success and nest-site fidelity in a long-lived seabird, Leach’s storm-petrel. Biological Conservation 116:141–148.

Boyd I, Wanless S, Camphuysen C (2006) Introduction. In: Boyd I, Wanless S, Camphuysen CJ, editors. Top Predators in Marine Ecosystems: Their Role in Monitoring and Management. Cambridge University Press, pp 1–10.

Bradstreet MSW, Cross WE (1982) Trophic relationships at high Arctic ice edges. Arctic 35:1–12.

Burger AE, Shaffer SA (2008) Application of tracking and data-logging technology in research and conservation of seabirds. Auk 125:253–264.

Butler RG, Buckley DE (2002) Black guillemot (Cepphus grylle). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/675. Accessed 2014 Apr 25.

CLS (2014) Argos user’s manual. CLS Argos, Toulouse, France. 68 p.

Conover MR, Miller DE (1979) Reaction of ring-billed gulls to predators and human disturbances at their breeding colonies. Proceedings of the Colonial Waterbird Group 2:41–47.

COSEWIC (2006) Assessment and update status report ivory gull Pagophila eburnea. 42 + vi p.

Day RH, Stenhouse IJ, Gilchrist GH (2001) Sabine’s gull (Xema sabini). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/593. Accessed 2014 Apr 25.

Elliott KH, Gaston AJ (2005) Flight speeds of two seabirds: a test of Norberg’s hypothesis. Ibis 147:783–789.

Environment Canada (2014) Recovery strategy for the ivory gull (Pagophila eburnea) in Canada. Species at Risk Act Recovery Strategy Series 26 + iv p.

Falk K, Møller S (1995) Satellite tracking of high-arctic northern fulmars. Polar Biology 15:495–502.

Fort J, Porter WP, Grémillet D (2009) Thermodynamic modelling predicts energetic bottleneck for seabirds wintering in the northwest Atlantic. Journal of Experimental Biology 212:2483–90.

Furness RW, Crane JE, Bearhop S, Garthe S, Kakela A, Kakela R, Kelly A, Kubetzki U, Botier S, Waldron S (2006) Techniques to link individual migration patterns of seabirds with diet specialization, condition and breeding performance. ARDEA 94:631–638.

Gaston AJ (2004) Seabirds: A Natural History. Yale University Press. 222p.

Gaston AJ, Hipfner M (2000) Thick-billed Murre (Uria lomvia). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/497. Accessed 2014 Apr 25.

Gaston AJ, Mallory ML, Gilchrist G (2012) Populations and trends of Canadian Arctic seabirds. Polar Biology 35:1221–1232.

Gehrold A, Bauer H, Fiedler W, Wikelski M (2014) Great flexibility in autumn movement patterns of European gadwalls Anas strepera. Journal of Avian Biology 45:131–139.

Gilchrist G, Mallory ML (2005) Declines in abundance and distribution of the ivory gull (Pagophila eburnea) in Arctic Canada. Biological Conservation 121:303–309.

Gilg O, Boertmann D, Merkel F, Aebischer A, Sabard B (2009) Status of the endangered ivory gull, Pagophila eburnea, in Greenland. Polar Biology 32:1275–1286.

González-Solís J, Smyrli M, Militão T, Grémillet D, Tveraa T, Phillips RA, Boulinier T (2011) Combining stable isotope analyses and geolocation to reveal kittiwake migration. Marine Ecological Progress Series 435:251–261.

Hatch SA, Gill VA, Mulcahy DM (2010) Individual and colony-specific wintering areas of Pacific northern fulmars (Fulmarus glacialis). Canadian Journal of Fisheries and Aquatic Science 67:386–400.

Hatch SA, Robertson GJ, Baird PH (2009) Black-legged kittiwake. In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/092. Accessed 2014 Apr 25.

Hedenström A (1998) Flight speed of ross’s gull Rhodostethia rosea and Sabine’s gull Larus sabini. Arctic 51: 283-285

Holm S (1979) A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics 6:65–70.

Howell SN (2001) Molt of the ivory gull. Waterbirds 24:438–442.

Karnovsky N, Hobson K, Brown ZW, Hunt GJ (2009) Distribution and diet of ivory gulls (Pagophila eburnea) in the North Water Polynya. Arctic 62:65–74.

Klaassen RHG, Ens BJ, Shamoun-Baranes J, Exo K-M, Bairlein F (2011) Migration strategy of a flight generalist, the Lesser Black-backed Gull Larus fuscus. Behavioral Ecology 23:58–68.

Latour P, Leger J, Hines J, Mallory ML, Mulders D, Gilchrist G, Smith P, Dickson D (2008) NU Site 4- Seymour Island. In: Key Migratory Bird Terrestrial Habitat Sites in Northwest Territories Nunavut, 3rd ed. Canadian Wildlife Service, Yellowknife. Habitat Management Section Technical Report No. 114. 120 p.

Lindsay R, Zhang J (2005) The thinning of Arctic sea ice, 1988-2003: have we passed a tipping point? Journal of Climate 18:4879–4894.

Lowther, PE, Diamond AW, Kress SW, Robertson, GJ, Russell K (2002) Atlantic Puffin (Fractercula arctica). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/709. Accessed 2014 Apr 25.

Mallory ML, Allard KA, Braune BM, Gilchrist G, Thomas VG (2012b) New longevity record for ivory gulls (Pagophila eburnea) and evidence of natal philopatry. Arctic 65:98–101.

Mallory ML, Akearok J, Gaston A (2009) Status of high Arctic black-legged kittiwake (Rissa tridactyla) colonies in Barrow Strait, Nunavut, Canada. Arctic 62:96–101.

Mallory ML, Gilbert CD (2008) Leg-loop harness design for attaching external transmitters to seabirds. Marine Ornithology 36:183–188.

Mallory ML, Gilchrist G (2003) Marine birds breeding in Penny Strait and Queens Channel, Nunavut, Canada. Polar Research 22:399–403.

Mallory M, Gilchrist G, Fontaine A, Akearok J (2003) Local ecological knowledge of ivory gull declines in Arctic Canada. Arctic 56:293-298.

Mallory ML, Hatch SA, Nettleship DN (2012a) Northern Fulmar (Fulmarus glacialis). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/361. Accessed 2014 Apr 25.

Maslanik J, Stroeve J (1999) Updated daily. Near-Real-Time DMSP SSM/I Daily Polar Gridded Sea Ice Concentrations, [Jul 2010- Jul 2013]. Boulder, Colorado USA: National Snow and Ice Data Center. Digital media.

McLaren PL (1982) Spring migration and habitat use by seabirds in eastern Lancaster Sound and western Baffin Bay. Arctic 35:88–111.

Mundy C, Gosselin M, Ehn JK, Gratton Y, Rossnagel A, Barber DG, Martin J, Tremblay J-É, Palmer M, Arrigo KR, Darnis G, Fortier L, Else B, Papakyriakou T (2009) Contribution of under-ice primary production to an ice-edge upwelling phytoplankton bloom in the Canadian Beaufort Sea. Geophysical Research Letters 36:L17601.

Naef-Daenzer B (2007) An allometric function to fit leg-loop harnesses to terrestrial birds. Journal of Avian Biology 38:404–407.

Naves L, Monnat JY, Cam E (2006) Breeding performance, mate fidelity, and nest site fidelity in a long-lived seabird: behaving against the current? Oikos 115:263–276.

Nicholls D, Robertson C (2007) Validating locations from CLS: Argos satellite telemetry. Notornis 54:121–136.

Oldén B, Peterz M (1985) A statistical method for determination of flight speeds of migrating birds. Journal of Field Ornithology 56:17–22.

Orr C, Parsons J (1982) Ivory gulls, Pagophila eburnea, and ice edges in Davis Strait and the Labrador Sea. Canadian Field-Naturalist 96:323–328.

Pennycuick CJ (1987) Flight of auks (Alcidae) and other northern seabirds compared with southern Procellariiformes: ornithodolite observations. Journal of Experimental Biology 128:335–347.

Phillips RA, Xavier JC, Croxall JP (2003) Effects of satellite transmitters on albatrosses and petrels. Auk 120:1082–1090.

R Core Team (2012) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna Austria. Available from: http://www.R-project.org/.

Renaud WE, McLaren PL (1982) Ivory gull (Pagophila eburnea) distribution in late summer and autumn in eastern Lancaster Sound and western Baffin Bay. Arctic 35:141–148.

Roberts JJ, Best BD, Dunn DC, Treml EA, Halpin PN (2010) Marine Geospatial Ecology Tools: An integrated framework for ecological geoprocessing with ArcGIS, Python, R, MATLAB, and C++. Environmental Modelling & Softwware 25:1197–1207.

Robertson GJ, Gilchrist G, Mallory ML (2007) Colony dynamics and persistence of ivory gull breeding in Canada. Avian Conservation Ecology 2:8.

Salyer JW (1962) A bow-net trap for ducks. Journal of Wildlife Management 26:219– 221.

Sénéchal É, Bêty J, Gilchrist G, Hobson K, Jamieson S (2011) Do purely capital layers exist among flying birds? Evidence of exogenous contribution to arctic-nesting common eider eggs. Oecologia 165:593-604.

Sittler B, Aebischer A, Gilg O (2011) Post-breeding migration of four long-tailed skuas (Stercorarius longicaudus) from north and east Greenland to west Africa. Journal of Ornithology 152:375–381.

Stenhouse IJ, Egevang C, Phillips RA (2012) Trans-equatorial migration, staging sites and wintering area of Sabine’s gulls Larus sabini in the Atlantic Ocean. Ibis 154:42– 51.

Stenhouse IJ, Robertson GJ, Gilchrist G (2004) Recoveries and survival rate of ivory gulls banded in Nunavut, Canada, 1971-1999. Waterbirds: International Journal of Waterbird Biology 27:486–492.

Stirling I (1997) The importance of polynyas, ice edges, and leads to marine mammals and birds. Journal of Marine Systems 10:9–21.

Stirling I, Parkinson CL (2006) Possible effects of climate warming on selected populations of polar bears (Ursus maritimus) in the Canadian Arctic. Arctic 59:261– 275.

Switzer P (1993) Site fidelity in predictable and unpredictable habitats. Evolutionary Ecology 7:533–555.

Thomas D, Fogg G, Convey P, Fritsen C, Gili J-M, Gradinger R, Laybourn-Parry J, Reid K, Walton D (2008) The Biology of Polar Regions. 394 p.

Thomas VG, MacDonald SD (1987) The breeding distribution and current population status of the ivory gull in Canada. Arctic 40:211–218.

Tracey DM, Schamel D, Dale J (2002) Red phalarope (Phalaropus fulicarius). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/698. Accessed 2014 Apr 25.

Chapter three

Critical Winter Habitat of the Endangered Ivory Gull (Pagophila eburnea) in the Canadian Arctic

Introduction

Arctic sea ice

The Arctic marine environment is a dynamic habitat for marine birds, with sea ice covering ~ 15 million km2 at its maximum extent annually in March and ~ 6 million km2 at its minimum in September (1981-2010 average; National Snow and Ice Data Center,

2014). This area forms the foraging habitat for ~ 10 000 000 seabirds in Arctic Canada

(Mallory and Fontaine 2004) and the spatial and temporal extent of foraging opportunities changes through the year as sea ice recedes and forms (Gaston et al. 2005).

In addition to annual variation due to seasonality, sea ice extent has decreased in the past

20 years at a rate of 3% per decade presumably because of global warming, with current climate models showing that by mid-century most Arctic regions will be ice free (Heide-

Jørgensen and Laidre 2004; Stephenson et al. 2011). This long-term, directional change in ice thickness and extent is a significant concern for the sustainability of pagophilic organisms like polar bear (Ursus maritimus; Stirling and Parkinson 2006), ringed seal

(Pusa hispida; Laidre et al. 2008) and the ivory gull (Pagophila eburnea; Gilchrist and

Mallory 2005).

Migratory movements are influenced by natural variation in seasons; however the warming trend has the potential to alter the timing of these cues for seabirds (Marshall

1949; Newton 2003). Earlier ice break up in the Arctic can cause an early pulse of productivity and may benefit some migrating seabirds, as is occurring in Hudson Bay

(Mallory et al. 2010). However, later break up causes a delay in prey availability, lowering breeding success and potentially affecting chick growth (Gaston et al. 2005;

Mallory et al. 2010). Surface feeders and small seabirds are particularly affected by prey

abundance and distribution as they typically have a smaller, more two-dimensional foraging range and less flexible activity budgets, respectively (Camphuysen et al. 2006;

Hamer et al. 2006). This means that when abundance of a typical prey species is reduced, these seabirds may have to increase foraging times or forage for alternate prey, as has occurred with great skuas (Catharacta skua) in Shetland during the late 1980s (Hamer et al. 1991; Hamer et al. 2006). This often has deleterious consequences on subsequent breeding success (e.g. northern fulmars Fulmarus glacialis; Gaston et al. 2005). Food webs are complex and poorly understood and therefore mechanisms influencing changes in diet or foraging effort may not always be obvious (Daunt et al. 2006; Ainley and

Hyrenbach 2010). What is known is that the distributions of polar species are expected to shift northwards as climate change continues, potentially introducing new competitors for

Arctic seabirds and simultaneously adjusting the spatio-temporal predictability of their wintering habitat (Gilg et al. 2012). This includes one of the most poorly known seabirds in the world, the endangered ivory gull.

The ivory gull

The ivory gull is a small all-white colonial nesting gull that spends its entire year in the Canadian Arctic. The species is also found in Svalbard, Greenland and Russia

(Gilg et al. 2010). It is unique among seabirds being the only species, other than the black guillemot (Cepphus grylle), where some individuals remain north of 70º in the winter (Gaston 2004a). It has a small distribution throughout the year compared to other

Arctic breeding seabirds, such as the black-legged kittiwake (Rissa tridactyla) or

Sabine’s gull (Xema sabini), breeding in the high Arctic and migrating to Davis Strait and

Labrador Sea (Day et al. 2001; COSEWIC 2006; Hatch et al. 2009; Chapter 2). There are

49 known extant nesting colonies in Canada distributed on Seymour, Ellesmere, Devon,

Cornwallis islands, as well as the Brodeur Peninsula of Baffin Island. Ivory gulls are known for their affinity to ice year-round (Robertson et al. 2007; Chapter 2). During the winter months, observational work has shown that they are often located near hooded seal

(Crystophora cristata) whelping patches or following polar bears, likely to exploit foraging opportunities (Orr and Parsons 1982; COSEWIC 2006) . They are rarely seen over open water > 5 km away from the ice edge (Mallory et al. 2008). Beyond this, little is known of the ivory gull’s behaviour and distribution during the winter (Environment

Canada 2014). Year round the ivory gull`s habitat is extremely remote and logistically difficult to access, but through satellite telemetry, ivory gulls can be monitored over long periods of time to determine and protect winter distribution and critical habitat (Gaston et al. 2009).

Protected areas

Only in the last 10 years have protected areas moved from terrestrial environments to marine offshore areas (Garthe et al. 2012). One technique for identifying important marine habitats is to determine habitat use by large predators, such as seabirds, because they feed among many trophic levels including the top of marine food webs (Boyd et al 2006). As a consequence, the key areas in which they feed

(“hotspots”) are often areas of high productivity and important for a suite of organisms

(Falk and Møller 1995) and thus protecting key marine sites or seabirds will also help protect lower trophic levels that they use (as reviewed in Hooker and Gerber 2004).

However, it is difficult to clearly define and protect important seabird habitats because birds are highly mobile (Hyrenbach et al. 2000; Camphuysen et al. 2012; Lascelles et al.

2012). Because of this, marine protected areas may need to be larger than terrestrial ones

(Hooker and Gerber 2004).

Although there is no clear best practice to help select and design a protected area, one commonly used method is kernel density estimation, (KDE; Amstrup et al. 2004;

Hooker and Gerber 2004; Lascelles et al. 2012; O’Brien et al. 2012). This can be done by defining a home range of an animal (the area in which it spends most of its time), estimated by how often the animal uses that area (Worton 1995; Kie et al. 2010). This is known as the utilization distribution (UD). The KDE is an estimate of the UD, using a kernel (to estimate the probability density) and a smoothing parameter that influences the overall size of the density estimate.

Objectives and predictions

The ivory gull has experienced a rapid population decline in the Canadian Arctic

(Gaston et al. 2012), concurrent with changing sea ice conditions. Current estimates of population numbers in Canada are < 1000 breeding pairs, suggesting a > 70 % decline in

30 years (Gilchrist and Mallory 2005). The national Recovery Strategy for this species identifies several threats for the ivory gull, including climate change and industrial development and it also lists as a priority using satellite telemetry to assess critical habitat away from colonies (Environment Canada 2014). With threats of climate change creating new possibilities in the near future for resource exploitation due to melting sea ice, the likelihood of increased stress on wintering birds, in addition to that presumably underway due to changing ice conditions, makes identification and conservation of key habitats of paramount importance for this endangered species (Stephenson et al. 2011). In this chapter, I assess winter distribution and recommend sites for consideration as critical

winter habitat for the ivory gull using KDEs. I anticipated that two hotspots would be identified based on historical evidence from Orr and Parsons (1982): one in Davis Strait and one in Labrador Sea. I also expected that ivory gulls would use dense pack ice as their primary habitat (Mallory et al. 2008).

Methods

Study area

Seymour Island (78.80° N, 101.27° W), located in Nunavut, is a Migratory Bird

Sanctuary established in 1975 (Fig 3.1). This low lying rocky island is approximately

3km long and an example of a nesting location that is subject to predation by Arctic fox

(Vulpes lagopus) and polar bears (Mallory et al. 2012). Seymour Island has supported over 300 ivory gulls in the past; however, despite population declines of approximately

This study site was selected because ivory gulls are present in most years and is one of the few locations where ivory gulls can be captured easily on flat ground (Robertson et al.

2007).

The geographic study range extended from the bird collection area, Seymour

Island, to Davis Strait and Labrador Sea in the north Atlantic Ocean, based on previous banding work and at-sea surveys (Orr and Parsons 1982; Stenhouse et al. 2004).

Figure 3.1: Bird capture and tagging of 12 ivory gulls took place on Seymour Island,

NU; a small low-lying island (3 km long), supporting the largest breeding colony for ivory gulls in Canada and a Migratory Bird Sanctuary in Canada.

Bird capture and tagging (2010)

Twelve ivory gulls were captured using a modified version of a bownet trap

North Star Technologies (King Georges, Virginia). The remaining seven individuals were tagged with 15g solar powered PTTs (a customized PTT-100 12g model in a larger case to fit a larger solar chip) by Microwave Telemetry, Inc. (Columbia, Maryland).

Individuals were caught during incubation to ensure they were actively breeding in

Canada and a leg loop harness design was used to attach the transmitters, leaving flight muscles and major fat deposits free (Mallory and Gilbert, 2008). The transmitter plus the harness represented approximately 3% of ivory gull body mass; the preferred load to minimize deleterious effects on individuals (Phillips et al. 2003). Ivory gulls were released and observed for approximately 15 min (unless they flew out of sight) to ensure birds were accepting the harnesses. All transmitter attachments on birds appeared to be successful.

The PTTs were compatible with the Argos satellite positioning system (CLS

2014). The duty cycle of battery-powered PTTs was programmed to send signals within an 8 h period and shut off for 72 h. Solar powered PTTs had 10 h on and 48 h off with customized modifications to the voltage output done by the manufacturer to accommodate the low incident light conditions of the Arctic fall and winter. Each message received from Argos was given an accuracy of the location estimate if four or more messages were sent to the satellite: location class (LC) 0= greater than 1500 m; LC

1= better than 1500 m; LC2= better than 500 m; LC3= better than 250 m. Location classes A and B did not have accuracies (i.e. only two or three messages were received) calculated and LC Z described a failed accuracy calculation.

Messages began transmitting on 1 July 2010. As of July 2013, two of the PTTs continued to provide locations for this project, while eight others lasted between four and

12 months and two others lasted until August and September 2012. A black out period occurred each year for approximately 8-10 weeks between November and January where the solar powered PTTs were not able to transmit data because of a lack of sunlight and consequently, insufficient power to send signals. Nonetheless, 59 439 locations were available for analyses. There was a trade-off between the low number of individual birds tracked and the long periods of tracking (repeat observations) of the same individual birds.

Processing of location data

The statistical program R (R Core Team 2012) and ArcMap 10.1 (ArcMap;

Environmental Systems Research Institute, Redlands, CA) were used for the analysis.

I chose to include the highest quality location data for analyses to be used in a rate of travel filter, classifying which locations were realistic and deleting those that weren’t.

I excluded positions with location classes “A”, “B”, “C”, “Z”, those that were blank as well as any with erroneous dates (before 1 July 2010). One of the sensors on the transmitters detected movement by birds and indicated that birds were still active if the sensor value changed each time a message was received. If the motion sensor numbers remained the same, it indicated that the harness had been discarded or the bird had died.

Any data collected after the point at which motion sensor values stopped changing were excluded from further analyses.

Classifying the winter period

To describe the winter season for the ivory gull, I estimated dates of arrival and departure to and from the wintering grounds (Chapter 1). However, individual rates of travel and distances flown varied within and across years, and therefore defining a single date to begin and end the winter period was not intuitive. Arrival to the wintering area was defined as the date that a bird had obviously slowed down and was no longer making large directional movements. I defined the start of migration periods as the dates when a bird began to fly long distances in an obvious linear pattern, generally away from the breeding or wintering area. Using the range of arrival and departure dates for individual ivory gulls, a median date was given to the beginning of each season to standardize analyses across birds and years, as in Gilg et al. (2013). The median date of arrival to the wintering area was 19 December (range 20 November – 17 January) and the median date of departure was 15 May (range (02 May – 28 May).

A second challenge included the blackout period of 24 h darkness between

November and January when the tags transmitted few signals because they were solar powered. Only two birds arrived at their wintering location before the blackout period in late November (44517 in 2010; 44523 in 2012). One bird arrived at the wintering location in mid-January after the tags began transmitting again. The remaining birds had a data gap of one to three months, arriving at the wintering location at some point within the blackout period. Therefore the median date of arrival at the wintering location was taken to minimize bias for those ivory gulls.

Duplicated data

Duplicated observations sharing location, date and time (making up approximately 10% of the overall data) were deleted from the dataset. Duplicate observations either had differing location classes or latitude and longitude. For those records, I retained the observation with a more accurate location class, or if these were the same, I retained the record with the smaller error radius (i.e. I always kept the most accurate record). No cases arose where all variables were the same.

Filtering unrealistic locations

(Rissa tridactyla; Mallory et al. 2008). Oldén and Peterz (1985) recorded black-legged kittiwake ground speeds of 96km/h, so I used 100km/h as the threshold rate of travel for ivory gulls. This was probably a liberal estimate; however, a lower threshold may have removed satellite messages that were real bird movements, as wind-assisted flights could have resulted in true, high rates of travel. Wind data were not incorporated in this study.

To filter the data, distance was calculated as the orthodromic (great circle route) distance and was calculated in kilometres (km) between consecutive locations for each individual (Gilg et al. 2013). Rate of travel was then calculated from the distance and the time difference between two consecutive locations in km/h. However, as indicated above, wintering at high latitudes resulted in periods where the satellite transmitters did not receive enough sunlight to recharge and transmit (“blackout periods”). This was typically one to three months in length. Using the last point before the blackout period

and the first point after the blackout period led to the calculation of unrealistic time differences and distances (as a linear pattern was assumed). Thus, I zeroed the beginning of each year for these variables, eliminating the inflation of future calculations.

Similarly, since the difference in time between two consecutive points was often greater than 24 h, the number of days (24 h period) between consecutive points was calculated.

If the number of days was greater than one, the distance was divided by the number of days (then divided by 24 h to get km/h) to provide a more realistic representation of how far an ivory gull could travel within 24 h. Rates of travel were adjusted accordingly for the filtration process. The first filter pass showed records with rates greater than 100 km/h still present after re-calculating the distance and time for deleted records. The filter was repeated until there were no longer any speeds greater than 100 km/h.

Analysis

Kernel density estimates

Few guidelines exist as to what the best method of estimating habitat size or home range size. The most widely used method is kernel density estimates (KDEs; Worton

1995; Laver and Kelly 2008; as reviewed in Pebsworth et al. 2012). To implement a

KDE, a smoothing parameter (h) is chosen and is deemed the most influential aspect of estimation (Worton 1995; Kie et al. 2010). Least-squares cross validation (hlscv) has been suggested as the best estimator of bandwidth size (Worton 1989; as reviewed in Gitzen et al. 2006). However this method often fails with datasets comprised of clusters of points and when number of records exceed 1000 locations (as reviewed in Walter et al. 2011).

The reference bandwidth (href) is another estimator of optimal bandwidth size, but this method often over-smooths the data, creating a UD that is too large (Worton 1995).

Therefore, the reference bandwidth can be reduced iteratively just before the polygons begin to fragment (as reviewed in Kie et al. 2010). This subjective method of bandwidth selection may be realistic and appropriate for large datasets with multiple centres of activity (as reviewed in Calenge 2011; Kertson and Marzluff 2011; Vormwald et al. 2011; Pebsworth et al. 2012).

Five groups of KDEs were created to assess the sources of variation in the wintering sites. The first group of KDEs pooled all individuals in each year of the study period (2010, 2011 and 2012) to identify critical wintering habitat across years. The second group pooled individuals across each winter month (excluding May, as individuals begin their migration in this month) and year (2010, 2011 and 2012) to assess variation of wintering habitat between months and years. These KDEs were compared with monthly sea ice extent lines (drawn from the last day of each month’s daily sea ice chart provided by National Snow and Ice Data Center; NSIDC) to provide a coarse estimate of what proportion of the kernels were over ice. The third group assessed variation between individuals, by creating KDEs for each individual that provided at least one or more years of data (44523, 44524, 44525, 44526, 44530 and 44531). The fourth group looked at inter-annual variation within each of the six individuals. That is, a KDE was created for each year and individual transmitted data. The fifth and final group pooled all data from the study period to identify the areas most used by ivory gulls in the winter. The analysis was conducted in program R using the adehabitatHR package.

Ivory gull records between the median arrival date at the wintering area (19 December) and the median departure date from the wintering area (15 May) of all years were used for analysis (n= 3707). This included only location classes 1, 2 and 3. In all cases, the

hlscv method did not converge; therefore KDEs used href estimation for ivory gull UDs, reducing the bandwidth by 10% to the point just before the 50% contour(s) began to fragment. The 95% and 50% contours of the UD were used to determine home range and core areas (km2), respectively (Worton 1995; Vormwald et al. 2011; Ludynia et al. 2012;

Montevecchi et al. 2012; Loring et al. 2014). All means are reported ± standard deviation

(SD).

Historical dataset

A subset of the PIROP (Programme Intégré des Recherches sur les Oiseaux

Pélagiques) dataset, a large database of georeferenced observations of seabirds taken from ships at sea, containing only ivory gull records was provided by Dr. Karel Allard

(Canadian Wildlife Service- Atlantic Region). From these data, observations that were recorded between the median arrival date at the wintering area (19 December) and the median departure date from the wintering area (15 May) were retained. All observations that had zero counts were deleted. Each observation (consisting of a 10 min survey period) recorded a total count of ivory gulls seen within the survey period, therefore each observation was replicated according to the total count (i.e., if the total count for a given observation was 10, it was replicated 10 times) in Statistical program R to represent all birds counted (n= 4905). The observations spanned 41 years (range 1969 - 2010). A KDE was created from the data acquired from ships at sea to compare percent overlap with the

KDE of pooled ivory gull data developed from recent Canadian satellite transmitter data

(Chapter 2) to determine if the historical winter distribution may have changed.

Norwegian satellite data

Raw data from nine satellite-tagged ivory gulls using the same transmitters as my

Canadian data were provided by Dr. Hallvard Strøm (Norwegian Polar Institute,

Norway). Gulls originated from two colonies in Barentsøya, Svalbard (Auga and

Freeman) and I used data from the period which corresponded to those defining the winter period for the Canadian ivory gulls (19 December and 15 May, 2010-2013).

Methods to process and filter the raw data were followed as above for the Canadian satellite transmitter data. Only location classes 1, 2 and 3 were kept for creating a KDE.

The KDE was used to compare percent overlap with the KDE of pooled ivory gull data developed from the recent Canadian satellite transmitter study and the KDE of the PIROP dataset to assess differences in winter habitat between Norwegian and Canadian individuals as well as gulls that could have come from any population. The percent of core area shared between the three data sources was calculated according to Hedd et al.

(2014).

Results

Variation of winter habitat across years

In winter of 2010/2011, 1851 records (n = 8 individuals) made up the pooled

KDE, of which 1681 (91%) of those locations were over ice (Fig. 3.2a). The href bandwidth was reduced by 20%. Home range (95%) and core (50%) areas were 386 801 km2 and 55 304 km2, respectively. In 2011/2012, 1390 records (n = 4) made up the KDE of which 1372 records (99%) were recorded over ice. The href bandwidth was reduced by

40%. The 95% contour area was 444 023 km2 and the 50% area was 73 062 km2 (Fig.

3.2b). Finally, 466 records (n = 2) made up the KDE in winter 2012/2013 and 450 (96%)

of those locations were over ice. The href bandwidth was reduced by 40%. The home range was 145 564 km2 and the core area was 31 455 km2 (Fig. 3.2c). In all years ivory gulls used similar core areas in Davis Strait; however, home range area in 2010/2011 and

2011/2012 included offshore Labrador coast. Consistently across all three years, ivory gulls used an overlapping area of approximately 8050 km2.

Ivory gull winter core areas were typically within 172 km ± 72.3 km of each other

(as measured between centroids) across all years of the study. Distance between ivory gull core areas in 2010/2011 and 2011/2012 was 121 km. Ivory gull core areas in

2010/2011 and 2012/2013 were 255 km apart and core areas in 2011/2012 and 2012/2013 were 141 km apart.

Variation of winter habitat across years and months

Mean core area for pooled individuals across years and months was 29 215 ± 14

179 km2 (range 14 808 km2 – 192 130 km2) and a mean home range size of 138 567 ± 66

269 km2 (range 79 008 km2 – 668 149 km2). Among individual ivory gulls and years, ivory gull core habitats extended farther west over the sea ice than east out over the open water. The mean maximum distance of the core area west of the ice edge in any month

(January, February, March and April) was 128 ± 40.2 km (range 55 km– 363 km; see Fig.

3.3a, b, c) compared with the mean maximum distance that the core area extended east over water (64 ± 38.3 km; range 7 km – 130 km; Table 3.1). Note that in the winter of

2010/2011, one core area of two ivory gulls was found only over water.

2010/2011 (n=8); b) 2011/2012 (n=4); and c) 2012/2013(n=2) in Davis Strait, NU. The white area is comprised of cells with 625 km2 area representing the maximum sea ice extent of the respective year.

Table 3.1 : The maximum distance (km) of 50% kernels (core area) west of the sea ice edge and the maximum distance (km) of core areas east of the ice edge in January,

February, March and April over the study period, 2010-2013. Two distances provided for one month indicate that there were two centres of activity.

Winter Month Max Distance Over Ice (km) Max Distance Over Water (km) 2010/2011 Jan 363 44 Feb 146 44 Feb 129 7 Mar 143 127 Mar* - 120 Apr 150 - 2011/2012 Jan 188 95 Jan 110 33 Feb 140 71 Mar 140 68 Apr 106 - Apr 85 75 2012/2013 Jan 96 - Feb 125 20 Feb 123 - Mar 55 25 Mar 174 74 Apr 194 - *This core area, representing two ivory gulls, was almost exclusively over open water.

Individual variation of wintering habitat

Individually, ivory gulls traveled a great deal throughout the winter, represented by their large home range and core areas. Kernels pooling all years across individual gulls showed a mean home range size of 446 758 ± 244 405 km2 (n=6; range 146 299 km2 – 624 579 km2) and a core area of 87 495 ± 52 533 km2 (range 34 913 km2 – 90 202 km2). Two of the six ivory gulls (44524 and 44530) used the southern tip of Labrador as a secondary wintering area as indicated by their 50% core areas. The 95% home ranges showed that one gull (44531) traveled down to the Labrador Sea. Across years, individuals exhibited high site fidelity to their wintering locations (see Appendix B).

Winter habitat of all ivory gulls from Canadian satellite data

Collectively, ivory gulls equipped with satellite transmitters from all years of the study period had a combined home range size of 521 975 km2 and a core area of 70 352

2 km (href bandwidth reduced by 50%). Bounds of the core area were approximately 60ºN to 64º N and 61º W to 59º W (Fig. 3.4).

Winter habitat of ivory gulls from PIROP dataset

Over a 40 year period, ivory gull observations from the PIROP dataset estimated

2 2 an overall home range of 543 026 km and a core area of 92 743 km (href bandwidth reduced by 20%). The bounds of the two centres of activity were approximately 61º N to

64º N and 59º W to 56º W and 57ºN to 58º N and 60º W to 57º W (Fig. 3.5).

Winter habitat of ivory gulls from Norwegian satellite data

Satellite-tagged ivory gulls from Norwegian colonies used a home range size of 1

2 2 228 741 km and core area of 182 837 km (href decreased by 20%). Bounds of the core area were 58ºN to 64º N and 63º W to 58º W (Fig. 3.6).

Figure 3.4: UD (95%) area and core (50%) area of all Canadian satellite-tagged ivory gulls (n=8) throughout all years of the study (2010-2013), in Davis Strait, NU and

Labrador Sea, NL.

Figure 3.5: UD (95%) area and core (50%) area of wintering ivory gulls sighted from ship-based surveys (PIROP database) over 41 years (1969-2010) in Davis Strait, NU and

Labrador Sea, NL.

Figure 3.6: UD (95%) area and core (50%) area of all Norwegian satellite-tagged ivory gulls (n=9) throughout all years of the study (2010-2013), in Davis Strait, NU, Labrador

Sea, NL and east of Greenland.

Overlap between ivory gulls from Canadian satellite transmitter data, PIROP data and Norwegian data

The core areas of Canadian satellite-tagged ivory gulls, Norwegian satellite- tagged ivory gulls and ivory gulls observed from ship surveys over 41 years showed that all data sources indicated very similar areas used as winter habitat in Davis Strait (Fig.

3.7 and Fig. 3.8). Overlap between the three sources of data was 45 930 km2. The area of overlap that was shared between Norwegian, PIROP and Canadian birds was 18%

(8267 km2). The overlap represented 25% of the core area used by the Norwegian ivory gulls; 65% of the core area used by PIROP ivory gulls and; 50% of the core area used by

Canadian ivory gulls. Approximate bounds of the area were 61ºN to 64º N and 61º W to

57º W (Fig. 3.8).

Figure 3.7: All observations from Canadian satellite-tagged ivory gulls over three years

(2010-2013; black outlined circles); Norwegian satellite-tagged ivory gulls over three years (2010-2013; black circles); and PIROP at-sea observations of ivory gulls (1969-

2010; white circles).

100

Discussion

This is the first detailed summary of ivory gull winter distribution in Canada, with the only previous winter research being an aerial survey conducted in 1978-1979 (Orr and

Parsons 1982). Wintering ivory gulls occupy extremely remote habitats which are difficult to access in winter, a main challenge in attempting to describe their core wintering area, which is a key management priority for the species (Environment Canada

2014). Without advances in satellite telemetry that permit remote tracking of these birds, data collection on winter habitat use would be extremely labour intensive and cost prohibitive. Further, carrying out the analyses of selecting realistic observations and defining migratory periods appropriately required incorporating relevant biological information with scientific techniques for filtering. Selection of a suitable kernel method among literature where few guidelines exist and techniques vary across studies was also a challenge (reviewed in Walter et al. 2011).

Surviving winter in Davis Strait

Collectively, the distributions of the 50% core areas of kernels for both the pooled individuals as well as pooled years showed consistency. Sizes of kernels varied as expected, given that conditions differ each winter and that individual behaviour varies based on availability of foraging resources. Gulls originating from the Seymour Island population clearly favoured Davis Strait as a wintering ground. On an individual scale, however, two ivory gulls used the area off the Labrador coast, which is not surprising as they are known to occur in this area during winter (COSEWIC 2006; Chubbs and Phillips

2007). Both Davis Strait and Labrador Sea likely have consistent sources of food for the ivory gulls, including hooded seals and polar bear kills.

101

Ivory gulls are opportunistic feeders, so the offal of wintering hooded seals could provide a predictable food source if they winter in similar locations (Mallory et al. 2008), in addition to myctophid lanternfishes (Orr and Parsons 1982) . Hooded seal breeding locations in Orr and Parsons (1982) were found mostly near the ice edge, but could be seen as far west of the ice edge as 100km, which may explain why ivory gulls occurred so far from the ice edge in the winter. This is not a surprising finding, as ivory gulls are known to prefer dense pack ice between 70-90% concentrations (Orr and Parsons 1982;

Mosbech and Johnson 1999). One core area for two ivory gulls located off the Labrador coast in March 2010/2011, however, was found over water. I suspect that in this case, the two ivory gulls (44524 and 44530) were using icebergs or ice floes to perch while foraging.

Ivory gulls are often also associated with polar bears (Mallory et al., 2008). Diets of polar bears have the highest proportion of hooded seal numbers in northern Davis

Strait (as reviewed in Stirling and Parkinson 2006), an area commonly exploited by the ivory gull as suggested by the historical PIROP data and Canadian/Norwegian satellite data. Davis Strait polar bears use a total area over sea ice of approximately 420 000 km2, where females use Davis Strait more often than Labrador Sea (Taylor et al. 2001;

Peacock et al. 2013). The 95% home range contours for female polar bears in Taylor et al. (2001) had the greatest amount of overlap with ivory gull ranges, where their 50% core areas appear closer to the coast of Baffin Island and Labrador. The current estimate of bear population numbers in Davis Strait is 2158, an increase from original estimates in the 1970s of 900 (IUCN/PBSG 2014). This increase may reflect the increase in harp

(Phoca groenlandica) and hooded seal numbers (Stirling and Parkinson 2006). In

102

December, hooded seals move south from Baffin Bay into Davis Strait and Labrador shelf which are considered high use areas (Andersen et al. 2009). Given these migratory movements and locations (some only approximate) of hooded seal whelping patches from

Sergeant (1974), Orr and Parson (1982), Bowen et al. (1987), Stenson et al. (1996) and

Frie et al. (2012) as well as the locations of ivory gull core areas in Davis Strait and

Labrador coast (Fig. 3.9), I suspect that ivory gulls are relying on hooded seals (dead pups, placentae, dead adults, perhaps faeces) as a main food source throughout the winter.

By association, ivory gulls also likely follow polar bears to scavenge carcasses of prey that bears kill. Finally, ivory gulls have very high mercury (Hg) levels, suggesting that they feed near the top of the Arctic marine food chain (Braune et al. 2006). The exceptionally high tissue Hg, markedly higher than species at similar trophic levels, might be attributable to their scavenging marine mammal carcasses (e.g., polar bear kills and hooded seals in the winter) year-round, whereas other species that compete for these prey during the breeding season (e.g., glaucous gulls Larus hyperboreus) may feed at lower trophic levels in the winter, such as exploiting dumps and human food sources

(Weiser and Gilchrist 2012).

Area of international significance

Few studies have been conducted on seabird wintering distribution in the

Northwest Atlantic and of those, most only include the northernmost area of Davis Strait

(Orr and Parsons 1982; Mosbech and Johnson 1999). Studies from Greenland note that northern Davis Strait is an internationally significant wintering area for seabirds because of the continuing productivity of the Greenland Open Water polynya throughout most winters (Mosbech and Johnson 1999; Merkel et al. 2002; Boertmann et al. 2004). Gilg et

103

al. (2010) provided evidence that this area is used by the ivory gull, however, the species may not be a frequent visitor to this area except during heavy ice years (Mosbech and

Johnson 1999). My analyses suggest that the southern Davis Strait is also an area of international significance for the ivory gull, given the large overlap between the historical

PIROP data, the Canadian satellite data and Norwegian satellite data. This also suggests that ivory gull winter distribution has not changed greatly since the late 1960s. The

PIROP data had two core areas, one in Davis Strait and the second east of the Labrador coast, while the Canadian and Norwegian satellite transmitter data only had one in Davis

Strait. The lack of gulls from both tracking studies visiting the Labrador Sea likely shows that gulls from Seymour Island and Svalbard prefer Davis Strait. Overall, these results suggest that ivory gulls exhibit site fidelity to this wintering location, across years and individuals. Presumably the winter food availability of this region is a key reason why so many birds are there, and continuing research on the wintering distribution and behaviour of seabirds in the Northwest Atlantic will help us understand how ivory gulls and other seabirds may be affected by anthropogenic disturbances of this region (Gaston et al. 2009; Orben et al. 2012).

Future threats

Although industrial activity is increasing in this region, disturbances such as ships and potential hydrocarbon exploitation in northern Davis Strait / southern Baffin Bay are unlikely risks in the winter for ivory gulls and offshore oil platforms off Newfoundland and Labrador are out of the species’ distribution range (Wiese et al. 2001; Gregersen and

Bidstrup 2008; Arctic Council 2009). However, climate change models show that sea ice is decreasing rapidly and with a ~2-4ºC increase in temperatures predicted by the end of

104

the century, this could have a deleterious effect on the extent and thickness of the sea ice

(Maslanik et al. 2007; Irons et al. 2008; Stephenson et al. 2011). With increased interest in resource exploitation in the Arctic as warming progresses, more opportunities will be available for extended shipping routes and for longer periods of the year (Humphries and

Huettmann 2014). It is predicted that Canada will have almost year-round access to its

Exclusive Economic Zone (EEZ) by mid-century, making resource exploitation and shipping achievable during the winter months (Stephenson et al. 2011). With a reduction in the extent and duration of winter sea ice, the ivory gull will likely have decreased its winter distribution to adjust to the extent of the remaining sea ice and could be negatively influenced by anthropogenic activities.

Critical habitat and marine protected areas

As defined in the Species at Risk Act (SARA; subsection 2(1)), critical habitat is

“the habitat that is necessary for the survival or recovery of a listed wildlife species and that is identified as the species’ critical habitat in the recovery strategy or in an action plan for the species”. My results suggest that ivory gulls spent 90% of their winter over sea ice, where 30% of daily gull positions were over relatively dense ice (≥50% ice concentration). This suggests that the sea ice and perhaps the sea ice-water interface, are key components of their critical winter habitat; most likely within the range of the winter distribution of hooded seal and polar bear in Davis Strait and Labrador Sea. When taken alone, the Canadian satellite data, tracking individuals from one colony (thus only representing a portion of the Canadian breeding population, although it is Canada’s largest colony), is strongly suggestive that Davis Strait is a particularly important wintering area for the species. However, the historical PIROP surveys were unbiased

105

towards any single colony and showed strong spatial overlap with the Seymour Island wintering birds. Furthermore, telemetry data from two Norwegian subpopulations also had high overlap with Seymour Island birds, while evidence from Gilg et al. (2010) on

Greenlandic, Russian and Norwegian ivory gulls also pointed to Davis Strait as being a significant wintering area. Consequently, the combination of data from ivory gulls of

Canadian and international origins all consistently wintering in a similar region provide very strong and convincing evidence to suggest that Davis Strait is critical winter habitat for not just the Canadian but for the global ivory gull population and worthy of protection.

However, assigning geo-referenced bounds for habitat protection for the ivory gull based on sea ice in a highly dynamic environment is challenging. For example, my data showed that individual variation among birds and annual variation within and among birds (Appendix B), can lead to annual variation in core habitats, presumably in response to variable ice conditions. Incorporating multiple sources of data (as in my study) such as tracking studies and ship surveys, can help capture as much of this variation as possible, distinguishing overlap between the sources (Camphuysen et al. 2012). One solution may be to develop adaptive marine protected areas (MPAs), which can use flexible boundaries that are defined by water column properties, species’ distribution, life history, ocean temperature, annual ice conditions and chlorophyll (Hyrenbach et al. 2000; Polovina et al.

2000). For example, predictive modelling using ocean temperature has been used to help reduce bycatch of the southern bluefin tuna (Thunnus maccoyii) in their winter habitat, where boundaries of the protected area are updated regularly to reflect the current oceanographic state (Hobday and Hartmann 2006). A similar approach could also be

106

applied for the ivory gull. Importantly, despite annual and individual variation, across three years of data from at least four individual birds, a key, consistent core area that I recommend be considered for protection for the conservation and recovery of the ivory gull comprises ~ 90 885 km2 with approximate bounds of 61.36ºN to 64.0ºN and

62.34ºW to 56.64ºW, where the current satellite data and PIROP data overlap.

107

Figure 3.9: Approximate historical hooded seal whelping locations (small circles) and bounds (large hollow circles) as described in sergeant (1974), Orr and Parson (1982),

Bowen et al., 1987 and Stenson et al. (1996) and the overlapping area of wintering

Canadian and Norwegian satellite-tagged ivory gulls and wintering gulls sighted from ship-based surveys (PIROP database) in Davis Strait, NU.

108

References

Ainley DG, Hyrenbach KD (2010) Top-down and bottom-up factors affecting seabird population trends in the California current system (1985–2006). Progress in Oceanography 84:242–254.

Amstrup S., Mcdonald TL, Durner GM (2004) Using satellite radiotelemetry data to delineate and manage wildlife populations. Wildlife Society Bulletin 32:661–679.

Andersen JM, Wiersma YF, Stenson GB, Hammill MO, Rosing-Asvid A (2009) Movement patterns of hooded seals (Cystophora cristata) in the Northwest Atlantic Ocean during the post-moult and pre-breed seasons. Journal of Northwest Atlantic Fisheries Science 42:1–11.

Arctic Council (2009) Marine shipping assessment 2009 report. http://www.pame.is/index.php/amsa-2009-report. Accessed 20 Apr 2014.

Boertmann D, Lyngs P, Merkel FR, Mosbech A (2004) The significance of Southwest Greenland as winter quarters for seabirds. Bird Conservation International 14:87– 112.

Bogdanova MI, Daunt F, Newell M, Phillips RA, Harris MP, Wanless S (2011) Seasonal interactions in the black-legged kittiwake, Rissa tridactyla: links between breeding performance and winter distribution. Proceedings of the Royal Society B 278:2412– 8.

Bowen W, Myers R, Hay K (1987) Abundance estimation of a dispersed, dynamic population: hooded seals (Crystophora cristata) in the northwest Atlantic. Canadian Journal of Fisheries Aquatic Sciences 44:282–295.

Calenge C (2011) Home Range Estimation in R : the adehabitatHR Package. 61p.

Camphuysen CJ, Scott B, Wanless S (2006) Distribution and foraging interactions of seabirds and marine mammals in the North Sea: multispecies foraging assemblages and habitat-specific feeding strategies. In: Boyd I, Wanless S, Camphuysen C, editors. Top Predators in Marine Ecosystems: Their Role in Monitoring and Management. Cambridge University Press, pp 82–97.

Camphuysen CJ, Shamoun-Baranes J, Bouten W, Garthe S (2012) Identifying ecologically important marine areas for seabirds using behavioural information in combination with distribution patterns. Biological Conservation 156:22–29.

109

Chubbs T, Phillips F (2007) Winter occurrences of ivory gulls, Pagophila eburnea, in inland Labrador. Canadian Field-Naturalist 121:327–328.

CLS (2014) Argos user’s manual. CLS Argos, Toulouse, France. 68p.

COSEWIC (2006) Assessment and update status report ivory gull Pagophila eburnea. 42 + vip.

Daunt F, Wanless S, Peters G, Benvenuti S, Sharples J, Grémillet D, Scott B (2006) Impacts of oceanography on the foraging dynamics of seabirds in the North Sea. In: Boyd I, Wanless S, Camphuysen C, editors. Top Predators in Marine Ecosystems: Their Role in Monitoring and Management Cambridge University Press, pp 177– 190.

Environment Canada (2014) Recovery strategy for the ivory gull (Pagophila eburnea) in Canada. Species at Risk Act Recovery Strategy Series 26 + iv p.

Frie AK, Stenson GB, Haug T (2012) Long-term trends in reproductive and demographic parameters of female northwest Atlantic hooded seals (Cystophora cristata): population responses to ecosystem change? Canadian Journal of Zoology 90:376– 392.

Garthe S, Markones N, Mendel B, Sonntag N, Krause JC (2012) Protected areas for seabirds in German offshore waters: designation, retrospective consideration and current perspectives. Biological Conservation 156:126–135.

Gaston AJ (2004) Seabirds: a natural history. Yale University Press. 222p.

Gaston AJ, Bertram DF, Boyne AW, Chardine JW, Davoren G, Diamond A., Hedd A, Montevecchi WA, Hipfner JM, Lemon MJF, Mallory ML, Rail J-F, Robertson GJ (2009) Changes in Canadian seabird populations and ecology since 1970 in relation to changes in oceanography and food webs. Environmental Review 17:267–286.

Gaston AJ, Gilchrist H, Mallory M (2005) Variation in ice conditions has strong effects on the breeding of marine birds at Prince Leopold Island, Nunavut. Ecography 3:331–344.

Gaston AJ, Mallory ML, Gilchrist G (2012) Populations and trends of Canadian Arctic seabirds. Polar Biology 35:1221–1232.

Gilchrist G, Mallory ML (2005) Declines in abundance and distribution of the ivory gull (Pagophila eburnea) in Arctic Canada. Biological Conservation 121:303–309.

110

Gilg O, Moe B, Hanssen SA, Schmidt NM, Sittler B, Hansen J, Reneerkens J, Sabard B, Chastel O, Moreau J, Phillips RA, Oudman T, Biersma EM, Fenstad AA, Lang J, Bollache L (2013) Trans-equatorial migration routes, staging sites and wintering areas of a high Arctic avian predator: the long-tailed skua (Stercorarius longicaudus). PLoS ONE 8:1–10.

Gitzen RA, Millspaugh JJ, Kernohan BJ (2006) Bandwidth selection for fixed-kernel analysis of animal distributions. Journal of Wildlife Management 70:1334–1344.

Gregersen U, Bidstrup T (2008) Structures and hydrocarbon prospectivity in the northern Davis Strait area, offshore West Greenland. Petroleum Geoscience 14:151–166.

Hamer KC, Furness RW, Caldow RW. (1991) The effects of changes in food availability on the breeding ecology of great skuas Catharacta skua in Shetland. Journal of Zoology 223:175–188.

Hamer KC, Lewis S, Wanless S, Phillips R, Sherratt T, Humphreys E, Hennicke J, Garther S (2006) Use of gannets to monitor prey availability in the northeast Atlantic Ocean: colony size, diet and foraging behaviour. In: Boyd I, Wanless S, Camphuysen C, editors. Top Predators in Marine Ecosystems: Their Role in Monitoring and Management. Cambridge University Press, pp 236–248.

Hatch SA, Robertson GJ, Baird PH (2009) Black-legged kittiwake Rissa tridactyla. In: Poole A, Birds of North America Online, No. 92. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/092. Accessed 2014 Apr 25.

Hedd A, Montevecchi R, Phillips R, Fifield D (2014) Seasonal sexual segregation by monomorphic sooty shearwaters Puffinus griseus reflects different reproductive roles during the pre-laying period. PLos ONE 9:e85572.

Heide-Jørgensen MP, Laidre KL (2004) Declining extent of open-water refugia for top predators in Baffin Bay and adjacent waters. Ambio 33:488–95.

Hobday A, Hartmann K (2006) Near real-time spatial management based on habitat predictions for a longline bycatch species. Fisheries Management Ecology 13:365– 380.

111

Hooker SK, Gerber LR (2004) Marine reserves as a tool for ecosystem-based management: the potential importance of megafauna. Bioscience 54:27–39.

Humphries GRW, Huettmann F (2014) Putting models to a good use: a rapid assessment of Arctic seabird biodiversity indicates potential conflicts with shipping lanes and human activity. Diversity and Distribution 20:478–490.

Hyrenbach K, Forney K, Dayton P (2000) Marine protected areas and ocean basin management. Aquatic Conservation: Marine Freshwater Ecosystems 10:437–458.

IUCN/PBSG (2014) Summary of polar bear population status per 2013. http://pbsg.npolar.no/en/status/populations/davis-strait.html. Accessed 4 Apr 2014.

Kertson BN, Marzluff JM (2011) Improving studies of resource selection by understanding resource use. Environmental Conservation 38:18–27.

Kie JG, Matthiopoulos J, Fieberg J, Powell R a, Cagnacci F, Mitchell MS, Gaillard J-M, Moorcroft PR (2010) The home-range concept: are traditional estimators still relevant with modern telemetry technology? Philosophical Transactions of the Royal Society B 365:2221–2231.

Laidre KL, Stirling I, Lowry LF, Wiig O, Heide-Jørgensen MP, Ferguson SH (2008) Quantifying the sensitivity of Arctic marine mammals to climate-induced habitat change. Ecological Applications 18:S97–S125.

Lascelles BG, Langham GM, Ronconi RA, Reid JB (2012) From hotspots to site protection: identifying Marine Protected Areas for seabirds around the globe. Biological Conservation 156:5–14.

Latour P, Leger J, Hines J, Mallory ML, Mulders D, Gilchrist G, Smith P, Dickson D (2008) NU Site 4- Seymour Island. Key Migratory Bird Terrestrial Habitat Sites in Northwest Territories, Nunavut, 3rd ed. Canadian Wildlife Service, Yellowknife. Habitat Management Section Technical Report No. 114. 120p.

Laver P, Kelly MJ (2008) A critical review of home range studies. Journal of Wildlife Management 72:290–298.

Loring PH, Paton PWC, Osenkowski JE, Gilliland SG, Savard J-PL, Mcwilliams SR (2014) Habitat use and selection of black scoters in southern New England and siting of offshore wind energy facilities. Journal of Wildlife Management 78:645– 656.

112

Ludynia K, Kemper J, Roux J-P (2012) The Namibian Islands’ Marine Protected Area: using seabird tracking data to define boundaries and assess their adequacy. Biological Conservation 156:136–145.

Mallory ML, Allard KA, Braune BM, Gilchrist G, Thomas VG (2012) New longevity record for ivory gulls (Pagophila eburnea) and evidence of natal philopatry. Arctic 65:98–101.

Mallory ML, Gaston AJ, Gilchrist G, Robertson GJ, Braune BM (2010) Effects of climate change, altered sea-ice distribution and seasonal phenology on marine birds. In: Ferguson SH, Loseto LL, Mallory ML, editors. A Little Less Arctic: Top Predators of the World’s Largest Northern Inland Sea, Hudson Bay. Springer, Dordrecht, pp 179– 195.

Mallory ML, Gilchrist G (2003) Marine birds breeding in Penny Strait and Queens Channel, Nunavut, Canada. Polar Research 22:399–403.

Mallory M, Stenhouse I, Gilchrist H, Robertson G, Haney C, MacDonald S (2008) Ivory gull (Pagophila eburnea). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/175. Accessed 2013 September 05.

Marshall A (1949) Weather factors and spermatogenesis in birds. Proceedings of the Zoological Society of London 119:711–716.

Merkel FR, Mosbech A, Boertmann D, Grøndahl L (2002) Winter seabird distribution and abundance off Greenland, 1999. Polar Research 21:17–36.

Montevecchi WA, Hedd A, McFarlane Tranquilla L, Fifield D, Burke C, Regular P, Davoren G, Garthe S, Robertson GJ, Phillips RA (2012) Tracking seabirds to identify ecologically important and high risk marine areas in the western North Atlantic. Biological Conservation 156:62–71.

Mosbech A, Johnson SR (1999) Late winter distribution and abundance of sea-associated birds in south-western Greenland, the Davis Strait and southern Baffin Bay. Polar Research 18:1–17.

Newton I (2003) Geographical patterns in bird migration. In: Berthold P, Gwinner E, Sonnenschein E, editor. Avian Migration. Springer- Berlin Heidelberg, pp 211–236.

113

O’Brien SH, Webb A, Brewer MJ, Reid JB (2012) Use of kernel density estimation and maximum curvature to set Marine Protected Area boundaries: identifying a Special Protection Area for wintering red-throated divers in the UK. Biological Conservation 156:15–21.

Oldén B, Peterz M (1985) A statistical method for determination of flight speeds of migrating birds. Journal of Field Ornithology 56:17–22.

Orben RA, Irons DB, Shaffer SA, Roby DD (2012) Winter distribution and ecology of black-legged kittiwakes (Rissa tridactyla) and thick-billed murres (Uria lomvia) breeding at two Bering Sea colonies with differing population trends. North Pacific Research Board Project Final Report, 911. 86+ii p.

Orr C, Parsons J (1982) Ivory gulls, Pagophila eburnea, and ice edges in Davis Strait and the Labrador Sea. Canadian Field-Naturalist 96:323–328.

Peacock E, Taylor MK, Laake J, Stirling I (2013) Population ecology of polar bears in Davis Strait, Canada and Greenland. Journal of Wildlife Management 77:463–476.

Pebsworth PA, Morgan HR, Huffman MA (2012) Evaluating home range techniques: use of Global Positioning System (GPS) collar data from chacma baboons. Primates 53:345–55.

Phillips RA, Xavier JC, Croxall JP (2003) Effects of satellite transmitters on albatrosses and petrels. Auk 120:1082–1090.

Polovina J, Kobayashi D, Parker D, Seki M, Balazs G (2000) Turtles on the edge: movement of loggerhead turtles (Caretta caretta) along oceanic fronts, spanning longline fishing grounds in the central North Pacific, 1997–1998. Fisheries Oceanography 9:71–82.

R Core Team (2012) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna Austria. Available from: http://www.R-project.org/.

Robertson GJ, Gilchrist G, Mallory ML (2007) Colony dynamics and persistence of ivory gull breeding in Canada. Avian Conservation and Ecology 2:8.

Salyer JW (1962) A bow-net trap for ducks. Journal of Wildlife Management 26:219– 221.

Sergeant D (1974) A rediscovered whelping population of hooded seals Cystophora cristata Erxleben and its possible relationship to other populations. Polarforschung 44:1–7.

114

Stenhouse IJ, Robertson GJ, Gilchrist G (2004) Recoveries and survival rate of ivory gulls banded in Nunavut , Canada , 1971-1999. Waterbirds: 27:486–492.

Stenson G, Myers R, Warren W, Ni I-H (1996) Pup production of hooded seals (Cystophora cristata) in the northwest Atlantic. NAFO Science Council Studies 26:105–114.

Stephenson SR, Smith LC, Agnew JA (2011) Divergent long-term trajectories of human access to the Arctic. Nature Climate Change 1:156–160.

Stirling I, Parkinson CL (2006) Possible effects of climate warming on selected populations of polar bears (Ursus maritimus) in the Canadian Arctic. Arctic 59:261– 275.

Taylor MK, Akeeagok S, Andriashek D, Barbour W, Born EW, Calvert W, Cluff HD, Ferguson S, Laake J, Rosing-Asvid A, Stirling I, Messier F (2001) Delineating Canadian and Greenland polar bear (Ursus maritimus) populations by cluster analysis of movements. Canadian Journal of Zoology 79:690–709.

Vormwald LM, Morrison ML, Mathewson HA, Cocimano MC, Collier BA (2011) Survival and movements of fledgling willow and dusky flycatchers. Condor 113:834–842.

Walter WD, Fischer JW, Baruch-mordo S, VerCauteren KC (2011) What is the proper method to delineate home range of an animal using today’s advanced GPS telemetry systems: the initial step. In: Krejcar O, editor. Modern Telemetry. InTech, pp 249– 268.

Weiser E, Gilchrist HG (2012) Glaucous gull (Larus hyperboreus). In: Poole A, Birds of North America Online. Ithaca. Available: http://bna.birds.cornell.edu/bna/species/573. Accessed 2014 June 2.

Wiese F, Montevecchi W, Davoren G, Huettmann F, Diamond A, Linke J (2001) Seabirds at risk around offshore oil platforms in the North-west Atlantic. Marine Pollution Bulletin 42:1285–1290.

Worton BJ (1989) Kernel methods for estimating the utilization in home-range studies. Ecology 70:164–168.

Worton BJ (1995) Using Monte Carlo simulation to evaluate kernel-based home range estimators. Journal of Wildlife Management 59:794–800.

115

Chapter four

Future recovery and management of the ivory gull

116

Arctic seabirds live in a dynamic and harsh environment that is changing drastically by climate change. The possibility of extended shipping periods and resource exploitation from receding sea ice in the Arctic could potentially add to the list of threats that the ivory gull currently faces (Gilchrist and Mallory 2005; Stephenson et al. 2011).

Many of these threats can be mitigated, such as illegal shooting and disturbances near the breeding colonies; however, contaminants and climate change will be the dominant threats in the future. The ivory gull is one of the most poorly known seabirds in the world, where distribution, migratory movements and winter habitat have been based on observation or banding recovery (Orr and Parsons 1982; Robertson et al. 2007; Mallory et al. 2008). My study is the first quantitative, multi-year tracking study conducted for the ivory gull in Canada, analyzing annual movements, distribution and critical habitat.

Many results were consistent with predictions based on earlier banding or observational work, while some results markedly changed our perception of ivory gull behaviour and migration. They will contribute to the knowledge required to put recovery measures in place for this species.

A number of key issues relating to the future recovery and management of the ivory gull were addressed in this thesis. Ivory gulls were found to vary greatly in their breeding and migratory behaviour. Their lack of breeding site fidelity may suggest that population estimates are actually higher than what has been suggested from recent studies

(COSEWIC 2006; Robertson et al. 2007); however, Seymour Island is often subject to disruption by predators leaving this colony more vulnerable than others nesting on nunataks or inland plateaus. Second, their rate of travel during migration may help describe where an ivory gull is throughout the year and what factors may influence their

117

behaviour; having implications on management decisions. For example, they migrated slowly in the fall and were not found in Davis Strait until December at the earliest, while they can be expected to fly back to the breeding area quickly in the spring. This information could help direct the possible timing of regulatory restrictions that might be put in place to protect ivory gulls and what areas can be used for industrial activities.

A common factor influencing the initiation of the pre- and post-breeding migratory periods is sea ice. The ivory gull appears to use the ice formation as a cue in the fall to begin their migration, exhibiting a relatively slow migration punctuated by foraging stops along the way. In the spring, ice recession is a cue to begin a relatively rapid migration back to the breeding area. The variable route selection among ivory gulls during fall migration was a surprising finding, as was the avoidance of the Greenland coast. This result may represent a shift in the migratory routes of the ivory gulls from historical hunting pressure along the Greenland coast or may suggest that most ivory gulls did/do not migrate near Greenland (Stenhouse et al. 2004). The lack of variation during spring migration suggests that there are fewer areas to forage in the spring.

A main management goal of the ivory gull Recovery Strategy (Environment

Canada 2014) is to determine if there is other critical winter habitat for the ivory gull and this was the primary undertaking of Chapter 3. The data strongly support the identification of critical habitat for this species in the winter based on my satellite data, similar data for birds arriving from Norway and data from years of at-sea observations in the area. This is the first time the wintering area has been quantified for the Canadian ivory gull and compared with international subpopulations. Results confirmed the regular use of Davis Strait as a winter ground for the ivory gull. Both Norwegian

118

subpopulations and the Seymour Island subpopulation used Labrador Sea less than predicted; however this may be due to individual preference as the area hosts the largest number of hooded seals in Canada (Stenson et al. 1994). Ivory gulls exhibited larger spatial distributions than expected throughout the winter. As discussed in Chapter 3,

Canadian satellite tagged ivory gulls were likely searching constantly for hooded seal whelping patches, and may explain why ivory gulls preferred ≤50% ice concentration in the winter (see Fig. 3.10; Orr and Parsons 1982). However, they were also found further west of the ice edge than expected. Gulls found over highly concentrated ice could be explained by the presence of hooded seals (Orr and Parsons 1982), or they may have been following and scavenging from polar bears, frequenting Davis Strait (Mallory et al. 2008;

Taylor et al. 2001). The importance of sea ice to not only ivory gull winter habitat, but also route selection during migration and likely breeding habitat should be emphasized.

It is a unique characteristic among seabirds, as described in Chapter 2, and appeared to have a large influence on individual behaviour and will be essential to incorporate into a management plan for recovery of the species.

The number of birds that I tracked from one colony of Canadian satellite-tagged ivory gulls was small (n=12) and might not represent all populations, and consequently some may question the credibility of using key wintering areas developed from those birds to represent ivory gull critical winter habitat in Canada. However, I also provided evidence from at-sea surveys of ivory gulls (which could include birds from Canada,

Greenland, Norway or elsewhere) and Norwegian satellite-tagged ivory gulls. These other data sources showed strong agreement that Davis Strait was and is critical winter habitat for the ivory gull. Gilg et al (2010) has also shown that this area is a significant

119

wintering habitat for populations from Greenland, Norway and Russia. Thus, I contend that Davis Strait should be considered an internationally important area for the ivory gull and parts of Davis Strait merit consideration for protection of marine habitat during the winter, given increasing risk of climate change and the inevitable future of expansion of industrial activities in the north.

Establishing a protected area and monitoring the recovery of the ivory gull in

Davis Strait will be difficult for a number of reasons. This includes the challenges of establishing a protected area in such a dynamic environment, choosing bounds for a possible protected area and physical monitoring / censusing birds as part of the recovery of the ivory gull. As described in Chapter 3 there are solutions to creating management areas in dynamic environments. For example, predicting the extent of sea ice in Davis

Strait each year would be a good starting point to define a protected area for the ivory gull, while using the approximate bounds suggested in Chapter 3. The actual area and bounds of the overlapping core area between all three sources of ivory gull data was irregularly shaped, therefore difficult to describe simply for management purposes. The bounds suggested as the proposed area were chosen taking half of the length of the widest part of the overlapping area to be the radius of a circle. I considered this to be the simplest and clearest method to describe and defend the bounds of a possible protected area, but it includes areas that are typically not of high use by the ivory gull and would therefore put unnecessary restrictions on possible industrial activity where ivory gulls are not (Fig. 4.1). However, using lines of latitude and longitude would also create an area too large for protection, due to the orientation of the overlap and the natural curvature of

120

the earth (Fig. 4.2). Further consultation will be required to establish dimensions for a protected area.

My thesis provides a baseline for understanding the annual spatial ecology of the

Canadian ivory gull; however, more research will be imperative to future management and recovery of the species. A winter ecology study will be necessary to evaluate the importance of hooded seals and polar bear kills to the ivory gull, particularly since one study (Orr and Parsons 1982) documented myctophid fish in the diet of wintering birds.

A winter study would also provide a basis for comparison of the proposed management area with actual locations of hooded seals and ivory gulls. Outside of the winter season, other research required to better understand and manage this species includes the development of a habitat suitability model to determine potential sites to target surveys for unknown breeding colonies (to better estimate ivory gull numbers), as well as new surveys of known breeding colonies (the last complete set conducted in 2009 on this endangered species) should be conducted.

121

Figure 4.1: The proposed protected area for the ivory gull selected by taking half the length of the widest part of the overlap to be the radius of a circle. The area is 90 885 km2.

122

123

References

COSEWIC (2006) Assessment and update status report ivory gull Pagophila eburnea. 42 + vi p.

Environment Canada (2014) Recovery strategy for the ivory gull (Pagophila eburnea) in Canada. Species at Risk Act Recovery Strategy Series 26 + iv p.

Gilchrist G, Mallory ML (2005) Declines in abundance and distribution of the ivory gull (Pagophila eburnea) in Arctic Canada. Biological Conservation 121:303–309.

Orr C, Parsons J (1982) Ivory gulls, Pagophila eburnea, and ice edges in Davis Strait and the Labrador Sea. Canadian Field-Naturalist 96:323–328.

Robertson GJ, Gilchrist G, Mallory ML (2007) Colony dynamics and persistence of ivory gull breeding in Canada. Avian Conservation and Ecology 2:8.

Stenhouse IJ, Robertson GJ, Gilchrist G (2004) Recoveries and survival rate of ivory gulls banded in Nunavut, Canada, 1971-1999. Waterbirds: 27:486–492.

Stenson G, Myers R, Warren W, Ni I-H (1996) Pup production of hooded seals (Cystophora cristata) in the northwest Atlantic. NAFO Science Council Studies 26:105–114.

Stephenson SR, Smith LC, Agnew JA (2011) Divergent long-term trajectories of human access to the Arctic. Nature Climate Change 1:156–160.

124

Appendix A

Annual distances traveled by ivory gulls broken down by season

125

Table A1: Sum of distances (km) traveled by 12 satellite-tagged ivory gulls each season throughout each calendar year of the study period (July 2010- July 2013) using the median start dates (Table 2.2). Bolded years indicate a full year of data was collected

(breeding season to breeding season). Brackets indicate the bird transmitted signals for only part of the season and were not included in calculating means and Standard deviations (SD).

Wintering Area to Bird Year Wintering (km) Breeding (km) Breeding Area (km) 44509 2010 - - (607) 2011 (33) - - 44516 2010 - - (407) 44517 2010 (160) - (750) 2011 (1849) - - 44519 2010 (338) - (251) 2011 (1161) - - 44522 2010 - - (192) 44523 2010 - - (934) 2011 8259 1869 9187 2012 7360 2212 6508 2013 7739 2005 8185 44524 2010 - - (1005) 2011 14706 3810 3672 2012 10700 2232 5695 44525 2010 - - (767) 2011 13379 2861 10365 2012 10861 1806 (1396) 44526 2010 - - (2063) 2011 20878 4541 3823 2012 22406 4139 6522 2013 19808 3402 - 44529 2010 - - (1497) 44530 2010 - - (1616) 2011 16355 3702 (200) 44531 2010 - - (1462) 2011 13278 (1113) - ̅ ± SD 13811 ± 5180 2962 ± 994 6745 ± 2400

126

Table A2: Sum of distances (km) traveled by 12 satellite-tagged ivory gulls each season throughout each calendar year of the study period (July 2010- July 2013) using the median start dates (Table 2.2). Bolded years indicate a full year of data was collected

(breeding season to breeding season). Brackets indicate the bird transmitted signals for only part of the season and were not included in calculating means and Standard deviations (SD).

Non-Breeding/ Breeding Area to Total (km) Bird Year Abandoned Breeding Wintering Area (km) (km) 44509 2010 1905 (883) 3395 2011 - - 33 44516 2010 1863 (785) 3055 44517 2010 2482 3912 7304 2011 - - 1849 44519 2010 2553 3411 6553 2011 - - 1161 44522 2010 1682 (420) 2294 44523 2010 4725 3567 9226 2011 - 1115 20430 2012 5186 2733 23999 2013 (610) - 18539 44524 2010 4245 1991 7241 2011 4136 1967 28291 2012 3952 (76) 22655 44525 2010 6695 2074 9536 2011 - 3902 30507 2012 - - 14063 44526 2010 8625 7787 18475 2011 7448 6180 42870 2012 6439 8152 47658 2013 - - 23210 44529 2010 10067 (3510) 15074 44530 2010 6905 2778 11299 2011 - - 20257 44531 2010 6137 1869 9468 2011 - - 14391 ̅ ± SD 5003 ± 2499 3674 ± 2206

127

Appendix B

Individual variation of habitat use among wintering Canadian ivory gulls

128

Individual gulls used similar winter habitats across years, traveling large distances throughout the winter. Two ivory gulls

(44524, Fig. B3 and 44530, Fig. B6) used the Labrador Coast. All others (Figs. B1, 2, 4 and 5) primarily used Davis Strait. Ivory gull

44524 had the greatest change in winter habitat between 2010/2011 and 2011/2012.

Figure B1: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44523); a) over all years of the study

(2010-2013); b) in winter 2010/2011; c) in winter 2011/2012; and d) in winter 2012/2013 in Davis Strait, NU.

129

Figure B2: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44526); a) over all years of the study

(2010-2013); b) in winter 2010/2011; c) in winter 2011/2012; and d) in winter 2012/2013 in Davis Strait, NU.

130

Figure B3: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44524); a) over all years the transmitter sent signals (2010-2012); b) in winter 2010/2011; and c) in winter 2011/2012 in Davis Strait, NU.

131

Figure B4: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gulls (44525); a) over all years the transmitter sent signals (2010-2013); b) in winter 2010/2011; and c) in winter 2011/2012 in Davis Strait, NU.

132

Figure B5: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gull

(44530) over the period the transmitter was sending signals (2010-2011).

Figure B6: UD (95%) area and core (50%) area of a wintering satellite-tagged ivory gull

(44531) over the period the transmitter was sending signals (2010-2011).

133