Behavioural Consistency and Foraging Specialisations in the Australasian Gannet (Morus Serrator)

Behavioural consistency and foraging specialisations in the Australasian gannet (Morus serrator)

Marlenne Adriana Rodríguez Malagón B.Sc. Biology, M.Sc. Ecology

Submitted in fulfilment of the requirements for the degree of Doctor of Philosophy (Life & Env)

Deakin University October 2018

To MM and JMB, for your unconditional love and your faith in me,

thanks to you I fulfilled this dream.

I love you

ABSTRACT

For decades conspecifics were considered as equivalent in ecological studies, but recent science now recognises the presence and importance of inter-individual differences. These differences can be driven by intrinsic factors such as sex, phenotype, and/or personality, and are known to influence individual foraging decisions. The concept of ‘individual foraging specialist’ refers to the use of a specific proportion of the full range of available resources (or foraging strategies) used by a subset of a population and involves the repetition of specific behaviours over time. Furthermore, behavioural consistency and/or individual specialisation can arise within different aspects of a species’ ecological niche. The presence of both phenomena may vary over spatial and temporal scales due to environmental stochasticity, but because these phenomena can have major implications for the ecology of individuals, it is important to identify and quantify the presence individual foraging specialisation and behavioural consistency in animal populations.

Seabirds are major marine predators and traits such as colonial breeding, central-place foraging during the breeding season, and high levels of nest-site fidelity make them good models to investigate behavioural consistency. The Australasian gannet (Morus serrator) is a large pelagic seabird endemic to Australia and New Zealand. It is considered a generalist predator with high levels of foraging plasticity. However, little is known about the factors influencing behavioural consistency in this species. Thus, the objectives of this research project were to determine: 1) the level of inter- and intra-individual variation within foraging behaviour, diet and winter dispersal patterns in Australasian gannets, 2) how these attributes vary over different timescales and habitats, and 3) the factors influencing these same attributes.

Data were collected at two breeding colonies in southeastern Australia, Point Danger and Pope’s Eye, which are subject to contrasting oceanographic conditions and foraging habitats. A combination of bio-logging, stable isotope analysis, and diet analysis was used to v investigate behavioural consistency. Breeding individuals and in most cases, their nest partners, were sampled over two breeding seasons. The present study found that colony, breeding stage, year, and sex were factors influencing the foraging behaviour of Australasian gannets and the isotopic values (δ13C and δ15N) measured in blood plasma of individual adults. Behavioural consistency, measured as the contribution of the individual to the observed variance based on all study birds, was low to moderate for the five foraging metrics tested (maximum distance from the colony, bearing from the colony to the most distal point, tortuosity index, total number of dives, and mean Vectorial Dynamic Body Acceleration [VeDBA]), with greater consistency values occurring over shorter timescales. Similarly, intra-individual diet variation was lower over medium timescales (breeding stage-to-breeding stage), compared to long timescales (year- to-year). In addition, behavioural consistency in foraging activity and the average degree of individual diet specialisation were both higher in individuals foraging in inshore, shallow water habitats compared to those foraging in pelagic habitats, supporting the notion that consistency is favoured in more stable environments. Finally, similarity between nest partners in trophic levels (δ15N values) was found, although no reproductive advantage was observed in relation to the degree of similarity.

Regarding winter dispersion behaviour, Australasian gannets exhibited differences between the two sampled colonies and between years, as well as between sexes. Study individuals could be separated into three distinctive winter dispersal strategies based mainly on location, distance from colony, and duration. Intra-individual similarity in winter dispersal behaviour between two years was found in bearing from the colony to the most distal point and in return dates to the colony. In addition, nest partners showed similar values in departure dates, bearing, and return dates, suggesting coordinated behaviour between members of a mated pair.

The findings of the present study highlight the presence of behavioural consistency in

Australian gannets, and give insights about its potential drivers. More importantly, this study

vi reveals ecological opportunity as a major component for the development of foraging consistency and foraging specialisation in wildlife populations. Additionally, this research contribute with the understanding of the level of resource partitioning among the individuals of the two study populations, which is crucial information for their conservation and management.

vii

PREFACE

The animal handling procedures conducted for the present study complied with the ethical guidelines of Deakin University Animal Ethics Committee under Approval B20-2013. The fieldwork was conducted in accordance with the regulations of Parks Victoria and Department of Environment, Land, Water and Planning (Wildlife Research Permit No. 10006878, File No.

FF383295), issued to Professor John Arnould. The student conducted all leg banding of birds under an Australian Bird and Bat Banding Scheme R Class Banding Licence (No. 3168).

Two chapters of this thesis (Chapters 2 & 3) have been submitted for publication in peer- reviewed journals. Chapter 1 serves as an introduction to the main body of work while Chapter

6 discusses the main findings of the study within the context of the larger field of study. I am the primary contributor to all aspects of the work presented in this thesis, with the exception of the minor additional data collection for Chapters 3, 4, and 5 and the data analysis of Chapters

2 and 5.

John P.Y. Arnould is a co-author on all submitted publications for providing logistic and financial support, contributing to the study designs and required permits, providing equipment, editorial advice, guidance, and general support in all aspects of my PhD.

Chapter 2 - Elodie C.M. Camprasse and Lauren P. Angel contributed substantially to the data analysis about linear mixed effect models and accelerometry data, respectively.

Chapter 3 - Lauren P. Angel contributed to fieldwork and collection of marine prey samples.

viii

Chapter 4 - Lauren P. Angel contributed to fieldwork and collection of bird blood samples.

Chapter 5 – Mark A. Hindell and Melanie R. Wells contributed substantially to the analysis of the geolocation data, and with fieldwork activities for the recovery of the geolocators loggers from the birds, respectively.

Publications arising from this thesis

Chapter 2

Marlenne A. Rodríguez-Malagón, Elodie C.M. Camprasse, Lauren P. Angel and John P.Y.

Arnould (in review). Influences on foraging behaviour and consistency in a marine aerial predator. Royal Society of Open Science.

Chapter 3

Marlenne A. Rodríguez-Malagón, Lauren P. Angel and John P.Y. Arnould (in review).

Temporal and spatial variability in isotopic niche of marine prey species in south-eastern

Australia: potential implications for predator diet studies. Marine Biology.

ACKNOWLEDGEMENTS

Empezaré por agradecerle a mi familia, la cual fue, ha sido y seguirá siendo siempre mi mayor soporte de vida. A mi madre hermosa que siempre ha creído y disfrutado de mis sueños tanto como yo, a mis queridas hermanas Martha y Queta, a mis sobrinas Renné y

Luna a quienes extraño todos los días y por la cuales lamento no estar cerca para disfrutarlas y verlas crecer. Ustedes saben que las amo. Al resto de mi familia, tíos, tías, primos y sobrinos, de los que siempre he recibido apoyo y cariño y quienes me recuerdan todo el tiempo que la familia es el tesoro más grande que la vida te puede brindar. Gracias a todos ustedes por estar siempre a mi lado. A mi querida familia australiana: Edgar, Laura,

Jorge, Mayte, Mayela, Paco, Karina, Tere, Pablo, Débora, Pablo Juliano, Lisa, Leo, Sue,

Michael, Fernanda, Sergio, Luis, Nora y Fernanda V., quienes me adoptaron desde mi llegada y me han hecho sentir querida y apoyada en estas tierras lejanas. Les agradezco mucho esa hermandad y todos los buenos momentos. A mis queridas amigas y hermanas de vida Darlene Guerra y Alejandra Portillo, quienes en los últimos meses de este viaje me ayudaron a seguir de pie, me enseñaron a conocerme a mí misma y a ser una mejor persona.

Por esto, les estaré a ambas eternamente agradecida. Finalmente, a ‘mi héroe’, ese ser de noble corazón que me ha acompañado en muchos buenos y malos momentos, que me ama, me cuida y me cocina, y que siempre encuentra la manera de sacarme una sonrisa. A ti, todo mi ser y todo mi amor! Gracias por esta hermosa y divertida aventura que es vivir!

I will start by thanking my family which was, has been and always will be my greatest support in life. To my beautiful mother who has always believed in me and has enjoyed my dreams as much as I do; to my beloved sisters Martha and Queta; to my nieces Renné and

Luna, whom I miss every day and regret not being close to cherish and see them grow, you

xi know I love you. To the rest of my family, my uncles, aunts, cousins and nephews, who have always loved and supported me, and constantly remind me that family is the greatest treasure in life. Thank you for always being by my side. To my dear Australian family: Edgar, Laura,

Jorge, Mayte, Mayela, Pablo Juliano, Paco, Karina, Tere, Pablo, Débora, Lisa, Leo, Sue,

Michael, Fernanda, Sergio, Luis, Nora and Fernanda V., who adopted me since my arrival to the country and have made me feel loved and supported in these distant lands. Thank you very much for your brotherhood and all the good times. I also want to give special thanks to

Jenny, Molly and Max Cummings, who were at first my landlords and soon after became my friends and family. I love you and I thank you for your friendship and support. To my dear friends and sisters of life, Darlene Guerra and Alejandra Portillo, who have helped me to remain strong these last months, and taught me how to know myself more deeply and become a better person. For all of this, and more, I will be forever grateful.

My most honest thanks to my supervisor, Professor John P.Y. Arnould, who believed in me from the beginning and showed me his trust and patience in many ways throughout these years. Thanks to him I had the opportunity to come to Australia and start a new life in this wonderful country, and for that, I will be eternally grateful. John, thank you for being my mentor throughout this journey, I feel very fulfilled with everything I have learned and all the experiences I have gained. Thank you for the freedom you gave me to make mistakes and to explore, and for the extraordinary support you offered in difficult moments of my life, I owe you a lot.

There are many people from Deakin University who I also want to thank for all of their support. My labmates Lauren Angel, Maud Berlincourt, Laetitia Kernaleguen, Travis

Knox, Elodie Camprasse, Timothee Poupart, Katherine Brownlie, Grace Sutton, Melanie

Wells and Naomi Wells, who shared many hours of fieldwork and studies with me, and helped me throughout this student journey by sharing their knowledge and good company. I

xii also want to thank Ms Wendy Pump, the Grants Officer from the Research department at

Deakin. Many thanks as well to the wonderful staff of the School of Life and Environmental

Sciences: Michael Holmes, Clorinda Schofield, Jessica Bywater, Thomas Schneider, Leanne

Fitton from the Technical Services Department; Higo Jasser and Fawzi Elfaidi from the IT

Services; Loveleen Kumar, Linda Moon, Cuong Van Huynh excellent laboratory technicians and good friends; and finally to Dr Sharon La Fontaine for her support as the HDR

Coordinator at the Burwood Campus. Thank you all for your valuable support given to me during my field and laboratory activities. Finally, I’d also like to give thanks to Dr Peter Biro for your advice on some of the statistical analyses required for this thesis.

I’d also like to give a special thank you to the donors of this research who made this project possible, the Holsworth Wildlife Research Endowment and the Centre for Integrative

Ecology of Deakin University. Both of these institutions supported me over three consecutive years by covering field activities, equipment and sample analysis. Many thanks as well to

Birdlife Australia who helped to cover the remaining sample analysis by the end of my research.

I’d also like to thank the generous support from the Victorian Marine Science

Consortium (Queenscliff, VIC) for the logistic assistance offered during my field and laboratory work performed over there. Special thanks to Roderick Watson, Elizabeth

McGrath and Yvonne Gilbert who always offered me a helping hand and a smiling face over the two years that I spent visiting the Pope’s Eye gannet colony. Thank you as well to the

Main family at Point Lonsdale, particularly to Robby and Jenny, who kindly hosted me on all my trips and gave me a nice, comfortable and friendly place to stay. Thanks to the personnel of Parks Victoria―Port Phillip Heads Marine National Park, particularly to Steve Tuohy, who were always mindful of my safety and allowed me to visit the gannet colony countless times.

xiii

At Portland, I got invaluable support from Ewen Lovell and Phillip King, members of the Point Danger Committee of Management, who facilitated the field activities and always offered me a helping hand. Many thanks as well to Andy Govanstone and family, and to

Susan Miller who were my angels in Portland and offered a nice and cosy place to stay.

Thanks to Brian ‘Sand’ Duro for his help on the field and his passion for the gannets.

Lastly, thank you to the numerous volunteers who participated in field and laboratory activities, to the “French girls” who were there with me at the field (dealing with the cold or super-hot weather), sometimes even against their will. Very special thanks to Roderick

Watson and Nicole Chapple for providing aid with fish identification. Thanks to Dr Hilary

Stuart-Williams from the Farquhar laboratory of the Research School of Biology of the

Australian National University in Canberra, for his valuable help and guidance during the sample preparation and analyses.

Thanks again to everyone for sharing your passion for life with me!

xiv xv

TABLE OF CONTENTS

CHAPTER 1 General introduction ...... 1

Factors influencing the foraging behaviour of animals ...... 2

Behavioural consistency and individual specialisation ...... 4

Foraging behaviour and specialisation in seabirds ...... 6

Study species ...... 8

Study sites ...... 14

Research aims and thesis structure ...... 16

CHAPTER 2 Influences on foraging behaviour and consistency in a marine aerial predator ...... 18

Introduction ...... 20

Materials and methods ...... 23

Study sites and animal handling ...... 23

Data processing and statistical analysis ...... 26

Results ...... 30

Factors influencing foraging behaviour ...... 30

Influence of timescales and habitats on individual consistency of foraging behaviours . 35

Discussion ...... 39

Factors influencing foraging behaviour ...... 40

Influence of timescales and habitats on individual consistency of foraging behaviours . 43

Supporting information ...... 47

CHAPTER 3 Temporal isotopic variability of marine prey species in south-eastern

Australia: potential implications for predator diet studies ...... 54

Introduction ...... 56

Materials and methods ...... 59

xvi

Sample collection ...... 59

Stable isotope analysis...... 63

Statistical analysis ...... 64

Results ...... 64

Discussion ...... 74

Supporting information ...... 79

CHAPTER 4 Inter- and intra-individual variation in the diet of Australasian gannets

(Morus serrator) ...... 83

Introduction ...... 85

Methods ...... 88

Study sites and sample collection ...... 88

Stable isotope analysis...... 91

Regurgitation analysis and diet reconstruction ...... 92

Statistical analysis ...... 94

Results ...... 96

Discussion ...... 110

Factors influencing inter- and intra-individual isotopic variation ...... 110

Inter- and intra-individual diet variation ...... 114

Diet similarity between nest partners ...... 115

Supporting information ...... 118

CHAPTER 5 Factors influencing winter distribution and habitat use in Australasian gannets (Morus serrator) ...... 124

Introduction ...... 126

Methods ...... 129

Study sites and animal handling ...... 129

xvii

Data processing and statistical analysis ...... 130

Results ...... 134

Winter distribution and factors influencing dispersal behaviour ...... 134

Intra-individual variation and within-pair behavioural similarity ...... 139

Discussion ...... 144

Winter distribution and factors influencing dispersal behaviour ...... 144

Intra-individual and within-pairs behavioural similarity ...... 149

Supporting information ...... 151

CHAPTER 6 General discussion ...... 156

Introduction ...... 157

Drivers of behaviour...... 161

Intrinsic drivers ...... 161

Extrinsic drivers ...... 165

Inter- and intra-individual variation in behaviour ...... 166

Limitations and recommendations for future work ...... 169

LITERATURE CITED ...... 172

xviii

LIST OF FIGURES

Fig. 1.1: Band recovery records (blue dots) from adult Australasian gannets (Morus serrator) found during the non-breeding season (austral winters from 1965 to 2013) on the

Australia and New Zealand coasts (Australian Bird and Bat Banding Scheme unpublished data)...... 13

Fig 1.2: Location of study sites (orange) and other Australasian gannet breeding colonies

(yellow) in south-eastern Australia: Point Danger (PD), Pope’s Eye (PE), Lawrence Rocks

(LR), Port Phillip Bay (PPB), Black Pyramid (BP), Pedra Branca (PB), and Eddystone Rock

(ER). Bathymetric contours (every 20 m) within the continental shelf are shown...... 15

Fig. 2.1: Location of study sites: Point Danger (left) and Pope’s Eye (right). The 200 m bathymetric contour is given to indicate the edge of the continental shelf...... 25

Fig. 2.2: Examples of the time-scale comparisons analysed: T-to-T (Trip-to-Trip, left column); S-to-S (Stage-to-Stage, middle column); and Y-to-Y (Year-to-Year, right column).

Each row represents one instrumented bird: first row, male from PD-pelagic; second row, a

PE-pelagic female; third row, a PE-mixed male; and fourth row, a PE-inshore male.

Breeding stages include incubation (INC), early chick-rearing (ECR), and late chick-rearing

(LCR)...... 37

Fig. 3.1: Location of the Point Danger (PD, left) and Popes Eye (PE, right) breeding colonies (black crosses). The South Australian Current (SAC, winter) and East Australian

Current (EAC, winter and summer) bring warm and low nutrient waters into the marine region, while the Sub-Antarctic Surface Water (SASW, summer) drives cold and productive waters from the south (Sandery & Kämpf 2007). PD (green) and PE (orange) Australasian

xix gannet (Morus serrator) foraging ranges obtained from GPS data (Chapter 2, Wells et al.

2016). Bathymetric contours (every 20 m) within the continental shelf are shown...... 61

Fig. 3.2: Stable isotope biplot indicating the mean ± SD of δ13C and δ15N positions of the

18 prey species collected from Australasian gannet (Morus serrator) regurgitates at the

Point Danger and Pope’s Eye breeding colonies. Numbers in parentheses represent total number of prey individuals analysed. Asterisks on species name identify benthic species, all others are considered pelagic...... 68

Fig. 3.3: Stable isotope biplot indicating the mean (± SD) of δ13C and δ15N values for different prey species. A. Inter-annual comparisons with significant results: barracouta

(Thyrsites atun), Gould’s squid (Nototodarus gouldi), and king gar (Scomberesox saurus) in both δ13C and δ15N values, redbait (Emmelichthys nitidus) in δ15N values only. B. Inter- annual and geographic (Point Danger = PD and Popes Eye = PE) comparison with significant results: jack mackerel (Trachurus declivis) for δ15N values only...... 72

Fig. 3.4: Coefficients of variation in δ13C and δ15N values within each year and colony for all prey with ≥ 3 samples collected (n = 12). Asterisks on species name identify benthic species...... 73

Fig. 4.1: Location of study sites (black stars): Point Danger (PD, left) and Pope’s Eye (PE, right). Bathymetric contours (every 20 m) within the continental are shown...... 89

Fig. 4.2: Stable isotope ratios (δ13C and δ15N, mean ± SD) of adult Australasian gannets

(Morus serrator) blood plasma nesting at Point Danger (PD) and at Pope’s Eye (PE) breeding colonies over a four-year period (2012-2015). Three breeding stages: incubation

(INC), early chick-rearing (ECR), and late chick-rearing (LCR); and sexes: female (F) and male (M), were considered. The sample size for each group is listed in Table 4.1...... 102

Fig. 4.3: Proportional similarity index (PS) of Australasian gannets (Morus serrator) from

Point Danger (PD) and Pope’s Eye (PE) breeding colonies over a four-year period (2012-

2015). PS values are estimated as the diet overlap between each individual and the population, where values close to 1 represent generalist individuals and values close to 0 specialist individuals (Bolnick et al. 2002)...... 105

Fig. 4.4: Correlations of blood plasma δ13C (top) and δ15N values (bottom) values from

Australasian gannet (Morus serrator) breeding partners (n= 51 pairs) sampled at Point

Danger (purple) and Pope’s Eye (green) breeding colonies. A significant and positive

2 relationship is shown for nitrogen values (R = 0.22, F1,100 = 28.57, P < 0.0001)...... 109

Fig S4.1: Medium-term (left panels) and long-term (right panels) correlations between two breeding stages of the same year, or the same breeding stage of two different years. Values represent nitrogen isotopic ratios (δ 15N ‰) from blood plasma of adult Australasian gannets (Morus serrator) sampled at two breeding colonies Point Danger (squares) and

Pope’s Eye (circles). Males are represented in blue, females in red...... 120

(Morus serrator) sampled at two breeding colonies Point Danger (squares) and Pope’s Eye

(circles). Males are represented in blue, females in red...... 121

Fig 5.1: Percentage of total time spent in pre-defined grid cells for adult Australasian gannets (Morus serrator) from Point Danger during the winter non-breeding dispersal period (March to September). A) Females 2015 (n = 13), B) Females 2016 (n = 19), C)

Males 2015 (n = 15), D) Males 2016 (n = 15). The breeding colony location is defined by a black star...... 137 xxi

Fig 5.2: Percentage of total time spent in pre-defined grid cells for adult Australasian gannets (Morus serrator) from Pope’s Eye during the winter non-breeding dispersal period

(March to September). A) Females 2015 (n = 13), B) Females 2016 (n = 14), C) Males 2015

(n = 14), D) Males 2016 (n = 20). The breeding colony location is defined by a black star...... 138

Fig. 5.3: Identification of winter dispersal strategies. A) Hierarchical clustering of dispersal metrics in which listed numbers represent individual winter tracks of adult Australasian gannets (Morus serrator). B-D) Representative individuals of each of the three winter dispersal strategies identified. Stars denote location of breeding colonies: Point Danger

(left) and Pope’s Eye (right). Legend on each map represents percentage of total time spent in pre-defined grid cells by each individual...... 142

Fig. S5.1: Differences in the winter dispersal metrics values between the clusters identified by hierarchical method for Australasian gannets (Morus serrator)...... 155

xxii

LIST OF TABLES

Table 2.1: Means ± SE of the foraging trip parameters collected from instrumented

Australasian gannets (Morus serrator) during two years 2014 and 2015 at Point Danger and

Pope’s Eye colonies in Victoria, Australia. Instrumentation on birds lasted 10-12 consecutive days. Data represent the dataset acquired separated by sex, breeding stages

(incubation, early chick-rearing, late chick-rearing), year, and colony-habitat. Samples sizes of the metrics estimated from the GPS (maximum distances from the colony, bearing and tortuosity index; n1) and accelerometer data loggers (mean Vectorial Dynamic Body

Acceleration and number of dives, n2), are shown...... 32

Table 2.2: Most parsimonious models after model averaging, and their corresponding estimated regression parameters, for five foraging metrics obtain from free-ranging

Australasian gannets (Morus serrator)...... 33

Table 2.3: Variance component analysis of instrumented Australasian gannets (Morus serrator). Short-term (Trip-to-Trip), medium-term (Stage-to-Stage) and long-term (Year-to-

Year) comparisons are shown. Sample sizes (number of trips/number of individuals) are presented for each final model. The significant fixed components of the models in which the coefficients of variation were used as a response variable are shown as the factors influencing the individual variation on each case...... 36

Table 2.4: Variance component analysis of instrumented Australasian gannets (Morus serrator). Short timescale comparison (T-to-T) results are shown for the models split by colony and habitat. Sample sizes (number of trips/number of individuals) are presented for each final model...... 38

xxiii

Table S2.1: AICc based model selection (∆ < 4) for factors that influence the foraging metrics in adult free-ranging Australasian gannets (Morus serrator). Model variables: Body

Condition Index; Body Size Index; Wing Length Index; colony; stage: breeding stage; year; sex...... 47

Table S2.2: Average model coefficients and relative importance of variables included in top model set (∆AICc ≤ 4) explaining individual variation in Australasian gannets (Morus serrator) foraging metrics...... 50

Table S2.3: Factors influencing individual variation (measured as the coefficient of variation within deployments for each foraging metric), at short- term scale (T-to-T) in

Australasian gannets (Morus serrator). Most parsimonious models after model averaging and their corresponding estimated regression parameters are shown. The most parsimonious model was selected with using Akaike Information Criterion (AICc, ∆ < 4). Model variables: Body Condition Index; Body Size Index; Wing Length Index; stage: breeding stage; colony; year; sex...... 52

Table 3.1: Sample sizes of all the prey species found shown by year and colony. An asterisk indicates those species whose isotopic values were statistically tested for temporal and spatial differences. Habitat and diet for each prey species is listed ...... 66

Table 3.2: Two-way ANOVA test results for temporal and spatial differences in δ13C values of different prey species (mean ± SE). Significant results of the ANOVA test are shown. Means with the same superscript denote homogenous subsets (P > 0.05)...... 70

Table 3.3: Two-way ANOVA test results for temporal and spatial differences in δ15N values of different prey species (mean ± SE). Significant results of the ANOVA test are shown. Means with the same superscript denote homogenous subsets (P > 0.05)...... 71

xxiv

Table S3.1: Means (± SE) of the standard length and body mass for prey species by year and colony (where whole specimens or otoliths could be measured)...... 79

Table S3.2: Examples of Australasian gannet (Morus serrator) prey species analysed in this study that are shared with other marine predators...... 81

Table 4.1: Samples sizes of the blood plasma collected from individual Australasian gannets (Morus serrator) at both breeding colonies: Point Danger and Pope’s Eye. Data represents the dataset acquired separated by sex (females, males), breeding stages

(incubation, early chick-rearing, late chick-rearing), and years (2012 to 2015)...... 99

Table 4.2: Linear mixed effect model outputs evaluating the effect of colony, sex, breeding stage and year on the carbon (δ13C) and nitrogen (δ15N) stable isotope values quantified from adult Australasian gannets (Morus serrator) blood plasma. Model selection is based on AICc values. All models include individual identity as random factor. Model variables: colony (Point Danger and Pope’s Eye); breeding stage (incubation, early chick-rearing, and late chick-rearing); year (2012, 2013, 2014, 2015) and sex (male and female)...... 100

Bayesian isotopic mixing models (MixSIAR). The “small schooling pelagics” category represents a posteriori group formed by Australian anchovy, Australian sardine and king gar which according to the model diagnostics were unable to be discerned from each other. The number of individuals analysed by year of sampling is indicated in parenthesis. The degree of individual specialisation of the population and the proportional similarity index range estimated by breeding season are shown...... 106

xxv

Table 4.4: Carbon and nitrogen stable isotopes values (mean ± SE) of Australasian gannet

(Morus serrator) blood plasma paired samples and correlation parameters at different medium- (INC-ECR, ECR=LCR, INC-LCR) and long-term (INC-INC, ECR-ECR, LCR-

LCR) time scales...... 108

Table S4.1: Diet composition analysis from regurgitate samples of Australasian gannets

(Morus serrator) collected at Point Danger and Pope’s Eye breeding colonies over a four year period (2012-2015). Diet is described as frequency of occurrence (F%), numerical abundance (or number of individuals per species, N) and mass percentage (M%). An asterisk indicates species that uses the benthos as preferred habitat, those without it are considered pelagic. Sample size (n) and total mass collected (kg) for F% and M% estimation, respectively, are shown...... 118

Table S4.2: AICc based model selection (∆ < 4) for factors that influence the reproductive output (binomial: successful = 1 (chick fledged), unsuccessful = 0), in Australasian gannets

(Morus serrator). Model variables: the absolute difference between female and male δ13C values; the absolute difference between female and male δ15N values; colony; year. All models share the same random component (nest identity)...... 122

Relative variable importance based on AIC weights. Model variables: the absolute difference between female and male δ13C values; the absolute difference between female and male δ15N values...... 123

xxvi

Table 5.1: Means ± SE of winter dispersal metrics for adult Australasian gannets (Morus serrator) during austral winters (March to September) in 2015 and 2016 at Point Danger and Pope’s Eye breeding colonies in south-eastern Australia...... 136

Table 5.2: Most parsimonious models after model averaging for five winter dispersal metrics for adult Australasian gannets (Morus serrator). Their corresponding estimated regression parameters are shown...... 141

Table 5.3: Winter dispersal metrics (mean ± SE) of Australasian gannets (Morus serrator) paired samples, correlation parameters and paired t-test results. Two levels of similarity comparisons, by individuals and by nest pairs, are shown...... 143

Table S5.1: AICc based model selection (∆ < 4) for factors influencing the winter dispersal metrics in adult Australasian gannets (Morus serrator). Model variables: colony; breeding success; departure date; year; sex...... 151

Table S5.2: Average model coefficients and relative importance of variables included in top model set (∆AICc ≤ 4) explaining winter dispersal metrics in Australasian gannets

(Morus serrator) ...... 153

xxvii

CHAPTER 1 General introduction

Factors influencing the foraging behaviour of animals

The main concern of every animal is to gather resources for survival and reproduction, which implies a constant decision-making process. A set of hierarchical decisions is needed every day which leads animals to choose among feeding habitats, prey items, food patches, foraging strategies, and time allocation (Stephens 2008). Such decisions are influenced by inter- and intra-specific competition and predation avoidance (Lima et al.

1985). The theory of optimal foraging addresses these decisions and postulates that animals use decision rules in order to maximise their net energy intake per unit of time while adjusting their foraging strategies in order to reach maximum efficiency (Pyke et al. 1977,

Stephens & Krebs 1986). According to this theory, optimal decision rules are followed by individuals within populations, rules that guide them in several aspects of their foraging behaviour such as ideal prey size, ideal feeding habitats, or ideal searching methods (Hughes

1980). Correspondingly, as individuals may differ in their physical, physiological, or cognitive abilities, their perception and their ability to exploit their surrounding environment could be individually biased (Svanbäck & Bolnick 2007). In addition, environmental constraints that influence food availability or local weather, and the presence of predators or competitors, may result in animals altering their foraging behaviour to fit the current conditions (Gibson et al. 1998, Svanbäck & Bolnick 2007). For example, the lesser kestrel

(Falco naumanni) adjusts its flight patterns in response to increasing solar radiation

(Hernández-Pliego et al. 2017). Similarly, the gerbil (Gerbillus andersoni allenbyi) manages predation risk by allocating foraging time in dangerous food patches or under bright moonlight (Embar et al. 2011).

Intrinsic factors are considered to be the main drivers of inter-individual variation

(Svanbäck & Bolnick 2007). Age is a developmental constraint in which young animals experience body growth and maturation of physiological functions, learn abilities, and may

2 have specific nutritional requirements which lead them to forage differently than adults

(Marchetti & Price 1989, Senior et al. 2015). Differences in the proportion of food items consumed (Navarro et al. 2010), prey size and capture rates (Gochfeld & Burger 1984), prey detection, and prey handling times (Heise & Moore 2003) have been reported as some of the consequences of age as a biological constraint.

Likewise, strong physiological, physical, and social differences may exist between individuals of different sexes (Bull 1983). Some females for example, face high energetic and nutritional demands during the breeding season which can influence their diet, foraging behaviour and strategies (Catry et al. 2005). Also, in mammals, differences in the proportions of consumed food items have been reported between males and pregnant and/or lactating females (Bernard & Hohn 1989, Kunz et al. 1995). Some carabid beetle females often consume more prey types than males, which is believed to help them with egg production

(Lövei & Sunderland 1996). Sexual size dimorphism (one sex larger than the other), also has strong implications for the way animals forage. A greater body size may confer greater physical endurance allowing animals to be faster, or to have a large home range and, consequently, consume a different set of food items (Perry & Garland 2002, Edwards et al.

2011). In addition, the larger sex could have higher energetic requirements, leading to the consumption of different types and amounts of food (Key & Ross 1999).

Within the same context, different phenotypes within a species can also be adapted to consume different prey items (Maret & Collins 1997). Physiological conditions may restrict certain individuals to handle or digest specific prey items (Page et al. 1995, Dall et al. 2012).

Differences in the development of a search image (the selective attention to few visual clues) could increase the feeding rate and competitive ability of individuals (Reid & Shettleworth

1992, Bernays & Funk 1999). Moreover, it has been demonstrated that individuals with different nutritional optima personalise food intake in order to maximise their own fitness

(Raubenheimer & Simpson 1997, Senior et al. 2015, Machovsky-Capuska et al. 2016b).

Finally, different personality traits such as boldness or aggressiveness can influence individual feeding decisions by encouraging individuals to take greater risks while foraging, feeding in different habitats, or to prefer solitary feeding versus feeding in groups (Aplin et al. 2014, Toscano et al. 2016). All these intrinsic factors potentially have a great impact on the way animals forage and, therefore, it is important to recognise their effects because foraging is a primary activity for all animals.

Behavioural consistency and individual specialisation

Foraging is generally an activity involving high-energy expenditure rates and foragers, while looking for food, constantly search for ways to minimise these energetic costs

(Schoener 1971). With time, foragers are able to learn which food items or foraging locations are profitable following the predictions of optimal foraging theory (Stephens & Krebs 1986).

Numerous foraging strategies are possible for every individual, but those with greater rewards will more likely be repeated over time, favouring the development of behavioural consistency (Reader 2015). Behavioural consistency in foraging activities leads to the evolutionary development of foraging specialisation within animal populations, with potential implications for the ecology of the individuals who exhibit these specialisations (Dall et al.

2012).

Foraging specialists are those individuals that make use of a specific proportion of the full range of available resources (or foraging strategies) for reasons not attributable to sex, age, or discrete a priori morphological group, resulting in inter-individual niche variation

(Bolnick et al. 2003, Sargeant 2007). This phenomenon is thought to mainly arise in stable environments in which resources are predictable, allowing the development of behavioural differences between individuals and favouring the reduction of niche overlap with

4 conspecifics, and thus, minimising competition (Araújo et al. 2011). Foraging specialisation has been demonstrated in a wide variety of taxa (Bolnick et al. 2003). In the long term, these behavioural variations can influence both the evolution and ecology of populations (Araújo et al. 2011). This is true because resource competition and selection will operate very differently when inter‐individual variations became permanent within populations (Bolnick et al. 2003, Dall et al. 2012); as with time, food network structure, phenotype abundance, and risk of predation or parasitism can be altered as a result of individual specialisations and survival of the more efficient individuals (Bolnick et al. 2011).

Different methods have been used to address the level of individual foraging specialisation in animals, which require longitudinal sampling in order to estimate individual consistency (Bolnick et al. 2003, Carneiro et al. 2017). Traditional methods of gut content analysis or direct feeding observations (Woo et al. 2008, Ceia et al. 2012), stable isotope values from single or multiple animal tissues (e.g. δ13C and δ15N; Bearhop et al. 2006, Araújo et al. 2007, Robertson et al. 2014), or data obtained from electronic devices such as GPS and behavioural activity recorders, have recently been used to analyse individual consistency

(Woo et al. 2008, Masello et al. 2013, Amelineau et al. 2014, Granadeiro et al. 2014).

Furthermore, some recent studies have used combinations of two or more methods for complementarity or cross-validation (Masello et al. 2013, Granadeiro et al. 2014, Robertson et al. 2014). Depending on the nature of the data (e.g. Gaussian or non-Gaussian), several statistical approaches can be used to estimate individual consistency such as correlations,

ANOVAs, linear mixed effects models, or ranges overlap (Nakagawa & Schielzeth 2010,

Carneiro et al. 2017). Statistically, to detect behavioural consistency in traits it is necessary to compare the amount of intra-individual variation with inter-individual variation. In the context of specialisation, consistency relates to intra-individual variation (Nakagawa &

Schielzeth 2010). These methods and statistical approaches are of importance in ecological

5 studies as they enable investigation of how individual variability differs among environments and ecological contexts.

Foraging behaviour and specialisation in seabirds

Seabirds are major top predators in the world’s oceans. A total of 365 seabird species are currently recognised (BirdLife International 2018b), with an estimated global population of 0.7 billion individuals and an annual food consumption rate of 69.8 million tonnes (de L.

Brooke 2004). Seabirds occur in all seas and oceans, and their role as potential indicators of marine conditions is widely acknowledged (Croxall et al. 2012). The ecology of seabirds is highly influenced directly or indirectly by the environmental phenomena that influence the oceans on temporal scales from days to decades (Chambers et al. 2014). Their ability to find and capture prey, their reproductive success, and their own survival can be determined by external factors such as weather conditions, oceanic fronts, or prey availability (Baird 1990,

Chastel et al. 1993, Weimerskirch et al. 2005a). Therefore, it is important to understand the factors that influence their foraging ecology and success in order to determine their role in the ecosystem and how they may respond to environmental variability.

Most seabird species are characterised by being colonial breeders, central-place foragers during the breeding season, and displaying high levels of nest-site fidelity (Schreiber

& Burger 2001, Gaston 2004). These traits facilitate the collection of repeated samples from individuals under similar environmental conditions and make seabirds good models to investigate behavioural consistency (Camprasse et al. 2017a). Consequently, many seabird species have been subject of recent studies analysing behavioural consistency and the presence of individual specialisation (Ceia & Ramos 2015, Phillips et al. 2017).

Several factors have been shown to influence the level of inter-individual variation in foraging behaviour of seabirds. Sex-specific differences in foraging and migration strategies

6 have been analysed in numerous species of seabirds from the orders Charadriiformes,

Pelecaniformes, Procellariiformes and Sphenisciformes (Phillips et al. 2017). Likewise, age- related differences in foraging behaviour have also been widely observed (e.g. Orians 1969,

Greig et al. 1983, Jaeger et al. 2014, Froy et al. 2015). The variation in the energetic requirements between adults and offspring, and between the various stages of the breeding cycle (i.e. incubation, early chick-rearing, and late chick-rearing) also causes divergence in foraging strategies of seabirds (Jaeger et al. 2014, Jakubas et al. 2014, Leal et al. 2017).

Behavioural consistency has been reported in several aspects of the foraging behaviour of seabirds. For example, individual consistency in foraging trip departure bearings, diving activities and patterns, foraging site selection, and total distances travelled have been found (e.g. Hamer et al. 2001, Mackley et al. 2010, Patrick et al. 2014a). All of these foraging trip characteristics have been measured during the breeding season through the use of behavioural data loggers. During the non-breeding season, seabirds have also been shown to be consistent regarding their use of non-breeding sites, particularly at the regional scale (Phillips et al. 2017). In terms of diet, conventional stomach contents and stable isotope studies have revealed significant consistency within individuals from many seabird species in habitat use, prey type, or trophic level in both, the short- and long-term (days to weeks, between the breeding and non-breeding seasons or annually; Phillips et al. 2017).

Individual foraging specialisation has been analysed in several species of seabirds

(Ceia & Ramos 2015) and for some, associated benefits have been reported. For example,

Golet et al. (2000) found that breeding pairs of pigeon guillemots (Cepphus columba) specialised on certain prey types, and those that specialised had higher reproductive success because they delivered prey items of higher energy content to chicks. Votier et al. (2004) also found prey specialisation in great skuas (Stercorarius skua), documenting that a small proportion of the skua population fed on other seabirds while the rest fed on fish discards.

The specialised skuas had smaller home-ranges and more consistent behaviour over breeding seasons, and they were able to hatch their eggs earlier and lay larger clutch sizes than non- specialised birds. However, other studies have reported no specific fitness or breeding benefits for specialised individuals compared to their un-specialised counterparts. Woo et al.

(2008), after analysing a 15-year dataset on the Brünnich’s guillemot (Uria lomvia), found no difference in survival or reproductive success between specialist and generalist foragers.

Similarly, Dehnhard et al. (2016) after seven years of studying southern rockhopper penguins

(Eudyptes chrysocome) were unable to link individual consistency in diet with several measures of reproductive output. Hence, the benefits of specialisation for individuals remain equivocal. Despite this uncertainty, the study of individual specialisations within animal populations allows us to understand the degree of resource partitioning that exists among individuals, which in turn helps to acquire information about susceptibility to anthropogenic threats and environmental change (Carneiro et al. 2017). Such information is crucial to inform management plans that aim to protect the resource base necessary for the survival of species. In a larger context, the accumulation of information about individual specialisations will allow better inferences about the benefits of this phenomenon and its implications for evolutionary processes (Phillips et al. 2017).

Study species

The Australasian gannet (Morus serrator), endemic to Australia and New Zealand, represents one of the major marine top predators in south-eastern Australia (Norman &

Menkhorst 1995). The species is a member of the Sulidae family (Aves: Pelecaniformes) in which there are nine other species of boobies and gannets. Its closest relatives are the Cape gannet (M. capensis) and the northern gannet (M. bassanus) with which it shares strong physical and behavioural similarities (Nelson 1978, Patterson et al. 2011). The Australasian

8 gannet, after the Abbot’s booby (Papasula abbotti), is considered to be the rarest sulid in the world (Nelson 1978), with an estimated global population of 53,000 breeding pairs (Bunce

2001). Its distribution ranges from southwest to northeast Australia, and to the coasts of New

Zealand (Australian Bird and Bat Banding Scheme [ABBBS] unpublished data, BirdLife

International 2016). In recent decades, the population of Australasian gannets in south- eastern Australia has reportedly grown and new colonies have become stablished (Norman et al. 1998, Bunce et al. 2002, Pyk et al. 2013), a trend that has been shared with other breeding areas in New Zealand (Greene 1999, Brown & Wilson 2004, Sawyer & Fogle 2013)

Presently, the conservation status of this species is considered stable or increasing by local and international authorities (Robertson et al. 2017, BirdLife International 2018a, Department of the Environment 2019), although no census has been recently published and the population trend of certain breeding colonies remains unclear (Frost 2017).

The Australasian gannet breeds during the austral summer. Peak dates for laying, hatching, and fledging in some of the Australian colonies occur between August-September,

October-November and February-March, respectively. Adult birds usually leave the breeding grounds in April of each year (Nelson 1978). The incubation period lasts 37-50 days, the early chick-rearing stage is considered to be from hatching until the first black primaries appear on the chicks, and lastly, the late chick-rearing stage involves the development of juvenile plumage until fledging (Wingham 1982). In Australasian gannets only one chick is reared at a time and both parents contribute to chick provisioning (Nelson 1978, Wingham

1984). In previous studies, bird age has been strongly related to chick provisioning efficiency and, consequently, to chick survival during development or post-fledging (Gibbs et al. 2000).

These birds were previously considered monogamous (Nelson 1978), but recent resarch has reported divorce rates between breeding seasons as high as 40-44% (Ismar et al. 2010a).

However, little is known regarding pair retention or within-pair behavioural similarity and its potential effects on the reproductive output of this species (Ismar et al. 2010a).

The Australasian gannet is considered an inshore species, with foraging ranges restricted to the continental shelf (Nelson 1978). Its main foraging strategy is vertical plunge- diving, a highly specialised technique (Machovsky-Capuska et al. 2013), in which association with heterospecifics (such as seals, dolphins, sharks, and other seabirds) has been proven beneficial to find and capture prey (Machovsky Capuska et al. 2011, Wells et al. 2016).

Angel et al. (2015b) reported reverse sexual-dimorphism (females larger than males) for

Australasian gannets found in two breeding colonies of Australia. However, this physical trait appears to be colony-specific as evidence has been found supporting monomorphism in other breeding sites within the distribution of this species (Daniel et al. 2007, Ismar 2010, Krull et al. 2012). Despite this, recent studies suggest Australasian gannets exhibit sex-related differences in habitat use (Angel et al. 2016, Wells et al. 2016, Besel et al. 2018) and prey selection (Machovsky-Capuska et al. 2016a). Furthermore, inter-colony differences in diving behaviour and habitat selection have been documented throughout the species range

(Machovsky-Capuska et al. 2014b, Angel et al. 2016).

Like other members of the Sulidae, the Australasian gannet is considered a generalist forager and has been shown to be adaptable in its feeding habits (Bunce 2001, Schuckard et al. 2012, Machovsky-Capuska et al. 2016a). The diet of Australasian gannets within south- eastern Australian colonies has been reported to consist of at least 37 demersal/reef- associated and pelagic/oceanic species of fish and squid (Norman & Menkhorst 1995, Bunce

& Norman 2000, Bunce 2001, Pyk et al. 2007, Barker 2012, Tait et al. 2014). Significant differences in isotopic values (δ13C and δ15N), used as a proxy for diet, have been shown in relation to the collection site (breeding colony), sex, or reproductive stage of the birds (Angel et al. 2016, Ismar et al. 2017). Although the diet of this species has previously been

10 documented, all previous studies have assumed conspecifics are ecologically equivalent and ignore potential inter-individual variability. Investigating the inter- and intra-individual variation in diet is necessary as it could potentially have important ecological implications, such as a reduction in the intraspecific competition or individual differences in habitat use

(Ceia & Ramos 2015, Carneiro et al. 2017).

The Australasian gannet is a long-lived seabird with a maximum age reported from banding data of 30 years and 8.2 months (ABBBS unpublished data). It becomes a central- place forager during the breeding season and displays high levels of nest-site fidelity (Nelson

1978). These traits combined make this marine predator a good model for investigating behavioural consistency and for obtaining repeated samples from several individuals under similar conditions. In closely related species, like the northern gannet, behavioural consistency has been reported for some foraging parameters such as paths, departure bearings, average speed, and dive locations (Hamer et al. 2007, Patrick et al. 2014), as well as for dispersal patterns during the winter non-breeding season (Fort et al. 2012). However, there is no information about individual behavioural consistency in Australasian gannets or the factors that could potentially influence inter-and intra-individual behavioural differences within their populations.

Little is known about the movement patterns of Australasian gannets during the winter non-breeding period. Historical records report juveniles over-winter around New

South Wales (Sydney) and along the New Zealand coast (Stein 1962, Norman 1991), while adults have been reported off South Australia (Cox 1978). Norman (1991) reports banded gannets seen in Western Australia, and Dann et al. (2004) mention some records of

Australasian gannets at Western Port in Victoria during the autumn and winter. Data from the

ABBBS indicate birds banded at Tasmanian, Victorian, and New Zealand colonies winter in southwest and southeast Australia during the non-breeding season (Fig. 1.1, ABBBS

11 unpublished data). It has been suggested for other species of seabirds that wintering strategies

(e.g. departure-returning dates or distribution) may vary between the sexes and individual behavioural consistency in wintering areas and dispersal routes can be present between years

(Yamamoto et al. 2014, Carneiro et al. 2016). The winter season has also been recognised as a period of relatively high mortality rates for seabirds as they face a reduction in food resources (McCutcheon et al. 2011). Thus, it becomes important to analyse the winter ecology of Australasian gannets in order to understand the distribution of this species and the potential for inter- and intra-individual differences in dispersal strategies.

Fig. 1.1: Band recovery records (blue dots) from adult Australasian gannets (Morus serrator) found during the non-breeding season (austral winters from 1965 to 2013) on the Australia and New Zealand coasts (Australian Bird and Bat Banding Scheme unpublished data).

Study sites

The present study focuses on breeding colonies located at Point Danger (PD, 38º 23’

36.09” S, 141º 38’ 55.94” E) and at Pope’s Eye (PE; 38º 16’ 35.88” S, 144º 41’ 56.21” E), two colonies specifically chosen because of different oceanographic conditions and potential foraging habitats, plus their accessibility for research activities. PD is the only mainland colony of Australasian gannets in Australia. This colony is located close to the seasonally strong and productive Bonney Upwelling System that occurs from Robe (South Australia) to

Portland (Victoria). Upwelling in this systems occurs during the austral summer (November to May), when seasonal winds blowing from the southeast run parallel to the continental shelf and produce upwelling of high nutrient, low-temperature water masses (Lewis 1981). This highly productive area is also heavily used by other marine predators such as penguins, fur seals and whales (Butler et al. 2002).

The PE colony is located at the entrance of Port Phillip Bay, < 3 km from the coast.

Oceanographically, Port Phillip Bay is an area of low primary productivity, but an important habitat for small fish (Norman & Menkhorst 1995, Norman et al. 1998). It covers an area of

1,930 km2 and is characterised by shallow water habitats (average depth < 13.6 m; Walker

1999). Water movement inside the bay is mainly driven by tidal action, wind, and salinity/temperature gradients that lead to high sea surface temperatures (SST) during the summer (16-22ºC), and low chlorophyll-a levels (1-20 mg/m3; Harris et al. 1996). The PD colony (660 breeding pairs) and the PE colony (180 breeding pairs) share foraging habitats with other nearby Australasian gannet colonies at Lawrence Rocks (3100 breeding pairs) and in Port Phillip Bay (310 breeding pairs), respectively (Bunce et al. 2002, Angel et al. 2016).

Other Australasian gannet colonies present in the region are Black Pyramid, Pedra Branca and Eddystone Rock (Fig. 1.2; Brothers et al. 1993). Currently, established colonies in south-

14 eastern Australia are considered to be at full capacity due to nesting space limitations (Pyk et al. 2013).

Fig 1.2: Location of study sites (orange) and other Australasian gannet breeding colonies

(yellow) in south-eastern Australia: Point Danger (PD), Pope’s Eye (PE), Lawrence Rocks

(LR), Port Phillip Bay (PPB), Black Pyramid (BP), Pedra Branca (PB), and Eddystone Rock

(ER). Bathymetric contours (every 20 m) within the continental shelf are shown.

Research aims and thesis structure

The objectives of this research project were to determine: 1) the level of inter- and intra-individual variation within foraging behaviour, diet and winter dispersal patterns in

Australasian gannets in order to quantify the presence of behavioural consistency in this species; 2) how these attributes vary over different timescales and habitats in order to address the prevalence of behavioural consistency; and 3) the factors influencing these same attributes in order to identify the potential drives of behavioural consistency. In accordance with the optimal foraging theory and differential adaptation (either colony- or sex-specific) found in Australasian gannets and other close related species, some level of individual behavioural consistency in foraging, diet, and winter dispersal patterns of this species are expected. The results of this research will contribute with the understanding of the level of resource partitioning among individuals of the two study populations, which is crucial information to address their susceptibility to anthropogenic threats and environmental change. The central chapters of this thesis represent specific studies that have been or will be submitted for publication in peer-reviewed scientific journals.

Specifically,

 In Chapter 2, I analyse the factors influencing the consistency of foraging behaviour in

Australasian gannets; I use five different foraging metrics extracted from bio-logging

techniques to determine the magnitude of behavioural consistency and the main factors

influencing consistency. Also, I assess how individual consistency varies over different

timescales and foraging habitats.

 In Chapter 3, I analyse the temporal variability in the stable isotope ratios of marine prey

species in south-eastern Australia. In this chapter, I state the importance of analysing

isotopically the prey base on which marine top predators depend, as isotopic values of

prey species can change due to external factors and can bias our behavioural consistency

estimations if these changes are not considered.

 In Chapter 4, I investigate the factors affecting variation in the diet of Australasian

gannets by estimating the variation in isotopic ratios (δ13C and δ15N) in birds’ blood

plasma, the level of inter- and intra-individual variation in diet over different timescales,

and the similarities between nest partners in diet.

 In Chapter 5, I describe the winter distribution and of Australasian gannets, and I analyse

the factors influencing their winter dispersal behaviour. I also analyse five dispersal

metrics extracted from bio-logging techniques to determine the main factors influencing

dispersal, identify different dispersal strategies, quantify the levels of intra-individual

variation within the population, and investigate levels of within-pair similarity.

CHAPTER 2

Influences on foraging behaviour and consistency in a marine

aerial predator

A version of this chapter has been submitted as:

Marlenne A. Rodríguez-Malagón, Elodie C.M. Camprasse, Lauren P. Angel and John P.Y. Arnould (in review). Influences on foraging behaviour and consistency in a marine aerial predator. Royal Society of Open Science.

Abstract

Foraging is a behaviour that can be influenced by multiple factors, to which animals can respond in various ways depending on their behavioural plasticity. In recent decades, consistency in individual foraging behaviour has been described and quantified in several taxa and is now recognised to have ecological and evolutionary implications within species and populations. Such information is crucial if predictions about the responses of natural populations to changing environments are to be made. The Australasian gannet (Morus serrator) is a generalist marine top predator that adopts a central-place foraging strategy during the breeding season. Five foraging metrics (maximum distance from the colony, bearing from the colony to the most distal point, tortuosity index, total number of dives, and mean Vectorial Dynamic Body Acceleration [VeDBA]) were obtained using GPS tracking and accelerometry data in adult birds from two colonies in south-eastern Australia with divergent oceanographic conditions. Individuals were sampled over two breeding seasons to assess factors influencing foraging behaviour and behavioural consistency over multiple timescales (consecutive trips, breeding stages and years) and habitats (pelagic and inshore).

Colony, breeding stage and year were the factors that had the greatest influence on foraging behaviour, followed by sex. Behavioural consistency within individuals, measured as the contribution of the individual to the total observed variance in foraging behaviour, was low to moderate for all foraging metrics (9.4 to 28.6%), with the higher values occurring over shorter timescales. In addition, behavioural consistency was higher in individuals foraging in inshore as compared to pelagic habitats. This study supports the notion that behavioural consistency is favoured in more stable environments and shows that foragers from the same site can present different levels of behavioural consistency depending on their preferred foraging habitat.

Introduction

Foraging is a primary activity of animals which can be highly influenced by intrinsic factors such as age, sex, or genotype (Lewis et al. 2002, Pankiw et al. 2002, McGraw et al.

2011), extrinsic factors such as geographic location, local weather, or predation risk (Holmes

1984, Staniland et al. 2004, Peat & Goulson 2005), and by reproductive constraints such as breeding stage or brood size (Soanes et al. 2014, Lewis et al. 2015). However, such factors do not necessarily affect all animals in the same way, with one or multiple influential factors potentially acting in different directions at a particular time upon individuals of the same population (Galef & Giraldeau 2001). Furthermore, as foraging is generally an activity involving relatively high rates of energy expenditure, there is strong selection for animals to develop foraging strategies to minimise their energy costs (Schoener 1971). Such strategies can include variation in foraging time and effort in particular habitats, the choice of specific search methods, and/or the choice of food types consumed (Cohen 1993). If a particular foraging strategy provides greater rewards, it is likely that this strategy will be repeated over time, favouring the development of behavioural consistency (Reader 2015). Behavioural consistency in foraging activities leads to the evolutionary development of foraging specialisation within animal populations (Dall et al. 2012), but the information on the persistence of this phenomenon over different timescales and habitats is limited.

Foraging specialisation refers to the use of a specific proportion of the full range of available resources (or foraging strategies) used by a subset of a population, resulting in inter- individual niche variation (Sargeant 2007). This phenomenon has been demonstrated in a wide variety of taxa (Bolnick et al. 2003). Foraging specialisation is thought to primarily arise in stable environments in which resources are predictable and diverse, enabling individuals to develop behavioural differences to reduce niche overlap with conspecifics and, thus, minimise competition (Araújo et al. 2011). Such behavioural consistency may,

20 therefore, have significant ecological consequences not only at the individual level but, during the breeding season, also on the development of offspring (Provencher et al. 2013).

Consequently, knowledge of temporal and spatial variation in foraging specialisations is important to fully understand their ecological implications within species (Bolnick et al.

2011, Dingemanse & Dochtermann 2013).

The marine environment is complex and dynamic, and the foraging behaviour of marine animals is highly influenced by environmental variability (Ballance et al. 2006). At a global scale, oceans display clear patterns of water circulation and climate (Macdonald &

Wunsch 1996). At local scales, physical features such as bathymetry, tidal regimes, and nutrient fluxes determine the structure of marine and coastal ecosystems and influence the behaviour of marine fauna (Butler et al. 2002). Marine environments are rich in ecosystem types and diversity which can lead to the development of a wide range of foraging techniques and/or strategies (higher ecological opportunity; Araújo et al. 2011, Yurkowski et al. 2016) even within the same species (e.g. Paiva et al. 2010). Behavioural consistency in foraging activities have been found within different animal groups in the marine environment

(Cummings & Mollaghan 2006, Matich et al. 2011), and it is expected to occur more commonly at the upper trophic levels that are regulated by bottom-up processes and often experience high levels of resource competition (Estes et al. 2003, Baylis et al. 2015).

Seabirds are important upper level predators in marine food webs (Shealer 2001,

Smith 2011). They are long-lived animals and, during the breeding season, adopt a central place foraging strategy that can lead to high levels of resource competition (Ashmole 1971,

Lewis et al. 2001). These attributes have been shown to favour the development of behavioural consistency within individuals and, combined with other factors such as age, sex, stage of the annual cycle, or breeding status, influence the development of individual behavioural differences (Phillips et al. 2017). However, the degree to which species and

21 populations develop individual behavioural consistency in foraging activities can vary (Ceia

& Ramos 2015). Intra- and inter-population differences may be related to temporal changes in resources availability, environmental stochasticity, or habitat accessibility (Hamer et al.

2007, Woo et al. 2008, Phillips et al. 2017, Machovsky-Capuska et al. 2018), but the mechanisms influencing individual foraging consistency across populations or habitats are poorly understood (Devictor et al. 2010). Such information is crucial to enable predictions about the response of natural populations to changing environments (Colles et al. 2009,

Bolnick et al. 2011).

The Australasian gannet (Morus serrator) is an important marine predator in south- eastern Australia and New Zealand (Bunce et al. 2002, Srinivasan et al. 2015), with an estimated annual consumption of 228.2 tons of schooling pelagic fish (e.g. Australian sardine

Sardinops sagax, barracouta Thyrsites atun, and blue mackerel Scomber australasicus) in

Australian waters alone (Bunce 2001). This region is one of the fastest warming oceanic areas and significant changes to ocean currents are predicted to occur (Ridgway 2007, Lough

& Hobday 2011). Such changes are likely to alter the distribution and abundance of marine species (Hobday & Pecl 2014, Pecl et al. 2014). Indeed, expansions in the species ranges of fish and invertebrates have already been documented in south-eastern Australia (Johnson et al. 2011). Therefore, knowledge of the factors influencing foraging activity and behavioural consistency in Australasian gannets is necessary to predict how their populations may respond to anticipated climatic changes.

Like other members of the Sulidae family, the Australasian gannet is considered a generalist forager and has been shown to be adaptable in its feeding habits (Bunce 2001,

Schuckard et al. 2012, Machovsky-Capuska et al. 2016a, Machovsky-Capuska et al. 2018).

Colony-specific reverse sexual-dimorphism (females larger than males) has been found on the study sites (Angel et al. 2015b), and recent studies suggest individuals exhibit sex-related

22 differences in habitat use (Wells et al. 2016, Besel et al. 2018) and prey selection

(Machovsky-Capuska et al. 2016a). Furthermore, inter-colony differences in diving behaviour and habitat selection have been documented (Machovsky-Capuska et al. 2014b,

Angel et al. 2016). Consequently, it can be expected that the contrasting oceanographic regimes adjacent to the study sites and the sex of individuals could potentially influence the foraging behaviour of birds within this study. However, little is known of other factors influencing the foraging behaviour and behavioural consistency of individuals in this species.

The aims of the present study, therefore, were to investigate in Australasian gannets: 1) the factors influencing foraging behaviour; 2) the degree of behavioural consistency in foraging activities and; 3) how individual foraging consistency varies over different timescales and foraging habitats.

Materials and methods

Study sites and animal handling

The study was conducted at two Australasian gannet breeding colonies in northern

Bass Strait (BS), south-eastern Australia, that experience divergent oceanographic conditions and may present differences in resource availability or habitat accessibility (Fig. 2.1). Point

Danger (PD, 38º 23’ 36.09” S, 141º 38’ 55.94” E) is located at the western edge of BS near the seasonally active (austral summer) Bonney Upwelling, an important source of primary productivity for the Bass Strait region (Lewis 1981, Butler et al. 2002). Previous studies have shown that individuals from this colony range up to 238 km north-west and south-east, remaining over the narrow (~40 km wide) continental shelf to forage on schooling fish and cephalopods (Butler et al. 2002, Angel et al. 2016). Bass Strait is a pelagic habitat with constant input from three major water masses (Middleton & Bye 2007), where the local

23 distribution of fish prey species is influenced by sea surface temperature regimes (Hoskins AJ et al. 2008, Kirkwood et al. 2008).

Pope’s Eye (PE; 38º 16’ 35.88” S, 144º 41’ 56.21” E) is located at the entrance of Port

Phillip Bay (PPB) on an artificial structure. Previous studies have shown that individuals from this colony forage within the shallow waters of PPB (average depth < 13.6 m; Walker,

1999) primarily on benthic/demersal fish, outside of PPB within northern BS on schooling fish and cephalopods, or in both habitats (Angel et al. 2016, Wells et al. 2016). The PPB area is an important site for fish reproduction (Parry et al. 1995, Jenkins et al. 1997a), and a nursery site for some pelagic species such as Australian anchovy (Engraulis australis) and

Australian sardine (Sardinops sagax; Neira et al. 1999, Dimmlich et al. 2004) . The main habitat types within the PPB are reefs, seagrasses, and soft sediments (Parry et al. 1995), and it has been repeatedly reported as a reliable source of food for marine predators (Chiaradia et al. 2002, McCutcheon et al. 2011, Filby 2016).

Fig. 2.1: Location of study sites: Point Danger (left) and Pope’s Eye (right). The 200 m bathymetric contour is given to indicate the edge of the continental shelf.

Data were collected during the 2014/15 and 2015/16 breeding seasons (October-

March) in each of three breeding stages: incubation; early chick-rearing (chick age 0-50 d); and late chick-rearing (chick age >50 d; Wingham 1982). Individuals were captured at the nest by hand or with the aid of a noose-pole (Garthe et al. 2014) and weighed in a cloth bag with a suspension scale (± 25 g, Salter Australia Pty Ltd, Australia). A GPS data logger

(programmed to record location every 2 min; I-gotU GT-600, Mobile Action Technologies

Inc., Taiwan; ±10 m error) and a tri‐axis accelerometer data logger (sampling rate of 25 Hz;

X8M-3mini, Gulf Coast Data Concepts LLC, USA), encapsulated together in heat shrink plastic (total package 53.7 g, <3% body mass), were then attached to central tail feathers using water-proof tape (Tesa® 4651, Beiersdorf AG, Germany). Individuals were then returned to the nest and resumed natural behaviours within 10 min of capture.

After 10-12 d, individuals were recaptured as previously described and the data loggers were removed by peeling the tape from the feathers. Body mass was again recorded and morphometric measures of culmen length and bill depth, and tarsus length and ulna length, were recorded using Vernier callipers (± 0.1 mm) and a metal ruler (± 1 mm), respectively. A blood sample (0.1 mL) was then obtained by venipuncture of a tarsal vein for genetic sexing (DNA Solutions, Wantirna South, VIC, Australia) before the bird was returned to the nest. Where possible, both partners on nests were tracked simultaneously or within

< 5 d of each other, and individuals were sampled in multiple breeding stages and across years.

Data processing and statistical analysis

Except when mentioned, all data processing and statistical analyses were conducted in

R version 3.3.2 (R Core Team 2017). Deployment data were split into individual foraging trips by visual inspection of the raw GPS tracks. Trips were then filtered using the McConnell

26 et al. (1992) speed filter to remove erroneous locations in the trip package (Sumner 2016), applying a maximum average speed of 55 km·h-1 as suggested by Hamer et al. (2007) and

Grémillet et al. (2004) for northern and Cape gannets (M. bassanus & M. capensis) respectively, after detailed studies of travel speeds during foraging trips. Subsequently, for each foraging trip, the behavioural metrics of maximum distance from the colony (km), total distance travelled (km), and bearing (0-360º, from the colony to the most distal point) were calculated using the adehabitatHR package (Calenge 2006). A tortuosity index, a measure of an animal’s searching behaviour, was also estimated by dividing the maximum distance reached from the colony with the total distance travelled during the trip (Benhamou 2004,

Calenge et al. 2009).

At-sea behaviours throughout the foraging trip were inferred from the tri-axis accelerometer data loggers. Data were initially inspected visually to assign foraging behaviours (plunge diving and surface foraging; Warwick-Evans et al. 2015) in IGOR Pro

(Version 6.37, WaveMetrics, USA), based on the acceleration profiles given by Ropert-

Coudert et al. (2004), Weimerskirch et al. (2005b), and Ropert-Coudert et al. (2009) for other species of gannets and boobies. The Ethographer package was then used to perform a k- means clustering analysis to identify these behaviours using an unsupervised continuous wavelet transformation (1 s window), following Sakamoto et al. (2009). From these data, the total number of dives (plunge diving and surface foraging) was estimated for each foraging trip. In addition, the accelerometry data were used to calculate the average Vectorial

Dynamic Body Acceleration (VeDBA) for each trip as a proxy for the energy expenditure rate per activity which allows comparison of the rate of energy expenditure across foraging trips (Gleiss et al. 2011, Qasem et al. 2012, Angel 2015, Angel et al. 2015a).

A Body Condition Index (BCI) was calculated for each bird at each deployment, as a proxy for total body fat (%) content, using body mass (kg), wing ulna (mm), and tarsus (mm)

27 measurements (Angel et al. 2015b). As in Australasian gannets there are suggestions of differences in size between adult females and males of this species, body size indices were calculated to investigate the effect of size on foraging behaviour independent of sex. A Body

Size Index (BSI) and Wing Length Index (WLI) were estimated using the deviation of each individual’s body mass (kg) and wing length (mm) from the means for their respective sex.

To analyse the factors influencing the foraging behaviour of instrumented individuals, linear mixed effects models were created using the nlme package (Pinheiro et al. 2014). The foraging metrics (maximum distance from the colony, average bearing, and tortuosity index), the total number of dives, and mean VeDBA were used separately as response variables.

Fixed factors such as colony (PD, PE), year (2014-15, 2015-16), breeding stage (incubation, early chick-rearing, late chick-rearing), and sex (male, female), were used as explanatory variables in combination with the BCI, BSI, and WLI. Because a focus of the present study was to investigate broad scale temporal and geographic influences on foraging behaviour and consistency, specific environmental variables were not included in any analyses. Rather, influences of colony and year were considered to reflect the differences in resource availability and environmental variation, respectively.

Where appropriate, variables were cube-root-transformed to fit model assumptions of constant variance and normal distribution (Cox 2011). Model assumptions were checked by plotting residuals and using quantile-quantile plots. Following modelling recommendations by Zuur et al. (2010), collinearity among all the explanatory variables was checked before conducting each model using pairplots, boxplots, and the Variance Inflation Factor (cutting value used = 2). The initial models were then fitted with restricted maximum likelihood

(REML), and models with and without the random structure (nest identity, due to the use of nest partners, and individual identity) were compared using the anova function. Variance structure for the explanatory variables was included when the residuals inspection suggested

28 it was necessary. The best fixed structure was found using the dredge function of the MuMIn package based on the AICc (Barton 2016), using models refitted with maximum likelihood

(ML). Where multiple models had ΔAICc ≤ 4 and no single model had an AICc weight above

0.90, model averaging was used to calculate the relative importance of each explanatory variable using the MuMIn package (Burnham & Anderson 2002, Symonds & Moussalli 2011,

Barton 2016).

To quantify the magnitude of the individual behavioural consistency in each metric of foraging behaviour, variance component analyses were conducted using the models containing the parameters defined as influential after model averaging. The ape package

(Paradis et al. 2004) was used to calculate the variance, standard deviation, and the proportion of total variance occurring at the individual level, as well as the residual variation.

The variance explained by the individual is considered an estimate of the individual specialisation within a population (Bolnick et al. 2003, Dingemanse & Dochtermann 2013).

A second set of models using the coefficients of variation of the foraging metrics (standard deviation in the case of the bearing as it is a circular variable), calculated per deployment and used as response variable were then developed to investigate the factors influencing individual variation (Camprasse et al. 2017c). The same set of explanatory variables and modelling recommendations as previously described were used.

As multiple logger deployments were performed on most individuals (mean ± SE: 1.9

± 0.1 deployments per bird), the dataset acquired allowed comparisons at different timescales: trip-to-trip (T-to-T, data from consecutive trips obtained within the same deployment), breeding stage-to-breeding stage (S-to-S, data obtained from different breeding stages within the same year), and year-to-year (Y-to-Y, data obtained from the same breeding stage in two different years). This partition allowed the assessment of the timescales over which individual behavioural consistency was maintained. The dataset acquired was then

29 modified to match each timescale tested, using in each case three or more foraging trips per deployment.

For the T-to-T comparison, the dataset acquired was divided into colonies (PD and

PE) to quantify the level of individual behavioural consistency separately at each site. The

PD data was analysed in their entirety, reflecting the relatively uniform foraging habitat used by these individuals, whereas the PE data was split according to the predominant habitat individuals foraged in. Individuals that spent >70% of trips during a deployment in PPB or

BS were classified as PE-inshore and PE-pelagic, respectively, while individuals that displayed no preference were defined as PE-mixed. This partition allowed the assessment of the habitats, related to the sites, over which individual behavioural consistency was maintained. Models based on these further divisions of the dataset were made using the same set of response and explanatory variables and followed all considerations described previously. Unless otherwise stated, results are presented as mean ± SE.

Results

Factors influencing foraging behaviour

The GPS and accelerometry data loggers were deployed on 142 adult birds (260 deployments obtained through longitudinal sampling) from which data on 3-50 foraging trips were obtained (18.10 ± 0.86 trips per individual). From the GPS data loggers, a total of 2,493 foraging trips were recorded but, due to battery life restrictions, accelerometry data were recorded for only 1,284 trips. Consequently, the sample size of the different foraging metrics measured varied depending on the device from which these were derived. A summary of the calculated foraging metrics is presented in Table 2.1.

After model averaging, the best explanatory variables for the response variable of maximum distances from the colony were colony, breeding stage, year, and sex. Namely, individuals from PD, individuals during incubation, individuals during 2015, and females commuted greater distances from the colony during foraging trips. For the response variables of bearing and tortuosity index, colony, breeding stage, and year were the most influential explanatory variables. Specifically, individuals from PD, individuals during late chick-rearing and individuals during 2015 showed westerly bearings and higher tortuosity indices. Lastly, the response variables mean VeDBA and number of dives per foraging trip were both influenced the most by the explanatory variables breeding stage and sex. Individuals during early chick-rearing and males had higher VeDBA values during foraging trips, and individuals during incubation and females displayed a higher number of dives during foraging trips (Tables 2.2, S2.1 and S2.2).

Table 2.1: Means ± SE of the foraging trip parameters collected from instrumented Australasian gannets (Morus serrator) during two years 2014

and 2015 at Point Danger (PD) and Pope’s Eye (PE) colonies in Victoria, Australia. Instrumentation on birds lasted 10-12 consecutive days. Data

represent the dataset acquired separated by sex, breeding stages (incubation = INC, early chick-rearing = ECR, late chick-rearing = LCR), year,

and colony-habitat. Samples sizes of the metrics estimated from the GPS (maximum distances from the colony, bearing and tortuosity index; n1)

and accelerometer data loggers (mean Vectorial Dynamic Body Acceleration and number of dives, n2), are shown.

n1 Distances from colony (km) Bearing (º) Tortuosity index n2 Mean VeDBA (g) Number of dives

Sex ♂ 1,424 60.41 ± 1.76 192.5 ± 2.0 0.29 ± 0.01 666 0.63 ± 0.21 345.7 ± 11.8

♀ 1,056 77.22 ± 1.98 201.4 ± 2.0 0.30 ± 0.01 618 0.57 ± 0.17 390.6 ± 12.8

Breeding INC 452 95.47 ± 4.08 186.2 ± 4.1 0.26 ± 0.01 205 0.54 ± 0.16 584.0 ± 28.4 stage ECR 1,143 65.58 ± 1.78 193.9 ± 2.0 0.30 ± 0.01 619 0.61 ± 0.19 313.6 ± 9.5

LCR 885 55.88 ± 1.90 204.6 ± 2.3 0.30 ± 0.01 460 0.61 ± 0.20 342.8 ± 14.6

Year 2014 1,224 64.18 ± 2.09 187.8 ± 2.2 0.27 ± 0.01 390 0.62 ± 0.21 398.1 ± 19.4

2015 1,256 70.86 ± 1.64 204.8 ± 1.9 0.31 ± 0.01 894 0.59 ± 0.19 353.7 ± 9.2

Colony- PD–pelagic 1,162 93.31 ± 2.32 234.1 ± 1.9 0.32 ± 0.01 686 0.59 ± 0.18 369.8 ± 12.5 habitat PE–pelagic 699 56.77 ± 1.61 181.0 ± 1.6 0.29 ± 0.01 375 0.58 ± 0.17 369.1 ± 14.8

PE–mixed 355 39.46 ± 1.75 150.8 ± 4.1 0.26 ± 0.01 145 0.62 ± 0.22 401.4 ± 26.2

PE–inshore 264 20.63 ± 1.68 131.1 ± 3.6 0.22 ± 0.01 78 0.74 ± 0.29 273.4 ± 29.6

Table 2.2: Most parsimonious models after model averaging, and their corresponding estimated regression parameters, for five foraging metrics obtain from free-ranging Australasian gannets (Morus serrator).

Foraging metric Model fixed effects Fixed effect Estimate SE df t-value P-value

Distance from colony (km) colony + stage + year + sex (Intercept) 4.25 0.09 2340 47.58 <0.0001 Colony (PE) -0.79 0.09 134 -8.65 <0.0001 Stage (INC) 0.51 0.06 2340 8.18 <0.0001 Stage (LCR) -0.21 0.06 2340 -3.48 0.0005 Sex (male) -0.32 0.09 134 -3.54 0.0005 Year (2015) 0.13 0.04 2340 3.08 0.002 Bearing (º) colony + stage + year (Intercept) 226.72 4.71 2353 48.15 <0.0001 Colony (PE) -71.88 6.22 135 -11.55 <0.0001 Stage (INC) -16.91 3.98 2353 -4.24 <0.0001 Stage (LCR) 11.73 3.81 2353 3.07 0.002 Year (2015) 14.82 2.97 2353 4.97 <0.0001 Tortuosity Index colony + stage + year (Intercept) 0.31 0.005 2353 66.70 <0.0001 Colony (PE) -0.05 0.006 135 -8.18 <0.0001 Stage (INC) -0.05 0.005 2353 -10.07 <0.0001 Stage (LCR) 0.01 0.005 2353 -0.55 0.58 Year (2015) 0.04 0.004 2353 11.20 <0.0001

Mean VeDBA stage + sex (Intercept) 0.83 0.01 1175 140.90 <0.0001 Stage (INC) -0.04 0.01 1175 -5.55 <0.0001 Stage (LCR) -0.01 0.01 1175 -0.07 0.95 Sex (male) 0.03 0.01 108 3.70 0.0003 Number of dives stage + sex (Intercept) 6.62 0.14 1127 47.76 <0.0001 Stage (INC) 1.48 0.16 1127 9.12 <0.0001 Stage (LCR) 0.19 0.14 1127 1.40 0.16 Sex (male) -0.36 0.17 106 -2.04 0.04

Influence of timescales and habitats on individual consistency of foraging behaviours

Variance component analyses were performed to determine the proportion of variance explained by the individual for each of five foraging metrics. As 86% (n=226) of the deployments were conducted simultaneously on nest partners, nest identity was tested during the modelling as a random component. However, in no model was nest identity a significant factor (P > 0.05 in all cases). Conversely, the individual random component was significant

(P < 0.05) in all but two of the models developed. For short-term comparisons (T-to-T comparison level), the variance associated with the individual component ranged from low to moderate (11.07 to 27.05%) and consistently decreased as the timescale comparison increased to mid-term (S-to-S: 9.45 to 22.9%) and long-term (Y-to-Y: 0 to 28.63%, Table

2.3).

For PE, 127 of the 260 deployments were classified according to the predominant habitat in which each bird foraged. From them, 70 were classified as PE-pelagic, 33 as PE- mixed and 24 as PE-inshore (see examples in Fig. 2.2). The proportion of females for each classification was 0.69, 0.27, and 0.04, respectively, with relatively higher representation of males in PE-inshore and PE-mixed habitats. The proportion of variance explained by the individual between habitats ranged from low to moderate values (3.17 to 50.37%). In general,

PD-pelagic and PE-pelagic showed similar consistency values in bearing, tortuosity index, and mean VeDBA, but differed in their maximum distance from the colony and number of dives in which PD-pelagic presented higher and lower values than PE-pelagic, respectively.

With the exception of bearing to most distal point and mean VeDBA, the PE-mixed individual variation values were higher than their PE-pelagic counterpart (9.73 to 26.03%), while variation values for the PE-inshore individuals were even higher (12.21 to 50.37%,

Table 2.4).

Table 2.3: Variance component analysis (variance [σ2], standard deviation [Σ] and the proportion of total variance occurring at the levels of individual [σ2%]) of instrumented Australasian gannets (Morus serrator). Short-term (Trip-to-Trip), medium-term (Stage-to-Stage) and long- term (Year-to-Year) comparisons are shown. Sample sizes (number of trips/number of individuals) are presented for each final model. The significant fixed components of the models in which the coefficients of variation were used as a response variable are shown as the factors influencing the individual variation on each case.

Foraging trip parameter Timescale σ2 Σ σ2% n Influences on individual variation Distance from colony (km)* T-to-T 0.21 0.46 27.05 2480/137 Colony, year S-to-S 0.14 0.37 13.00 1166/56 Colony, breeding stage Y-to-Y 0.166 0.40 16.64 1069/53 Sex Bearing (º) T-to-T 1090.48 33.02 26.32 2490/137 Breeding stage, year S-to-S 944.32 30.72 22.90 1184/57 Breeding stage Y-to-Y 1195.88 34.58 28.63 1069/53 None Tortuosity index T-to-T 0.0006 0.026 11.07 2490/137 Colony, sex, year, WLI S-to-S 0.0005 0.024 9.45 1184/57 Colony, BSI, sex Y-to-Y 0.0005 0.023 8.70 1069/53 WLI, year, sex Mean VeDBA (g)* T-to-T 0.001 0.005 16.91 1237/108 Stage, sex S-to-S 0.001 0.005 14.66 641/51 None Y-to-Y - - 0 190/15 NA Number of dives* T-to-T 0.57 0.75 18.89 1237/108 Colony, breeding stage, year S-to-S 0.55 0.74 16.25 641/51 None Y-to-Y - - 0 190/15 NA *Transformed variable

Fig. 2.2: Examples of the time-scale comparisons analysed: T-to-T (Trip-to-Trip, left column); S-to-S (Stage-to-Stage, middle column); and Y-to-Y (Year-to-Year, right column).

Each row represents one instrumented bird: first row, male from PD-pelagic; second row, a

PE-pelagic female; third row, a PE-mixed male; and fourth row, a PE-inshore male. Breeding stages include incubation (INC), early chick-rearing (ECR), and late chick-rearing (LCR).

Table 2.4: Variance component analysis (variance [σ2], standard deviation [Σ] and the proportion of total variance occurring at the levels of individual [σ2%]) of instrumented

Australasian gannets (Morus serrator). Short timescale comparison (Trip-to-Trip) results are shown for the models split by colony and habitat. Sample sizes (number of trips/number of individuals) are presented for each final model.

Foraging trip parameter Colony-habitat σ2 Σ σ2% n Distances from colony (km)* PD-pelagic 0.24 0.49 15.93 1170/76 PE-pelagic 0.019 0.13 3.17 704/41 PE-mixed 0.227 0.47 20.22 355/25 PE-inshore 0.28 0.53 50.37 264/16 Bearing (º) PD-pelagic 1171.45 34.22 26.32 1170/76 PE-pelagic 443.88 21.06 23.48 704/41 PE-mixed 1032.96 32.13 14.74 355/25 PE-inshore 2200.57 46.91 49.18 264/16 Tortuosity index PD-pelagic 0.0005 0.02 8.92 1170/76 PE-pelagic 0.0004 0.02 8.53 704/41 PE-mixed 0.0015 0.038 10.84 355/25 PE-inshore 0.0010 0.031 12.21 264/16 Mean VeDBA (g)* PD-pelagic 0.001 0.03 14.81 668/54 PE-pelagic 0.001 0.03 15.48 353/36 PE-mixed 0.001 0.02 9.73 149/18 PE-inshore 0.01 0.04 12.56 76/10 Number of dives* PD-pelagic 0.24 0.49 8.00 668/54 PE-pelagic 0.49 0.70 20.17 353/36 PE-mixed 0.98 0.99 26.03 149/18 PE-inshore 0.96 0.98 41.62 76/10 *Transformed variable

Using the coefficient of variation (or the standard deviation) of foraging metrics within deployments as a measure of an individual’s consistency, the factors influencing individual variation were investigated. The five foraging metrics examined (maximum distance from the colony, bearing from the colony to the most distal point, tortuosity index, mean VeDBA, and total number of dives) required model averaging due to the lack of a single best model from the candidate set of models. After model averaging, the most influential factors affecting individual variation were year, colony, and breeding stage for the

T-to-T comparison level, with individuals sampled during year 2014, at PD, and in the late chick-rearing stage having the highest consistency levels (Table S2.3).

Discussion

In the present study, the foraging behaviour and behavioural consistency of individual

Australasian gannets were quantified across different timescales and foraging habitats.

Foraging metrics were influenced by colony, breeding stage, and year. Consistency, expressed by the proportion of variation explained by the individual, was higher over shorter

(T-to-T) than longer (Y-to-Y) timescales, consistent with previous studies investigating the persistence of behavioural consistency in seabirds (Harris et al. 2014, Camprasse et al.

2017c), and the repeatability of behaviours in several taxa (Bell et al. 2009). In addition, individual consistency in foraging behaviour was found to be higher for birds foraging in inshore habitats compared to those foraging in pelagic habitats, supporting the hypothesis that consistency is favoured in more stable environments (Wolf & Weissing 2012, Harris et al.

2014). As the focus of the study was to investigate broad-scale temporal and geographic influences on foraging behaviour and consistency, and the contrasting oceanographic regimes around the study colonies and inter-annual variation have been shown to reflect fluctuations

39 in prey abundance (Angel et al. 2015a, Angel et al. 2016), specific environmental variables were not analysed as explanatory variables.

Factors influencing foraging behaviour

The use of metrics in foraging ecology research to detect and quantify behavioural consistency in movement and foraging strategies of marine predators, particularly seabirds and marine mammals, is a common practice (Carneiro et al. 2017). For the foraging metrics analysed in the present study (i.e. maximum distance from the colony, bearing, tortuosity index, mean VeDBA, and number of dives), colony, year, and breeding stage were found to have the most consistent influence, while sex and BSI were less influential. Geographic variation in foraging behaviour has previously been reported in gannets (Moseley et al. 2012,

Pettex et al. 2012, Machovsky-Capuska et al. 2014b, Angel et al. 2016) and other marine predators, reflecting spatial differences in resource availability or habitat accessibility (Falk et al. 2002, Baylis et al. 2008, Phillips et al. 2017). The results of the present study are consistent with these findings and are probably the result of the substantial differences in oceanographic regimes and habitats available to individuals from the PD and PE gannet colonies (Angel et al. 2016, Wells et al. 2016). In particular, the individuals from PD, which forage within the Bonney Upwelling System, had longer foraging trips and higher tortuosity index than individuals sampled at PE. Similar differences have previously been reported before for these two colonies (Angel et al. 2016), suggesting significant differences in foraging conditions encountered at each location.

Year of sampling was an important factor influencing the foraging behaviour of

Australasian gannets, with individuals travelling less, having a lower tortuosity index, higher energy expenditure rate, and diving more often during the 2014 breeding season compared to the 2015 breeding season. Interestingly, breeding success (proportion of chicks fledged) was

40 significantly lower in 2014 (25% and 48%) than in 2015 (50% and 79%) at PE and PD, respectively (Rodríguez-Malagón unpublished. data). This suggests that both sites experienced similar environmental variation influencing both foraging behaviour and reproductive success to the same degree. Previous studies at PE have reported an increased foraging effort in years of low local marine productivity (Angel et al. 2015a, Machovsky-

Capuska et al. 2018), and inter-annual variation in foraging behaviour in response to environmental perturbations have been observed in Australasian gannets (Machovsky-

Capuska et al. 2018), and other gannet species (Moseley et al. 2012, Kai et al. 2013, Cleasby et al. 2015). Indeed, primary productivity (as measured by chlorophyll-a concentration) was substantially higher in 2014 than in 2015 (Evans et al. 2017), coinciding with a strong El

Niño/Southern Oscillation (ENSO) event with sea surface temperatures above average

(www.bom.gov.au) in 2015.

Temporal variation in foraging metrics was also evident in relation to breeding stage.

In particular, individuals conducted longer foraging trips, had a higher tortuosity index, lower energy consumption rate, and dived more during incubation than during either chick-rearing stage. Greater foraging trip duration and average speed during incubation compared to chick- rearing have previously been reported for Australasian gannets (Bunce 2001, Angel et al.

2015a). Similar observations have been made in other seabirds and have been considered to reflect a shift from self-feeding during incubation to chick-provisioning during chick-rearing

(Phillips et al. 2017). However, other studies have related changes in foraging behaviour between incubation and chick-rearing to be a response to temporal variation in prey availability due to environmental changes around colonies throughout the breeding period

(Ito et al. 2009, Jakubas et al. 2014, Shoji et al. 2016). While several studies have analysed the diet (Pyk et al. 2007, Schuckard et al. 2012) and the nutrient intake (Machovsky-Capuska et al. 2016a, Machovsky-Capuska et al. 2018) of Australasian gannets during the breeding

41 season, none has reported significant dietary shifts that could be related to the environmental variability within the breeding period. Thus, the reason for differences in the foraging behaviour of Australasian gannets observed in relation to their breeding stage still remains to be established.

Maximum distance from the colony and the number of dives were shown to be influenced by sex, with females exhibiting higher values than males. Previous research at the two study colonies has shown sex differences in core foraging areas, with only 4.2% and

18.4% of overlap between sexes at PD and PE, respectively (Angel et al. 2016). Australasian gannets have been previously reported to present reverse sexual dimorphism within these two breeding colonies, with females being significantly heavier and larger than males (Angel et al. 2015b). These findings are consistent with observations in other Sulidae species with reserve sexual dimorphism such as the red-footed booby (Sula sula) in which males forage closer inshore than females, the larger sex (Weimerskirch et al. 2006). In species with sexual size dimorphism, trophic or spatial segregation can function to reduce intra-specific competition, particularly during periods of intense resource competition (Cleasby et al. 2015,

Phillips et al. 2017). Surprisingly, despite the greater foraging range and higher dive rate, females in the present study had lower mean VeDBA. This suggests females may be more efficient in some aspects of their foraging behaviour. Indeed, females from the study colonies have been previously reported to spend a greater proportion of their foraging trips in gliding rather than flapping flight (Angel et al. 2015a, Angel et al. 2016).

Using animal-borne cameras and GPS data loggers on single trips, previous research found individuals from PE displayed habitat-specific behaviour (Wells et al. 2016). In the pelagic environments of Bass Strait, individuals displayed group foraging typical of gannets

(Tremblay et al. 2014), while within Port Phillip Bay individuals displayed solitary (in the absence of conspecifics and heterospecifics) foraging along shallow sandbanks and

42 shorelines. In the present study, sampling over multiple trips revealed some individuals from

PE displayed both behaviours. Similar to the findings of Wells et al. (2016), the overwhelming proportion (96%) of individuals adopting the inshore (and mixed) foraging strategy were males.

Influence of timescales and habitats on individual consistency of foraging behaviours

Moderate levels of behavioural consistency (>20%; Bell et al. 2009) in maximum distances from the colony and bearing were observed over the short-term (T-to-T) by individuals in the present study. These results suggest Australasian gannets from the two study colonies displayed some degree of foraging-site fidelity, potentially exploiting the same resource patches, over several days. However, low levels of short-term consistency (<20%) were observed in tortuosity index, mean VeDBA, and number of dives, indicating these foraging metrics may be more influenced by current environmental conditions rather than individual foraging strategies or abilities (Bell et al. 2009).

Behavioural consistency in foraging implies individuals learn, remember, and select specific resources and foraging strategies (Hamer et al. 2001). It requires reliability in the abundance and location of the exploited resources so that the strategies can be maintained in the population (Devictor et al. 2010). However, environmental variability can lead to fluctuations in exploited resources, influencing the level and permanence of behavioural consistency over time (Araújo et al. 2011). All the foraging metrics examined in the present study, with the exception of bearing to the most distal location, decreased in consistency with increasing timescales. Similar findings have been made in other seabirds (Harris et al. 2014,

Camprasse et al. 2017c), emphasising that behavioural consistency can only persist as long as stability in environmental conditions prevails (Ceia et al. 2014). Importantly, as seabird foraging conditions are highly susceptible to fluctuations in the environment (Schreiber &

Burger 2001, Ballance et al. 2006, Bost et al. 2009), it is likely advantageous for individuals to maintain a certain level of behavioural plasticity in order to respond to such changes

(Wakefield et al. 2015).

In contrast to the other foraging metrics, bearing to the most distal location had a moderate level of individual consistency through all timescales analysed (23-28%) at both study sites, suggesting fidelity to broad scale foraging locations. High consistency in departure directions from the colony over multiple timescales has been previously shown in

Australasian gannets (Machovsky-Capuska et al. 2014a), northern gannets (Wakefield et al.

2015), and other seabirds (Phillips et al. 2017). These findings suggest a degree of spatial and temporal resource predictability that birds are able to remember and exploit based on directional references (Harris et al. 2014).

Individual consistency in foraging behaviour is thought to be promoted in stable environments (Weimerskirch 2007, Wolf & Weissing 2012, Patrick & Weimerskirch 2014).

Inshore environments are considered refuges for fish and marine invertebrate communities, as they provide stable habitats and nutrients (Jenkins et al. 1997b, Sampson et al. 2014, Griffiths et al. 2017). Port Phillip Bay is a shallow inshore environment with extensive coverage of seagrass beds and sandy bottoms that represent important habitats for marine invertebrates and fish in south-eastern Australia (Blandon & zu Ermgassen 2014). Correspondingly, in the present study the highest behavioural consistency values were found within the foraging metrics of the birds that fed in this habitat (PE-inshore) in comparison to the other two foraging strategies observed in individuals from PE. High behavioural consistency in foraging site and distance travelled has also been observed in other seabirds foraging in benthic habitats, such as cormorants and shags (Phalacrocorax georgianus, P. atriceps P. verrucosus; Ratcliffe et al. 2013, Harris et al. 2014, Camprasse et al. 2017a), and is thought to be due to the ability of individuals to memorise bathymetric features as cues for resource

44 availability and aids for navigation (Phillips et al. 2017). Indeed, Australasian gannets foraging within Port Phillip Bay have been previously reported targeting bathymetric features such as shallow sand banks and coastlines, potentially using prey silhouettes as hunting cues

(Wells et al. 2016).

In contrast, individuals in the present study that foraged in pelagic habitats (PD- pelagic and PE-pelagic) are likely to exploit schooling fish (Wells et al. 2016), a temporally and spatially variable prey resource. Bass Strait, where these individuals foraged, is influenced by numerous oceanographic processes such as the Bonney Upwelling System and three major currents (the Subantarctic Surface Water from the south, the South Australian

Current from the northwest and the East Australia Current from the northeast; Middleton &

Bye 2007). These oceanographic features are subject to both intra- and inter-annual variations in strength that affect the abundance, diversity, and distribution of prey species within the region (Sandery & Kämpf 2007). Correspondingly, individuals foraging in this habitat displayed lower consistency values in their foraging metrics.

Finally, some individuals from PE in the present study adopted a strategy of consistently foraging in both pelagic (BS) and inshore (PPB) habitats, either within the same or successive foraging trips, suggesting a degree of behavioural versatility. Similar findings have been reported for Gentoo penguins (Pygoscelis papua) in which some individuals switched between pelagic and benthic strategies on successive foraging trips (Camprasse et al. 2017a). While it is not known whether this mixed foraging strategy has specific benefits, it has been suggested that spatial and temporal environmental variation and resource competition can promote different adaptive responses in individuals, giving rise to different levels of behavioural plasticity (Wolf et al. 2008). South-eastern Australia is one of the fastest warming marine regions in the world and the anticipated oceanographic changes are likely to strongly affect the distribution, abundance, and diversity of prey species (Ridgway 2007,

Lough & Hobday 2011). Inter-individual differences in foraging behaviour and behavioural plasticity in Australasian gannets, therefore, could affect how the population responds to changing environmental conditions (Dingemanse & Wolf 2013).

In summary, the present study found foraging behaviour in Australasian gannets to be influenced primarily by colony, breeding stage and year (reflecting spatial and temporal variation in resources around breeding colonies) and, to a lesser degree, by sex. Moderate levels of behavioural consistency were observed at short timescales, but decreasing with increasing timescales; behavioural consistency was also higher in more stable inshore environments. These findings have important implications for population dynamics as individuals will not be uniformly affected by environmental variability. Future studies should further investigate the links between specific environmental conditions and consistency of foraging behaviour, and whether individual foraging strategies confer specific reproductive advantages.

Supporting information

Table S2.1: AICc based model selection (∆ < 4) for factors that influence the foraging metrics

in adult free-ranging Australasian gannets (Morus serrator). Model variables: BCI: Body

Condition Index; BSI: Body Size Index; WLI: Wing Length Index; cln: colony; stage: breeding

stage; yr: year; sex.

Foraging Model fixed effects df AICc ∆AIC AIC metric Weight Maximum WLI + cln + stage + yr + sex 12 6723.60 0 0.24 distance from cln + stage + yr + sex 11 6724.10 0.50 0.19 the colony BCI + cln + stage + yr + sex 12 6724.50 0.92 0.15 (km) BCI + WLI + cln + stage + yr + sex 13 6724.80 1.25 0.13 BSI + WLI + cln + stage + yr + sex 13 6725.60 2.02 0.09 BSI + cln + stage + yr + sex 12 6726.00 2.44 0.07 BCI + BSI + cln + stage + yr + sex 13 6726.50 2.91 0.06 BCI + BSI + WLI + cln + stage + yr + sex 14 6726.80 3.27 0.05 Bearing (º) BSI + cln + stage + yr 8 27346.30 0 0.16 cln + stage + yr 7 27346.50 0.25 0.14 BCI + cln + stage + yr 8 27347.10 0.80 0.11 BCI + BSI + cln + stage + yr 9 27347.10 0.83 0.11 BSI + WLI + cln + stage + yr 9 27348.20 1.89 0.06 WLI + cln + stage + yr 8 27348.20 1.92 0.06 BSI + cln + stage + yr + sex 9 27348.20 1.98 0.06 cln + stage + yr + sex 8 27348.50 2.23 0.05 BCI + WLI + cln + stage + yr 9 27349.00 2.74 0.04 BCI + cln + stage + yr + sex 9 27349.00 2.78 0.04 BCI + BSI + cln + stage + yr + sex 10 27349.10 2.82 0.04 BCI + BSI + WLI + cln + stage + yr 10 27349.10 2.84 0.04 BSI + WLI + cln + stage + yr + sex 10 27350.10 3.88 0.02 WLI + cln + stage + yr + sex 9 27350.20 3.91 0.02

Tortuosity cln + stage + yr 7 -5741.90 0 0.20 Index WLI + cln + stage + yr 8 -5741.70 0.28 0.18 BCI + cln + stage + yr 8 -5740.20 1.79 0.08 cln + stage + yr + sex 8 -5740.00 1.95 0.08 BSI + cln + stage + yr 8 -5739.90 2.01 0.08 WLI + cln + stage + yr + sex 9 -5739.70 2.22 0.07 BSI + WLI + cln + stage + yr 9 -5739.70 2.25 0.07 BCI + WLI + cln + stage + yr 9 -5739.70 2.28 0.07 BCI + cln + stage + yr + sex 9 -5738.20 3.74 0.03 BCI + BSI + cln + stage + yr 9 -5738.10 3.80 0.03 BSI + cln + stage + yr + sex 9 -5738.00 3.96 0.03 Mean stage + sex 7 -2754 0 0.14 VeDBA cln + stage + sex 8 -2753.9 0.14 0.13 stage + year + sex 8 -2752.1 1.93 0.05 stage + BSI + sex 8 -2752 2.02 0.05 stage + WLI + sex 8 -2752 2.02 0.05 BCI + stage + sex 8 -2752 2.02 0.05 cln + stage + year + sex 9 -2751.9 2.06 0.05 BCI + cln + stage + sex 9 -2751.9 2.12 0.05 cln + stage + BSI + sex 9 -2751.8 2.16 0.05 cln + stage + WLI + sex 9 -2751.8 2.17 0.05 stage + BSI + year + sex 9 -2750.1 3.94 0.02 BCI + stage + year + sex 9 -2750 3.95 0.02 stage + WLI + year + sex 9 -2750 3.96 0.02 Number of stage + sex 6 5068.70 0 0.09 dives stage + yr + sex 7 5068.80 0.09 0.08 stage 5 5069.90 1.2 0.05 stage + yr 6 5070.20 1.46 0.04 BSI + stage + sex 7 5070.20 1.49 0.04 WLI + stage + sex 7 5070.70 1.99 0.03 BCI + stage + sex 7 5070.80 2.02 0.03 cln + stage + sex 7 5070.80 2.02 0.03 WLI + stage + yr + sex 8 5070.80 2.04 0.03

BSI + stage + yr + sex 8 5070.80 2.09 0.03 BCI + stage + yr + sex 8 5070.90 2.11 0.03 cln + stage + yr + sex 8 5070.90 2.11 0.03 BSI + stage 6 5071.40 2.7 0.02 WLI + stage 6 5071.90 3.18 0.02 BCI + stage 6 5072.00 3.22 0.02 cln + stage 6 5072.00 3.22 0.02 WLI + stage + yr 7 5072.20 3.42 0.02 BSI + stage + yr 7 5072.20 3.46 0.02 BCI + stage + yr 7 5072.20 3.49 0.02 cln + stage + yr 7 5072.20 3.49 0.02 BCI + BSI + stage + sex 8 5072.30 3.51 0.02 BSI + cln + stage + sex 8 5072.30 3.51 0.02 BSI + WLI + stage + sex 8 5072.30 3.51 0.02

Table S2.2: Average model coefficients and relative importance of variables included in top

model set (∆AICc ≤ 4) explaining individual variation in Australasian gannets (Morus

serrator) foraging metrics.

Foraging metric Parameter Estimate SE 5 % 95 % Relative CI CI importance Maximum distance (Intercept) 4.06 0.40 3.42 4.72 - from the colony Colony (PE) -0.81 0.09 -0.96 -0.66 1.00 (km) Breeding stage (INC) 0.51 0.06 0.41 0.62 1.00 Breeding stage (LCR) -0.21 0.06 -0.31 -0.11 1.00 Year (2015) 0.14 0.05 0.06 0.21 1.00 Sex (male) -0.33 0.09 -0.48 -0.18 1.00 WLI 0.02 0.01 -0.01 0.03 0.52 BCI 0.05 0.05 -0.02 0.12 0.40 BSI 0.01 0.01 -0.01 0.01 0.27 Bearing (º) (Intercept) 213.12 27.04 167.93 257.68 - Colony (PE) -72.44 6.23 -82.76 -62.13 1.00 Breeding stage (INC) -16.96 3.99 -23.52 -10.40 1.00 Breeding stage (LCR) 11.74 3.82 5.45 18.03 1.00 Year (2015) 13.69 3.32 8.19 19.12 1.00 BSI 0.01 0.01 -0.01 0.03 0.51 BCI 3.38 2.99 -1.55 8.30 0.39 WLI 0.26 0.73 -0.95 1.45 0.26 Sex (male) 1.06 6.18 -9.16 11.30 0.25 Tortuosity Index (Intercept) 0.31 0.02 0.28 0.33 - Colony (PE) -0.05 0.01 -0.06 -0.04 1.00 Breeding stage (INC) -0.05 0.01 -0.06 -0.04 1.00 Breeding stage (LCR) -0.01 0.01 -0.01 0.01 1.00 Year (2015) 0.04 0.01 0.04 0.05 1.00 WLI 0.01 0.01 -0.01 0.01 0.42 BCI 0.01 0.01 -0.01 0.01 0.23 Sex (male) -0.01 0.01 -0.01 0.01 0.23 BSI -0.01 0.01 -0.01 0.01 0.22

Mean VeDBA (Intercept) 0.83 0.02 0.79 0.86 - Breeding stage (INC) -0.04 0.01 -0.05 -0.03 1.00 Breeding stage (LCR) 0.01 0.01 -0.01 0.01 1.00 Sex (male) 0.03 0.01 0.02 0.04 1.00 Colony (PE) 0.01 0.01 -0.00 0.02 0.44 Year (2015) -0.01 0.01 -0.01 0.01 0.22 BSI 0.01 0.01 -0.01 0.01 0.16 WLI 0.01 0.01 -0.01 0.01 0.16 BCI -0.01 0.01 -0.01 0.01 0.16 Number of dives (Intercept) 6.63 0.43 5.77 7.48 - Breeding stage (INC) 1.53 0.18 1.24 1.83 1.00 Breeding stage (LCR) 0.28 0.15 0.02 0.53 1.00 Sex (male) -0.33 0.18 -0.63 -0.03 0.68 Year (2015) -0.21 0.16 -0.47 0.05 0.44 BSI -0.01 0.01 -0.01 0.01 0.22 WLI -0.01 0.02 -0.04 0.03 0.16 BCI -0.01 0.09 -0.15 0.15 0.16 Colony (PE) 0.01 0.18 -0.30 0.30 0.16

Table S2.3: Factors influencing individual variation (measured as the coefficient of variation within deployments for each foraging metric), at short- term scale (T-to-T) in Australasian gannets (Morus serrator). Most parsimonious models after model averaging and their corresponding estimated regression parameters are shown. The most parsimonious model was selected with using Akaike Information Criterion (AICc, ∆ < 4).

Model variables: BCI: Body Condition Index; BSI: Body Size Index; WLI: Wing Length Index; stage: breeding stage; colony; year; sex.

Response Most parsimonious model Fixed effect Estimate SE t-value P-value Distance from colony (km)* colony + year (Intercept) 0.28 0.01 24.32 <0.0001 Colony (PE) -0.05 0.01 -4.49 <0.0001 Year (2015) -0.06 0.01 -4.25 <0.0001 Bearing (º) stage + year (Intercept) 1.60 0.04 37.19 <0.0001 Stage (INC) -0.26 0.06 -4.24 <0.0001 Stage (LCR) 0.07 0.06 1.10 0.27 Year 2015 0.11 0.05 2.06 0.04 Tortuosity index colony + WLI + year + sex (Intercept) 0.22 0.03 22.26 <0.0001 Colony (PE) 0.04 0.01 3.68 <0.0001 Year (2015) -0.04 0.01 -3.96 <0.0001 Sex (male) 0.02 0.01 2.54 0.01 WLI -0.01 0.01 -2.70 0.01 Sex (male) 0.03 0.01 1.83 0.07

Mean VeDBA (g)* BSI (Intercept) 0.08 0.01 21.82 <0.0001 BSI 0.01 0.01 2.70 0.01 Number of dives* colony + stage + year (Intercept) 0.25 0.01 16.92 <0.0001 Colony (PE) -0.04 0.01 -2.66 0.01 Year (2015) -0.06 0.02 -3.98 0.00 Stage (INC) 0.01 0.02 0.35 0.73 Stage (LCR) 0.06 0.02 3.57 0.001 *Transformed variable

CHAPTER 3

Temporal isotopic variability of marine prey species in south- eastern Australia: potential implications for predator diet studies

A version of this chapter has been submitted as:

Marlenne A. Rodríguez-Malagón, Lauren P. Angel and John P.Y. Arnould (in review). Temporal and spatial variability in isotopic niche of marine prey species in south-eastern Australia: potential implications for predator diet studies. Marine Biology.

Abstract

Stable isotope analyses, particularly of carbon (δ13C) and nitrogen (δ15N), of marine animal tissues are used to investigate ecological relationships among species. For marine predators, the main factors influencing intra-specific and intra-individual isotopic variation are geographical movements and changes in diet the composition over time. Because differences in stable isotope ratios may be the result of changes in the prey items consumed, a change in feeding location, or a combination of both, knowledge of the temporal and spatial consistency in the isotopic values of prey becomes crucial for making accurate inferences about predator diets. This study used an abundant marine predator, the Australasian gannet

(Morus serrator), as a prey sampler to investigate the annual variation in prey isotope values over a 4 year period (2012-2015), and the geographic variation between two sites with contrasting oceanographic conditions. Significant inter-annual variation was found in δ13C and/or δ15N values in five of the eight prey species analysed, suggesting temporal fluctuations in their geographic source or in the origin of their nutrients. These results could potentially be related to a major climatic event (El Niño/Southern Oscillation) that occurred during the study period. The predictive power of isotopic mixing models used to investigate predator diets can be compromised if the isotopic values of specific prey vary substantially across the temporal scales or spatial scales over which the predators move. Therefore, this study demonstrates the importance of considering the potential significant differences in prey species isotopic values within the assemblages that predators consume, in studies that use isotope analyses, to better understand the influence of environmental variability on diet composition.

Introduction

Stable isotope analyses are a powerful method to address biological questions and are widely used in ecology (Boecklen et al. 2011). In particular, the stable isotope ratios of carbon (13C/12C) and nitrogen (15N/14N) are commonly used to describe the trophic niche of study animals (Bearhop et al. 2004), to reconstruct their diets (DeNiro & Epstein 1978,

1981), and to make inferences about their temporal variability and/or consistency in diet

(Dalerum & Angerbjörn 2005, MacNeil et al. 2005, Jaeger et al. 2010). These inferences are possible due to the variation in δ13C values, for example, between plant species with different photosynthetic pathways (i.e. C3 or C4 plants; Kelly 2000) or along horizontal and vertical gradients in the marine environment (Hobson et al. 1994, Cherel & Hobson 2007) that serve to determine the primary sources of food. Similarly, the 15N enrichment of consumers relative to their food items serves as an indicator of the consumer’s trophic position (Vanderklift &

Ponsard 2003). These differences are due to the fractionation of the heavier isotope (15N) from the isotopically lighter isotope (14N) during amino acid synthesis, resulting in the retention of the heavier isotope and the excretion of the lighter (Minagawa and Wada 1984).

In recent years, the use of isotopic analysis in studies of predator diets has increased in response to improved statistical tools such as mixing models (which consider the isotopic composition of consumers and their foods to make inferences about the composition of the animal’s assimilated diet; Phillips et al. 2014) and knowledge associated with isotopic enrichment processes (Vanderklift & Ponsard 2003). Such research has shown that the main factors influencing intra-specific and inter-individual variation in stable isotope values are the geographical movement of predators (Graham et al. 2010, Hobson et al. 2010) and/or changes in the composition of their diet over time (Crawford et al. 2008, Blight et al. 2015). In addition, stable isotope analyses have been used in studies addressing the level behavioural consistency and individual diet specialisation in animals (e.g. Newsome et al. 2009, Jaeger et

56 al. 2010). Such studies require longitudinal sampling from individuals and the ability to quantify individual diet preferences accurately (Coblentz et al. 2017). However, without knowledge of the isotopic signatures of potential prey and how these vary spatially and temporally, interpreting intra-specific, inter- and intra-individual differences in predator isotopic values is problematic (Inger & Bearhop 2008, Barnes & Jennings 2009). This is because isotopic differences seen in predators may be the result of changes in the prey items consumed, a change in feeding location, or a combination of both. In addition, prey isotopic values within the same location may change over time if the food base on which the prey depends also varies as a consequence of biogeochemical processes (Goericke & Fry 1994,

Lorrain et al. 2015). Hence, knowledge of the temporal and spatial consistency in the isotopic values of prey is crucial for making accurate inferences about predator diets and their level of individual specialisation (Cherel et al. 2007, Bond & Jones 2009, Quillfeldt et al. 2015).

Within marine environments, top predators play an important role as top-down controllers of prey species, nutrient cyclers, and ecosystem engineers (Heithaus et al. 2008,

Baum & Worm 2009). Marine environments are complex and dynamic, and their temporal and spatial variation influences the ecology of marine life (Ballance et al. 2006). At local and regional scales, physical features such as water currents, bathymetry, tide regimes, and nutrient fluxes determine the structure of marine and coastal ecosystems and influence the behaviour and distribution of marine fauna (Butler et al. 2002). Concurrently, naturally occurring stable isotopes are influenced by water temperature and phytoplankton photosynthetic pathways (Rau et al. 1982, Graham et al. 2010), and by N2 fixation processes in the surface ocean, terrestrial runoff, and atmospheric precipitation (Sigman & Casciotti

2001) for δ13C and δ15N, respectively. These factors influence the isotopic values of marine plankton that could potentially produce spatial isotopic variation at the base of marine food chains (Kurle & McWhorter 2017). Previous studies suggest isotopic values vary in relation

57 to latitude and longitude, distance from the coast, and water depth (Quillfeldt et al. 2015,

Trueman et al. 2017). In addition, particularly for δ13C, local temporal variation in isotope values has also been reported in relation to changes in currents and nutrient availability influencing primary productivity of marine food chains (Radabaugh et al. 2013, Quillfeldt et al. 2015, Kurle & McWhorter 2017). Consequently, knowledge of the temporal and spatial variation in prey isotopic values is especially important when examining marine predator diets.

Despite being a region of relatively low primary productivity, with dominant high salinity and dense waters (Gibbs et al. 1986), Bass Strait in south-eastern Australia supports a large diversity of seabirds, marine mammals and predatory fish (Lewis 1981, Gibbs 1992,

Butler et al. 2002). The region is influenced by the seasonally wind-driven Bonney

Upwelling System and three major currents (the Subantarctic Surface Water from the south, the South Australian Current from the northwest and the East Australia Current from the northeast; Middleton & Bye 2007). These oceanographic features are subject to both intra- and inter-annual variations in strength, affecting the abundance, diversity, and source of prey species within Bass Strait (Sandery & Kämpf 2007). In addition, numerous bays and shallow inlets, influenced by variable terrestrial run-off, represent important fish habitats (Jenkins et al. 1997b) that are exploited by some major predators (Montague & Cullen 1988, Wells et al.

2016, Filby et al. 2017). The region is also one of the fastest warming marine areas in the world and the anticipated oceanographic changes are likely to affect the distribution, abundance, and diversity of prey species (Ridgway 2007, Lough & Hobday 2011).

Previous studies have documented significant inter-annual and geographic differences in the stable isotope values of marine predators within Bass Strait (e.g. Australian fur seals

Arctocephalus pusillus doriferus, little penguins Eudyptula minor, and Australasian gannets

Morus serrator), suggesting variation in their diets in relation to temporal and spatial factors

(Arnould et al. 2011, Chiaradia et al. 2012, Kernaléguen et al. 2015, Angel et al. 2016).

However, isotopic information on the many potential prey species in the Bass Strait region is limited (Davenport & Bax 2002, Chiaradia et al. 2010). In addition, there is currently no information on the temporal or spatial variation in the region’s isoscape (Bowen et al. 2009,

Graham et al. 2010). Consequently, it is not possible to ascertain whether variation in predator isotopic values reflect changes in diet species composition, foraging areas, isoscape, or a combination of these. Information about the isotopic variation of regional marine predators is crucial for understanding how current oceanic variability influences their diets and predicting how their populations may respond to future ecosystem changes. Therefore, the primary objective of the present study was to investigate the temporal and spatial variation in δ13C and δ15N isotopic values in an assemblage of common prey species found in south-eastern Australia. Specifically, using an abundant marine predator as a sample collection agent to investigate (1) annual variation in prey isotope values over a 4 year period

(2012-2015), and (2) geographic variation between two sites with contrasting oceanographic conditions. Given the environmental variability existing within the south-eastern Australian region and the marine habitat diversity found around the study sites, isotopic variation was expected within the prey component to reflect the isotopic variation at the base of the food chain associated with these physical and dynamic processes.

Materials and methods

Sample collection

The Australasian gannet (Morus serrator), a top marine predator, was used as a means of prey sample collection in northern Bass Strait. Samples were collected from individuals at the Point Danger (PD, 38º 23’ 36.09” S, 141º 38’ 55.94” E) and Pope’s Eye (PE, 38º 16’

35.88” S, 144º 41’ 56.21” E) breeding colonies (Fig. 3.1). Adult birds from PD have been shown to forage over the continental shelf, ranging up to 238 km northwest and southeast from the colony, and hunting within the limits of the Bonney Upwelling System (Butler et al.

2002, Angel et al. 2016). In contrast, birds from PE forage within the shallow waters of Port

Phillip Bay (average depth < 13.6 m; Walker 1999), outside the bay within central Bass

Strait, or in both habitats (Angel et al. 2016, Wells et al. 2016). The diet of Australasian gannets within Bass Strait has been reported to consist of at least 37 species of demersal/reef- associated and pelagic/oceanic species of fish and squid (Norman & Menkhorst 1995, Bunce

& Norman 2000, Bunce 2001, Pyk et al. 2007, Barker 2012). Hence, the location of these colonies, the foraging range of this predator, and its broad diet allowed numerous prey species representative of a wide range of habitats to be sampled.

Fig. 3.1: Location of the Point Danger (PD, left) and Popes Eye (PE, right) breeding colonies

(black crosses). The South Australian Current (SAC, winter) and East Australian Current

(EAC, winter and summer) bring warm and low nutrient waters into the marine region, while the Sub-Antarctic Surface Water (SASW, summer) drives cold and productive waters from the south (Sandery & Kämpf 2007). PD (green) and PE (orange) Australasian gannet (Morus serrator) foraging ranges obtained from GPS data (Chapter 2, Wells et al. 2016). Bathymetric contours (every 20 m) within the continental shelf are shown.

As part of concurrent studies on the foraging ecology of Australasian gannets, during the 2012-2015 breeding seasons (October-March) and in each of three breeding stages: incubation, early chick-rearing (chick age 0-50 d), and late chick-rearing (chick age >50 d;

Wingham 1982), voluntary regurgitations by birds upon handling were collected in plastic bags and stored frozen (-20º C) until analysis in the laboratory. In the laboratory, regurgitate samples were thawed and prey specimens were identified to the lowest possible taxonomic level using published guides (Gomon et al. 2008). From complete prey items, standard length and body mass of individual specimens were recorded using Vernier callipers (± 0.1 mm) or a metal ruler (± 1 mm), and a top-loading balance (± 0.01 g, Ohaus Corporation, Parsippany

USA), respectively. From incomplete prey items, fresh otoliths were extracted where possible in order to confirm fish identification and estimate standard length from published otolith-fish length regressions (Furlani et al. 2007). Muscle tissue was collected from representative individual prey in all samples. For fish, muscle tissue was sampled posterior to the anus, above the lateral line, on one side of the vertebral column. For squid, muscle tissue was taken from the base of the mantle. Previous studies have reported fish muscle has an isotopic turnover that varies depending upon the species and fish age, suggesting that younger fish with higher growth rates have faster isotope turnover times than adult fish (Sakano et al.

2005, Weidel et al. 2011). For the muscle of young fish, the isotopic turnover rate ranges from 8 to18 days for carbon and from 4 to 16 days for nitrogen. For adult fish, isotope turnover rate ranges from 116 to 173 days for carbon and from 76 to 122 days for nitrogen

(Bosley et al. 2002, Church et al. 2008, Weidel et al. 2011). In squid, young individuals have an estimated turnover rate of 35 days, while older life stages have a turnover rate up to 80 days (Ruiz-Cooley et al. 2006).

Stable isotope analysis

All tissue samples were oven-dried at 50ºC for at least 24 h and then ground into powder using a mortar and pestle. Lipids were extracted using 1 mL of a 2:1 chloroform– methanol mixture added to powdered samples. Samples were then shaken using a vortex mixer and centrifuged for 10 min at 10ºC (2500 rpm). The supernatant was discarded and the procedure repeated at least once or until the supernatant was clear and colourless after centrifugation (Ehrich et al. 2011). Once dry, 2.0 mg of each sample was loaded into separate tin capsules. Stable isotope analyses of carbon and nitrogen were conducted at the Farquhar laboratory of the Research School of Biology, Australian National University (Canberra,

Australia). Samples were combusted in a CHN elemental analyser (CE1110, Carlo Erba) and resulting gases were analysed using an interfaced isoprime continuous-flow isotope ratio mass spectrometer (Micromass Instruments). Quality control samples were run before and after each sequence using laboratory standards of sucrose ANU (-10.45 ‰) and BEET (-

24.62 ‰) for δ13C and the amino acids alanine, glycine and cysteine used for δ15N, which provided replicate measurement errors of ± 0.1‰ and ± 0.3 ‰, respectively. Stable isotope values were expressed in δ-notation as the deviation from standards in parts per mil (‰) according to the following equation:

δX = [(Rsample / Rstandard) – 1]

13 15 13 12 15 14 Where, X is C or N and R is the corresponding ratio of C/ C or N/ N. Rstandard values were based on international standards Vienna Pee Dee Belemnite (VPDB) for δ13C,

15 and atmospheric nitrogen (N2) in air for δ N. The mean C:N mass ratio for all samples was estimated as 3.17 (± 0.15 SD), indicating lipid concentrations were uniformly low and no data normalization was needed (Post et al. 2007).

Statistical analysis

All data processing and statistical analysis were conducted in R version 3.4.1 (R Core

Team 2017). To simultaneously evaluate the effect of years (from 2012 to 2015) and colonies

(PD and PE) on the prey isotopic niches, a two-way ANOVA test with interaction terms was used considering δ13C and δ15N as response variables separately. As sample sizes varied considerably among years and colonies for each prey species, an unbalanced design with

Type-III sums of squares, was considered when running this statistical analysis. The function

Anova of the car package version 2.1-5 was used for such purposes (Fox & Weisberg 2018).

Significant terms were tested using a posteriori multiple comparison tests with the TukeyHSD function of the stats 3.4.1 package (R Core Team 2017).

To investigate whether time and habitat (benthic and pelagic) influence variation in stable isotope values within and in between prey species, the coefficient of variation of δ13C and δ15N absolute values was calculated as a proxy for variation. The coefficients of variation were estimated by year and colonies for species with at least three individuals per year and colony. The equality of the estimated coefficients of variation was tested using the asymptotic_test2 function of the cvequality package for summary statistics (Marwick &

Krishnamoorthy 2018). Each prey species was classified by habitat according to its biological information available on the database fishesofaustralia.net. Unless stated otherwise, results are presented as mean ± standard error (SE).

Results

A total of 298 individual adult gannets nesting at the PD or PE colonies (143 and 145, respectively) were captured on multiple occasions as part of a foraging behaviour study over the 2012-2015 breeding seasons. From these birds, 404 regurgitated food samples (207 from

PD and 197 from PE) were obtained. Food samples contained from 1 to 4 prey species each, from which 1,181 individual prey items were collected (704 from PD and 477 from PE) and

427 muscle samples were analysed for stable isotope ratios (Table 3.1). A total 18 different prey species were identified in food samples. Eight prey species were exclusively collected from birds at PE, while five species were exclusively collected from birds at PD. The means

(± SE) of the standard length and body mass for each prey species in each year and at each colony can be found in the Supporting information Table S3.1.

A wide range in δ13C and δ15N values was observed among the species identified (Fig.

3.2). A clear distribution in isotopic space was observed among the prey species collected, from the prey species with the most depleted values in δ 13C, the pelagic Australian anchovy

(Engraulis australis, −20.60 ± 0.42 ‰ SD), to the prey species with the most enriched, the benthic blue weed-whiting (Haletta semifasciata, −15.20 ± 2.63 ‰ SD). The δ13C mean range for all species was 1.73 ± 1.28 ‰ SD. The δ15N values ranged from the velvet leatherjacket (Meuschenia scaber, 9.51 ± 0.89 ‰ SD), a mostly benthic invertebrate feeder, to the yellowfin goby (Acanthogobius flavimanus, 21.21 ‰), a crustacean and fish predator.

The δ15N mean range for all species was 4.46 ± 3.88 ‰ SD.

Eight prey species were collected with sufficient sample sizes for investigating temporal and spatial variability. Of these eight, there was significant temporal intraspecific variability in mean δ13C values among three prey species. In all three species, δ13C values in

2015 were significantly different from the other years. For barracouta (Thyrsites atun), δ13C values in 2015 and 2012 were lower compared to 2014. In contrast, for Gould’s squid

(Nototodarus gouldi) δ13C values were higher in 2015 than in 2014, and for king gar

(Scomberesox sauri) δ13C values were higher in 2015 than in 2012 and 2014. There were, however, no significant differences in δ13C values of prey species between colonies (Table

3.2, Fig. 3.3).

Species Habitat Diet 2012 2013 2014 2015

PD PE PD PE PD PE PD PE

Australian anchovy* Pelagic Zooplankton c 1 8 19 15

Australian sardine* Pelagic Zooplankton and phytoplankton c 1 2 32 21 12 19

Barracouta* Pelagic Cephalopods, pelagic fish and invertebrates a 3 1 5 21 16 12 18

Blue mackerel Pelagic Small fish and squid, pelagic invertebrates b 1 7 2

Blue sprat Pelagic Zooplankton b 3

Blue weed-whiting Benthic Invertebrates and plant matter b 1 1

Bluespotted goatfish* Benthic Invertebrates and small fish b 2 8 15 10

Eastern Australian salmon Pelagic Fish a 2 1

Flathead Benthic Fish and crustaceans b 2 1

Gould's squid* Pelagic Fish, crustaceans and cephalopods a 2 1 10 2 15

Jack mackerel* Pelagic Zooplankton, crustaceans and invertebrates b 8 2 7 21 3 4

King gar* Pelagic Zooplankton and fish larvae a 6 12 9

Longsnout boarfish Pelagic Polychaete worms, sea stars and algae b 1

Redbait* Pelagic Zooplankton, crustaceans and invertebrates b 2 29 20

Snook Pelagic Small fish and invertebrates a 1

Southern garfish Pelagic Invertebrates and plant matter a 5 1

Velvet leatherjacket Benthic Invertebrates and plant matter c 3

Yellowfin goby Benthic Crustacean and fish a 1

Species: Australian anchovy (Engraulis australis), Australian sardine (Sardinops sagax), barracouta (Thyrsites atun), blue mackerel (Scomber australasicus), blue sprat (Spratelloides robustus), blue weed-whiting (Haletta semifasciata), bluespotted goatfish (Upeneichthys vlamingii), Eastern Australian salmon (Arripis trutta), flathead (Platycephalus sp.), Gould’s squid (Nototodarus gouldi), jack mackerel (Trachurus declivis), king gar (Scomberesox saurus), longsnout boarfish (Pentaceropsis recurvirostris), redbait (Emmelichthys nitidus), snook (Sphyraena novaehollandiae), southern garfish (Hyporhamphus melanochir), velvet leatherjacket (Meuschenia scaber),yellowfin goby (Acanthogobius flavimanus). a Migratory/highly mobile species. b non-migratory species. c age differences in habitat use: young = inshore, adults = open sea.

Fig. 3.2: Stable isotope biplot indicating the mean ± SD of δ13C and δ15N positions of the 18 prey species collected from Australasian gannet (Morus serrator) regurgitates at the Point

Danger and Pope’s Eye breeding colonies. Numbers in parentheses represent total number of prey individuals analysed. Asterisks on species name identify benthic species, all others are considered pelagic.

Of the eight prey species with sufficient sample sizes for investigating temporal and spatial variability, there was significant temporal intraspecific variability in mean δ15N values for five prey species. Similar to δ13C values, δ15N values from specimens collected in 2015 were significantly different from those collected in the other years sampled. For king gar,

δ15N values in 2015 were higher than in 2012 or 2014, while for barracouta, Gould’s squid, and redbait (Emmelichthys nitidus) δ15N values in 2015 were lower than those from the other years sampled. Jack mackerel (Trachurus declivis) was the only prey species where significant differences in δ15N values were detected between colonies, as well as an interaction between year and colony. For this prey species, however, the between colony differences was only due to higher δ15N values for fish collected at PD colony in 2014 (Table

3.3, Fig. 3.3).

The coefficients of variation for stable isotope values of each prey species within each year and colony ranged from 1% to 8% and from 2% to 29% for δ13C and δ15N, respectively.

Samples collected at PE, in particular for benthic species, generally presented a higher relative degree of variability in both stable isotope ratios than those samples collected at PD

(Fig. 3.4). The equality test showed significant results between the species partition (D’AD =

163.78, P < 0.0001), indicating that the species-specific isotope values were significantly more variable in benthic environments.

Table 3.2: Two-way ANOVA test results for temporal and spatial differences in δ13C values of different prey species (mean ± SE). Significant

results of the ANOVA test are shown. Means with the same superscript denote homogenous subsets (P > 0.05). Sample sizes indicated in Table

3.1.

Species Two-way 2012 2013 2014 2015

ANOVA test PD (‰) PE (‰) PD (‰) PE (‰) PD (‰) PE (‰) PD (‰) PE (‰)

Australian anchovy -20.09 -20.36 ± 0.2 -20.56 ± 0.4 -20.81 ± 0.4

Australian sardine -20.12 ± 0.4 -20.30 ± 0.3 -20.33 ± 0.5 -20.46 ± 0.5

a a b b a a Barracouta Year, F2,65 = 3.90, P = 0.025 -19.55 ± 0.1 -20.04 -19.03 ± 0.5 -19.03 ± 0.5 -19.54 ± 0.5 -19.58 ± 0.6

Bluespotted goatfish -17.59 ± 0.007 -17.16 ± 0.8 -17.64 ± 0.7 -17.33 ± 1.3

a b Gould's squid Year, F1,23 = 9.14, P = 0.006 -18.94 ± 0.4 -18.35 ± 0.4

Jack mackerel -19.31 ± 0.3 -19.68 ± 0.8 -19.33 ± 0.7 -20.55 ± 0.1

a a b King gar Year, F2,24 = 9.31, P < 0.001 -20.36 ± 0.3 -20.15 ± 0.3 -19.76 ± 0.1

Redbait -19.45 ± 0.3 -19.26 ± 0.4

Breeding colonies: PD = Point Danger, PE = Pope’s Eye.

3.1.

Species Two-way ANOVA test 2012 2013 2014 2015

PD (‰) PE (‰) PD (‰) PE (‰) PD (‰) PE (‰) PD (‰) PE (‰)

Australian anchovy 13.73 13.41 ± 1.7 12.31 ± 0.7 13.22 ± 3.8

Australian sardine 12.13 ± 0.5 12.44 ± 1.9 12.20 ± 0.6 11.55 ± 1.8

a a a a b b Barracouta Year, F2,65 = 10.48, P < 0.0001 14.36 ± 0.2 14.03 14.48 ± 0.5 15.10 ± 1.3 12.69 ± 1.1 12.97 ± 1.2

Bluespotted goatfish 18.10 ± 4.1 16.29 ± 2.5 15.57 ± 3.0 17.10 ± 3.9

a b Gould's squid Year, F1,23 = 21.90, P < 0.0001 13.33 ± 0.6 11.52 ± 1.0

a b b b Jack mackerel Year, F1,31 = 16.07, P = 0.0003 14.49 ± 0.5 13.52 ± 0.4 13.32 ± 0.2 13.20 ± 0.3

Colony, F1,31 = 27.51, P < 0.0001

Interaction, F1,31 = 5.20, P = 0.03

a b b King gar Year, F2,24 = 4.1, P = 0.02 11.18 ± 0.9 12.24 ± 1.1 12.59 ± 0.7

a b Redbait Year, F1,47 = 23.89, P < 0.0001 13.65 ± 0.4 12.96 ± 0.5

Breeding colonies: PD = Point Danger, PE = Pope’s Eye.

Fig. 3.3: Stable isotope biplot indicating the mean (± SD) of δ13C and δ15N values for different prey species. A. Inter-annual comparisons with significant results: barracouta

Fig. 3.4: Coefficients of variation in δ13C and δ15N values within each year and colony for all prey with ≥ 3 samples collected (n = 12). Asterisks on species name identify benthic species.

Discussion

The findings of the present study indicated significant inter-annual variation in both the δ13C and δ15N values of several prey species, suggesting temporal fluctuations in their geographic source or in the origin of their nutrients. This has significant implications for the interpretation of predator diets from analyses of stable isotopes in tissues and for the potential estimations of their individual diet specialisations. This is especially so as several of the prey species analysed (e.g. barracouta, jack mackerels, redbait, and Gould’s squid) have been shown to be important food items for a range of other predators in the region such as southern blue-fin tuna, Australian fur seal, little penguin, and sharks (Ward et al. 2006, Pyk et al. 2007,

Braccini 2008, Deagle et al. 2009, Chiaradia et al. 2012). More examples of overlap between the Australasian gannet and other marine predators prey items can be found in Supporting information Table S3.2.

The δ13C values of lower trophic level prey species can be influenced by temporal and spatial variations in the carbon isotope composition of phytoplankton, which in turn are influenced by variations in sea surface temperatures (SST) and the concentration of dissolved

CO2, which influence the degree of carbon isotope fractionation that occurs during photosynthesis (Magozzi et al. 2017). Additionally, the spatial differences between coastal/inshore environments and pelagic/offshore communities related to the distinct carbon fixation processes of phytoplankton and algal communities can also potentially have an effect on the isotopic values of the prey species analysed (France 1995b, Newsome et al. 2007).

Although small, the significant inter-annual differences in δ13C values observed in the present study for barracouta, Gould’s squid, and king gar are consistent with previous expectations and studies of fish and squid that have documented similar temporal variation (Kurle et al.

2011, Albo-Puigserver et al. 2016). As there was no evidence that the sampled Australasian gannets changed their foraging areas during the study (Chapter 2, Angel et al. 2015, 2016,

Wells et al. 2016), this suggests the isotopic values of these prey (or their nutrients) do not necessarily reflect the area in which they were consumed and the observed variation may be due to fluctuations in the major currents influencing the area (Middleton & Bye 2007,

Sandery & Kämpf 2007). Indeed, previous studies (Blackburn and Gartner 1954, Gomon et al. 2008, Stark 2008, Agüera and Brophy 2011) have shown that barracouta, Gould’s squid, and king gar are highly mobile in the south-east Australia region.

The δ15N values of prey species reflect their trophic position (Vanderklift & Ponsard

2003). In the present study, as expected, five species displayed significant inter-annual differences in their δ15N values: king gar, a zooplankton and fish larvae consumer (Gomon et al. 2008); jack mackerel and redbait, which consume krill, fish larvae, crustaceans, and other marine invertebrates (Ward et al. 2008, Goldsworthy et al. 2011); barracouta, a predator of cephalopods, pelagic fish, and invertebrates (Bulman et al. 2001, Ward et al. 2008); and

Gould’s squid, which eats mostly pelagic fish and crustaceans (O'Sullivan & Cullen 1983).

Barracouta and Gould’s squid have also been reported to prey on each other (O'Sullivan &

Cullen 1983). The observed temporal variation in δ15N values of these prey species could reflect changes in diet composition of the prey or the oceanic source of their nutrients (Kurle et al. 2011). Indeed, while jack mackerel and redbait are not considered migratory (Webb &

Grant 1979, Neira et al. 2009), their abundance has been closely linked to the availability of

Australian krill (Nyctiphanes australis). The abundance of this euphausiid is influenced by regional oceanography on both seasonal and inter-annual scales (Young et al. 1993),

15 potentially resulting in δ N variation due to variability in biogeochemical processes (e.g. N2 fixation processes in the surface ocean, terrestrial runoff, or atmospheric precipitation;

Sigman and Casciotti 2001). This variation highlights the potential problems for inferring changes to the diet of higher predators from tissue stable isotope values without concurrent information on isotope values of the prey base (Bond & Jones 2009, Quillfeldt et al. 2015).

Interestingly, most of the significant inter-annual variation in both δ13C and δ15N values occurred in 2015, which coincided with a strong El Niño/Southern Oscillation (ENSO) event with widespread below average rainfall and higher (≥ 2º C) SST across south-eastern

Australia (www.bom.gov.au). These extreme conditions could potentially alter the range and spawning areas of the sampled species (Holbrook et al. 1997, Sugimoto et al. 2001, Harley et al. 2006), their growth rates (Hernández-Miranda and Ojeda 2006), their migration routes

(Kimura et al. 1997), and even their nutritional composition (Machovsky-Capuska et al.

2018). Additionally, the physical changes associated with ENSO events have been shown to influence the dynamics of primary producers and, consequently, alter the rest of the food chain (Barber & Chavez 1983, Chavez et al. 2011). In addition, sample sizes of prey tissue from 2012 and 2013 were smaller than from 2014 and 2015, and the lack of significant temporal differences in some prey species may have been related to sample size issues.

In contrast, five prey species did not display significant temporal or spatial variability in their δ13C values. With the exception of the bluespotted goatfish, an inhabitant of the sandy sea floor of Port Phillip Bay (Currie & Sorokin 2010), the species with no temporal or spatial variability in δ13C values (Australian anchovy, Australian sardine, jack mackerel, and redbait) are pelagic and abundant in inshore and shelf waters of eastern Australia during summer and autumn (Ward et al. 2006, Gomon et al. 2008). Similarly, the Australian anchovy and

Australian sardine, two planktivorous species, plus the bluespotted goatfish, a consumer of benthic invertebrates, showed no significant temporal or spatial variation in δ15N values.

These findings suggest that the feeding behaviour of these prey species was not isotopically influenced by external factors, or they did not experience a significant diet change, during the study.

There was a substantial range in the coefficients of variation of isotope values within sampling periods and sites. Samples collected at PE, most notably those of benthic species,

76 generally displayed higher relative variation in isotope values than those collected at PD.

Gannets at the PE breeding colony feed in both Port Phillip Bay, a shallow water body adjacent to the city of Melbourne with constant freshwater input from rainfall, rivers, creeks, and drains (Gearing 1988b), and in northern Bass Strait, an area influenced by multiple currents over short time scales (Sandery & Kämpf 2005, Angel et al. 2016). The results of the present study suggest these features may lead to fine-scale spatial isoscape variation for the region’s fish prey species (Radabaugh et al. 2013, Kurle & McWhorter 2017). In contrast, gannets from PD forage over the continental shelf area of western Bass Strait with relatively uniform oceanographic influences during the breeding season (Lewis 1981, Stark 2008).

Such factors should be considered when inferring diet composition from tissue stable isotope values in predators.

Stable isotope analyses are powerful tools for qualitatively and quantitatively assessing animal diets (DeNiro & Epstein 1978, 1981, Dalerum & Angerbjörn 2005) and to estimate the level of individual diet specialisation within animal populations (Araújo et al.

2007, Jaeger et al. 2010, Phillips et al. 2017). It has been shown that the isotopic differences seen in a predator diet across time may be the result of a change in prey species composition

(Crawford et al. 2008, Blight et al. 2015), a change in the feeding location, or a combination of both (Graham et al. 2010, Hobson et al. 2010). As has been shown elsewhere in previous studies (Kurle et al. 2011, Quillfeldt et al. 2015, Rumolo et al. 2017), the present study has demonstrated temporal variation in the isotopic values of some common marine prey species within south-eastern Australia. While the differences among the annual means of the prey isotope values were small, if these prey species constitute a large proportion of a predator’s diet, it could substantially affect interpretation of predator isotope values and subsequent assessment of individual diet specialisation. Similarly, geographic variation in prey isotopic values can affect interpretations of predator diets if these prey occur within their foraging

77 range (Radabaugh et al. 2013). For example, the predictive power of isotopic mixing models can be compromised if the isotopic values of specific prey species vary substantially over temporal scales of a study or spatial scales over which the predators move (Phillips et al.

2014) and the subsequent evaluation of the changes in foraging behaviour of animals could be biased. Therefore, it is advisable to use isotopic data from marine prey collected within the same location and time period of the predators under study (Bond & Jones 2009, Phillips et al. 2014). In conclusion, it is important to consider the potentially significant differences in isotopic values of prey species within the assemblages that predators consume in order to correctly interpret their isotopic variability and better understand the influence of the environment on their diets.

Supporting information

Table S3.1: Means (± SE) of the standard length and body mass for prey species by year and colony (where whole specimens or otoliths could be measured).

Species Year Colony n Standard length (mm) Body mass (g)

Australian anchovy 2014 PD 1 87.2 6.2

Australian anchovy 2014 PE 14 96.6 ± 2.5 9.1 ± 0.8

Australian anchovy 2015 PD 24 99.3 ± 1.8 10.8 ± 0.7

Australian anchovy 2015 PE 23 98.4 ± 2.1 9.5 ± 0.7

Australian sardine 2013 PE 3 116.7 ± 21.2 16.03 ± 6.5

Australian sardine 2014 PD 162 101.7 ± 3.2 17.3 ± 1.6

Australian sardine 2014 PE 42 123.3 ± 3.7 23.0 ± 2.2

Australian sardine 2015 PD 20 147.1 ± 8.9 39.6 ± 4.5

Australian sardine 2015 PE 52 123.1 ± 5.5 25.5 ± 2.8

Barracouta 2014 PD 5 199.9 ± 64.6 66.8 ± 57.1

Barracouta 2014 PE 9 321.6 ± 51.7 299.8 ± 73.5

Barracouta 2015 PD 16 187.7 ± 24.8 52.3 ± 24.8

Barracouta 2015 PE 17 236.1 ± 19.1 73.21 ± 29.9

Blue mackerel 2014 PE 2 291.5 ± 25.5 307.7 ± 100.4

Blue mackerel 2015 PE 1 310.0 461

Bluespotted goatfish 2013 PE 3 181.6 ± 18.3 118.4 ± 32.8

Bluespotted goatfish 2014 PE 1 172.0 121.4

Bluespotted goatfish 2015 PE 4 141.5 ± 6.4 56.3 ± 3.7

Flathead 2013 PE 1 211.0 81.8

Gould's squid 2012 PD 1 425.0 305.0

Gould's squid 2014 PD 5 117.2 ± 20.0 44.2 ± 18.2

Gould's squid 2014 PE 2 128.4 ± 11.5 103.9 ± 64.0

Gould's squid 2015 PD 11 120.4 ± 9.0 10.7 ± 1.5

Jack mackerel 2012 PE 7 225.6 ± 6.7 131.5 ± 10.9

Jack mackerel 2014 PD 4 259.7 ± 2.4 183.7 ± 23.2

Jack mackerel 2014 PE 10 241.4 ± 7.9 173.0 ± 17.6

Jack mackerel 2015 PD 2 256.7 ± 9.7 195.6 ± 15.6

Jack mackerel 2015 PE 3 244.6 ± 4.8 161.18 ± 8.3

King gar 2014 PD 2 200.0 ± 7.0 21.3 ± 1.9

Redbait 2014 PD 9 199.8 ± 2.1 94.6 ± 2.94

Redbait 2015 PD 16 137.4 ± 3.9 36.9 ± 3.3

Southern garfish 2014 PE 3 214.8 ± 26.9 26.4 ± 12.2

Southern garfish 2015 PE 1 300.0 89.0

Velvet leatherjacket 2014 PD 1 162.0 83.2

Yellowfin goby 2015 PE 1 180.0 38.0

Table S3.2: Examples of Australasian gannet (Morus serrator) prey species analysed in this study that are shared with other marine predators.

Prey species Other known predators Reference Australian anchovy Australian fur seal (Arctocephalus pusillus doriferus) Kirkwood et al. (2008), Deagle et al. (2009) (Engraulis australis) Australian salmon (Arripis trutta) Stewart et al. (2011) barracouta (Thyrsites atun) Blackburn and Gartner (1954) common dolphin (Delphinus sp.) Meynier et al. (2008) crested tern (Sterna bergii) Chiaradia et al. (2002) Gould’s squid (Nototodarus gouldi) Braley et al. (2010) little penguin (Eudyptula minor) Montague and Cullen (1988), Chiaradia et al. (2012) Australian sardine Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) (Sardinops sagax) Australian salmon Stewart et al. (2011) blue mackerel Robert et al. (2008) common dolphin Meynier et al. (2008) Gould’s squid O'Sullivan and Cullen (1983), Braley et al. (2010) little penguin Montague and Cullen (1988), Chiaradia et al. (2012) Barracouta (Thyrsites atun) Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) Australian salmon Stewart et al. (2011) Burrunan dolphin (Tursiops australis) Filby (2016) common dolphin Meynier et al. (2008) crested tern Chiaradia et al. (2002) Gould’s squid Braley et al. (2010) little penguin Chiaradia et al. (2003) Blue mackerel (Scomber Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) australasicus) Australian salmon Stewart et al. (2011) Gould’s squid O'Sullivan and Cullen (1983), Braley et al. (2010) Blue sprat (Spratelloides Australian salmon Stewart et al. (2011) robustus) little penguin Chiaradia et al. (2012) Bluespotted goatfish Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) (Upeneichthys vlamingii)

Prey species Other known predators Reference Eastern Australian salmon little penguin Montague and Cullen (1988), Chiaradia et al. (2012) (Arripis trutta) Flathead (Platycephalus sp.) Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) black-faced cormorant (Phalacrocorax fuscescens) Taylor et al. (2013) little penguin Montague and Cullen (1988), Chiaradia et al. (2012) Gould’s squid (Nototodarus Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) gouldi) common dolphin Meynier et al. (2008) crested tern Chiaradia et al. (2002) little penguin Montague and Cullen (1988), Chiaradia et al. (2012) Jack mackerel (Trachurus Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) declivis) Australian salmon Stewart et al. (2011) common dolphin Meynier et al. (2008) crested tern Chiaradia et al. (2002) Gould’s squid Braley et al. (2010) little penguin Montague and Cullen (1988), Chiaradia et al. (2012) King gar (Scomberesox sharks Klarian et al. (2018) saurus) Redbait (Emmelichthys Australian fur seal Kirkwood et al. (2008), Deagle et al. (2009) nitidus) Australian salmon Stewart et al. (2011) common dolphin Meynier et al. (2008) Gould’s squid Braley et al. (2010) Southern garfish Burrunan dolphin Filby (2016) (Hyporhamphus melanochir) Gould’s squid Braley et al. (2010) little penguin Chiaradia et al. (2012) Velvet leatherjacket Australian fur seal Deagle et al. (2009) (Meuschenia scaber)

CHAPTER 4

Inter- and intra-individual variation in the diet of Australasian

gannets (Morus serrator)

Abstract

Diets are defined by the foods that contain the nutrients that every organism consumes and have been shown to vary due to the influence of factors such as age, sex, learned abilities, and/or individual preferences. When these feeding differences are maintained over time, behavioural consistency is developed and individual diet specialisations within animal populations emerge. Furthermore, behavioural similarity within breeding pairs can result in assortative mating by diet, which can have major, long-term effects on the reproductive success of a species. The Australasian gannet (Morus serrator), an important marine predator considered a generalist forager with colony-specific reverse sexual-dimorphism and biparental care, was used as a model species to test some of these assumptions. Voluntary regurgitations and blood plasma were collected from breeding adults (mostly from nests partners) over four breeding seasons (2012 to 2015) and at two breeding sites. These samples were used to investigate the factors influencing the inter- and intra-individual variation in tissue δ13C and δ15N isotopic values, the inter- and intra-individual variation in diet composition, and the diet similarity between nest partners. Significant differences among individuals were found in δ13C and δ15N values from blood plasma and the differences were associated with as colony, year, breeding stage, and sex. The magnitude of inter-individual variation in diet and the average degree of individual specialisation were greater at the Point

Danger colony (generalists individuals), where Australasian gannets forage within a relatively uniform habitat, compared to the Pope’s Eye colony, where birds had access to a greater diversity of marine habitats for foraging. The intra-individual variation in isotopic values was lower over medium- (breeding stage-to-breeding stage) than over long-term timescales (year-to-year). Lastly, the trophic levels (δ15N values) of nest partners were more similar on average than those of non-nest partners, although the degree of trophic level similarity did not confer a reproductive advantage to nest partners. This study describes the influence of foraging habitats within diet composition and isotope values of marine predators, as well as within the inter-individual diet variation. Finally, it suggests the occurrence of individual diet specialisation could depend on the predictability of food resources.

Introduction

The diet of any organism is a central feature of its ecological niche (Bearhop et al.

2004, Newsome et al. 2007). Diets are defined by the foods that contain the nutrients that every organism consumes and have been shown to vary with intrinsic factors such as age or sex (Kidawa & Kowalczyk 2011, McGraw et al. 2011, Machovsky-Capuska et al. 2016a,

Machovsky-Capuska et al. 2018). Geographic variation in diet composition within species has also been described (e.g. Simpfendorfer et al. 2001, Tait et al. 2014, Davis et al. 2015), and it is known that the relative abundance of potential food items influences trophic diversity in generalist species (Halaj et al. 1998, Baudrot et al. 2016). Furthermore, recent research has recognised the importance of individual preferences in the selection of food items and, consequently, in the foraging behaviour of species (Dall et al. 2004).

In recent decades, the use of stable isotopes of carbon and nitrogen in diet analysis has contributed greatly towards the understanding of feeding ecology because it provides good proxies for investigating food sources, foraging locations and resource use (through 13C/12C differences; DeNiro & Epstein 1978, Araujo et al. 2007), and trophic positions (through

15N/14N differences; Dalerum & Angerbjörn 2005, Phillips 2012). Traditional methods of diet analysis, such as the examination of stomach contents or regurgitations, are now often used in combination with stable isotope analyses, and together, they have been shown to be a powerful tool for trophic ecology research (Miller et al. 2010, Davis et al. 2012, Franco-

Trecu et al. 2013). Improved statistical tools, such as mixing models, allow inferences to be made about the composition of an animal’s assimilated diet by converting the isotopic data into estimates of food source contributions (Phillips et al. 2014). Recently developed

Bayesian mixing models are able to incorporate a considerable number of sources into this analysis, as well as uncertainties, concentration dependence, and a priori knowledge of the animal’s diet (Parnell et al. 2013, Chiaradia et al. 2014, Phillips et al. 2014). These new

85 techniques provide a clear methodological framework for the analysis of diet compositions that enables the contributions of different food items to be investigated and allows inferences about interspecific relationships to be made.

In many diet studies, conspecifics are considered to be ecologically equivalent

(Araújo et al. 2011). However, the diet of conspecifics may differ intrinsically by their age, sex, morphology, learned abilities, or even preferences (Bolnick et al. 2002, Dall et al. 2004).

If these differences are maintained over time, the development of behavioural consistency can be favoured (Reader 2015). Behavioural consistency in feeding activities or diet can lead to the evolutionary development of individual specialisations within animal populations (Dall et al. 2012), a phenomenon that can be attributed to variation in habitat type and available prey items (Rosenblatt et al. 2015), to the level of intraspecific competition (Lewis et al.

2001), or to the presence of ecological opportunity (Araújo et al. 2011). Given the potential for considerable effects on fitness of individual specialisation, it is important to understand both, the strength (incidence and magnitude) and the temporal consistency of individual diet specialisation within animal populations, particularly under the existent scenarios of human- induced environmental and climate changes (Devictor et al. 2010).

In species with biparental care, the similarity in behavioural features within breeding pairs can have major, long-term effects on reproductive success and fitness (Spoon et al.

2006, Schuett et al. 2011). Assortative mating by diet or trophic level has been shown to occur in some species (Snowberg & Bolnick 2008, Martin 2013). It is believed to take place as a consequence of mate evaluation between individuals when selecting for certain morphological or physiological traits that reflect on similar diets (Snowberg & Bolnick

2008). This within-pair diet similarity has been shown to lead an enhancement of the development and survival of offspring in some species (Forero et al. 2001). However, the presence of this phenomenon in a variety of taxa is poorly understood.

Within marine environments, seabirds play an important role as top predators (Baum

& Worm 2009, Heithaus et al. 2012). The feeding ecology of seabirds, as in other taxa, can be influenced by factors such as age and sex, with differences in the competitive ability of individuals largely affected by these factors, which could determine at the same time the quality and quantity of food resources exploited by each individual (Forero et al. 2002,

Navarro et al. 2010, Phillips et al. 2017). The sex differences in morphology of dimorphic seabird species have also been shown to influence the capture of different prey types, resulting in a reduction of food competition between the sexes (Mancini et al. 2013). The reproductive constraints associated with different stages of the breeding cycle (e.g. incubation vs. chick-rearing), may be another factor contributing to diet variation between and within individuals (Scioscia et al. 2014, Phillips et al. 2017). Because seabirds are effective samplers of prey populations (Fossi et al. 2012), and their diets can provide information about lower trophic levels over a range of spatial and temporal scales (Iverson et al. 2007), knowledge about the factors influencing diet variation is an important part of marine ecology studies and ecosystems models (Fulton et al. 2003).

The Australasian gannet (Morus serrator) is considered an important marine predator in south-eastern Australia and New Zealand, where a population of >50,000 breeding pairs resides (Bunce et al. 2002, Srinivasan et al. 2015). Like other members of the Sulidae family, the Australasian gannet is considered a generalist forager with a reported diet consisting of at least 47 demersal/reef-associated and pelagic/oceanic species throughout its range (Norman

& Menkhorst 1995, Bunce & Norman 2000, Pyk et al. 2007, Schuckard et al. 2012,

Machovsky-Capuska et al. 2016a, Machovsky-Capuska et al. 2018). Australasian gannets display reverse sexual-dimorphism (females larger than males, evident only within certain breeding colonies; Krull et al. 2012, Angel et al. 2015, Ismar et al. 2014) and biparental care, with recent studies suggesting individuals exhibit sex-related differences in habitat use

(Angel et al. 2016, Wells et al. 2016, Besel et al. 2018) and prey selection (Machovsky-

Capuska et al. 2016a). All these traits make this species a good model to investigate individual diet variation and since high levels of inter-individual competition and ecological opportunity persist around the study sites, high levels of individual diet variation can be expected. However, little is known about other factors influencing its diet composition.

Therefore, the aims of the present study were to determine in Australasian gannets: 1) factors influencing inter- and intra-individual variation in tissue δ13C and δ15N isotopic values; 2) inter- and intra-individual variation in diet composition based on analysis of regurgitations and stable isotopes mixing models; and 3) similarity in diet between nest partners and whether the degree of similarity influences reproductive outcomes.

Methods

Study sites and sample collection

The study was conducted over four breeding seasons (October-March, 2012/13 to

2015/16) at Point Danger (PD, 38º 23’ 36.09” S, 141º 38’ 55.94” E) and Pope’s Eye (PE; 38º

16’ 35.88” S, 144º 41’ 56.21” E), two breeding colonies of Australasian gannet in south- eastern Australia (Fig. 4.1). PD is located at the western edge of Bass Strait (BS), near the seasonally active (austral summer) Bonney Upwelling System an important source of primary productivity for the Bass Strait region (Lewis 1981, Butler et al. 2002). PE is located at the entrance of Port Phillip Bay (PPB) on an artificial structure where the birds have access to the shallow waters of PPB (average depth < 13.6 m; Walker, 1999) and to central BS (Angel et al. 2016, Wells et al. 2016).

Fig. 4.1: Location of study sites (black stars): Point Danger (PD, left) and Pope’s Eye (PE, right). Bathymetric contours (every 20 m) within the continental are shown.

As part of concurrent studies on the foraging ecology of Australasian gannets

(Chapter 2), voluntary regurgitations obtained from adult birds during handling and, 1 mL of blood from the tarsal vein (by venipuncture) were collected. Samples were collected during three stages of the breeding cycle: incubation, early chick-rearing (chick age 0-50 d), and late chick-rearing (chick age >50 d; Wingham 1982), with some individuals sampled in more than one breeding stage or year.

Individuals were genetically sexed using a subsample (0.1 mL) of the collected blood

(DNA Solutions, Wantirna South, VIC, Australia). Both partners on nests were sampled and, in most cases, were captured on the same day or within < 5 d of each other. The reproductive output of each nest by breeding season was determined by the single chick reaching 90 days of age (Pyk et al. 2007). Australasian gannets raise at most only one chick per breeding season (Wingham 1982). Chick age was estimated base on regular nest checks every 10 to 14 days, from the egg stage until fledging or chick’s death (Pyk et al. 2013).

Regurgitate samples were collected and stored in plastic bags, while collected blood was centrifuged in the field to separate plasma from red blood cells. All samples were stored frozen (–20º C) until analysis. In the laboratory, regurgitate samples were thawed and prey specimens were identified to the lowest possible taxonomic level using published guides

(Gomon et al. 2008). To establish the isotopic prey base, muscle and mantle tissue (from fish and squid, respectively) were collected from representative individual prey items of all samples analysed (see Chapter 3 for details on prey sample collection). To reconstruct individual diets, stable isotope values of blood plasma were analysed.

Stable isotopes analyses of different body tissues provide information on diet at various timescales due to among-tissue differences in metabolism and protein turnover rates

(Tieszen et al. 1983). Although isotopic turnover rates have not been established for

Australasian gannets, blood plasma in other seabird species has been shown to have a

90 turnover rate of approximately 7 d (Barquete et al. 2013). The duration of a foraging trip in

Australasian gannets is between 25 to 55 hours at colonies in south-eastern Australia (Angel et al. 2016). Consequently, the assimilated diet estimated from stable isotope values analysed in this study would represent food consumption during about 3 to 7 foraging trips prior to taking the sample of plasma.

Stable isotope analysis

Bird blood plasma and regurgitate samples were oven-dried at 50º C for at least 24 h and then ground into powder using mortar and pestle. Lipids were extracted using 1 mL of a

2:1 chloroform–methanol mixture added to the powdered samples. Samples were then shaken using a vortex mixer and centrifuged for 10 min at 10º C (2500 rpm). The supernatant was discarded and the procedure repeated at least once or until the supernatant was clear and colourless after centrifugation (Ehrich et al. 2011). Once dry, 2.0 mg of each sample were loaded into separate tin capsules.

Stable isotope analysis of carbon and nitrogen were conducted at the Farquhar

Laboratory of the Research School of Biology, Australian National University (Canberra,

(–24.62 ‰) for δ13C, and the amino acids alanine, glycine, and cysteine used for δ15N. Runs of laboratory standards provided replicate measurement errors of ± 0.1 ‰ for carbohydrates and 0.3 ‰ for amino acids. Stable isotope values were expressed in δ-notation as the deviation from standards in parts per mil (‰) according to the following equation (Bond &

Hobson 2012): δX = [(Rsample / Rstandard) – 1]

13 15 13 12 15 14 Where, X is C or N and R is the corresponding ratio of C/ C or N/ N. Rstandard values were based on international standards Vienna Pee Dee Belemnite (VPDB) for δ13C,

15 and atmospheric nitrogen (N2) in air for δ N. The mean C:N mass ratio for the prey and bird blood samples was estimated to be 3.17 (± 0.15 SD) and 3.42 (± 0.15 SD), respectively, indicating that lipid concentrations were uniformly low and no data normalization was needed (Post et al. 2007).

Regurgitation analysis and diet reconstruction

For each prey species found in the regurgitation analysis (n = 18), the following parameters were determined for each breeding season: frequency of occurrence (F%), as the percentage of the number of samples containing a particular prey species; number (N); and mass (M%), as the percentage of mass that each species represents of the total amount analysed (Barrett et al. 2007). These general diet estimations and the prey isotopic values were subsequently employed as informative priors and as prey base, respectively, within the individual diet reconstruction process.

Prey isotopic values were previously analysed in Chapter 3 to investigate differences related to the collection year and site. Significant inter-annual differences in either carbon or nitrogen isotopic values were found in four species (i.e. barracouta Thyrsites atun, Gould’s squid Nototodarus gouldi, king gar Scomberesox saurus, and redbait Emmelichthys nitidus).

Significant inter-annual and geographic differences in nitrogen isotopic values were observed in jack mackerel (Trachurus declivis). Consequently, where significant differences existed between groups, the annual or annual-colony means for the species were used for the individual diet reconstruction calculations. For the remainder, the overall means were used.

The Bayesian stable isotope mixing model MixSIAR version 3.1 (Stock & Semmens

2016b) was used to estimate proportional dietary contribution (mass) of assimilated sources to each individual’s diet. Several models were created to consider the inter-annual and inter- colony differences of the prey component. To account for the trophic enrichment between prey items and bird plasma, discrimination factors of 2.4 ± 0.7 ‰ and 0.7 ± 1.6 ‰ for δ13C and δ15N values, respectively, were used. These values follow Lavoie et al. (2012), who summarised discrimination factors from eight studies where birds were raised on controlled diets. Organic concentrations of carbon and nitrogen (%; mean ± SD) were included in the model to compensate for potential differences in elemental composition of the sources

(Phillips & Koch 2002). Informative priors were also included in each model as the seasonal prey frequency of occurrence observed during the regurgitation analysis (Ward et al. 2010).

A multiplicative error structure was included in each model (Stock & Semmens 2016a) and the individual identity was used as a fixed effect to estimate dietary proportions for each individual sampled (Stock & Semmens 2016b). Each model was run with three Markov

Chain Monte Carlo simulations, and a burn-in of 5 x 105 draws, followed by 1 x 106 draws to calculate the posterior distribution and mean proportions. An a posteriori approach to group similar sources was implemented following the recommendations of Phillips et al. (2005a) and the steps given by Stock and Semmens (2016b). In 2013 at PD, blood samples were collected from individuals but no regurgitates were collected. To reconstruct the diet of individuals sampled in 2013, the inter-annual mean isotopic values were used as sources and uninformative priors were included in the model. The Gelman-Rubin criteria, a diagnostic test to determine the adequate model convergence within the Bayesian analysis, was used to evaluate all the models performed (Gelman et al. 2013). Mean prey consumption estimates from MixSIAR were selected for subsequent analyses of inter-individual diet variation.

Statistical analysis

All data processing and statistical analysis were conducted in R version 3.4.1 (R Core

Team 2018). To analyse the factors influencing individual isotopic variation (δ13C and δ15N) of blood plasma from breeding adults, linear mixed effects models considering colony (PD,

PE), year (2012, 2013, 2014 and 2015), breeding stage (incubation, early chick-rearing, late chick-rearing) and sex (male, female) as fixed factors (including all two-way and three-way interactions) were run using the nlme package (Pinheiro et al. 2014). The carbon and nitrogen stable isotope values from bird blood plasma were used separately as response variables. The analyses included consideration of nest and individual identities as random factors to control for longitudinal sampling. Model assumptions were checked by plotting residuals and via quantile-quantile plots. Following modelling recommendations by Zuur et al. (2010), collinearity among all the explanatory variables was checked before running each model using the Variance Inflation Factor (cutting value = 2). The best fixed structure was found using the dredge function of the MuMIn package based on the Akaike Information Criterion corrected for small sample sizes (AICc), with the best supported model considered to be the one with the lowest AICc (Burnham & Anderson 2002, Barton 2016).

The level of inter-individual diet variation in Australasian gannets was calculated using the reconstructed diet from each individual estimated through mixing models (similar to Coblentz et al. 2017 and Resano-Mayor et al. 2014). The diet overlap between each individual and the population was measured by the proportional similarity index (PS; Bolnick et al. 2002). This index considers the proportion of the jth resource category in the individual i’s (pij) and the population’s diet (qj) as follow:

푃푆푖 = 1 − 0.5 ∑ | 푝푖푗 − 푞푗 | 푗

Low PS values (close to 0) indicate individuals who forage on prey that are scarce in the population resource use distribution (specialist individuals). High PS values (close to 1)

94 denote individuals who consume resources in the same proportion as the population as a whole (generalist individuals). The degree of individual specialisation of the population (IS) was then calculated averaging the PS scores across all individuals (Bolnick et al. 2002). Both indices were estimated using the RInSp R package (Zaccarelli et al. 2013). The significance of IS was assessed using a non-parametric Monte Carlo technique to generate 10,000 replicate datasets under the null hypothesis that all individuals are generalists, from which P- values were calculated (Bolnick et al. 2002).

As multiple sampling was performed on most individuals (mean ± SE: 1.8 ± 0.07 samples per bird), the isotopic dataset allowed for comparisons of individual diets over medium-term timescales (breeding stage-to-breeding stage, data obtained from different breeding stages within the same year) and long-term timescales (year-to-year, data obtained from the same breeding stage in two different years). Correspondingly, three medium-term data subsets were formed: incubation to early chick-rearing (INC-ECR); early chick-rearing to late chick-rearing (ECR-LCR); and incubation to late chick-rearing (INC-LCR). In addition, three long-term data subsets were formed: incubation to incubation (INC-INC); early chick-rearing to early chick-rearing (ECR-ECR); and late chick-rearing to late chick- rearing (LCR-LCR). These subsets were used to investigate the intra-individual variation in diet over time through the creation of linear models where the isotope values of the first blood sample were used as response variables, and the isotope values of the second blood sample were used as explanatory variables (Nakagawa & Schielzeth 2010, Carneiro et al.

2017).

To investigate whether nest partners exhibit similar diets, linear mixed effect models were used with paired blood isotopic values from nest partners sampled simultaneously.

Isotope values from females, both carbon and nitrogen, were used separately as response variables, while isotope values from males were used as explanatory variables. Nest and

95 individual identities were included in each model as random components (individual nested within nest). To test whether the isotopic similarity between nest partners confers a reproductive advantage, a generalised linear mixed model with a binomial error distribution and logit link was used. The reproductive output of each nest by breeding season was set as a response variable (i.e. successful = 1 chick fledged, unsuccessful = 0), and the absolute difference between female and male δ13C and δ15N values, colony, and year were used as fixed factors. Nest identity was considered as a random component. This model was fit using the function glmer from lme4 package (Bates et al. 2011). Best models were selected based on their Akaike Information Criterion (Bolker et al. 2008). Model averaging was used to calculate the relative importance of each explanatory variable using the MuMIn package

(Burnham & Anderson 2002, Symonds & Moussalli 2011, Barton 2016). Unless stated otherwise, values presented are mean ± SE.

Results

A total of 298 individual birds belonging to either the PD or PE colonies (143 and

145, respectively) were captured on multiple occasions as part of a foraging behaviour study over the 2012-2015 breeding seasons. From these adults, 404 regurgitated samples (207 from

PD and 197 from PE) were obtained (containing from 1 to 4 prey species each). A total of 18 prey species were identified and used as inputs for the individual diet reconstruction process.

Eight prey species were exclusively collected from PE birds, while five species were exclusively collected from PD birds. Schooling pelagic fish (e.g. Australian sardine

Sardinops sagax, redbait Emmelichthys nitidus and jack mackerel Trachurus declivis) was the most important dietary prey type at both colonies by numerical abundance. Cephalopods (i.e.

Gould’s squid Nototodarus gouldi) were found at both colonies, but were more prevalent at

PD than at PE. With the exception of the bluespotted goatfish (Upneichthys vlamingii), which was the most common benthic species in samples from PE, the remainder of the benthic species collected were present in small numbers and only in particular years. The total list of species and resulting parameters from the regurgitation analysis are presented in the

Supporting information Table S4.1.

Blood plasma samples were obtained from 185 adult individuals (89 females and 96 males), with repeated samples (1-5) acquired from individuals in different years and breeding stages (336 blood samples in total). As both mated individuals in a nest were sampled in most cases, both sexes were well represented among colonies, years, and breeding stages (Table

4.1). The overall mean isotopic values for the blood plasma of adult birds and the range of individual variation were –18.89 ± 0.03 ‰ (–20.64 to –15.90 ‰) for δ13C and +15.27 ± 0.08

‰ (+12.37 to +21.24 ‰) for δ15N.

When investigating the factors influencing individual variation in δ13C and δ15N isotopic values of plasma, significant differences between colonies, sex, sampling years, and breeding stages were found according to model regression parameters. Both carbon and nitrogen stable isotope values indicated that colony and sex were influential factors, with values from PE and from males being higher than those from PD or from females. Year of sampling had a distinct influence on values of both isotopes, with δ13C values in 2013 being lower than in 2012, while those in 2014 and 2015 being highest. For δ15N, 2013 and 2015 values were lower than 2012, while 2014 values were the highest. Breeding stage was also found to be influential for both isotopes, with δ13C values during incubation higher than for those in either early or late chick-rearing stages, and δ15N values during late chick-rearing higher than either incubation or early chick-rearing. The interaction between colony and sex was significant for the combination of PE-males, which had higher values for both isotopes compared to the remainder of the combinations. In addition, the interaction between colony

97 and year of sampling was significant for δ13C values only, with PE values from 2013 higher than the other years of sampling. Lastly, there was a significant interaction between colony and breeding stage for δ15N values, with PE values during incubation higher than the remaining breeding stages (Table 4.2; Fig. 4.2).

Table 4.1: Samples sizes of the blood plasma collected from individual Australasian gannets

(Morus serrator) at both breeding colonies: Point Danger and Pope’s Eye. Data represents the dataset acquired separated by sex (F = females, M = males), breeding stages (INC = incubation, ECR = early chick-rearing, LCR = late chick-rearing), and years (2012 to 2015).

2012 2013 2014 2015 Sex ECR LCR INC ECR INC ECR LCR INC ECR LCR

Point Danger

F 7 2 9 12 7 10 10 9

M 7 5 8 12 11 11 7 10

Pope’s Eye

F 8 4 2 10 10 12 11 14 9 10

M 7 4 7 17 11 11 22 9 9 12

Table 4.2: Linear mixed effect model outputs evaluating the effect of colony, sex, breeding stage, and year on the carbon (δ13C) and nitrogen

(δ15N) stable isotope values quantified from adult Australasian gannets (Morus serrator) blood plasma. Model selection is based on AICc values.

All models include individual identity as random factor. Model variables: colony (Point Danger: PD and Pope’s Eye: PE); sex (male: M and female: F); year (2012, 2013, 2014, 2015); and breeding stage (incubation: INC, early chick-rearing: ECR, and late chick-rearing: LCR).

Response Model Fixed effect Estimate SE df t P df AICc ∆AICc AIC Weight

δ13C colony + sex + year + stage + colony:sex + colony:year* 14 559 0.00 0.85

colony + sex + year + colony:sex + colony:year 12 564 4.53 0.08

(Intercept) -19.54 0.16 183 -119.11 <0.0001

colPE 0.12 0.20 183 0.58 0.56

sexM 0.05 0.12 139 0.46 0.65

year2013 -0.80 0.27 139 -2.95 0.004

year2014 0.47 0.18 139 2.65 0.01

year2015 0.62 0.18 139 3.50 0.001

stageINC 0.21 0.07 139 2.89 0.005

stageLCR 0.04 0.07 139 0.63 0.53

sexM:colPE 0.47 0.15 139 3.11 0.002

colPE:year2013 0.50 0.30 139 1.66 0.10

colPE:year2014 -0.004 0.21 139 -0.02 0.99

colPE:year2015 -0.31 0.21 139 -1.43 0.16

100

Response Model Fixed effect Estimate SE df t P df AICc ∆AICc AIC Weight

δ15N colony + sex + year + stage + colony:sex + colony:stage* 13 1072 0.00 0.93

colony + sex + year + stage + colony:stage 12 1092 5.08 0.073

(Intercept) 14.41 0.25 183 57.523 <0.0001

colonyPE 1.21 0.26 183 4.577 <0.0001

sexM 0.14 0.23 140 0.592 0.555

year2013 -0.37 0.28 140 -1.304 0.195

year2014 0.01 0.22 140 0.024 0.981

year2015 -0.71 0.22 140 -3.236 0.002

stageINC 0.25 0.23 140 1.090 0.278

stageLCR 0.51 0.22 140 2.288 0.024

sexM:colonyPE 0.89 0.31 140 2.875 0.005

colonyPE:stageINC 0.44 0.29 140 1.525 0.129

colonyPE:stageLCR -1.28 0.29 140 -4.382 <0.0001

* Most parsimonious model as shown by AIC values (Burnham & Anderson 2002, Barton 2016).

101

Fig. 4.2: Stable isotope ratios (δ13C and δ15N, mean ± SD) of adult Australasian gannets

(Morus serrator) blood plasma nesting at Point Danger (PD) and at Pope’s Eye (PE) breeding colonies over a four-year period (2012-2015). Three breeding stages: incubation (INC), early chick-rearing (ECR), and late chick-rearing (LCR); and sexes: female (F) and male (M), were considered. The sample size for each group is listed in Table 4.1.

102

During the individual diet reconstruction process using MixSIAR, the Gelman-Rubin criteria met the recommended values (<1.1) in all models performed (Gelman et al. 2013). At

PD, the model diagnostics indicated an inability to discern between three prey types (i.e.

Australian anchovy Engraulis australis, Australian sardine, and king gar) and, consequently, they were pooled to form a single group called “small schooling pelagics”. The posterior density estimates indicated that at PD, small schooling pelagics constituted the majority of the diet in 2012-14, while in 2015 both small schooling pelagics and redbait contributed similar proportions. Other prey species such as barracouta, Gould's squid, jack mackerel and blue mackerel each represented a similar average contribution over the four-year study period

(Table 4.3).

At PE, no a posteriori prey grouping was suggested by the model diagnostics. At this colony, there was greater variation in the prey contributions to individual diets. In 2012, only one species, the jack mackerel, contributed the highest proportion of the diet while, in other years, several prey such as Australian sardine, barracouta, blue weed-whiting, and bluespotted goatfish contributed the greatest proportion of the diet of some individuals (Table

4.3). The degree of inter-individual variation in diet, measured as the dietary overlap between each individual and the population through the proportional similarity index (PS), varied greatly ranging from 0.64 to 0.97 at PD, and from 0.33 to 0.82 at PE (Fig. 4.3). The average degree of individual specialisation (IS) for the sampled colonies was more variable at PE than at PD over the four year period, with an overall average of 0.65 ±0.01 and 0.78 ±0.01, respectively (Table 4.3).

In the regression analysis used to investigate the intra-individual variation in foraging area and trophic position over time, significant and positive relationships were found for both isotopes in all the medium-term scenarios analysed (i.e. INC-ECR, ECR-LCR, INC-LCR).

103

Weak but significant relationships were found in three of the six long-term scenarios for both

δ13C and δ15N examined (i.e. INC-INC ECR-ECR, LCR-LCR, Table 4.4).

Individual nests were sampled on average 1.8 ± 0.1 times during the study. Linear mixed effect models were constructed to test for isotopic similarity in blood plasma of nest partners sampled on the same day (n = 27 and n = 28 nests at PD and PE, respectively).

Female blood plasma δ13C values were not influenced by male δ13C values (linear mixed

13 effects model: t43 = –0.45, P > 0.05). However, females had slightly lower δ C values than males (females: –18.97 ± 0.03 ‰, males: –18.65 ± 0.07 ‰, T test: df = 139, T value = -4.17,

P < 0.001). In contrast, female blood plasma δ15N values were positively influenced by their

15 partner’s blood plasma δ N values (linear mixed effects model: t43 = 5.18, P < 0.001). For

δ15N values, males had higher values than females (females: +14.7 ± 0.11 ‰, males: +15.5 ±

0.19 ‰, T test: df = 164, T value = -3.68, P < 0.001, Fig. 4.4).

Nest partner similarity in isotopic values was investigated as a potential influence for reproductive success. After model averaging, the influential variables found within this test were colony and year (binomial GLMM: χ2 = 21.96, df = 1, P <0.001; χ2 = 17.12, df = 1, P <

0.001, respectively). Neither the absolute difference in δ13C values or in δ15N values between females and males were found to influence breeding success in the nests of study individuals and, consequently, these explanatory variables were excluded from the final model during model averaging procedures (Supporting information Tables S4.2 and S4.3).

104

Fig. 4.3: Proportional similarity index (PS) of Australasian gannets (Morus serrator) from

Point Danger (PD) and Pope’s Eye (PE) breeding colonies over a four-year period (2012-

105

Table 4.3: Mean and range (maximum-minimum) of estimated proportional dietary contribution (% mass) individual Australasian gannets (Morus serrator) at each monitored colony (Point Danger: PD, Pope’s Eye: PE) and year (2012-2015). Proportions were estimated from Bayesian isotopic mixing models (MixSIAR). The “small schooling pelagics” category represents a posteriori group formed by Australian anchovy, Australian sardine and king gar which according to the model diagnostics were unable to be discerned from each other. The number of individuals analysed by year of sampling is indicated in parenthesis. The degree of individual specialisation of the population (IS) and the proportional similarity index

(PS) range estimated by breeding season are shown (Bolnick et al. 2002).

Prey species PD PE 2012 2013 2014 2015 2012 2013 2014 2015 (14) (7) (37) (42) (26) (34) (50) (48) School 76.7 69.8 56.0 27.5 (57.2 – 81.8) (56.4 – 76.1) (24.0 – 77.2) (16.6 – 42.7) Australian anchovy 0.5 1.5 13.7 16.3 (0.2 – 1.6) (0.9 – 2.1) (3.4 – 21.3) (11.0 – 30.4) Australian sardine 0.9 27.3 29.1 36.6 (0.3 – 2.2) (14.0 – 44.3) (6.1 – 49.6) (20.0 – 51.6) Barracouta 10.1 4.9 8.2 15.3 12.6 22.8 7.4 23.9 (7.8 – 17.3) (3.8 – 8.4) (4.4 – 18.1) (7.1 – 20.2) (7.0 – 18.0) (15.6 – 28.5) (2.1 – 11.2) (14.6 – 32.3) Blue mackerel 1.5 8.6 0.8 0.4 0.8 5.4 13.8 3.1 (0.2 – 17.0) (7.7 – 9.9 ) (0.1 – 1.8) (0.1 – 6.6) (0.3 – 1.4) (3.9 – 6.3) (3.2 – 20.5) (2.2 – 4.0) Blue weed-whiting 1.3 0.1 12.5 (1.1 – 1.7) (0 – 0.4) (4.1 – 40.3) 106

Bluespotted goatfish 11.8 16.5 25.2 2.3 (9.2 – 19.2) (8.8 – 38.4) (6.0 - 80) (1.1 – 5.4) Eastern Australian salmon 3.3 0.1 0.1 (2.6 – 4.9) (0.1 – 0.3) (0.1 – 0.2) Flathead 0.08 7.6 0.5 (0 – 0.1) (4.7 – 13.1) (0.3 – 0.9) Gould’s squid 6.1 5.2 11.2 13.0 0.2 3.2 1.2 0.1 (4.3 – 17.4) (4.1 – 7.3) (3.7 – 11.27) (4.4 – 37.4) (0.1 – 0.3) (2.4 – 3.9) (0.6 – 2.0) (0.1 – 0.1) Jack mackerel 1.3 5.7 3.4 3.5 69.4 13.7 8.5 4.1 (0 – 17.3) (4.2 – 9.5) (1.9 – 7.3) (2.0 – 17.2) (64.9 – 74.2) (9.3 – 16.7) (3.3 – 12.1) (3.1 – 5.4) Redbait 8.0 5.5 20.1 39.5 (5.7 – 21.8) (4.2 – 8.5) (10.1 – 32.7) (8.8 – 52.3) Southern garfish 0.1 0.1 0.5 0.3 (0.1 – 0.2) (0.1 – 0.1) (0.3 – 1.0) (0.2 – 0.7) IS 0.70 0.68 0.85 0.76 0.45 0.77 0.67 0.66

(PSmin – PSmax) (0.65 – 0.80) (0.64 – 0.76) (0.70 – 0.97) (0.66 – 0.89) (0.41 – 0.48) (0.62 – 0.82) (0.33 – 0.81) (0.60 – 0.71) Species: Australian anchovy (Engraulis australis), Australian sardine (Sardinops sagax), barracouta (Thyrsites atun), blue mackerel (Scomber australasicus), blue weed-whiting (Haletta semifasciata), bluespotted goatfish (Upeneichthys vlamingii), Eastern Australian salmon (Arripis trutta), flathead (Platycephalus sp.), Gould’s squid (Nototodarus gouldi), jack mackerel (Trachurus declivis), king gar (Scomberesox saurus), redbait (Emmelichthys nitidus), southern garfish (Hyporhamphus melanochir).

107

Table 4.4: Carbon and nitrogen stable isotopes values (mean ± SE) of Australasian gannet

(Morus serrator) blood plasma paired samples and correlation parameters at different medium- (INC-ECR, ECR=LCR, INC-LCR) and long-term (INC-INC, ECR-ECR, LCR-

LCR) time scales. INC = incubation, ECR = early chick-rearing and LCR = late chick-rearing.

δ13C

Time scale n 1st sample 2nd sample R2 Correlation

INC-ECR 28 - 18.82 ± 0.14 - 18.90 ± 0.07 0.73 F1,26 = 71.75, P<0.001

ECR-LCR 28 - 18.70 ± 0.11 - 18.78 ± 0.10 0.47 F1,26 = 23.30, P<0.001

INC-LCR 31 - 18.50 ± 0.18 - 18.74 ± 0.06 0.48 F1,29 = 27.57, P<0.001

INC-INC 14 - 18.91 ± 0.17 - 18.53 ± 0.23 0.46 F1,12 = 10.50, P=0.007

ECR-ECR 18 - 18.97 ± 0.09 - 19.04 ± 0.09 0.05 F1,16 = 0.94, P=0.346

LCR-LCR 15 - 18.32 ± 0.28 -18.52 ± 0.16 0.24 F1,13 = 4.12, P=0.063

δ15N

INC-ECR 28 15.76 ± 0.34 15.67 ± 0.26 0.68 F1,26 = 55.18, P<0.001

ECR-LCR 28 15.41 ± 0.29 15.06 ± 0.27 0.38 F1,26 = 16.51, P=0.003

INC-LCR 31 15.35 ± 0.38 14.80 ± 0.22 0.40 F1,29 = 20.09, P=0.001

INC-INC 14 16 10 ± 0.51 16.43 ± 0.54 0.32 F1,12 = 5.66, P=0.03

ECR-ECR 18 15.51 ± 0.30 14.77 ± 0.25 0.14 F1,16 = 2.80, P=0.113

LCR-LCR 15 15.67 ± 0.40 14.98 ± 0.39 0.38 F1,13 = 8.19, P=0.01

108

Fig. 4.4: Correlations of blood plasma δ13C (top) and δ15N values (bottom) values from

Australasian gannet (Morus serrator) breeding partners (n= 51 pairs) sampled at Point

Danger (purple) and Pope’s Eye (green) breeding colonies. A significant and positive

2 relationship is shown for nitrogen values (R = 0.22, F1,100 = 28.57, P < 0.0001).

109

Discussion

In the present study, inter- and intra-individual variation in the diet of Australasian gannets was investigated. The reconstructed diet composition of gannets nesting at PD suggested that schooling fish was the most preferred prey type foraged, while gannets nesting at PE consumed a diet consisting of both schooling and solitary fish that contributed similarly to the overall diet. Significant differences were found in δ13C and δ15N values from bird blood plasma associated with factors such as colony, year, breeding stage, and sex. The estimated level of inter-individual diet variation (PS) and the average degree of individual specialisation (IS) had higher values at PD (close to 1 = generalists individuals), where

Australasian gannets feed within a uniform habitat, and lower at PE, where birds face a greater diversity of marine environments. These last findings agree with previous studies investigating the diet and habitat use of Australasian gannets in south-eastern Australia in which birds from PE presented a more diverse diet and were able to forage in different environments (Pyk et al. 2007, Barker 2012, Wells et al. 2016). The intra-individual variation was lower over medium-term (breeding stage-to-breeding stage), than over long-term timescales (year-to-year), similar to what has been found for other aspects of the foraging behaviour in Australasian gannets (Chapter 2), or in other seabirds species (Harris et al. 2014,

Camprasse et al. 2017c). In addition, nest partner similarity in trophic levels (δ15N values) was found, although no reproductive advantage was observed in relation to this trait.

Factors influencing inter- and intra-individual isotopic variation

The different oceanographic conditions that prevail between the two study sites investigated showed to have a major influence within some results of this study. PD is near the Bonney Upwelling System in which the seasonal winds running parallel to the continental shelf create highly nutritive water masses and low water temperatures (Lewis 1981). In

110 contrast, PE is located between the entrance of the PPB, a shallow environment with low primary productivity and water movement mainly driven by tidal activity and winds (Harris et al. 1996); and next to northern BS, an area influenced by multiple currents over short timescales (Sandery & Kämpf 2005). Consequently, the prey components found within each study site reflect the nearby environments in which the Australasian gannets forage. For example, at PD, schooling pelagic prey species were identified as the primary food source.

Contrastingly at PE, an assemblage of benthic/inshore species also contributed substantially to the total population diet. In addition, the blood plasma values of individuals differed isotopically between sites, with PE birds presenting higher values for both elements. Higher

δ13C values in the marine environments are associated with coastal/inshore environments due to the distinct carbon fixation processes of phytoplankton and algal communities between these and the pelagic/offshore environments (France 1995a, Newsome et al. 2007). At PE, as the colony is located at the entrance of the semi-enclosed and shallow embayment of PPB

(Walker 1999), individuals from here and particularly males, feed constantly within its waters

(Chapter 2, Angel et al. 2016, Wells et al. 2016). In contrast, the observed δ13C values coming from individuals at PD agree with reported values from isotopic analyses in marine offshore environments (Hobson et al. 1994, Cherel & Hobson 2007).

Concerning δ15N values, individuals from PE also presented higher values than those from PD which suggest they are feeding at a higher trophic position (Vanderklift & Ponsard

2003). For PE, some of the prey species were found exclusively in this colony during the regurgitation analysis (i.e. yellowfin goby Acanthogobius flavimanus, Eastern Australian salmon Arripis trutta and flathead Platycephalus sp.) and these are indeed predators of fish and crustacean species (Middleton 1982, Edgar & Shaw 1995, Bulman et al. 2001).

Additionally, inshore environments tend to have longer food-chains, and consequently, greater nitrogen absorption through the trophic levels (Hobson & Welch 1992, Zanden &

111

Rasmussen 2001). This aspect could potentially have an effect on δ15N values of other prey species collected exclusively from PE, such as the blue weed-whiting Haletta semifasciata, southern garfish Hyporhamphus melanochir and bluespotted goatfish Upeneichthys vlamingii, which are consumers of invertebrates and plant matter (Robertson & Klumpp

1983, MacArthur & Hyndes 2007, Currie & Sorokin 2010). Contrary at PD, where the majority of the diet constituted schooling fish (i.e. Australian sardine, Australian anchovy and king gar); δ15N values from individual gannets were lower as these pelagic preys are more likely to feed from zooplankton and other free-living organisms from the water column

(Dudley et al. 1985, Currie & Sorokin 2010).

Environmental variability has been shown to play an important role in the distribution and abundance of marine resources (Paiva et al. 2013). In this study, inter-annual variability was evident on the isotopic blood plasma values of the individuals sampled. Higher δ13C values were observed in 2015, while for δ15N the highest values were seen in 2014. As there was no evidence that the sampled Australasian gannets changed their foraging locations during the course of this study (Chapter 2, Angel et al. 2015, 2016, Wells et al. 2016), these differences could potentially be related to changes in the composition of their diet over time.

Inter-annual differences have been shown before in other dietary studies for this and others species of seabirds (Hedd et al. 2002, Schuckard et al. 2012, Besel et al. 2018). Indeed, during the regurgitation analysis performed, variation in the annual proportions of the consumed prey items were seen. For example, in 2015 at PD a fluctuation was observed in which redbait alternated with the Australian sardine and barracouta as the prey item most consumed of that year. Similar at PE, the jack mackerel shown to be the most consumed prey item in 2012, but its frequency of occurrence dropped considerable during the remaining years (Supporting information Table S4.1). These inter-annual effects were also reflected

112 within the individual diet reconstructions in which the proportions of consumed prey varied between years.

Sexual segregation in stable isotope values was also evident during this study. Both

δ13C and δ15N values were higher in males, and in particular, males from PE. These differences may be explained by the recently reported inshore feeding behaviour adopted by some males of this colony. Wells et al. (2016), using GPS and video data loggers, described an inshore foraging strategy in which male individuals use sandbanks of the shallow waters of PPB and, in the absence of conspecifics, they hunt large single prey items. Although not all the males from PE follow this strategy (Rodríguez-Malagón unpublished data), the majority of the individuals using this inshore foraging strategy were males. Contrary, the majority of the females from PE have been reported to forage almost exclusively in Bass

Strait (Angel et al. 2016). In this area, the carbon enrichment coming from inshore waters losses strength (Gearing 1988a) making female δ13C values lower and more similar to those from males and females from PD. These findings agree with previous studies of Australasian gannets in New Zealand suggesting males forage more consistently on prey at the highest possible trophic level (Machovsky-Capuska et al. 2016a, Ismar et al. 2017).

The breeding stage in which individuals were during sampling period also showed to have an influence in both isotopes. Higher δ13C values were present during the incubation period and higher δ15N were seen during the late chick-rearing period. Unfortunately, in this study, the sample size of the regurgitation samples collected did not allow for their temporal differentiation in relation to the breeding stage of the birds, limiting inferences that can be made. However Pyk et al. (2007), during a comprehensive diet analysis of adult birds from

PE, showed monthly fluctuations in the dominance of the prey species found, suggesting variation in prey availability due to environmental changes around the colony during the breeding period. Similar findings have been reported for other seabird colonies in which adult

113 birds switched from one prey item to another during the incubation period, and then switched prey item again during the chick-rearing period, and matching these behaviours with changes in the local sea surface temperatures (Ito et al. 2009).

Inter- and intra-individual diet variation

The diet overlap between each individual and the population (PS) and, consequently, the degree of individual diet specialisation of the population (IS), presented differences related to the colony in which each individual was sampled. In all four years (2012-2015), individuals from PD presented higher PS values than individuals from PE, indicating the later to be more specialised (Bolnick et al. 2002). Specifically, individuals from PE differ more in their proportions of consumed prey items, particularly benthic species, in respect to others.

These results concur with studies investigating the foraging behaviour of Australasian gannets in which higher levels of behavioural consistency were found in those individuals foraging inside the inshore environment of PPB (Chapter 2). This also agrees with the knowledge that dietary specialisations are more likely to take place within a population when individuals are able to feed upon a large diversity of prey as this opens the ecological opportunity for foraging strategy diversification (Araújo et al. 2011).

In terms of sexual differences, the average PS values found for each sex presented similar values when estimated by colonies and years. Nonetheless, males presented larger variation and reached the lowest numbers, in particular males from PE. This agrees with distinct foraging behaviour of the males from this colony previously described, which according to the individual specialisation theory, could be helping these individuals minimising intraspecific competition (Lewis et al. 2001, Araújo et al. 2011).

In this study, the intra-individual diet variation showed high significant correlations for both isotopic values, δ13C and δ15N, over medium-term scale (same tissue, two samples

114 from the same individual compared over two breeding stages) suggesting that individuals exhibit diet specialisation that persists throughout most of the year (Bearhop et al. 2006,

Wakefield et al. 2015). However, over the long-term timescales (same tissue, two samples from the same individual compared over two breeding seasons), isotopic correlations showed weaker results suggesting specialisations are stronger over the shorter periods. This is of interest as similar behavioural consistency patterns have been found in other aspects of the foraging behaviour of Australasian gannets, such as the foraging trip characteristics including maximum distances from the colony and tortuosity index (Chapter 2). In northern gannets, a similar study using red blood cells isotopic values found evidence of individual diet specialisation in δ13C values within years (medium-term scale), but not across years (long- term scale; Wakefield et al 2015). However, individual diet specialisation with respect to the trophic level (i.e. δ15N values) was found persisting over both medium- and long-term scales

(Wakefield et al. 2015). In other seabird species, a similar time decreasing trend has been shown regarding individual diet specialisation (e.g. Woo et al. 2008), a trend that is also shared with other marine animals (e.g. Novak & Tinker 2015). These results suggest that the occurrence of individual diet specialisation may depend on the predictability of resources and that under fluctuating prey resources, individual specialisation may only be beneficial over shorter periods of time (Dehnhard et al. 2016). It is important to consider that as δ13C values vary along horizontal and vertical gradients in the marine environment (Cherel & Hobson

2007), the correlated isotopic values in gannet tissues may also reflect not only dietary repeatability but also spatial consistency (Wakefield et al. 2015).

Diet similarity between nest partners

It has been proposed that successful reproduction or partner retention in species with complex mating relationships (e.g. long-term pairing species) requires well-coordinated

115 interactions between partners. Thus, correlated behavioural traits or diet could favour partners’ coordination and reproductive success in some species (Spoon et al. 2006, Schuett et al. 2011). For seabirds, partner correlation or similarity has been found in relation to the body size, foraging behaviour and trophic levels (e.g. Forero et al. 2001, Camprasse et al.

2017a). In Australasian gannets, within-pair similarity was found for δ15N values, suggesting that nest partners feed within similar trophic levels. For this species, recent research has shown individuals co-departing the colony are more likely to have similar foraging patches

(Jones et al. 2018). However, Australasian gannet nest partners do not hunt together during the breeding season as a strategy for nest attendance (Ismar 2010), and information transfer among nest partners has not been found during their social interactions at the colony

(Machovsky-Capuska et al. 2014a). Then, it remains unclear how individuals are able to assess mates’ diet. In other species, assortative diet is believed to occur as a consequence of mate evaluation between individuals when looking for certain morphological or physiological traits which eventually reflect on a similar diet (Camprasse et al. 2017b). Future research could investigate the spatial overlap in foraging range between nest partners that could potentially explain why partners tended to consume prey at similar trophic levels.

Lastly, no relationship was found between reproductive success of a pair and the isotopic similarity between nest partners, but colony and year were factors that influenced reproductive success. Australasian gannets have been shown to be highly adaptable in their feeding habits (Bunce 2001, Schuckard et al. 2012, Machovsky-Capuska et al. 2016a) and are able to modify their prey intake in years of low prey availability (Bunce & Norman 2000,

Machovsky-Capuska et al. 2018). Considerable links between prey availability and breeding success had been shown for several species of seabirds (e.g. Cairns 1988, Cury et al. 2011). It is therefore likely that variation in environmental conditions near particular colonies and between years at each colony had a stronger influence on the reproductive output. Across

116 years, the reproductive success (% of nests where chicks fledged) within these colonies has been highly variable. During the course of the present study the breeding success was 30.1% and 10.0% in 2012, 8.6% and 0% in 2013, 24.7% and 48.0% in 2014 and 50.5% and 79.3 % in 2015, at Pope’s Eye and Point Danger, respectively (Angel et al. 2015, Rodríguez-

Malagón unpublished data).

117

Supporting information

Table S4.1: Diet composition analysis from regurgitate samples of Australasian gannets (Morus serrator) collected at Point Danger and Pope’s

Eye breeding colonies over a four year period (2012-2015). Diet is described as frequency of occurrence (F%), numerical abundance (or number of individuals per species, N) and mass percentage (M%). An asterisk indicates species that uses the benthos as preferred habitat, those without it are considered pelagic (www.fishesofaustralia.net.au). Sample size (n) and total mass collected (kg) for F% and M% estimation, respectively, are shown.

Prey species 2012 2013 2014 2015 F% N M% F% N M% F% N M% F% N M% (6) (0.9) (93) (15.5) (73) (12.3) Point Danger Australian anchovy 33.3 2 11.7 2.2 2 0.1 13.7 32 6.0 Australian sardine 16.7 1 0.4 61.3 230 40.5 13.7 37 13.3 Barracouta 33.3 3 33.6 18.3 24 18.3 17.8 27 13.1 Blue sprat 1.1 4 0.01 Gould’s squid 16.7 1 33.2 4.3 8 2.0 6.8 18 2.1 Jack mackerel 8.6 8 7.1 4.1 4 5.2 King gar 16.7 7 13.0 5.4 18 3.3 4.1 9 1.7 Longsnout boarfish 1.1 1 0.2 Redbait 16.7 2 8.1 24.7 59 27.5 61.6 158 58.6 Velvet leatherjacket* 1.1 3 1.0

118

Prey species 2012 2013 2014 2015 F% N M% F% N M% F% N M% F% N M% (12) (1.6) (18) (2.0) (79) (12.9) (69) (12.2) Pope’s Eye Australian anchovy 13.9 22 4.7 17.4 40 9.2 Australian sardine 11.1 5 7.9 26.6 65 21.3 46.4 116 31.2 Barracouta 8.3 1 10.5 33.3 6 36.5 19.0 24 26.6 40.6 54 37.8 Blue mackerel 5.6 1 5.6 10.1 9 14.9 4.3 3 8.2 Blue weed whiting* 5.6 1 2.2 1.4 1 1.0 Bluespotted 16.7 3 9.2 27.8 8 34.4 16.5 16 8.2 10.1 10 5.7 goatfish* EA salmon 8.3 2 11.3 1.3 1 1.0 Flathead* 11.1 2 5.2 1.4 1 0.7 Gould’s squid 11.1 2 1.6 3.8 3 1.7 Jack mackerel 75.0 11 69.0 11.1 3 6.6 21.5 28 19.8 5.8 6 5.2 Snook 1.3 1 0.4 Southern garfish 2.5 6 1.5 1.4 1 0.7 Yellowfin goby* 1.4 1 0.3 Species: Australian anchovy (Engraulis australis), Australian sardine (Sardinops sagax), barracouta (Thyrsites atun), blue mackerel (Scomber australasicus), blue sprat (Spratelloides robustus), blue weed-whiting (Haletta semifasciata), bluespotted goatfish (Upeneichthys vlamingii), Eastern Australian salmon (Arripis trutta), flathead (Platycephalus sp.), Gould’s squid (Nototodarus gouldi), jack mackerel (Trachurus declivis), king gar (Scomberesox saurus), longsnout boarfish (Pentaceropsis recurvirostris), redbait (Emmelichthys nitidus), snook (Sphyraena novaehollandiae), southern garfish (Hyporhamphus melanochir), velvet leatherjacket (Meuschenia scaber),yellowfin goby (Acanthogobius flavimanus).

119

Fig S4.2: Medium-term (left panels) and long-term (right panels) correlations between two breeding stages of the same year, or the same breeding stage of two different years. Values represent carbon isotopic ratios (δ13C ‰) from blood plasma of adult Australasian gannets (Morus serrator) sampled at two breeding colonies Point Danger (squares) and Pope’s Eye (circles). Males are represented in blue, females in red.

121

Table S4.2: AICc based model selection (∆ < 4) for factors that influence the reproductive output

(binomial: successful = 1 (chick fledged), unsuccessful = 0), in Australasian gannets (Morus serrator). Model variables: Diff_dC: the absolute difference between female and male δ13C values; Diff_dN: the absolute difference between female and male δ15N values; colony; year. All models share the same random component (nest identity).

Model no. Model fixed effects df AICc ∆AIC AIC Weight

15 colony + year 4 60.1 0 0.44

11 Diff_dC + colony + year 5 62.1 2.0 0.16

16 Diff_dN + colony + year 5 62.3 2.1 0.15

12 year 3 62.5 2.3 0.14

122

Table S4.3: Average model coefficients and relative importance of variables included in top model set (∆AICc ≤ 4) explaining individual variation in Australasian gannets (Morus serrator) reproductive output (binomial: successful = 1 (chick fledged), unsuccessful = 0). Relative variable importance based on AIC weights. Model variables: Diff_dC: the absolute difference between female and male δ13C values; Diff_dN: the absolute difference between female and male δ15N values.

Parameter Estimate SE 5 % CI 95 % CI Relative importance

(Intercept) -6.97 8.10 -20.25 5.92 ---

year 20.54 4.66 12.72 28.28 1.00

colony 21.40 6.00 11.52 31.39 0.84

Diff_dC 1.59 3.95 -5.27 8.44 0.18

Diff_dN 0.27 1.15 -1.88 2.29 0.17

123

CHAPTER 5

Factors influencing winter distribution and habitat use in

Australasian gannets (Morus serrator)

124

Abstract

In order to survive the scarcity of the winter season, many animal species are force to move into more profitable places. It is known that factors such as age or sex influence winter dispersal strategies of individuals within a population, leading to diversification in resource use.

Additionally, certain individual decisions such as when to depart/return or where to disperse, appear to be flexible for individuals, although the factors influencing these decisions are poorly understood. The Australasian gannet (Morus serrator) is an important marine predator breeding in south-eastern Australia, but more information on its movement patterns during the winter non- breeding season is needed. The south-eastern Australia region has been identified as one of the fastest warming marine areas in the world. Thus, it is essential to understand the species distribution and the level of resource partitioning among individuals throughout the year, as predicted changes in marine currents happening within this region are able to alter the abundance and distribution of marine resources. In the present study, geolocation devices deployed on adult individuals from two breeding colonies were used to investigate five winter dispersal metrics

(spatial extent, bearing from the colony to the most distal point, total duration, departure date, and return date). Significant differences for these metrics were found between individuals associated with colony, year, and sex. Study individuals were classified into three distinctive winter dispersal strategies, distinguishable mainly by differences in habitat use. Low intra- individual variation among study years was found in relation to their spatial extent. Lastly, nest partners sampled within the same year did not showed to have similar values in the winter dispersal metrics analysed. This study describes the winter distribution of the Australasian gannets nesting in south-eastern Australia revelling some aspects that could potentially contribute to the behavioural consistency of individuals during this period.

125

Introduction

When vital resources become limited during winter in temperate and polar regions, many animal species leave their summer habitats as a strategy for survival (Bowlin et al. 2010). This winter dispersal (or in some cases migration) is considered to be the result of complex interactions between the animal’s needs and environmental factors such as weather, food availability, and/or geography (Åkesson & Hedenström 2007, Bowlin et al. 2010). Within populations, different winter dispersal strategies driven by age or sex can lead to diversification in resource use and a reduction in intra-specific competition (Chapman et al. 2011a). In addition, such differences can lead to segregation at local or regional scales (Craig et al. 2003, Bunnefeld et al. 2011) resulting in ecological processes such as predation, inclement weather, and food scarcity to act differently on various segments of the population (Cristol et al. 1999). Thus, differential winter dispersal could potentially affect individual survivorship and alter population demography (Sandercock et al. 2002, Jorge et al. 2011).

Winter dispersal has evolved independently in different animal groups (Alerstam et al.

2003) and different strategies (e.g. short- and long-distance and/or partial) have been described both between and within-species (Ketterson & Nolan 1983). Evidence of different winter dispersal strategies in closely related species living in similar environments suggest animals may use different cues to select their dispersal times and/or that different biological constraints may act upon them (Myers 1981, Francis & Cooke 1986, Emmenegger et al. 2014). Furthermore, individual variation in winter dispersal strategies not related to the sex or age have been reported for several species, with some studies suggesting personality traits may act as major drivers of these individual differences (Bêty et al. 2004, Chapman et al. 2011b). Consequently, knowledge

126 of the distribution and movements of animals during their winter dispersal is necessary for a better understanding of the interactions they have with their environment (Jorge et al. 2011).

In seabirds, winter post-breeding dispersal is necessary in many species as resource availability changes (Phillips et al. 2009). The extent of dispersal varies between species, ranging from long-distance travellers circumnavigating the globe to those that rarely leave the general vicinity of their breeding colonies (Schreiber & Burger 2001). It has been shown for several seabird species that winter dispersal strategies (e.g. departing and return dates or distribution) may vary between individuals of different sexes (Akesson & Weimerskirch 2014, Müller et al.

2014), or ages (Huettmann & Diamond 2000, Péron & Grémillet 2013). For some species, a link between the prior reproductive output and the winter dispersal strategy has also been found, with potential implications for the fitness of individuals and/or reproductive outcomes following dispersal (Catry et al. 2013, Bogdanova et al. 2017). However, decisions of when to depart/return or how far to disperse, appear to be flexible for individuals of certain species and populations

(Dias et al. 2011, Dias et al. 2013, Tranquilla et al. 2014), although the factors influencing these individual decisions are poorly understood. Besides, winter has been recognised as a period of high mortality for seabirds (Leat et al. 2013, Klaassen et al. 2014). Therefore, understanding the factors influencing winter dispersal behaviour in seabirds species is essential to accurately predict their distribution, habitat use, and ecosystem impacts throughout the annual cycle, especially under the current scenarios of changing environments (Sydeman et al. 2012, Phillips et al. 2017).

Most seabirds are long-lived species and display biparental care of offspring (Schreiber &

Burger 2001). Behavioural similarity within nest partners over the breeding season has been shown to have significant long-term effects on reproductive success (Spoon et al. 2006, Schuett

127 et al. 2011). Recent studies also suggest nesting pairs with similar winter dispersal dates develop stronger long-term bonds that help to increase reproductive success and tend to reduce divorce rates (Naves et al. 2007, Sánchez-Macouzet et al. 2014). However for many seabird species, information on their winter dispersal is limited and little is known of within-pair similarity and other factors affecting their winter behaviour.

The Australasian gannet (Morus serrator) is an endemic pelagic seabird of south-eastern

Australia and New Zealand (Norman & Menkhorst 1995). It breeds during the austral summer

(from August to March), with adult birds usually leaving the breeding grounds during March-

April of each year (Nelson 1978). Recent studies suggest individuals exhibit sex-related and inter-colony differences in habitat-use during the summer/breeding season (Angel et al. 2016,

Machovsky-Capuska et al. 2016a, Wells et al. 2016, Besel et al. 2018). However, little is known of their movement patterns or about potential inter- and intra-individual differences in behaviour during the winter non-breeding period (Ismar et al. 2010b, Ismar et al. 2011), particularly within the Australian range. The south-eastern Australia region is considered one of the fastest warming oceanic areas in the world with significant changes to ocean currents predicted to occur

(Ridgway 2007, Lough & Hobday 2011). Such changes are likely to alter the distribution and abundance of marine species which could have serious implications for the survival of top predators (Hobday & Pecl 2014, Pecl et al. 2014). Therefore, knowledge of the factors influencing the winter dispersal of this species is necessary to predict how their populations may respond to the anticipated changes.

The Australasian gannet is considered an inshore species, with foraging ranges restricted to the continental shelf (Nelson 1978). Within the study sites, this species has been shown to display reverse sexual-dimorphism (females larger than males; Angel et al. 2015b), and it is

128 considered socially monogamous (Nelson 1978), although recent resarch has reported high divorce rates (40-44%) between breeding seasons (Ismar et al. 2010a). Long-lived species with complex mating systems are considered to be particularly vulnerable to climate change as they are constrained by different biological conditions and require the permanence of vital resources in the long term (Croxall et al. 2012, Sydeman et al. 2012). Investigating the winter dynamics among nest partners can help to understand their interactions within this period and their potential association with the maintenance of the pair bond and, consequently, with the survival of the species (Griggio & Hoi 2011).

The south-eastern Australia region is oceanographically complex with specific water masses switching their influence in the region between the summer and winter periods

(Commonwealth of Australia 2007, 2015). In addition, Australasian gannet sex-related differences in size and foraging efficiency (Angel et al. 2015b, Chapter 2) may contribute to create variation within the winter dispersal strategies of this species. Therefore, the aims of the present study were to determine in adult Australasian gannets nesting in south-eastern Australia:

1) winter foraging areas; 2) factors influencing winter dispersal behaviour; and 3) intra- individual variation and within-pair similarity in winter dispersal behaviour.

Methods

Study sites and animal handling

The study was conducted over two austral winter non-breeding periods (April-September,

2015 and 2016) at two Australasian gannet colonies in south-eastern Australia: Point Danger

(PD, 38º 23’ 36.09” S, 141º 38’ 55.94” E) and Pope’s Eye (PE; 38º 16’ 35.88” S, 144º 41’56.21” 129

E). To obtain quantitative measures of individual winter dispersal behaviour, Global Location

Sensor (geolocation) data loggers (MK3005, sampling frequency 5 min, Biotrack, Wareham,

United Kingdom) were deployed on adult birds at both colonies during the summer (October-

March). Individuals were captured by hand or with the aid of a noose-pole (Garthe et al. 2014) when attending their nest during the chick-rearing period. Where possible, both nest partners were used in the study.

The geolocation loggers were attached with cable ties to a plastic leg band and placed on the left leg. A blood sample (0.1 mL) was then obtained by venipuncture of a tarsal vein for genetic sexing (DNA Solutions, Wantirna South, VIC, Australia) before the bird was returned to the nest. Geolocation loggers were collected in the following summers (2015/16 and 2016/17), providing data from a single winter or from two consecutive winters. As Australasian gannets raise only one chick per season (Wingham 1982), the reproductive outcome of each study nest was determined by regular (every 10-14 d) monitoring until fledging or chick death (Pyk et al.

2007, Pyk et al. 2013).

Data processing and statistical analysis

Unless stated otherwise, all analyses were conducted within the statistical environment R version 3.4.1 (R Core Team 2017). Data from recovered geolocation loggers were downloaded and decompressed using the M-Series BAStrack software suite (Biotrack, Wareham, United

Kingdom) with clock drift adjustment set using each logger’s start date/time. To obtain locations from the geolocation data the twilight-free method developed by Bindoff et al. (2017) was used.

This method uses a Hidden Markov Model in which the hidden states are the daily geographic locations and the measured response is the observed pattern of light and dark over the day. A

130 grid of cells (limits: -55º to -30º S, 110º to 170º E) of 0.25º x 0.25º was set as reference for the geolocation estimations (each 0.25° grid cell corresponding to approximately 25 km2). The obtained data was then divided in successive segments of 24 h in which each bird was assumed to maintain its geographical position. By using standard astronomical formulae (Meeus 1991), this method determined when each cell would be in daylight for any given day. This information was later used to compare the observed light record and the expected pattern of day and night at each cell using zenith angles calculated for the centre of each cell. Using a calibrated zenith angle of 96º, each cell was evaluated to estimate the likelihood of the bird's position being inside that cell assuming a Gamma distribution (1, 1/25). As it is known that Australasian gannets restrict their foraging ranges to the continental shelf (Nelson 1978), the shelf’s geographical limits were used as a prior within the Bayesian method of these calculations.

To correct the processed data for potential unrealistic locations, an iterative forward- backward averaging speed filter (McConnell et al. 1992) was applied. A maximum average speed of 55 km·h-1, as suggested by Hamer et al. (2007) for northern gannets, was employed using the trip package (Sumner 2016). Then, the geolocation data was visually inspected to define the departure and return dates from the colony area corresponding to the winter dispersal of each individual. A decision rule of 1º in longitude and 0.5º in latitude, based on data from breeding season tracking (Chapter 2, Angel et al. 2016), was set to define a threshold after which the birds were considered to be beyond the distance of an average foraging trip and had commenced their winter dispersal. Subsequently, for each dispersal track, the behavioural metrics of maximum distance from the colony (km), total distance travelled (km), total duration

(d) and bearing (0-360º, from the colony to the most distal point) were calculated using the adehabitatHR package (Calenge 2006).

131

As geolocation estimations can have considerable estimation errors (e.g. ± 169-186 km;

Phillips et al. 2004), data should be analysed in relation to its spatial distribution, such as core areas. In the present study, the time-in-area (time birds spent in each cell) was estimated using the function rasterize from the trip package (Sumner 2016), which adequately represents the foraging intensity in a given area (Warwick-Evans et al. 2015). For time-in-area calculations, the total area used by individuals was defined as the sum of all grid cells occupied. Cells were then ranked in order of time spent in each one, and separated into percentages of use by adding the cells that covered first 50%, 75% and 95% of the cumulative frequency distribution (Soanes et al. 2013a). The above gridding parameters were used for these calculations and the obtained distributions were plotted using ArcGIS 10.1 software (©1999-2012 Environmental Systems

Research Institute Inc., Redlands, CA). Additionally, the total number of grid cells occupied by individuals within the winter dispersal period was considered as a measure of each individual foraging range (Soanes et al. 2014).

Three of the dispersal metrics extracted from the geolocation data (maximum distance to colony, total distance travelled, and total number of grid cells occupied) were highly correlated

(Spearman’s rho > 0.8) and, consequently, only the total number of grid cells occupied was considered in further analyses. To analyse the factors influencing the winter dispersal behaviour, linear mixed effects models were created using the nlme package (Pinheiro et al. 2014). The dispersal metrics of total number of grid cells occupied (0.25º x 0.25º), total duration (d), average bearing (0-360º, from the colony to the most distal point), departure and return dates (Julian) were used separately as response variables, while the fixed factors of colony (PD, PE), year (2015,

2016) and sex (male, female), were used as explanatory variables. The reproductive performance of each nest during the reproductive season prior to the winter dispersal (i.e. successful = 1

132

(chick fledged), unsuccessful = 0) was also set as explanatory binary variable in the models realised for each dispersal metric. Within the models performed for the metrics of total number of grid cells occupied and bearing to most distal point, the departure date was also set as an explanatory variable.

Following modelling recommendations by Zuur et al. (2010), collinearity among all the explanatory variables was checked prior to modelling using a cutting Variance Inflation Factor level of 2. Additionally, model assumptions were checked by plotting residuals and using quantile-quantile plots. Initial models were then fitted with restricted maximum likelihood

(REML) and the random structure (individual identity nested in nest identity) was compared from models with and without it using the anova function. The dredge function of the MuMIn package (Barton 2016) was then used to find the best fixed structure using models refitted with maximum likelihood (ML). Where no single model had an AICc weight above 0.90, model averaging considering those models within ΔAICc ≤ 4 were used to calculate the relative importance of each explanatory variable (Burnham & Anderson 2002, Symonds & Moussalli

2011, Barton 2016).

To investigate the presence of different winter dispersal strategies, an agglomerative hierarchical clustering analysis with Euclidian distance and Ward’s linkage criterion (Kaufman

& Rousseeuw 2009) was used incorporating the winter dispersal metrics estimated for each individual within each year. The function HCPC of the FactoMineR package was used to determine the appropriate number of clusters (Lê et al. 2008). The degree of intra-individual variation between two consecutive years was estimated parametrically by correlation analysis using the observed value of the estimated winter dispersal metrics in years 2015 and 2016

(Nakagawa & Schielzeth 2010, Carneiro et al. 2017). To evaluate the robustness of these

133 correlations, non-parametric permutation test were run in which 2015 values were shuffled without replacement 10,000 times and the distribution of null values for the correlation coefficient and how often null values were more extreme than the observed one were determined (following

Snowberg & Bolnick 2008). Similarly, the degree of behavioural similarity between nest partners sampled within the same year was investigated through correlation and permutation analyses. A significant P-value in the permutation analysis was used to determine if 2015-2016 value or male and female values were more similar than expected by change (Higgins 2004). Unless otherwise stated, results are presented as mean ± SE.

Results

Due to device malfunction, devices being lost at sea or from individuals not being resighted, useful data were obtained from 121 (PD 61, PE 60) out of the 167 GLS loggers initially deployed. As some individuals were instrumented twice, these deployments represented only 91 individuals (PD 46, PE 45). A total of 32 individuals (PD 16, PE 16) were tracked in both winters and data were obtained from 37 nest pairs (PD 2015 and 2016, 9 and 10 pairs, respectively; PE 2015 and 2016, 7 and 11 pairs, respectively).

Winter distribution and factors influencing dispersal behaviour

During the winter non-breeding period, tracked individuals were widely distributed over southern Australian waters, ranging from the vicinity of their breeding colonies in Port Phillip

Bay and western Bass Strait, to mid-east coast of Australia at the east, and to south-west

Australia at the west. The core areas (50% of intensity use) of the distribution varied between 134 colonies, sexes, and years of study. In general for Point Danger, female birds ranged from central

Bass Strait to west Australia, with concentrated distribution in areas with high occurrence of submarine canyons and rich continental self-areas such as the Great Australian Bight and the

Spencer Gulf. In contrast, males from Point Danger concentrated their foraging activity mostly around Spencer Gulf and central Bass Strait. Females from the Pope’s Eye colony also reached west Australia and the mid-east coast of Australia, but concentrated their distribution in central

Bass Strait. While some males from this colony also reached the Great Australian Bight during the winter, this group displayed the least dispersal of all with the majority of birds concentrating their activity in central Bass Strait. In general, birds from both colonies reached further distances, either to the east or to the west, during winter dispersal in 2015 than in 2016 (Fig 5.1 and 5.2). A summary of the calculated winter dispersal metrics is presented in Table 5.1.

After model averaging, the variables to which influence on total number of grid cells occupied could be assigned were colony, year, and sex. Specifically, birds from PD, birds sampled in 2016, and females showed greater values. For bearing to the most distal point, colony was the only influential variable with individuals from PD having higher values reflecting their more westerly movements. Similar, total duration was influenced by colony and sex with individuals from PD and females displaying the larger values. Lastly, departure dates were influenced only by year (later departure in 2016), while return dates were influenced by colony

(PD later return) and sex (females later return; Tables 5.2, S5.1 and S5.2). The reproductive performance of each nest during the breeding season prior to the winter dispersal did not show influential values to any of the winter dispersal metrics analysed (Supporting information Table

S5.2).

135

Table 5.1: Means ± SE of winter dispersal metrics for adult Australasian gannets (Morus serrator) during

2015 and 2016 austral winters (March to September) at Point Danger and Pope’s Eye breeding colonies, south- eastern Australia.

Sex-year Females-2015 Females-2016 Males-2015 Males-2016

Point Danger

Birds tagged 13 19 15 15

Maximum distance from colony (km) 1,366 ± 177 1,114 ± 163 887 ± 110 788 ± 126

Total distance travelled (km) 7,286 ± 426 5,823 ± 477 5,714 ± 493 5,101 ± 373

Total duration (d) 180.8 ± 9.5 161.5 ± 9.1 145.6 ± 8.7 148.2 ± 11.4

Bearing (º) 274.3 ± 17.1 242.5 ± 19.5 243.1 ± 24.4 248.6 ± 22.2

Number of grid cells occupied (0.25ºx0.25º) 56.7 ± 3.5 38.2 ± 4.6 41.5 ± 4.0 29.4 ± 2.5

Departure date (Julian) 54.8 ± 7.0 79.7 ± 6.0 77.5 ± 3.9 74.7 ± 8.1

Return date (Julian) 234.6 ± 5.6 240.3 ± 4.9 222.1 ± 6.1 222.0 ± 6.8

Pope’s Eye

Birds tagged 13 14 14 20

Maximum distance from colony (km) 615 ± 171 464 ± 187 363 ± 150 261 ± 82

Total distance travelled (km) 4,670 ± 409 3,896 ± 394 3,726 ± 401 3,017 ± 285

Total duration (d) 137.3 ± 9.3 131.1 ± 6.1 137.5 ± 7.1 118.7 ± 9.3

Bearing (º) 242.5 ± 28.5 122.3 ± 32.1 175.3 ± 27.4 173.9 ± 22.3

Number of grid cells occupied (0.25ºx0.25º) 32.6 ± 3.2 26.4 ± 4.7 20.9 ± 4.2 16.3 ± 1.6

Departure date (Julian) 79.0 ± 9.1 80.3 ± 5.5 64.5 ± 6.6 91.2 ± 4.5

Return date (Julian) 215.3 ± 4.8 210.4 ± 5.3 204.7 ± 4.9 209.0 ± 8.5

136

Fig 5.1: Percentage of total time spent in pre-defined grid cells for adult Australasian gannets (Morus serrator) from Point Danger during the winter non- breeding dispersal period (March to September). A) Females 2015 (n = 13), B) Females 2016 (n = 19), C) Males 2015 (n = 15), D) Males 2016 (n = 15).

The breeding colony location is defined by a black star. 137

Fig 5.2: Percentage of total time spent in pre-defined grid cells for adult Australasian gannets (Morus serrator) from Pope’s Eye during the winter non- breeding dispersal period (March to September). A) Females 2015 (n = 13), B) Females 2016 (n = 14), C) Males 2015 (n = 14), D) Males 2016 (n = 20).

The breeding colony location is defined by a black star.

138

The hierarchical clustering analysis, which uses dispersal metrics as input, identified three main winter strategies. Group 1 (51 tracks from 43 individuals, 53.4% males) consisted of birds from both colonies (PE 73%) with an intense use of central Bass Strait as their main wintering area. This group occupied the least number (21.2 ± 1.6) of 0.25º x 0.25º grid cells, had the lowest mean for maximum distance from the colony (336.4 ± 45.6 km), and had intermediate values for total dispersal duration (138.7 ± 5.2 d, Figure 5.3D). Group 2 (41 tracks from 38 individuals,

65.8% males) consisted of birds from both colonies (PE 47%) dispersing both to eastern Bass Strait and to the west to the Great Australian Bight and Spencer Gulf. This group had intermediate values for number (32.9 ± 2.3) of occupied grid cells and mean maximum distance from the colony (815.8

± 99.4 km), but had the lowest values for total duration (121.6 ± 3.6 d, Figure 5.3B). Finally,

Group 3 (28 tracks from 25 individuals, 36.0% males) consisted of birds from both colonies (PE

15%) dispersing both east and west of the colonies but reaching further distances than the other two groups. This group occupied the largest number of grid cells (48.9 ± 3.5), had the greatest mean maximum distance from the colony (1,253.1 ± 131.1 km), and had the greatest total duration values (185.5 ± 5.1 d, Figure 5.3C and Supporting information S5.1).

Intra-individual variation and within-pair behavioural similarity

The intra-individual variation in winter dispersal metrics was analysed using information from 32 individuals tagged two consecutive years. In this analysis, the total number of grid cells occupied was the only metric showing a significant and positive correlation between the individuals’ values of year 2015 and year 2016, with the permutation analysis confirming these values were more similar than expected by change. Bearing and total duration showed moderate levels of similarity within the correlations, but permutations showed individuals values were not more similar or dissimilar than expected by change (Table 5.3). 139 When analysing nest partner similarity in winter dispersal metrics, the number of grid cells occupied and total duration had positive and moderate correlation values between paired females and males, although permutations showed paired individuals were not more similar or dissimilar than expected by change (Table 5.3).

140 Table 5.2: Most parsimonious models after model averaging for five winter dispersal metrics for adult Australasian gannets (Morus serrator). Their corresponding estimated regression parameters are shown.

Dispersal metric Model fixed effects Fixed effect Estimate SE df t-value P-value

Number of grid cells occupied colony + year + sex (Intercept) 10.76 0.43 87 24.86 <0.0001 (0.25 x 0.25º) Colony (PE) -3.15 0.43 87 -7.23 <0.0001

Sex (male) -1.23 0.43 87 -2.82 0.005

Year (2016) -0.79 0.36 29 -2.17 0.03

Bearing (º) colony (Intercept) 250.04 12.16 118 20.54 <0.0001

Colony (PE) -83.76 17.21 118 -4.86 <0.0001

Total duration (d) colony + sex (Intercept) 165.18 5.69 117 29.01 <0.0001

Colony (PE) -26.96 6.58 117 -4.10 <0.0001

Sex (male) -14.79 6.59 117 -2.24 0.03

Departure date (Julian) year (Intercept) 69.74 3.50 118 19.94 <0.0001

Year (2016) 12.43 4.68 118 2.65 0.01

Return date (Julian) colony + sex (Intercept) 236.58 3.44 116 68.78 <0.0001

Colony (PE) -21.78 3.99 116 -5.46 <0.0001

Sex (male) -12.85 3.99 116 -3.22 <0.0001

141

B) D)

A) n = 41 n = 28 n = 51

142 Table 5.3: Winter dispersal metrics (mean ± SE) of Australasian gannets (Morus serrator) paired samples, regression parameters and paired t-test results. Two levels of similarity comparisons, by individuals and by nest pairs, are shown.

Dispersal metric 2015 values 2016 values R2 P-value permutation

Number of grid cells occupied (0.25 x 0.25º) 36.53 ± 3.79 23.87 ± 2.45 0.76 < 0.0001

Bearing (º) 181.50 ± 18.78 210.49 ± 17.90 0.22 0.20

Total duration (d) 151.90 ± 6.98 130.53 ± 6.80 0.24 0.17

Departure date (Julian) 68.81 ± 4.84 83.56 ± 4.86 0.007 0.96

Return date (Julian) 219.71 ± 4.14 213.09 ± 4.74 0.13 0.45

Dispersal metric Females Males R2 P-value permutation

Number of grid cells occupied (0.25 x 0.25º) 38.56 ± 3.28 29.13 ± 2.65 0.28 0.08

Bearing (º) 197.67 ± 18.01 232.08 ± 15.17 0.03 0.89

Total duration (d) 158.40 ± 6.40 140.97 ± 6.76 0.20 0.24

Departure date (Julian) 68.83 ± 4.13 75.24 ± 4.32 0.03 0.84

Return date (Julian) 226.24 ± 3.98 216.64 ± 5.06 0.19 0.25

143 Discussion

In the present study, the winter distribution of Australasian gannets breeding in south- eastern Australia was described through the use of geolocation devices deployed on adult individuals. Information about the movement patterns of this species was necessary in order to understand the potential inter-individual differences that may be contributing to segregation in resource use during this period. Five different metrics were investigated in terms of influential factors and distinct dispersal strategies. Significant differences were found within these winter dispersal metrics associated with factors such as colony, year, and sex. Tagged individuals were classified in three distinctive winter dispersal strategies distinguishable mainly by differences in spatial distribution. In addition, low intra-individual variation of number of grid cells occupied was found, suggesting individuals tend to have similar spatial extent over the years. Finally, within-nest pair similarity was tested in the winter dispersal metrics of individuals sampled within the same year, in order to understand partners’ interactions within this period. Nest partners showed moderate values of similarity in number of grid cells occupied and total duration, although they were not more similar or dissimilar than non-paired individuals.

Winter distribution and factors influencing dispersal behaviour

Previous sightings made during at-sea surveys report that adult Australasian gannets winter off the coasts of Australia, from the west coast to the mid-eastern coast, plus the coast of Tasmania (Cox 1978, Dann et al. 2004). The results from this study support these reports by suggesting those birds in the past could possibly have their dispersion origin at colonies from south-eastern Australia, as it was demonstrated during this study.

Individuals in the present study displayed several distinct major hotspots of spatial- use during their winter dispersal period. For example, high intensity use areas were seen for

144 adults from Point Danger (females in particular) along the south-west coast of Australia. In this region, the presence of a complex network of submarine canyons and other seafloor features, together with the increased winter influence of the Leeuwin current running parallel and in close proximity to the coast, favour the periodic appearance of small and localised meso-scale eddies that contribute to local productivity (Commonwealth of Australia 2015).

These events occur at predictable locations, such as the Perth canyon, the Albany coast, and the Esperance and Eyre peninsulas (Schahinger 1987, Condie & Dunn 2006, Commonwealth of Australia 2015). Additionally in the inner shelf, where the Great Australian Bight is located, coastal currents move east due to wind-driven forces, and by the increased Leeuwin

Current over the shelf break. Biological productivity at this place is driven mainly through pulses of mixed water that irregularly wash through this system from the west (Richardson et al. 2005). The presence of all these marine physical processes coincides with the hotspot locations for some Australasian gannets during the winter period.

In contrast, birds from PE maintained their high intensity use areas during the winter closer to the vicinity of the breeding colony. In general, for both males and females, Bass

Strait was the major over-winter foraging hotspot during the two years of the study. During winter, the marine productivity of this area increases as a result of the diminished influence of the East Australian Current (warm equatorial waters moving southward along the east coast of Australia), and the elevated action of the Subantarctic Surface Water reaching its northern limit, the vicinity of Tasmania, and blocking the entrance of warm waters into central Bass

Strait (Commonwealth of Australia 2007). This phenomenon, in combination with other seafloor features and internal water circulation patterns, creates an “underwater waterfall” in which nutrient-rich waters rise to the surface, increasing the primary productivity of the area

(Croll et al. 1998, Sandery & Kämpf 2005). It is possible the improved marine conditions this

145 area experiences during the winter season favour the permanence of Australasian gannets within this region.

Three distinctive winter dispersal strategies were identified among adult Australasian gannets, distinguishable mainly by differences in spatial distribution. The three identified groups consisted of individuals belonging to both study colonies, but Groups 1 and 3 were dominated by individuals from PE and PD colonies, respectively. Additionally, Group 3 was dominated by females and the group had the highest values for distance travelled and total duration. Previous research performed on other bird species analysing partial migration strategies within populations have recognised the male tendency to remain closer to the breeding colony in concordance with the ‘arrival-time’ hypothesis (Bai et al. 2012, Fudickar et al. 2013, Pérez et al. 2014). This hypothesis states that reproductive pressures may act differently among individuals of the same population (usually contrasting males and females), and the sex more responsible for nest-site defence may find advantages to remaining closer to the breeding grounds and arriving earlier (Ketterson & Nolan 1983,

Chapman et al. 2011a). For Australasian gannets, however, the reasons why an individual would choose one strategy over the other remain unclear. Future research may contrast these winter dispersal strategy results with information about age or physical characteristics of individuals in an effort to identify potential links between them.

The oceanographic features around the breeding grounds described in previous sections agree with the results obtained from the analysis of the factors influencing the winter dispersal behaviour in Australasian gannets. Explicitly, the factor colony, which is used within this research as a proxy reflecting the differences in resource availability and/or habitat accessibility around the study sites, was found to influence the winter dispersal metrics of total number of grid cells occupied, bearing, total duration, and return dates. Indeed, birds from PD travelled further to the west and stayed for longer periods within the rich areas of

146 west Australia, while birds from PE travelled south and spent less time within Bass Strait.

The inter-colony differences in return dates could potentially be associated with different levels of intra-colony competition for nesting space, as has been suggested for other seabird species (Scott Forbes et al. 2000, Ainley et al. 2004). At PE (180 pairs), which is placed on top of a rock annulus formation and two artificial wooden platforms, nesting space is considered limited (Bunce et al. 2002, Pyk et al. 2007), with strong pressure to occupy any available nesting space (Rodríguez-Malagón pers. obs.). In contrast, Point Danger (660 pairs) is based on an elongated slope adjacent to the coast and surrounded by coastal grassland and scrub, where there are no limits to new nest sites and the pressure for occupying former nest sites is much lower (Norman et al. 1998, Bunce et al. 2002).

Year of sampling, used as a proxy of environmental variation, shown to have an influence on the winter dispersal behaviour of Australasian gannets with individuals departing earlier and travelling more in 2015 than in 2016. The earlier departures in 2015 could be understood by the differences in the reproductive performance of each colony within the studied years. During the breeding season prior to the 2015 winter dispersal, the average reproductive output estimated (proportion of chicks fledged) was lower at both colonies (PE:

25%, PD: 48%) than during the following breeding season prior to the winter dispersal of

2016 (PE: 50%, PD: 79%; Rodríguez-Malagón unpubl. data). This suggests both sites experienced similar environmental variation influencing both reproductive and winter dispersal behaviour to similar degree. In addition, during 2015 a strong El Niño/Southern

Oscillation (ENSO) event occurred. In Australia ENSO tends to begin in autumn and has its greatest impacts during the winter months (www.bom.gov.au). It is known that climatic events like this are able to alter the distribution of oceanic resources (Barber & Chavez 1983,

Chavez et al. 2011) which may help to explain the differences in the reproductive performance of each colony, the differences in the departure dates and the longer travelling

147 distances seen in the Australasian gannets that particular year. For other species of seabirds, longer travelling distances have been reported associated with ENSO events. For example,

McKnight et al. (2011) reported black-legged kittiwakes (Rissa tridactyla) traveling far offshore during a ENSO event which occurred in 2006 in comparison to their other years of study.

During winter, after the Australasian gannets are released from their parental responsibilities, becomes clear that sex-related differences in dispersal behaviour remain present, as they have been seen during summer seasons (Chapter 2, Machovsky-Capuska et al. 2014, Angel et al. 2016). In general, females from both colonies travelled further, had a longer winter dispersal period and arrived later to the breeding colonies. Similar sexual- related differences have been reported for the winter dispersal behaviour of northern gannets

(M. bassanus; Kubetzki et al. 2009), and other seabird species (Phillips et al. 2005b, Fifield et al. 2014, Müller et al. 2014, Orben et al. 2015). The early return of males to the breeding colony has been related before to the need of securing and defending the nesting space

(Kokko et al. 2006, Orben et al. 2015), but also to the physiological need of females to acquire sufficient reserves before the breeding season (Fayet et al. 2017). In Australasian gannets, males have been seen to spend more time guarding the nest than females during the breeding season (Ismar et al. 2010a). Also, females seem to experience greater foraging effort

(as indicated by deeper maximum dives) during early times of the incubation period (Ismar et al. 2017) or a higher dive rate during the breeding season (Chapter 2) than males, evidence that may help to support the sexual differences seen during their winter dispersal.

Additionally, females have shown to be more efficient in some aspects of their foraging behaviour (as indicated by lower mean Vectorial Dynamic Body Acceleration [VeDBA] values) than males (Chapter 2), a feature that may help them reach further distances with less effort.

148 Intra-individual and within-pairs behavioural similarity

For individuals sampled over the two study years, high levels of intra-individual variation were found for one of the winter dispersal metrics analysed. The total number of grid cells occupied, which was used as a proxy of the individual foraging-range size (Soanes et al. 2014), had a positive and significant correlation value implying intra-individual similarity between 2015 and 2016 values. Interestingly, this winter dispersal metric was previously found in this study to be influenced by the environmental variation related to the year of sampling, in particular due to the presence of an ENSO event in 2015. However, these results suggests that individuals can have strong preferences in relation to the habitat extent that they use over this period.

Bearing to most distal location and total duration showed moderate correlation values between individuals sampled in 2015 and 2016. For both metrics however, the permutation test showed non-significant results, meaning population-level consistency only. For the study sites, similar bearings may be the result of the geographical limitations that Australasian gannets face to find foraging locations given their restrictions to forage within the continental shelf (Nelson 1978). PD for example, is located along a narrow (~40 km) continental shelf which limit the birds to travel only in north-west and south-east directions (Butler et al. 2002,

Angel et al. 2016). Similarly, PE is located between the narrow entrance of the Port Phillip

Bay and Bass Strait, potentially limiting the birds to forage only among those habitats. These findings also suggest a degree of spatial and temporal resource predictability that birds are able to remember based on directional references (Harris et al. 2014).

Many long-lived species are able to maintain pair bonds in the long-term as a strategy to increase their reproductive success (Black 2001, Naves et al. 2007). In this context, it has been suggested for birds that similarity in behavioural traits within nesting pairs can lead to partner retention in the long-term (Spoon et al. 2006, Schuett et al. 2011). In the present

149 study, Australasian gannets did not showed within-pair similarity on any of the winter dispersal metrics analysed. Indeed, as mentioned in previous results, sex of individuals influences the value of some of the analysed metrics, suggesting males and females, independently of their pair status, may have different winter behavioural drivers (Kokko et al.

2006).

In conclusion, the present study describes the winter distribution of Australasian gannets breeding in south-eastern Australia, filling an important gap in the life history information for this species. The general winter distribution of this species can be explained by the physical and biological processes of the marine ecosystem that determine resource availability during this period. Three different winter dispersal strategies were identified, indicating segregation in habitat use by individuals that was not related to sex. Five different winter dispersal metrics were shown to be influenced by factors such as colony, year, and sex. Low intra-individual variation in one dispersal metric was documented and no within- pair similarity was found.

150 Supporting information

Table S5.1: AICc based model selection (∆ < 4) for factors influencing the winter dispersal metrics in adult Australasian gannets (Morus serrator). Model variables: cln: colony; b.scc: breeding success; dep: departure date; yr: year; sex.

Dispersal Model fixed effects df AICc ∆AIC AIC Weight

metrics

Number of grid cells cln + yr + sex 6 983.7 0 0.54

occupied (0.25º x 0.25º) cln + yr + sex + dep 7 985.7 1.92 0.21

cln + yr + sex + b.scc 7 986.0 2.25 0.17

Bearing (º) col 3 1435.8 0 0.26

b.scc + col 4 1437.1 1.25 0.139

col + yer 4 1437.6 1.84 0.103

col + sex 4 1437.9 2.05 0.093

col + dep 4 1437.9 2.1 0.091

b.scc + col + sex 5 1439.1 3.33 0.049

Total duration (d) col + sex + yr 5 1205.6 0 0.306

col + sex 4 1205.6 0.03 0.302

b.scc + col + sex 5 1207.7 2.1 0.107

b.scc + col + sex + yr 6 1207.8 2.18 0.103

col 3 1208.5 2.94 0.07

col + yr 4 1208.7 3.12 0.064

151 Departure date (Julian) col + yr 1121.6 0 0.236

yr 3 1121.7 0.07 0.228

sex + yr 4 1123.1 1.53 0.11

col + sex + yr 5 1123.2 1.64 0.104

b.scc + col + yr 5 1123.6 2.02 0.086

b.scc + yr 4 1123.7 2.11 0.082

b.scc + sex + yr 5 1125.2 3.6 0.039

b.scc + col + sex + yr 6 1125.3 3.69 0.037

Return date (Julian) col + sex 4 1075.4 0 0.525

b.scc + col + sex 5 1077.2 1.83 0.21

col + sex + yr 5 1077.6 2.18 0.176

152 Table S5.2: Average model coefficients and relative importance of variables included in top

model set (∆AICc ≤ 4) explaining winter dispersal metrics in Australasian gannets (Morus

serrator).

Dispersal metrics Parameter Estimate SE 5 % 95 % Relative

CI CI importance

Number of grid cells (Intercept) 52.15 3.43 46.45 57.84 --- occupied (0.25º x Colony (PE) -16.11 2.95 -21.01 -11.20 1.00

0.25º) Sex (male) -10.13 2.95 -15.03 -5.23 1.00

Year (2016) -10.55 2.22 -14.32 -6.77 1.00

Departure date -0.03 0.05 -0.11 0.05 0.22

Breeding success (1) -0.15 3.15 -5.50 5.20 0.19

Bearing (º) (Intercept) 243.22 19.86 209.14 277.28 ---

Colony (PE) -84.07 17.29 -112.77 -55.38 1.00

Breeding success (1) 18.67 20.66 -15.58 52.92 0.33

Year (2016) 8.17 17.84 -21.29 37.94 0.22

Sex (male) 5.32 17.37 -23.54 34.10 0.21

Departure date 0.06 0.34 -0.50 0.61 0.16

Total duration (d) (Intercept) 167.01 7.64 154.62 179.97 ---

Colony (PE) -27.17 6.60 -38.03 -16.19 1.00

Sex (male) -14.91 6.59 -25.83 -3.99 0.86

Year (2016) -9.71 6.70 -20.88 1.34 0.5

Breeding success (1) -0.54 8.28 -14.27 13.19 0.22

153 Departure date (Julian) (Intercept) 67.71 4.89 59.59 76.21 ---

Year (2016) 12.68 4.77 4.76 20.58 1.00

Colony (PE) 6.78 4.64 -0.91 14.47 0.5

Sex (male) 3.60 4.67 -4.14 11.35 0.31

Breeding success (1) -2.09 5.87 -11.81 7.64 0.27

Return date (Julian) (Intercept) 237.06 4.07 230.30 243.80 ---

Colony (PE) -21.76 3.99 -28.37 -15.13 1.00

Sex (male) -12.86 4.00 -19.48 -6.22 1.00

Breeding success (1) -2.75 4.71 -10.55 5.05 0.23

Year (2016) -0.03 4.02 -6.70 6.63 0.19

154

Fig. S5.1: Differences in the winter dispersal metrics values between the clusters identified by hierarchical method for Australasian gannets (Morus serrator).

155

CHAPTER 6 General discussion

156 Introduction

Behavioural studies promote understanding of the interactions between animal species and their ecological adaptations to the environment (Davies et al. 2012). This area of knowledge has grown rapidly due to the recent development of miniaturised loggers that allow remote recording of animal behaviour (Brown et al. 2013, Liechti et al. 2018). The combination of these modern data-collection devices with data mining methods for large databases has facilitated behaviour and/or movement pattern analyses under a wide variety of conditions (such as in flight or at sea), which is especially important for study species with great capacity for movement (Hooker et al. 2007, Burger & Shaffer 2008, López-López

2016).

Individual differences in foraging behaviour have been suggested to have important implications for the ecology of species (Soanes et al. 2013b, Robertson et al. 2014,

Camprasse et al. 2017a). These differences are fundamentally driven by biological factors such as age and sex, and by individual decisions governed by current conditions (including internal state); if individual differences are consistent over time (within or between generations) and/or are maintained over different contexts, they can also have major implications from an evolutionary perspective (Bolnick & Doebeli 2003, Dall et al. 2004,

Arnould et al. 2011). In seabirds, individual differences in foraging behaviour have been reported for several species (Ceia & Ramos 2015, Phillips et al. 2017). Although the benefits to specialised individuals associated with this phenomenon are not consistent (for example, see Golet et al. (2000), Votier et al. (2004) for positive results, or Woo et al. (2008),

Dehnhard et al. (2016) for negative results). Nonetheless, the study of individual foraging differences allows the understanding of the level of resource partitioning among individuals within a population, which is crucial information for understanding the susceptibility of the species to anthropogenic threats and environmental change (Carneiro et al. 2017).

157 The Australasian gannet, the study species for my dissertation research, has experienced population growth and the creation of new breeding colonies in south-eastern

Australia during recent decades (Norman & Menkhorst 1995, Norman et al. 1998, Bunce et al. 2002). Currently, its conservation status in within this range is considered stable or increasing (Department of the Environment 2019), although no recent census has been published. Certainty, the population’s interactions with local fisheries are a resource management conflict and could potentially represent a future threat to the survival of this species (Bunce 2001). Specifically, the Australian government has a ‘small pelagic fishery’ initiative that allows the harvest of four pelagic forage fish species throughout the year:

Australian sardine, blue mackerel, jack mackerel, and redbait. The extent of this fishery ranges from south-west Australia and Tasmania to the mid-eastern coast of Australia (Ward et al. 2015), similar to the general distribution of Australasian gannets in this country (Reid et al. 2002, BirdLife International 2016). Although the status of this fishery is currently considered sustainable (www.afma.gov.au), fishing quotas need to be monitored to ensure these resources are not over-fished. In particular these four forage fish species represent important prey types for the Australasian gannet (Chapter 3, Norman & Menkhorst 1995,

Bunce & Norman 2000, Pyt et al. 2007) and other marine predators of the region (Kirkwood et al. 2008, Chiaradia et al. 2012). In addition, viral diseases in the past have caused massive losses in stocks of Australian sardines throughout southern Australia, causing a reduction in recruitment at nearby Australasian gannet breeding colonies (Bunce & Norman 2000, Pyk et al. 2013, Angel et al. 2015a). The risk of similar events, together with the environmental changes predicted for this region in coming decades as a result of climate change

(Poloczanska et al. 2013), make this species vulnerable. For these reasons, it is important to study the foraging behaviour and movement patterns of Australasian gannets in an effort to

158 understand individual variability in resource use in the context of climate change and human population growth (Symonds & Moussalli 2011, Merino et al. 2012).

My use of complimentary bio-logging techniques, stable isotope analyses, and regurgitation analyses allowed the investigation of factors influencing foraging behaviour, diet, and winter dispersal patterns in Australasian gannets; as well as quantification of the levels of inter- and intra-individual variability within these attributes, and the understanding on how these attributes vary over timescales and habitats. In Chapter 2, behavioural data loggers deployed over the summer/breeding period were used to obtain information about foraging behaviour that allowed the estimation of spatial use metrics for each foraging trip conducted by individuals. Five metrics of foraging trips (maximum distance from the colony, bearing from the colony to the most distal point, tortuosity index, total number of dives, and mean Vectorial Dynamic Body Acceleration) were calculated and used in further analyses as response variables. Similarly in Chapter 5, several winter dispersal metrics estimated for each individual (total number of grid cells occupied, total duration, bearing from the colony to the most distal point, departure date, and return date) were studied. Linear mixed effects models were used in both studies to identify the influential factors for each metric, and to quantify, for the foraging metrics only, individual consistency in foraging behaviour (Nakagawa &

Schielzeth 2010, Dingemanse & Dochtermann 2013). In Chapter 2, variance component analyses complemented these analyses and allowed the estimation of the proportion of the total variance associated with the individual component in each of the summer foraging metrics studied. Additionally in Chapter 2, coefficients of variation estimated at the individual level for each foraging metric allowed the investigation of the links between individual consistency and intrinsic (morphometries, body mass, and sex) and extrinsic drivers (sampling location, breeding stage, and year of sampling). Furthermore in Chapter 5, because individual sample sizes did not allow estimation of individual consistency, intra-

159 individual variation analysis was carried out using winter dispersal metrics for those individuals with two consecutive years of data.

In Chapter 4, the combination of traditional regurgitation analysis with modern stable isotope mixing models allowed the reconstructions of individual diets by year and by colony of sampling. Later, the degree of inter-individual diet variation and the average degree of individual specialisation of each studied population were estimated using specialisation indices (Bolnick et al. 2002, Zaccarelli et al. 2013). Finally, the intra-individual variation in isotopic values was investigated by correlating individual blood plasma samples over medium-term timescales (breeding stage-to-breeding stage) and long-term timescales (year- to-year; Nakagawa & Schielzeth 2010, Carneiro et al. 2017).

Because the similarity in behavioural features within breeding pairs has been shown to have major, long-term effects on reproductive success and fitness of species (Spoon et al.

2006, Schuett et al. 2011), I investigated within-pair similarity for diet (Chapter 4) and winter dispersal metrics (Chapter 5) as part of my dissertation research. In Chapter 4, within-pair similarity in δ13C and δ15N isotopic values and its potential reproductive advantage were tested using linear (and generalised linear) mixed effect models. In Chapter 5, within-pair similarity in winter dispersal metrics was tested through linear models and paired t-tests.

A special chapter is included in this dissertation (Chapter 3). It was developed to specifically investigate the temporal and spatial variation in δ13C and δ15N isotopic values for common prey species found in the diet of Australasian gannets. Obtaining the answer to this specific question was necessary in order to further investigate variation in the birds’ diet using stable isotope analysis and, in particular, mixing models (Phillips et al. 2014). The importance of this particular study became more apparent when the lack of knowledge addressing this question in my study area became obvious. The findings of Chapter 3 were crucial for Chapter 4, and provided a reference frame for other research seeking to understand

160 the diet of marine predators through the analysis of stable isotopes at a regional scale.

Because the main focus of this dissertation is the analysis of animal behaviour, however, few references to Chapter 3 are included in this conclusions chapter. Only when reference is made to the influence of oceanographic differences on the marine environment and prey components near each study site are the results from Chapter 3 alluded to.

Drivers of behaviour

Intrinsic drivers

As populations consist of phenotypically diverse individuals, variability in resource use can be expected. Differences in morphology, age, sex, breeding status, physiology, learned experiences, and skills can lead to different responses to external stimuli, differences in basic needs, and/or differences in competitive abilities, causing niche differentiation within animal populations (Bolnick et al. 2002, Svanbäck & Bolnick 2007, Dall et al. 2012). Many studies recognise these intrinsic factors as primary causes of existing inter-individual variation because they provide the internal mechanisms from which certain differences originate (Becker & Bradley 2007, Phillips et al. 2017). In the present study, intrinsic factors such as body size indices, breeding stage, and sex were investigated to understand their influence on different metrics of foraging behaviour during the breeding season (Chapter 2), in δ13C and δ15N isotopic values of blood plasma sampled from individual adults (Chapter 4), and in metrics of winter dispersal behaviour (Chapter 5) obtained from adult Australasian gannet individuals.

Morphometry and body weight have been shown to influence individual foraging decisions in certain species. For example, Camprasse et al. (2017a) reported heavier individuals of gentoo penguins foraged more in the benthos and performed more distant and

161 longer foraging trips than lighter individuals. Similarly, (Price 1987) reported that individual variation in seed consumption was related to beak length and seed size in Darwin's medium ground finches (Geospiza fortis). In Australasian gannets, however, a body condition index

(Angel et al. 2015b) and two relative indices of body weight and wing length were tested as potential factors influencing foraging behaviour, but no relationships were found (Chapter 2).

This information suggests that factors other than morphometry and body weight have a greater influence on foraging behaviour in this species. In addition, previous studies in gannets have found no significant differences in body condition of individuals in relation to sex, breeding stage, year (Angel et al. 2015b), or colony of sampling (Moseley et al. 2012), supporting the notion that morphometry may not play a major role in the performance of foraging of individuals.

Australasian gannets have a long breeding season (October to March) in which the different stages of young development are well defined (i.e. incubation, early chick-rearing, and late chick-rearing; Wingham 1982). Because each of these stages places different demands or constraints on nesting pairs, investigating foraging behaviour separately during each stage of the nesting cycle may reveal aspects of the breeding biology of parents that influence foraging behaviour. In my research, stages of the breeding cycle were incorporated as explanatory variables in the analysis of potential influential factors on foraging behaviour

(Chapter 2) and δ13C and δ15N isotopic values in blood plasma from adults (Chapter 4).

Among these analyses, the incubation stage had the largest influence on the foraging behaviour of individuals, as the majority of the foraging metrics analysed had higher values during this stage (Chapter 2). The influence of breeding stage on foraging behaviour has been previously described for my study species (Bunce 2001, Angel et al. 2015a), and appears to be common among other species of seabirds as well (Phillips et al. 2017).

162 Regarding breeding stage as a potential factor influencing δ13C and δ15N isotopic values, the incubation stage was associated with higher δ13C values, while the late chick- rearing stage was associated with higher δ15N values (Chapter 4). For behavioural differences among foraging adults related to breeding stage, two potential explanations have been proposed. First, these differences likely reflect a shift from self-feeding during the incubation stage to chick-provisioning during the chick-rearing stage (Phillips et al. 2017). As most seabird species share parental responsibilities, during the incubation stage the foraging behaviour of the partner at sea is restricted only by the fasting ability of the incubating partner (Schreiber & Burger 2001). For some seabird species, this allows the off-duty partner to spend more time at sea (Weimerskirch 1995, Hedd et al. 2001, Shaffer et al. 2003).

Second, other studies have related differences in foraging behaviour between incubation and chick-rearing stages to be in response to temporal variation in prey availability due to environmental changes around colonies throughout the breeding period (Ito et al. 2009,

Jakubas et al. 2014, Shoji et al. 2016). The results of my research provide more support for the second explanation, as environmental changes are more likely to cause a shift in the diet as seen from the difference in stable isotope values between the incubation and chick-rearing stages. Further investigation is needed, however, to test between these competing hypotheses.

The sex of individuals was shown to have a strong influence on the individual foraging behaviour of Australasian gannets. For some of the metrics of foraging behaviour measured during the summer/breeding season (Chapter 2), and for some of the metrics of dispersal behaviour measured during winter (Chapter 5), females exhibited higher values than males. During the summer, females foraged at greater distances from the breeding colony and had higher dive rates compared to males. In the winter, females had greater dispersal distances, longer winter dispersal periods, and returned later to the breeding grounds than males. As a complement to these findings, sexual segregation in stable isotope values was

163 also evident from my research (Chapter 4); δ13C values were higher in males than in females, suggesting males foraged closer to the coast than females (Hobson et al. 1994, Cherel &

Hobson 2007). The Australasian gannet displays colony-specific reverse sexual-dimorphism, with females being heavier and having larger wing ulna than males (a dimorphism pattern found at PD and PE breeding colonies; Angel et al. 2015b). In species with sexual size dimorphism, trophic or spatial segregation has been proven to help reduce intra-specific competition, particularly during periods of intense resource competition (Cleasby et al. 2015,

Phillips et al. 2017). In other Sulid species (gannets and boobies), reverse sexual-dimorphism is common (Lewis et al. 2005, Weimerskirch et al. 2006, Castillo-Guerrero & Mellink 2011), and has been suggested to be the mainly cause of observed sexual differences in the movement capabilities of these species, resulting in the larger bodied sex traveling further from the colony (Lewis et al. 2005). In addition, the shorter foraging trips during breeding and the shorter dispersal distances during winter for males supports the hypothesis that competition for breeding sites among males is intense and selects for higher male attendance rates at the colony (Ismar et al. 2010b).

Age has been shown to be an intrinsic factor that influences the individual foraging behaviour of animals, with major differences described between juveniles and adults

(Baechler et al. 2002, Forero et al. 2005, McGraw et al. 2011). Among adults, age also has been shown to influence foraging behaviour, presumably because greater life experience confers enhanced abilities and/or knowledge of the presence or distribution of profitable resources (Phillips et al. 2017). For gannets nesting at Pope’s Eye, the Australian Bird and

Bat Banding Scheme (ABBBS) is able to provide information about the age of certain individuals. For my study, however, only a small percentage of individuals (26%) from this colony were known-age. The remaining birds were banded as adults during the study period, and because their age was unknown I did not include this factor in the statistical analyses.

164 This highlights the importance of banding programs that are capable of assigning a known age to marked individuals. Banding chicks and/or fledglings should be a priority over banding adults at breeding colonies where scientific hypotheses on factors affecting breeding behaviour are being tested. Future studies should consider increasing banding efforts of chicks/fledglings as a way to contribute to databases with known-age individuals.

Extrinsic drivers

This study was developed at two Australasian gannet breeding colonies specifically chosen due to the different oceanographic conditions prevailing in their vicinities throughout the year (Chapters 3 and 5). Consequently, the factor of ‘colony’ was included in the majority of the statistical analyses performed, because it is known that seabird populations are strongly regulated by food availability within commuting distance of the breeding colony (Schreiber

& Burger 2001). As a consequence of the habitat structure and prey abundance around the two study sites (Chapter 3), certain ecological aspects of Australasian gannets differed between colonies. For example, individuals from Point Danger travelled greater distances during both, their regular foraging activities in the breeding season (Chapter 2) and their dispersal behaviour in winter (Chapter 5). Furthermore, the diet composition of individuals also differed by colony. At Point Danger schooling pelagic prey were identified as the primary food source, while at Pope’s Eye an assemblage of benthic/inshore prey types contributed substantially to the diet (Chapter 3 and 4). Several studies have found inter- colony differences in seabird foraging behaviour, diet composition, and habitat use, providing evidence of a species’ adaptability to local environmental conditions (Tremblay & Cherel

2003, Ainley et al. 2004, Zavalaga et al. 2010).

Environmental variability has been shown to play an important role in the distribution and abundance of marine resources (Paiva et al. 2013) and, consequently, is a major factor

165 influencing seabird foraging behaviour (Grémillet & Boulinier 2009, Croxall et al. 2012,

Machovsky-Capuska et al. 2018). In my study, year of sampling was used as a proxy for environmental variation between years. In Australasian gannets inter-annual differences were seen in the metrics of foraging behaviour studied during the summer/breeding season

(Chapter 2), diet composition (Chapter 4), and winter dispersal behaviour (Chapter 5). A key factor influencing the inter-annual variability in these behavioural attributes was a strong El

Niño/Southern Oscillation (ENSO) event in 2015. In Australia, ENSO events tend to begin in autumn and have their greatest impact during the winter months (www.bom.gov.au). Major fluctuations in ocean conditions, such as ENSO events, have been associated with changes in the survival and reproductive success of seabirds (Frederiksen et al. 2007), changes in their foraging behaviour and diet (Ancona et al. 2012, Castillo-Guerrero et al. 2016, Machovsky-

Capuska et al. 2018), and changes in their winter dispersal strategies (McKnight et al. 2011).

These associations have been attributed to shifts in prey distribution and abundance due to fine-scale variation in sea surface temperature during ENSO events (Barber & Chavez 1983).

Although environmental variation is likely even in non-ENSO years, these major climatic events have shown to have a big impact on marine food chains (Barber & Chavez 1983,

Chavez et al. 2011).

Inter- and intra-individual variation in behaviour

Inter-individual differences in habitat use are widespread among animal populations

(Bolnick et al. 2003, Araújo et al. 2011). The implications that these individual differences within a population depend on the specialised response exhibited by individuals and the persistence of these differences over time (Bolnick et al. 2003, Bell et al. 2009). In general, seabirds possess certain attributes at the population level (such as being long-lived animals,

166 central-place foragers and experience high levels of resource competition) by which the presence of individual specialisations can be favoured (Svanbäck & Bolnick 2007, Ratcliffe et al. 2013). Indeed, a wide number of studies have reported inter-individual variation or specialisation in the foraging behaviour and diet of seabirds during the breeding season and also behavioural consistency in the migratory paths and over-winter strategies of individuals within this group (Ceia & Ramos 2015, Phillips et al. 2017).

In my study, the inter-individual variation in foraging behaviour of Australasian gannets during the summer/breeding season was analysed over the short-term (consecutive foraging trips obtained from one GPS deployment), medium-term (different stages of the breeding cycle within one breeding season), and long-term timescales (same breeding stage in two different years; Chapter 2). Moderate levels of behavioural consistency within individuals were observed for this species, showing a decreasing trend with increasing timescales. In addition, intra-individual comparisons made with paired isotope values (δ13C and δ15N from blood plasma) over medium-term (breeding stage-to-breeding stage) and long- term timescales (year-to-year) showed individual consistency values were lower over the long-term scales, suggesting specialisations are stronger over the shorter periods. Similar time decreasing trends have been shown regarding individual specialisation estimations in diet (e.g. Woo et al. 2008), foraging behaviour (e.g. Camprasse et al. 2017, Harris et al.

2014), and habitat use (e.g. Quillfeldt et al. 2008) of other seabird species, a trend that is shared with other marine animals as well (e.g. Novak & Tinker 2015).

Habitats differ in composition and structure of communities, and thus, in their diversity of available prey (Newsome et al. 2015). In addition, the level of environmental stability over time can vary among habitat types and, consequently, their resource reliability, which gives predators the option to choose between them to forage (Patrick & Weimerskirch

2014). In my study, two distinct habitat types (i.e. inshore and pelagic) were analysed as

167 potential sources of inter- and intra-individual variation, as adjacent to the Pope’s Eye colony these two habitats coincide. My results suggest higher levels of behavioural consistency in foraging behaviour among individuals feeding primarily in the inshore habitat (Chapter 2). In addition, the estimated levels of inter-individual diet variation and the average degree of individual diet specialisation were lower at Pope’s Eye suggesting the presence of more specialised individuals compared to the second study site, Point Danger, where Australasian gannets feed within a relatively uniform habitat. My findings provide support to the notion that specialisations are more likely to occur in populations where individuals are able to feed in more diverse environments as they provide ecological opportunity for diversification of foraging strategies (Araújo et al. 2011). Furthermore, specialisation are suggested to be promoted within stable environments as the presence of predictable resources help to decrease search and handling costs for individuals (Patrick et al. 2015) and facilitate the repetition of successful foraging attempts (Wolf & Weissing 2012). In general, inshore/benthic environments are considered more stable than offshore/pelagic habitats

(Griffiths et al. 2017), where the presence of geomorphic features such as the seafloor, reefs, sandbanks, or bedrocks and their associated prey species promote learning and memorisation of optimal feeding locations on marine predators (Cook et al. 2006, Woo et al. 2008).

In particular, the Port Phillip Bay area, adjacent to Pope’s Eye, provide important inshore/benthic habitat (i.e. reefs, seagrasses, and soft sediments; Parry et al. 1995) for fish reproduction (Parry et al. 1995, Jenkins et al. 1997a) and nursery space for some pelagic species (Neira et al. 1999, Dimmlich et al. 2004). Contrastingly, Bass Strait, adjacent to both

Pope’s Eye and Point Danger colonies, is an offshore/pelagic habitat with constant input from three major water masses and of complex patterns of water circulation (Middleton & Bye

2007) where the local distribution of fish prey species is influenced by sea surface temperature regimes (Hoskins et al. 2008, Kirkwood et al. 2008). These contrasting

168 oceanographic differences among study sites demonstrated, throughout my research, to have great influence on the environmental adaptations of Australasian gannets, including its behavioural consistency. This demonstrate the species ability to accommodate to different habitats and advice the presence of patterns for the development of individual specialisations in animal populations.

Limitations and recommendations for future work

The greatest limitation of this research was my inability to quantify the costs and benefits for those individuals displaying behavioural consistency. In other seabird species, benefits of higher reproductive output or higher fitness have been suggested for specialised individuals (Phillips et al. 2017). For example, Votier et al. (2004) found that individual great skuas (Stercorarius skua) that specialised in eating other seabirds had smaller home ranges during the breeding season, and were able to produce larger clutch sizes earlier in the nesting season compared with unspecialised individuals. Similarly, Masello et al. (2013) studying dolphin gulls (Leucophaeus scoresbii) found that females specialised in eating from mussel beds were heavier compared to unspecialised females, and in general, both males and females that used this specialised feeding strategy maximised net energy consumption over time by learning the optimal feeding locations and behaviours. However, reproductive metrics or parameters (e.g. egg laying date, length of incubation period, chick growth rate) and the reproductive consequences of individual specialisation (e.g. hatching success, fledging success) were not collected or available in my study. Future studies should include measurements of these parameters in order to answer this question.

A second limitation of my study was the lack of activity data from saltwater immersion sensors on the geolocation loggers deployed during winter dispersal. These

169 sensors were supposed to register the time elapsed when each device was submerged in sea water (wet stage) or airborne/on land (dry stage). Immersion data can provide profound insights into the bird’s activity over long timescales and allows the investigation of the requirement for roosting on land during winter dispersal (Mattern et al. 2015). Traditionally, the breeding habitat of a species, due to its importance for population maintenance, is the habitat considered for protection in order to conserve the species (Croxall et al. 2012). This is particularly true for seabirds, most of which are colonial breeders and highly susceptibility to predators and human disturbances at the colony site (Schreiber & Burger 2001). However, for species that disperse widely during winter, winter hotspot locations are equally important for conservation of the population (Yamamoto et al. 2011). In my research project, activity data from immersion sensors could not be collected due to methodological constraints and small sample sizes. Unfortunately, many deployed data loggers failed to collect as much immersion data as they collected data on light levels. Future research should incorporate reliable immersion sensors on data loggers in order to elucidate whether the winter habitat requirements of Australasian gannets include sites where individuals can roost on land.

Lastly, it is important to remember the different assumptions that are made with stable isotope analyses and, in particular, with mixing models when used to describe animal diet compositions. First, the isotopic composition of an animal's tissues equals the weighted average of the isotopic composition of its food (Gannes et al. 1997). This assumption is rarely true as animals may assimilate food items with differing efficiencies, allocate nutrients differentially to specific tissues, and animal tissues fractionate isotopes within (DeNiro &

Epstein 1978, Deniro & Epstein 1981, Gannes et al. 1997). Additionally for δ15 N, there are other internal processes involved that could affect the level of fractionation within an animal’s body. Diet type, excretion systems, growth, and the nutritional status of an animal have been proven to be sources of variation in nitrogen enrichment (Adams & Sterner 2000,

170 Vanderklift & Ponsard 2003, Sears et al. 2009). In my study, the stable isotope values (δ13C and δ15 N) of blood plasma from adult Australasian gannets were analysed, and although the sampling method was highly standardised among individuals, no attempt was made to match any of these processes with the individuals' isotopic composition. Regarding the use of

Bayesian mixing models, these methods have been identified to have great statistical power as allow users to incorporate variation in discrimination factors and prior information (Parnell et al. 2013, Stock et al. 2018). However, it has been demonstrated that Bayesian mixing models can be highly sensitive to the discrimination factors used in animal diet reconstruction

(Bond & Diamond 2011). This is of importance as species-specific discrimination factors are not available in many cases, which force studies to use discrimination factor based on similar species, either ecologically or taxonomically (Zanden & Rasmussen 2001, Martínez del Rio et al. 2009, Bond & Diamond 2011).In my study, the discrimination factors used are based on studies made with nine different species of aquatic birds (Lavoie et al. 2012), as no species- specific discrimination factors exist for Australasian gannets. When present, it could be valid to redo the analyses presented here to validate the suggestions made.

In essence, seabirds are considered the most endangered group among all avian species (Croxall et al. 2012). Different anthropogenic threats have been reducing seabird populations around the world, and climatic change increases these threats (Cherel et al.

2005). Information about resource partitioning and behavioural consistency is crucial if management plans are to protect the basic resources needed for species survival. In a larger context, the accumulation of research about individual foraging specialisations will allow better inferences about the benefits of this phenomenon for adaptation and resilience, with implications for evolutionary processes for seabird species in rapidly changing environments

(Phillips et al. 2017).

171

LITERATURE CITED

Adams TS, Sterner RW (2000) The effect of dietary nitrogen content on trophic level 15N

enrichment. Limnol Oceanogr 45:601-607

Agüera A, Brophy D (2011) Use of saggital otolith shape analysis to discriminate Northeast

Atlantic and Western Mediterranean stocks of Atlantic saury, Scomberesox saurus

saurus (Walbaum). Fish Res 110:465-471

Ainley DG, Ribic CA, Ballard G, Heath S and others (2004) Geographic structure of Adélie

penguin populations: overlap in colony-specific foraging areas. Ecol Monogr 74:159-

178

Akesson S, Weimerskirch H (2014) Evidence for sex-segregated ocean distributions of first-

winter wandering albatrosses at Crozet Islands. PLoS ONE 9:10

172 Åkesson S, Hedenström A (2007) How migrants get there: migratory performance and

orientation. Bioscience 57:123-133

Albo-Puigserver M, Navarro J, Coll M, Layman CA, Palomera I (2016) Trophic structure of

pelagic species in the northwestern Mediterranean Sea. J Sea Res 117:27-35

Alerstam T, Hedenström A, Åkesson S (2003) Long-distance migration: evolution and

determinants. Oikos 103:247-260

Amelineau F, Peron C, Lescroel A, Authier M, Provost P, Gremillet D (2014) Windscape and

tortuosity shape the flight costs of northern gannets. J Exp Biol 217:876-885

Ancona S, Calixto-Albarrán I, Drummond H (2012) Effect of El Niño on the diet of a

specialist seabird, Sula nebouxii, in the warm eastern tropical Pacific. Mar Ecol Prog

Ser 462:261-271

Angel LP (2015) Factors influencing foraging behaviour in the Australasian gannet. Deakin

University

Angel LP, Barker S, Berlincourt M, Tew E, Warwick-Evans V, Arnould JP (2015a) Eating

locally: Australasian gannets increase their foraging effort in a restricted range. Biol

Open 4:1298-1305

Angel LP, Wells MR, Rodríguez-Malagón MA, Tew E, Speakman JR, Arnould JPY (2015b)

Sexual size dimorphism and body condition in the Australasian gannet. PLoS ONE

10:e0142653

Angel LP, Berlincourt M, Arnould JPY (2016) Pronounced inter-colony variation in the

foraging ecology of Australasian gannets: influence of habitat differences. Mar Ecol

Prog Ser 556:261-272

Aplin LM, Farine DR, Mann RP, Sheldon BC (2014) Individual-level personality influences

social foraging and collective behaviour in wild birds. Proc R Soc B-Biol Sci 281

173 Araújo MS, Bolnick DI, Machado G, Giaretta AA, dos Reis SF (2007) Using δ13C stable

isotopes to quantify individual-level diet variation. Oecologia 152:643-654

Araújo MS, Bolnick DI, Layman CA (2011) The ecological causes of individual

specialisation. Ecol Lett 14:948-958

Arnould J, Cherel Y, Gibbens J, White J, Littnan C (2011) Stable isotopes reveal inter-annual

and inter-individual variation in the diet of female Australian fur seals. Mar Ecol Prog

Ser 422:291-302

Ashmole NP (1971) Seabird ecology and the marine environment. In: Farner DS, King JR

(eds) Avian biology. Academic Press, New York, N.Y., U.S.A.

Baechler J, Beck C, Bowen W (2002) Dive shapes reveal temporal changes in the foraging

behaviour of different age and sex classes of harbour seals (Phoca vitulina). Can J

Zool 80:1569-1577

Bai M-L, Severinghaus LL, Philippart MT (2012) Mechanisms underlying small-scale partial

migration of a subtropical owl. Behav Ecol 23:153-159

Baird PH (1990) Influence of abiotic factors and prey distribution on diet and reproductive

success of three seabird species in Alaska. Ornis Scand:224-235

Ballance LT, Pitman RL, Fiedler PC (2006) Oceanographic influences on seabirds and

cetaceans of the eastern tropical Pacific: A review. Prog Oceanogr 69:360-390

Barber RT, Chavez FP (1983) Biological Consequences of El Niño. Science 222:1203-1210

Barker S (2012) Geographic variation in the foraging ecology of the Australasian gannet

(Morus serrator). B of Env Sc (Hons) dissertation, Deakin University

Barnes C, Jennings JTBS (2009) Environmental correlates of large-scale spatial variation in

the δ13C of marine animals. Estuar Coast Shelf Sci 81:368-374

Barquete V, Strauss V, Ryan PG (2013) Stable isotope turnover in blood and claws: a case

study in captive African penguins. J Exp Mar Biol Ecol 448:121-127

174 Barrett RT, Camphuysen K, Anker-Nilssen T, Chardine JW and others (2007) Diet studies of

seabirds: a review and recommendations. ICES J Mar Sci 64:1675-1691

Barton KJ (2016) MuMIn: Multi-Model Inference. R package, version 1.15.6.

https://CRAN.R-project.org/package=MuMIn

Bates D, Maechler M, Bolker B (2011) lme4: linear mixed-effects models using S4 classes. R

package, version 1.1-6. https://cran.r-project.org/package=lme4.

Baudrot V, Perasso A, Fritsch C, Giraudoux P, Raoul F (2016) The adaptation of generalist

predators’ diet in a multi-prey context: insights from new functional responses.

Ecology 97:1832-1841

Baum JK, Worm B (2009) Cascading top-down effects of changing oceanic predator

abundances. J Anim Ecol 78:699-714

Baylis A, Orben R, Arnould J, Peters K, Knox T, Costa D, Staniland I (2015) Diving deeper

into individual foraging specializations of a large marine predator, the southern sea

lion. Oecologia 179:1053-1065

Baylis AMM, Page B, Goldsworthy SD (2008) Colony-specific foraging areas of lactating

New Zealand fur seals. Mar Ecol Prog Ser 361:279-290

Bearhop S, Adams CE, Waldron S, Fuller RA, MacLeod H (2004) Determining trophic niche

width: a novel approach using stable isotope analysis. J Anim Ecol 73:1007-1012

Bearhop S, Phillips RA, McGill R, Cherel Y, Dawson DA, Croxall JP (2006) Stable isotopes

indicate sex-specific and long-term individual foraging specialisation in diving

seabirds. Mar Ecol Prog Ser 311:157-164

Becker P, Bradley JS (2007) The role of intrinsic factors for the recruitment process in long-

lived birds. J Ornithol 148:377-384

Bell AM, Hankison SJ, Laskowski KL (2009) The repeatability of behaviour: a meta-

analysis. Anim Behav 77:771-783

175 Benhamou S (2004) How to reliably estimate the tortuosity of an animal's path: straightness,

sinuosity, or fractal dimension? J Theor Biol 229:209-220

Bernard HJ, Hohn AA (1989) Differences in feeding habits between pregnant and lactating

spotted dolphins (Stenella attenuata). J Mammal 70:211-215

Bernays EA, Funk DJ (1999) Specialists make faster decisions than generalists: experiments

with aphids. Proc R Soc Lond [Biol] 266:151-156

Besel D, Hauber ME, Hunter C, Ward-Smith T, Raubenheimer D, Millar CD, Ismar SMH

(2018) Multifactorial roles of interannual variability, season, and sex for foraging

patterns in a sexually size monomorphic seabird, the Australasian gannet (Morus

serrator). Mar Biol 165:72

Bêty J, Giroux J-F, Gauthier G (2004) Individual variation in timing of migration: causes and

reproductive consequences in greater snow geese (Anser caerulescens atlanticus).

Behav Ecol Sociobiol 57:1-8

Bindoff AD, Wotherspoon SJ, Guinet C, Hindell MA (2017) Twilight‐free geolocation from

noisy light data. Methods Ecol Evol 0:1-9

BirdLife International (2016) Species factsheet: Morus serrator. http://www.birdlife.org

(accessed 26/09/2016)

BirdLife International (2018a) Morus serrator. The IUCN Red List of Threatened Species

2018: e.T22696675A132588447 (accessed 15 April 2019)

BirdLife International (2018b) Data Zone: seabirds. http://datazone.birdlife.org (accessed

27/09/2018)

Black JM (2001) Fitness consequences of long-term pair bonds in barnacle geese: monogamy

in the extreme. Behav Ecol 12:640-645

Blackburn M, Gartner P (1954) Populations of barracouta, Thyrsites atun (Euphrasen), in

Australian waters. Mar Freshwater Res 5:411-468

176 Blandon A, zu Ermgassen PSE (2014) Quantitative estimate of commercial fish enhancement

by seagrass habitat in southern Australia. Estuar Coast Shelf Sci 141:1-8

Blight LK, Hobson KA, Kyser TK, Arcese P (2015) Changing gull diet in a changing world:

A 150-year stable isotope (δ13C, δ15N) record from feathers collected in the Pacific

Northwest of North America. Glob Change Biol 21:1497–1507

Boecklen WJ, Yarnes CT, Cook BA, James AC (2011) On the use of stable isotopes in

trophic ecology. Annu Rev Ecol Evol Syst 42:411-440

Bogdanova MI, Butler A, Wanless S, Moe B and others (2017) Multi-colony tracking reveals

spatio-temporal variation in carry-over effects between breeding success and winter

movements in a pelagic seabird. Mar Ecol Prog Ser 578:167-181

Bolker BM, Brooks ME, Clark CJ, Geange SW, Poulsen JR, Stevens MHH, White J-SS

(2008) Generalized linear mixed models: a practical guide for ecology and evolution.

Trends Ecol Evol 24:127-135

Bolnick DI, Yang LH, Fordyce JA, Davis JM, Svanbäck R (2002) Measuring individual-level

resource specialization. Ecology 83:2936-2941

Bolnick DI, Doebeli M (2003) Sexual dimorphism and adaptive speciation: two sides of the

same ecological coin. Evolution 57:2433-2449

Bolnick DI, Svanback R, Fordyce JA, Yang LH, Davis JM, Hulsey CD, Forister ML (2003)

The ecology of individuals: incidence and implications of individual specialization.

Am Nat 161:1-28

Bolnick DI, Amarasekare P, Araújo MS, Bürger R and others (2011) Why intraspecific trait

variation matters in community ecology. Trends Ecol Evol 26:183-192

Bond AL, Jones I (2009) A practical introduction to stable-isotope analysis for seabird

biologists: approaches, cautions and caveats. Mar Ornithol 37:e188

177 Bond AL, Diamond AW (2011) Recent Bayesian stable-isotope mixing models are highly

sensitive to variation in discrimination factors. Ecol Appl 21:1017-1023

Bond AL, Hobson KA (2012) Reportingstable-isotope ratios in ecology: recommended

terminology, guidelines and best practices. Waterbirds 35:324-331

Bosley KL, Witting DA, Chambers RC, Wainright SC (2002) Estimating turnover rates of

carbon and nitrogen in recently metamorphosed winter flounder Pseudopleuronectes

americanus with stable isotopes. Mar Ecol Prog Ser 236:233-240

Bost CA, Cotté C, Bailleul F, Cherel Y and others (2009) The importance of oceanographic

fronts to marine birds and mammals of the southern oceans. J Mar Syst 78:363-376

Bowen GJ, West JB, Hoogewerff J (2009) Isoscapes: isotope mapping and its applications. J

Geochem Explor 102:v-vii

Bowlin M, Bisson I-A, Shamoun-Baranes J, Reichard J and others (2010) Grand challenges

in migration biology. Integr Comp Biol 50:261-279

Braccini JM (2008) Feeding ecology of two high-order predators from south-eastern

Australia: the coastal broadnose and the deepwater sharpnose sevengill sharks. Mar

Ecol Prog Ser 371:273-284

Braley M, Goldsworthy SD, Page B, Steer M, Austin JJ (2010) Assessing morphological and

DNA-based diet analysis techniques in a generalist predator, the arrow squid

Nototodarus gouldi. Mol Ecol Resour 10:466-474

Bray DJ, Gomon MF (2018) Fishes of Australia. Museums Victoria and OzFishNet

http://fishesofaustralia.net.au/ (accessed 8 Feb 2018)

Brothers N, Gales R, Pemberton D (1993) Prey harvest of the Australasian gannet (Sula

serrator) in Tasmania. Wildl Res 20:777-783

178 Brown DA, Wilson PR (2004) Establishment and growth of an Australasian gannet colony at

Waimaru, Pelorus Sound, and a new colony at Arapawa Island, Queen Charlotte

Sound. Notornis 51:227-229

Brown DD, Kays R, Wikelski M, Wilson R, Klimley AP (2013) Observing the unwatchable

through acceleration logging of animal behavior. Anim Biotelem 1:20

Bull JJ (1983) Evolution of sex determining mechanisms. The Benjamin/Cummings

Publishing Company, Inc., Menlo Park, California

Bulman C, Althaus F, He X, Bax NJ, Williams A (2001) Diets and trophic guilds of demersal

fishes of the south-eastern Australian shelf. Mar Freshwater Res 52:537-548

Bunce A, Norman FI (2000) Changes in the diet of the Australasian gannet (Morus serrator)

in response to the 1998 mortality of pilchards (Sardinops sagax). Mar Freshwater Res

51:349-353

Bunce A (2001) Prey consumption of Australasian gannets (Morus serrator) breeding in Port

Phillip Bay, southeast Australia, and potential overlap with commercial fisheries.

ICES J Mar Sci 58:904-915

Bunce A, Norman FI, Brothers N, Gales R (2002) Long-term trends in the Australasian

gannet (Morus serrator) population in Australia: the effect of climate change and

commercial fisheries. Mar Biol 141:263-269

Bunnefeld N, Börger L, van Moorter B, Rolandsen CM, Dettki H, Solberg EJ, Ericsson G

(2011) A model-driven approach to quantify migration patterns: individual, regional

and yearly differences. J Anim Ecol 80:466-476

Burger AE, Shaffer SA (2008) Perspectives in ornithology application of tracking and data-

logging technology in research and conservation of seabirds. Auk 125:253-264

Burnham KP, Anderson DR (2002) Model selection and multimodel inference: a practical

information-theoretic approach. Springer - Science & Business Media, New York

179 Butler A, Althaus F, Furlani D, Ridgway K (2002) Assessment of the conservation values of

the Bonney Upwelling area: a component of the Commonwealth Marine Conservation

Assessment Program 2002–2004, (ed. D.o.t. Environment). Information Solution

Works, Hobart, Tasmania: CSIRO Marine Research, p 80

Cairns DK (1988) Seabirds as indicators of marine food supplies. Biol Oceanogr 5:261-271

Calenge C (2006) The package “adehabitat” for the R software: a tool for the analysis of

space and habitat use by animals. Ecol Modell 197:516-519

Calenge C, Dray S, Royer-Carenzi M (2009) The concept of animals' trajectories from a data

analysis perspective. Ecol Inform 4:34-41

Camprasse ECM, Cherel Y, Arnould JPY, Hoskins AJ, Bost CA (2017a) Combined bio-

logging and stable isotopes reveal individual specialisations in a benthic coastal

seabird, the Kerguelen shag. PLoS ONE 12:e0172278

Camprasse ECM, Cherel Y, Arnould JPY, Hoskins AJ, Bustamante P, Bost CA (2017b) Mate

similarity in foraging Kerguelen shags: a combined bio-logging and stable isotope

investigation. Mar Ecol Prog Ser 578:183-196

Camprasse ECM, Sutton GJ, Berlincourt M, Arnould JPY (2017c) Changing with the times:

little penguins exhibit flexibility in foraging behaviour and low behavioural

consistency. Mar Biol 164:169

Carneiro APB, Manica A, Clay TA, Silk JRD, King M, Phillips RA (2016) Consistency in

migration strategies and habitat preferences of brown skuas over two winters, a

decade apart. Mar Ecol Prog Ser 553:267-281

Carneiro APB, Bonnet-Lebrun AS, Manica A, Staniland IJ, Phillips RA (2017) Methods for

detecting and quantifying individual specialisation in movement and foraging

strategies of marine predators. Mar Ecol Prog Ser 578:151-166

180 Castillo-Guerrero JA, Mellink E (2011) Occasional inter-sex differences in diet and foraging

behavior of the blue-footed booby: maximizing chick rearing in a variable

environment? J Ornithol 152:269-277

Castillo-Guerrero JA, Lerma M, Mellink E, Suazo-Guillén E, Peñaloza-Padilla EA (2016)

Environmentally-mediated flexible foraging strategies in Brown Boobies in the Gulf

of California. Ardea 104:33-47

Catry P, Phillips RA, Croxall JP, Ruckstuhl K, Neuhaus P (2005) Sexual segregation in birds:

patterns, processes and implications for conservation. In: Ruckstuhl KE, Neuhaus P

(eds) Sexual segregation in vertebrates: ecology of the two sexes. Cambridge

University Press, Cambridge, UK, p 351-378

Catry P, Dias MP, Phillips RA, Granadeiro JP (2013) Carry-over effects from breeding

modulate the annual cycle of a long-distance migrant: an experimental demonstration.

Ecology 94:1230-1235

Ceia F, Ramos J (2015) Individual specialization in the foraging and feeding strategies of

seabirds: a review. Mar Biol 162:1923-1938

Ceia FR, Phillips RA, Ramos JA, Cherel Y, Vieira RP, Richard P, Xavier JC (2012) Short-

and long-term consistency in the foraging niche of wandering albatrosses. Mar Biol

159:1581-1591

Ceia FR, Paiva VH, Fidalgo V, Morais L and others (2014) Annual and seasonal consistency

in the feeding ecology of an opportunistic species, the yellow-legged gull Larus

michahellis. Mar Ecol Prog Ser 497:273-284

Chambers LE, Patterson T, Hobday AJ, Arnould JP, Tuck GN, Wilcox C, Dann P (2014)

Determining trends and environmental drivers from long-term marine mammal and

seabird data: examples from southern Australia. Reg Environ Change:1-13

181 Chapman BB, Brönmark C, Nilsson J-Å, Hansson L-A (2011a) The ecology and evolution of

partial migration. Oikos 120:1764-1775

Chapman BB, Hulthén K, Blomqvist DR, Hansson L-A and others (2011b) To boldly go:

individual differences in boldness influence migratory tendency. Ecol Lett 14:871-876

Chastel O, Weimerskirch H, Jouventin P (1993) High annual variability in reproductive

success and survival of an Antarctic seabird, the snow petrel Pagodroma nivea.

Oecologia 94:278-285

Chavez FP, Messié M, Pennington JT (2011) Marine primary production in relation to

climate variability and change. Annu Rev Mar Sci 3:227-260

Cherel Y, Hobson K, Weimerskirch H (2005) Using stable isotopes to study resource

acquisition and allocation in procellariiform seabirds. Oecologia 145

Cherel Y, Hobson KA (2007) Geographical variation in carbon stable isotope signatures of

marine predators: a tool to investigate their foraging areas in the Southern Ocean. Mar

Ecol Prog Ser 329:281-287

Cherel Y, Hobson KA, Guinet C, Vanpe C (2007) Stable isotopes document seasonal changes

in trophic niches and winter foraging individual specialization in diving predators

from the Southern Ocean. J Anim Ecol 76:826-836

Chiaradia A, Dann P, Jessop R, Collins P (2002) The diet of Crested Tern (Sterna bergii)

chicks on Phillip Island, Victoria, Australia. Emu 102:367-371

Chiaradia A, Costalunga A, Kerry K (2003) The diet of Little Penguins (Eudyptula minor) at

Phillip Island, Victoria, in the absence of a major prey—Pilchard (Sardinops sagax).

Emu 103:43-48

Chiaradia A, Forero MG, Hobson KA, Cullen JM (2010) Changes in diet and trophic position

of a top predator 10 years after a mass mortality of a key prey. ICES J Mar Sci

67:1710-1720

182 Chiaradia A, Forero MG, Hobson KA, Swearer SE, Hume F, Renwick L, Dann P (2012) Diet

segregation between two colonies of little penguins Eudyptula minor in southeast

Australia. Austral Ecol 37:610-619

Chiaradia A, Forero MG, McInnes JC, Ramírez F (2014) Searching for the true diet of marine

predators: incorporating bayesian priors into stable isotope mixing models. PLoS

ONE 9:e92665

Church MR, Ebersole JL, Rensmeyer KM, Couture RB, Barrows FT, Noakes DLG (2008)

Mucus: a new tissue fraction for rapid determination of fish diet switching using

stable isotope analysis. Can J Fish Aquat Sci 66:1-5

Cleasby IR, Wakefield ED, Bodey TW, Davies RD and others (2015) Sexual segregation in a

wide-ranging marine predator is a consequence of habitat selection. Mar Ecol Prog

Ser 518:1-12

Coblentz KE, Rosenblatt AE, Novak M (2017) The application of Bayesian hierarchical

models to quantify individual diet specialization. Ecology 98:1535-1547

Cohen D (1993) The equilibrium distribution of optimal search and sampling effort of

foraging animals in patchy environments. In: Yoshimura J, Clark CW (eds)

Adaptation in Stochastic Environments. Springer Berlin Heidelberg, Berlin,

Heidelberg, p 173-191

Colles A, Liow LH, Prinzing A (2009) Are specialists at risk under environmental change?

Neoecological, paleoecological and phylogenetic approaches. Ecol Lett 12:849-863

Commonwealth of Australia (2007) The south-west marine bioregional plan. Bioregional

profile: a description of the ecosystems, conservation values and uses of the south-

west marine region, Department of the Environment, Water, Heritage and the Arts,

Australian Government

183 Commonwealth of Australia (2015) South-east marine region profile: a description of the

ecosystems, conservation values and uses of the south-east marine region, Department

of Environment, Australian Government

Condie SA, Dunn JR (2006) Seasonal characteristics of the surface mixed layer in the

Australasian region: implications for primary production regimes and biogeography.

Mar Freshwater Res 57:569-590

Cook TR, Cherel Y, Tremblay Y (2006) Foraging tactics of chick-rearing Crozet shags:

individuals display repetitive activity and diving patterns over time. Polar Biol

29:562-569

Cox J (1978) The distribution and seasonal occurrence of gannets off the South Australian

coast. S Aust Ornithol 28:18-20

Cox NJ (2011) Stata tip 96: Cube roots. Stata Journal 11:149-154

Craig A, Herman L, Gabriele C, Pack A (2003) Migratory timing of humpback whales

(Megaptera novaeangliae) in the central-north Pacific varies with age, sex and

reproductive status. Behaviour 140:981-1001

Crawford K, McDonald RA, Bearhop S (2008) Applications of stable isotope techniques to

the ecology of mammals. Mammal Rev 38:87-107

Cristol DA, Baker MB, Carbone C (1999) Differential migration revisited. In: Nolan V,

Ketterson ED, Thompson CF (eds) Current Ornithology. Springer US, Boston, MA, p

33-88

Croll DA, Tershy BR, Hewitt RP, Demer DA and others (1998) An integrated approch to the

foraging ecology of marine birds and mammals. Deep-Sea Research Part II 45:1353-

1371

184 Croxall JP, Butchart SH, Lascelles B, Stattersfield AJ, Sullivan B, Symes A, Taylor P (2012)

Seabird conservation status, threats and priority actions: a global assessment. Bird

Conserv Int 22:1-34

Cummings M, Mollaghan D (2006) Repeatability and consistency of female preference

behaviours in a northern swordtail, Xiphophorus nigrensis. Anim Behav 72:217-224

Currie DR, Sorokin SJ (2010) A preliminary evaluation of the distribution and

trophodynamics of demersal fish from Spencer Gulf. Report to the South Australian

Department for Environment and Heritage. South Australian Research and

Development Institute (Aquatic Sciences), Adelaide. SARDI Publication No.

F2010/000088-1

Cury PM, Boyd IL, Bonhommeau S, Anker-Nilssen T and others (2011) Global seabird

response to forage fish depletion—one-third for the birds. Science 334:1703-1706

Dalerum F, Angerbjörn A (2005) Resolving temporal variation in vertebrate diets using

naturally occurring stable isotopes. Oecologia 144:647-658

Dall SR, Houston AI, McNamara JM (2004) The behavioural ecology of personality:

consistent individual differences from an adaptive perspective. Ecol Lett 7:734-739

Dall SR, Bell AM, Bolnick DI, Ratnieks FL (2012) An evolutionary ecology of individual

differences. Ecol Lett 15:1189-1198

Daniel C, Millar CD, Ismar SM, Stephenson BM, Hauber ME (2007) Evaluating molecular

and behavioural sexing methods for the Australasian gannet (Morus serrator). Aust J

Zool 55:377-382

Dann P, Arnould JP, Jessop R, Healy M (2004) Distribution and abundance of seabirds in

Western Port, Victoria. Emu 103:307-313

185 Davenport SR, Bax NJ (2002) A trophic study of a marine ecosystem off southeastern

Australia using stable isotopes of carbon and nitrogen. Can J Fish Aquat Sci 59:514-

530

Davies NB, Krebs JR, West SA (2012) An introduction to behavioural ecology. Wiley-

Blackwell, Oxford, UK

Davis AM, Blanchette ML, Pusey BJ, Jardine TD, Pearson RG (2012) Gut content and stable

isotope analyses provide complementary understanding of ontogenetic dietary shifts

and trophic relationships among fishes in a tropical river. Freshw Biol 57:2156-2172

Davis NE, Forsyth DM, Triggs B, Pascoe C and others (2015) Interspecific and geographic

variation in the diets of sympatric carnivores: dingoes/wild dogs and red foxes in

south-eastern Australia. PLoS ONE 10:e0120975 de L. Brooke M (2004) The food consumption of the world's seabirds. Proc R Soc Lond

[Biol] 271:S246-S248

Deagle BE, Kirkwood R, Jarman SN (2009) Analysis of Australian fur seal diet by

pyrosequencing prey DNA in faeces. Mol Ecol 18:2022-2038

Dehnhard N, Eens M, Sturaro N, Lepoint G, Demongin L, Quillfeldt P, Poisbleau M (2016)

Is individual consistency in body mass and reproductive decisions linked to individual

specialization in foraging behavior in a long‐lived seabird? Ecol Evol 6:4488-4501

DeNiro MJ, Epstein S (1978) Influence of diet on the distribution of carbon isotopes in

animals. Geochim Cosmochim Acta 42:495-506

Deniro MJ, Epstein S (1981) Influence of diet on the distribution of nitrogen isotopes in

animals. Geochim Cosmochim Acta 45:341-351

Department of the Environment (2019) Morus serrator in Species Profile and Threats

Database, Department of the Environment, Canberra.

http://www.environment.gov.au/sprat (accessed 16 Apr 2019)

186 Devictor V, Clavel J, Julliard R, Lavergne S and others (2010) Defining and measuring

ecological specialization. J Appl Ecol 47:15-25

Dias MP, Granadeiro JP, Phillips RA, Alonso H, Catry P (2011) Breaking the routine:

individual Cory's shearwaters shift winter destinations between hemispheres and

across ocean basins. Proc R Soc B-Biol Sci 278:1786-1793

Dias MP, Granadeiro JP, Catry P (2013) Individual variability in the migratory path and

stopovers of a long-distance pelagic migrant. Anim Behav 86:359-364

Dimmlich W, Breed W, Geddes M, Ward T (2004) Relative importance of gulf and shelf

waters for spawning and recruitment of Australian anchovy, Engraulis australis, in

South Australia. Fish Oceanogr 13:310-323

Dingemanse NJ, Dochtermann NA (2013) Quantifying individual variation in behaviour:

mixed-effect modelling approaches. J Anim Ecol 82:39-54

Dingemanse NJ, Wolf M (2013) Between-individual differences in behavioural plasticity

within populations: causes and consequences. Anim Behav 85:1031-1039

Dudley SFJ, Field JG, Shelton PA (1985) Distribution and abundance of eggs, larvae and

early juveniles of saury Scomberesox saurus scombroides (Richardson) off the south-

western Cape, South Africa, 1977/78. S Afr J Mar Sci 3:229-237

Edgar GJ, Shaw C (1995) The production and trophic ecology of shallow-water fish

assemblages in southern Australia II. Diets of fishes and trophic relationships between

fishes and benthos at Western Port, Victoria. J Exp Mar Biol Ecol 194:83-106

Edwards MA, Derocher AE, Hobson KA, Branigan M, Nagy JA (2011) Fast carnivores and

slow herbivores: differential foraging strategies among grizzly bears in the Canadian

Arctic. Oecologia 165:877-889

187 Ehrich D, Tarroux A, Stien J, Lecomte N, Killengreen S, Berteaux D, Yoccoz NG (2011)

Stable isotope analysis: modelling lipid normalization for muscle and eggs from arctic

mammals and birds. Methods Ecol Evol 2:66-76

Embar K, Kotler BP, Mukherjee S (2011) Risk management in optimal foragers: the effect of

sightlines and predator type on patch use, time allocation, and vigilance in gerbils.

Oikos 120:1657-1666

Emmenegger T, Hahn S, Bauer S (2014) Individual migration timing of common nightingales

is tuned with vegetation and prey phenology at breeding sites. BMC Ecol 14:9

Estes J, Riedman M, Staedler M, Tinker M, Lyon B (2003) Individual variation in prey

selection by sea otters: patterns, causes and implications. J Anim Ecol 72:144-155

Evans K, Bax N, Smith DC (2017) Australia state of the environment 2016: marine

environment, independent report to the Australian Government Minister for the

Environment and Energy, Australian Government Department of the Environment

and Energy, Canberra.

Falk K, Benvenuti S, Dall’Antonia L, Gilchrist G, Kampp K (2002) Foraging behaviour of

thick-billed murres breeding in different sectors of the North Water polynya: an inter-

colony comparison. Mar Ecol Prog Ser 231:293-302

Fayet AL, Shoji A, Freeman R, Perrins CM, Guilford T (2017) Within-pair similarity in

migration route and female winter foraging effort predict pair breeding performance

in a monogamous seabird. Mar Ecol Prog Ser 569:243-252

Fifield DA, Montevecchi WA, Garthe S, Robertson GJ, Kubetzki U, Rail JF (2014)

Migratory tactics and wintering areas of northern gannets (Morus bassanus) breeding

in North America. Ornithol Monogr 79:1-63

188 Filby NE (2016) Burrunan dolphin (Tursiops australis) tourism in Port Phillip Bay, Australia:

effects, implications and management. Doctor of Phylosophy Thesis, Victoria

University

Filby NE, Christiansen F, Scarpaci C, Stockin KA (2017) Effects of swim-with-dolphin

tourism on the behaviour of a threatened species, the Burrunan dolphin Tursiops

australis. Endang Species Res 32:479-490

Forero MG, Tella JL, Donázar JA, Blanco G, Bertellotti M, Ceballos O (2001) Phenotypic

assortative mating and within-pair sexual dimorphism and its influence on breeding

success and offspring quality in Magellanic penguins. Can J Zool 79:1414-1422

Forero MG, Hobson KA, Bortolotti GR, Donazar JA, Bertellotti M, Blanco G (2002) Food

resource utilisation by the Magellanic penguin evaluated through stable-isotope

analysis: segregation by sex and age and influence on offspring quality. Mar Ecol

Prog Ser 234:289-299

Forero MG, González-Solís J, Hobson KA, Donázar JA, Bertellotti M, Blanco G, Bortolotti

GR (2005) Stable isotopes reveal trophic segregation by sex and age in the southern

giant petrel in two different food webs. Mar Ecol Prog Ser 296:107-113

Fort J, Pettex E, Tremblay Y, Lorentsen S-H and others (2012) Meta-population evidence of

oriented chain migration in northern gannets (Morus bassanus). Front Ecol Environ

10:237-242

Fossi MC, Casini S, Caliani I, Panti C and others (2012) The role of large marine vertebrates

in the assessment of the quality of pelagic marine ecosystems. Mar Environ Res

77:156-158

Fox J, Weisberg S (2018) An R Companion to Applied Regression, Second Edition. Sage

Publications. http://socserv.socsci.mcmaster.ca/jfox/Books/Companion

189 France RL (1995a) Differentiation between littoral and pelagic food webs in lakes using

stable carbon isotopes. Limnol Oceanogr 40:1310-1313

France RL (1995b) Carbon-13 enrichment in benthic compared to planktonic algae: foodweb

implications. Mar Ecol Prog Ser 124:307-312

Francis CM, Cooke F (1986) Differential timing of spring migration in wood warblers

(Parulinae). Auk 103:548-556

Franco-Trecu V, Drago M, Riet-Sapriza FG, Parnell A, Frau R, Inchausti P (2013) Bias in

diet determination: incorporating traditional methods in Bayesian mixing models.

PLoS ONE 8:e80019

Frederiksen M, Mavor RA, Wanless S (2007) Seabirds as environmental indicators: the

advantages of combining data sets. Mar Ecol Prog Ser 352:205-211

Frost PGH (2017) Population status and trends of selected seabirds in northern New Zealand,

Conservation Services Programme - Department of Conservation, Wellington, NZ

Froy H, Lewis S, Catry P, Bishop CM and others (2015) Age-related variation in foraging

behaviour in the wandering albatross at South Georgia: no evidence for senescence.

PLoS ONE 10:e0116415

Fudickar AM, Schmidt A, Hau M, Quetting M, Partecke J (2013) Female-biased obligate

strategies in a partially migratory population. J Anim Ecol 82:863-871

Fulton E, A. , Smith A, D. M. , Johnson C, R. (2003) Effect of complexity on marine

ecosystem models. Mar Ecol Prog Ser 253:1-16

Furlani D, Gales R, Pemberton D (2007) Otoliths of common Australian temperate fish: a

photographic guide. CSIRO publishing, Australia

Galef BG, Giraldeau L-A (2001) Social influences on foraging in vertebrates: causal

mechanisms and adaptive functions. Anim Behav 61:3-15

190 Gannes LZ, O'Brien DM, del Rio CM (1997) Stable isotopes in animal ecology: assumptions,

caveats, and a call for more laboratory experiments. Ecology 78:1271-1276

Garthe S, Guse N, Montevecchi WA, Rail JF, Gregoire F (2014) The daily catch: flight

altitude and diving behavior of northern gannets feeding on Atlantic mackerel. J Sea

Res 85:456-462

Gaston AJ (2004) Seabirds: a natural history. Yale University Press, New Haven,

Connecticut

Gearing JN (1988a) The Use of Stable Isotope Ratios for Tracing the Nearshore-Offshore

Exchange of Organic Matter. In: Jansson B-O (ed) Coastal-Offshore Ecosystem

Interactions. Springer Berlin Heidelberg, Berlin, Heidelberg, p 69-101

Gearing JN (1988b) The use of stable isotope ratios for tracing the nearshore-offshore

exchange of organic matter. In: Coastal-offshore ecosystem interactions. Springer,

Berlin, Heidelberg, p 69-101

Gelman A, Carlin JB, Stern HS, Dunson DB, Vehtari A, Rubin DB (2013) Bayesian data

analysis. Chapman and Hall/CRC, Florida, US

Gibbs C, Tomczak Jr M, Longmore A (1986) The nutrient regime of Bass Strait. Mar

Freshwater Res 37:451-466

Gibbs C (1992) Oceanography of Bass Strait: implications for the food supply of little

penguins Eudyptula minor. Emu 91:395-401

Gibbs H, Norman F, Ward S (2000) Reproductive parameters, chick growth and adult 'age' in

Australasian Gannets Morus serrator breeding in Port Phillip Bay, Victoria, in 1994-

95. Emu 100:175-185

Gibson RN, Pihl L, Burrows MT, Modin J, Wennhage H, Nickell LA (1998) Diel movements

of juvenile plaice Pleuronectes platessa in relation to predators, competitors, food

191 availability and abiotic factors on a microtidal nursery ground. Mar Ecol Prog Ser

165:145-159

Gleiss AC, Wilson RP, Shepard ELC (2011) Making overall dynamic body acceleration

work: on the theory of acceleration as a proxy for energy expenditure. Methods Ecol

Evol 2:23-33

Gochfeld M, Burger J (1984) Age differences in foraging behavior of the American robin

(Turdus Migratorius). Behaviour 88:227-239

Goericke R, Fry B (1994) Variations of marine plankton δ13C with latitude, temperature, and

dissolved CO2 in the world ocean. Global Biogeochem Cycles 8:85-90

Goldsworthy SD, Page B, Rogers P, Ward T (2011) Establishing ecosystem-based

management for the South Australian Sardine Fishery: developing ecological

performance indicators and reference points to assess the need for ecological

allocations. Final Report to the Fisheries Research and Development Corporation.

South Australian Research and Development Institute (Aquatic Sciences), Adelaide.

SARDI Publication No. F2010/000863-1. SARDI Research Report Series No. 529.

173pp.

Golet GH, Kuletz KJ, Roby DD, Irons DB (2000) Adult prey choice affects chick growth and

reproductive success in pigeon guillemots. Auk 117:82-91

Gomon MF, Bray DJ, Kuiter RH (2008) Fishes of Australia’s southern coast. Reed New

Holland, Sydney

Graham BS, Koch PL, Newsome SD, McMahon KW, Aurioles D (2010) Using isoscapes to

trace the movements and foraging behavior of top predators in oceanic ecosystems.

In: Isoscapes. Springer, Netherlands, p 299-318

192 Granadeiro JP, Brickle P, Catry P (2014) Do individual seabirds specialize in fisheries'

waste? The case of black-browed albatrosses foraging over the Patagonian Shelf.

Anim Conserv 17:19-26

Greene BS (1999) Increase of gannets (Morus serrator) at Muriwai, Auckland. Notornis

46:423-433

Greig S, Coulson J, Monaghan P (1983) Age-related differences in foraging success in the

herring gull (Larus argentatus). Anim Behav 31:1237-1243

Grémillet D, Dell'Omo G, Ryan PG, Peters G, Ropert-Coudert Y, Weeks SJ (2004) Offshore

diplomacy, or how seabirds mitigate intra-specific competition: a case study based on

GPS tracking of Cape gannets from neighbouring colonies. Mar Ecol Prog Ser

268:265-279

Grémillet D, Boulinier T (2009) Spatial ecology and conservation of seabirds facing global

climate change: a review. Mar Ecol Prog Ser 391:121-137

Griffiths JR, Kadin M, Nascimento FJA, Tamelander T and others (2017) The importance of

benthic–pelagic coupling for marine ecosystem functioning in a changing world. Glob

Change Biol 23:2179-2196

Griggio M, Hoi H (2011) An experiment on the function of the long-term pair bond period in

the socially monogamous bearded reedling. Anim Behav 82:1329-1335

Halaj J, Ross DW, Moldenke AR (1998) Habitat structure and prey availability as predictors

of the abundance and community organization of spiders in western Oregon forest

canopies. J Arachnol 26:203-220

Hamer K, Phillips R, Hill J, Wanless S, Wood A (2001) Contrasting foraging strategies of

gannets Morus bassanus at two North Atlantic colonies: foraging trip duration and

foraging area fidelity. Mar Ecol Prog Ser 224:283-290

193 Hamer K, Humphreys E, Garthe S, Hennicke J and others (2007) Annual variation in diets,

feeding locations and foraging behaviour of gannets in the North Sea: flexibility,

consistency and constraint. Mar Ecol Prog Ser 338:5-305

Harley CDG, Hughes AR, Hultgren KM, Miner BG and others (2006) The impacts of climate

change in coastal marine systems. Ecol Lett 9:228-241

Harris G, Batley G, Fox D, Hall D and others (1996) Port Phillip Bay environmental study

final report, Collingwood, VIC. CSIRO Publishing

Harris S, Raya Rey A, Zavalaga C, Quintana F (2014) Strong temporal consistency in the

individual foraging behaviour of Imperial Shags Phalacrocorax atriceps. Ibis

156:523-533

Hedd A, Gales R, Brothers N (2001) Foraging strategies of shy albatross Thalassarche cauta

breeding at Albatross Island, Tasmania, Australia. Mar Ecol Prog Ser 224:267-282

Hedd A, Ryder JL, Cowen LL, Bertram DF (2002) Inter-annual variation in the diet,

provisioning and growth of Cassin's auklet at Triangle Island, British Columbia:

responses to variation in ocean climate. Mar Ecol Prog Ser 229:221-232

Heise CD, Moore FR (2003) Age-related differences in foraging efficiency, molt, and fat

deposition of gray catbirds prior to autumn migration. Condor 105:496-504

Heithaus MR, Frid A, Wirsing AJ, Worm B (2008) Predicting ecological consequences of

marine top predator declines. Trends Ecol Evol 23:202-210

Heithaus MR, Wirsing AJ, Dill LM (2012) The ecological importance of intact top-predator

populations: a synthesis of 15 years of research in a seagrass ecosystem. Mar

Freshwater Res 63:1039-1050

Hernández-Miranda E, Ojeda FP (2006) Inter-annual variability in somatic growth rates and

mortality of coastal fishes off central Chile: an ENSO driven process? Mar Biol

149:925-936

194 Hernández-Pliego J, Rodríguez C, Dell’Omo G, Bustamante J (2017) Combined use of tri-

axial accelerometers and GPS reveals the flexible foraging strategy of a bird in

relation to weather conditions. PLoS ONE 12:e0177892

Higgins J (2004) Introduction to modern nonparametric statistics: Pacific Grove, California,

Brooks/Cole-Thomson Learning. Inc

Hobday AJ, Pecl GT (2014) Identification of global marine hotspots: sentinels for change and

vanguards for adaptation action. Rev Fish Biol Fish 24:415-425

Hobson KA, Welch HE (1992) Determination of trophic relationships within a high Arctic

marine food web using δ13C and δ15N analysis. Mar Ecol Prog Ser 84:9-18

Hobson KA, Piatt JF, Pitocchelli J (1994) Using stable isotopes to determine seabird trophic

relationships. J Anim Ecol 63:786-798

Hobson KA, Barnett-Johnson R, Cerling T (2010) Using isoscapes to track animal migration.

In: West JB, Bowen GJ, Dawson TE, Tu KP (eds) Isoscapes: understanding

movement, pattern, and process on Earth through isotope mapping. Springer

Netherlands, Dordrecht, p 273-298

Holbrook SJ, Schmitt RJ, Stephens JS (1997) Changes in an assemblage of temperate reef

fishes associated with a climate shift. Ecol Appl 7:1299-1310

Holmes WG (1984) Predation risk and foraging behavior of the hoary marmot in Alaska.

Behav Ecol Sociobiol 15:293-301

Hooker SK, Biuw M, McConnell BJ, Miller PJO, Sparling CE (2007) Bio-logging science:

Logging and relaying physical and biological data using animal-attached tags. Deep

Sea Res Part 2 Top Stud Oceanogr 54:177-182

Hoskins AJ, Dann P, Ropert-Coudert Y, Kato A, Chiaradia A, Costa DP, Arnould JPY (2008)

Foraging behaviour and habitat selection of the little penguin Eudyptula minor during

early chick rearing in Bass Strait, Australia. Mar Ecol Prog Ser 366:293-303

195 Huettmann F, Diamond AW (2000) Seabird migration in the Canadian northwest Atlantic

Ocean: moulting locations and movement patterns of immature birds. Can J Zool

78:624-647

Hughes RN (1980) Optimal foraging theory in the marine context. Oceanogr Mar Biol Ann

Rev 18:423-481

Inger R, Bearhop S (2008) Applications of stable isotope analyses to avian ecology. Ibis

150:447-461

Ismar SM (2010) Foraging and breeding ecology of the Australasian gannet Morus serrator,

with applications for rare New Zealand seabirds. The University of Auckland,

Auckland, New Zealand.

Ismar SM, Daniel C, Stephenson BM, Hauber ME (2010a) Mate replacement entails a fitness

cost for a socially monogamous seabird. Naturwissenschaften 97:109-113

Ismar SM, Phillips RA, Rayner MJ, Hauber ME (2011) Geolocation tracking of the annual

migration of adult Australasian gannets (Morus serrator) breeding in New Zealand.

Wilson J Ornithol 123:121-125

Ismar SMH, Hunter C, Lay K, Ward-Smith T, Wilson PR, Hauber ME (2010b) A virgin

flight across the Tasman Sea? Satellite tracking of post-fledging movement in the

Australasian Gannet Morus serrator. J Ornithol 151:755-759

Ismar SMH, Daniel C, Igic B, Morrison-Whittle PK and others (2014) Sexual plumage

dichromatism in a size monomorphic seabird. J Ornithol 126:417-428

Ismar SMH, Raubenheimer D, Bury SJ, Millar CD, Hauber ME (2017) Sex-specific foraging

during parental care in a size-monomorphic seabird, the Australasian gannet (Morus

serrator). Wilson J Ornithol 129:139-147

196 Ito M, Minami H, Tanaka Y, Watanuki Y (2009) Seasonal and inter-annual oceanographic

changes induce diet switching in a piscivorous seabird. Mar Ecol Prog Ser 393:273-

284

Iverson SJ, Springer AM, Kitaysky AS (2007) Seabirds as indicators of food web structure

and ecosystem variability: qualitative and quantitative diet analyses using fatty acids.

Mar Ecol Prog Ser 352:235-244

Jaeger A, Connan M, Richard P, Cherel Y (2010) Use of stable isotopes to quantify seasonal

changes of trophic niche and levels of population and individual specialisation in

seabirds. Mar Ecol Prog Ser 401:269-277

Jaeger A, Goutte A, Lecomte VJ, Richard P and others (2014) Age, sex, and breeding status

shape a complex foraging pattern in an extremely long-lived seabird. Ecology

95:2324-2333

Jakubas D, Wojczulanis-Jakubas K, Iliszko L, Darecki M, Stempniewicz L (2014) Foraging

strategy of the little auk Alle alle throughout breeding season – switch from unimodal

to bimodal pattern. J Avian Biol 45:551-560

Jenkins G, May H, Wheatley M, Holloway M (1997a) Comparison of fish assemblages

associated with seagrass and adjacent unvegetated habitats of Port Phillip Bay and

Corner Inlet, Victoria, Australia, with emphasis on commercial species. Estuar Coast

Shelf Sci 44:569-588

Jenkins GP, May HMA, Wheatley MJ, Holloway MG (1997b) Comparison of fish

assemblages associated with seagrass and adjacent unvegetated habitats of Port Phillip

Bay and Corner Inlet, Victoria, Australia, with emphasis on commercial species.

Estuar Coast Shelf Sci 44:569-588

197 Johnson CR, Banks SC, Barrett NS, Cazassus F and others (2011) Climate change cascades:

shifts in oceanography, species' ranges and subtidal marine community dynamics in

eastern Tasmania. J Exp Mar Biol Ecol 400:17-32

Jones TB, Patrick SC, Arnould JPY, Rodríguez-Malagón MA, Wells MR, Green JA (2018)

Evidence of sociality in the timing and location of foraging in a colonial seabird. Biol

Lett 14

Jorge PE, Sowter D, Marques PAM (2011) Differential annual movement patterns in a

migratory species: effects of experience and sexual maturation. PLoS ONE 6:e22433

Kai ET, Benhamou S, van der Lingen CD, Coetzee JC, Pichegru L, Ryan PG, Gremillet D

(2013) Are Cape gannets dependent upon fishery waste? A multi-scale analysis using

seabird GPS-tracking, hydro-acoustic surveys of pelagic fish and vessel monitoring

systems. J Appl Ecol 50:659-670

Kaufman L, Rousseeuw PJ (2009) Finding groups in data: an introduction to cluster analysis.

John Wiley & Sons, New Jersey

Kelly JF (2000) Stable isotopes of carbon and nitrogen in the study of avian and mammalian

trophic ecology. Can J Zool 78:1-27

Kernaléguen L, Cherel Y, Knox TC, Baylis AMM, Arnould JPY (2015) Sexual niche

segregation and gender-specific individual specialisation in a highly dimorphic

marine mammal. PLoS ONE 10:e0133018

Ketterson ED, Nolan V (1983) The evolution of differential bird migration. In: Johnston RF

(ed) Current Ornithology. Springer, Boston, US, p 357-402

Key C, Ross C (1999) Sex differences in energy expenditure in non–human primates. Proc R

Soc Lond [Biol] 266:2479-2485

Kidawa D, Kowalczyk R (2011) The effects of sex, age, season and habitat on diet of the red

fox Vulpes vulpes in northeastern Poland. Acta Theriol 56:209-218

198 Kimura S, Nakai M, Sugimoto T (1997) Migration of albacore, Thunnus alalunga, in the

North Pacific Ocean in relation to large oceanic phenomena. Fish Oceanogr 6:51-57

Kirkwood R, Hume F, Hindell M (2008) Sea temperature variations mediate annual changes

in the diet of Australian fur seals in Bass Strait. Mar Ecol Prog Ser 369:297-309

Klaassen RHG, Hake M, Strandberg R, Koks BJ and others (2014) When and where does

mortality occur in migratory birds? Direct evidence from long-term satellite tracking

of raptors. J Anim Ecol 83:176-184

Klarian SA, Canales-Cerro C, Barría P, Zárate P and others (2018) New insights on the

trophic ecology of blue (Prionace glauca) and shortfin mako sharks (Isurus

oxyrinchus) from the oceanic eastern South Pacific. Mar Biol Res 14:173-182

Kokko H, Gunnarsson TG, Morrell LJ, Gill JA (2006) Why do female migratory birds arrive

later than males? J Anim Ecol 75:1293-1303

Krull CR, Ranjard L, Landers TJ, Ismar SMH, Matthews JL, Hauber ME (2012) Analyses of

sex and individual differences in vocalizations of Australasian gannets using a

dynamic time warping algorithm. Journal of the Acoustical Society of America

132:1189

Kubetzki U, Garthe S, Fifield D, Mendel B, Furness RW (2009) Individual migratory

schedules and wintering areas of northern gannets. Mar Ecol Prog Ser 391:257-265

Kunz TH, Whitaker JO, Wadanoli MD (1995) Dietary energetics of the insectivorous

Mexican free-tailed bat (Tadarida brasiliensis) during pregnancy and lactation.

Oecologia 101:407-415

Kurle CM, Sinclair EH, Edwards AE, Gudmundson CJ (2011) Temporal and spatial variation

in the δ15N and δ13C values of fish and squid from Alaskan waters. Mar Biol

158:2389-2404

199 Kurle CM, McWhorter JK (2017) Spatial and temporal variability within marine isoscapes:

implications for interpreting stable isotope data from marine systems. Mar Ecol Prog

Ser 568:31-45

Lavoie RA, Rail J-F, Lean DR (2012) Diet composition of seabirds from Corossol Island,

Canada, using direct dietary and stable isotope analyses. Waterbirds 35:402-419

Lê S, Josse J, Husson F (2008) FactoMineR: an R package for multivariate analysis. J Stat

Softw 25:1-18

Leal GR, Furness RW, McGill RAR, Santos RA, Bugoni L (2017) Feeding and foraging

ecology of Trindade petrels Pterodroma arminjoniana during the breeding period in

the South Atlantic Ocean. Mar Biol 164:211

Leat EHK, Bourgeon S, Magnusdottir E, Gabrielsen GW and others (2013) Influence of

wintering area on persistent organic pollutants in a breeding migratory seabird. Mar

Ecol Prog Ser 491:277-293

Lewis R (1981) Seasonal upwelling along the south-eastern coastline of South Australia. Mar

Freshwater Res 32:843-854

Lewis S, Sherratt T, Hamer K, Wanless S (2001) Evidence of intra-specific competition for

food in a pelagic seabird. Nature 412:816-819

Lewis S, Benvenuti S, Dall–Antonia L, Griffiths R and others (2002) Sex-specific foraging

behaviour in a monomorphic seabird. Proc R Soc Lond [Biol] 269:1687-1693

Lewis S, Phillips RA, Burthe SJ, Wanless S, Daunt F (2015) Contrasting responses of male

and female foraging effort to year-round wind conditions. J Anim Ecol 84:1490-1496

Lewis SUE, Schreiber EA, Daunt F, Schenk GA and others (2005) Sex-specific foraging

behaviour in tropical boobies: does size matter? Ibis 147:408-414

200 Liechti F, Bauer S, Dhanjal-Adams KL, Emmenegger T, Zehtindjiev P, Hahn S (2018)

Miniaturized multi-sensor loggers provide new insight into year-round flight

behaviour of small trans-Sahara avian migrants. Mov Ecol 6:19

Lima SL, Valone TJ, Caraco T (1985) Foraging-efficiency-predation-risk trade-off in the

grey squirrel. Anim Behav 33:155-165

López-López P (2016) Individual-based tracking systems in ornithology: welcome to the era

of big data. Ardeola 63:103-136

Lorrain A, Graham BS, Popp BN, Allain V and others (2015) Nitrogen isotopic baselines and

implications for estimating foraging habitat and trophic position of yellowfin tuna in

the Indian and Pacific Oceans. Deep Sea Res Part 2 Top Stud Oceanogr 113:188-198

Lough JM, Hobday AJ (2011) Observed climate change in Australian marine and freshwater

environments. Mar Freshwater Res 62:984-999

Lövei GL, Sunderland KD (1996) Ecology and behavior of ground beetles (Coleoptera:

Carabidae). Annu Rev Entomol 41:231-256

MacArthur LD, Hyndes GA (2007) Varying foraging strategies of Labridae in seagrass

habitats: Herbivory in temperate seagrass meadows? J Exp Mar Biol Ecol 340:247-

258

Macdonald AM, Wunsch C (1996) An estimate of global ocean circulation and heat fluxes.

Nature 382:436-439

Machovsky-Capuska G, Vaughn-Hirshorn R, Würsig B, Raubenheimer D (2013) Can

gannets define their diving profile prior to submergence? Notornis 60:255-257

Machovsky-Capuska G, Hauber M, Libby E, Amiot C, Raubenheimer D (2014a) The

contribution of private and public information in foraging by Australasian gannets.

Anim Cogn 19:849-858

201 Machovsky-Capuska GE, Hauber ME, Dassis M, Libby E and others (2014b) Foraging

behaviour and habitat use of chick-rearing Australasian gannets in New Zealand. J

Ornithol 155:379-387

Machovsky-Capuska GE, Senior AM, Benn EC, Tait AH and others (2016a) Sex-specific

macronutrient foraging strategies in a highly successful marine predator: the

Australasian gannet. Mar Biol 163:75

Machovsky-Capuska GE, Senior AM, Simpson SJ, Raubenheimer D (2016b) The

Multidimensional Nutritional Niche. Trends Ecol Evol 31:355-365

Machovsky-Capuska GE, Miller MGR, Silva FRO, Amiot C and others (2018) The

nutritional nexus: Linking niche, habitat variability and prey composition in a

generalist marine predator. J Anim Ecol 87:1286-1298

Machovsky Capuska GE, Vaughn RL, Würsig B, Katzir G, Raubenheimer D (2011) Dive

strategies and foraging effort in the Australasian gannet Morus serrator revealed by

underwater videography. Mar Ecol Prog Ser 442:255-261

Mackley EK, Phillips RA, Silk JR, Wakefield ED, Afanasyev V, Fox JW, Furness RW

(2010) Free as a bird? Activity patterns of albatrosses during the nonbreeding period.

Mar Ecol Prog Ser 406:291-303

MacNeil MA, Skomal G, B. , Fisk A, T. (2005) Stable isotopes from multiple tissues reveal

diet switching in sharks. Mar Ecol Prog Ser 302:199-206

Magozzi S, Yool A, Vander Zanden HB, Wunder MB, Trueman CN (2017) Using ocean

models to predict spatial and temporal variation in marine carbon isotopes. Ecosphere

8:e01763

Mancini PL, Bond AL, Hobson KA, Duarte LS, Bugoni L (2013) Foraging segregation in

tropical and polar seabirds: testing the intersexual competition hypothesis. J Exp Mar

Biol Ecol 449:186-193

202 Marchetti K, Price T (1989) Differences in the foraging of juvenile and adult birds: the

importance of developmental constraints. Biol Rev 64:51-70

Maret TJ, Collins JP (1997) Ecological origin of morphological diversity: a study of

alternative trophic phenotypes in larval salamanders. Evolution 51:898-905

Martin CH (2013) Strong assortative mating by diet, color, size, and morphology but limited

progress toward sympatric speciation in a classic example: Cameroon crater lake

cichlids. Evolution 67:2114-2123

Martínez del Rio C, Wolf N, Carleton SA, Gannes LZ (2009) Isotopic ecology ten years after

a call for more laboratory experiments. Biol Rev 84:91-111

Marwick B, Krishnamoorthy K (2018) cvequality: Tests for the equality of coefficients of

variation from multiple groups. R package, version 0.1.3

https://githubcom/benmarwick/cvequality

Masello JF, Wikelski M, Voigt CC, Quillfeldt P (2013) Distribution patterns predict

individual specialization in the diet of dolphin gulls. PLoS ONE 8:e67714

Matich P, Heithaus MR, Layman CA (2011) Contrasting patterns of individual specialization

and trophic coupling in two marine apex predators. J Anim Ecol 80:294-305

Mattern T, Masello JF, Ellenberg U, Quillfeldt P (2015) Actave.net – a web-based tool for

the analysis of seabird activity patterns from saltwater immersion geolocators.

Methods Ecol Evol 6:859-864

McConnell B, Chambers C, Fedak M (1992) Foraging ecology of southern elephant seals in

relation to the bathymetry and productivity of the Southern Ocean. Antarct Sci 4:393-

398

McCutcheon C, Dann P, Salton M, Renwick L, Hoskins AJ, Gormley AM, Arnould JPY

(2011) The foraging range of little penguins (Eudyptula minor) during winter. Emu

111:321-329

203 McGraw WS, Vick AE, Daegling DJ (2011) Sex and age differences in the diet and ingestive

behaviors of sooty mangabeys (Cercocebus atys) in the Tai Forest, Ivory Coast. Am J

Phys Anthropol 144:140-153

McKnight A, Irons DB, Allyn AJ, Sullivan KM, Suryan RM (2011) Winter dispersal and

activity patterns of post-breeding black-legged kittiwakes Rissa tridactyla from Prince

William Sound, Alaska. Mar Ecol Prog Ser 442:241-253

Meeus JH (1991) Astronomical algorithms. Willmann-Bell, Richmond, US

Merino G, Barange M, Blanchard JL, Harle J and others (2012) Can marine fisheries and

aquaculture meet fish demand from a growing human population in a changing

climate? Glob Environ Change 22:795-806

Meynier L, Stockin K, Bando M, Duignan P (2008) Stomach contents of common dolphin

(Deiphinus sp.) from New Zealand waters. N Z J Mar Freshw Res 42:257-268

Middleton JF, Bye JA (2007) A review of the shelf-slope circulation along Australia’s

southern shelves: Cape Leeuwin to Portland. Prog Oceanogr 75:1-41

Middleton MJ (1982) The Oriental Goby, Acanthogobius flavimanus (Temminck and

Schlegel), an introduced fish in the coastal waters of New South Wales, Australia. J

Fish Biol 21:513-523

Miller TW, Brodeur RD, Rau G, Omori K (2010) Prey dominance shapes trophic structure of

the northern California Current pelagic food web: evidence from stable isotopes and

diet analysis. Mar Ecol Prog Ser 420:15-26

Minagawa M, Wada E (1984) Stepwise enrichment of 15N along food chains: further

evidence and the relation between δ15N and animal age. Geochim Cosmochim Acta

48:1135-1140

Montague T, Cullen J (1988) The diet of the little penguin Eudyptula minor at Phillip Island,

Victoria. Emu 88:138-149

204 Moseley C, Grémillet D, Connan M, Ryan PG and others (2012) Foraging ecology and

ecophysiology of Cape gannets from colonies in contrasting feeding environments. J

Exp Mar Biol Ecol 422–423:29-38

Müller MS, Massa B, Phillips RA, Dell’omo G (2014) Individual consistency and sex

differences in migration strategies of Scopoli’s shearwaters Calonectris diomedea

despite year differences. Curr Zool 60:631-641

Myers JP (1981) A test of three hypotheses for latitudinal segregation of the sexes in

wintering birds. Can J Zool 59:1527-1534

Nakagawa S, Schielzeth H (2010) Repeatability for Gaussian and non‐Gaussian data: a

practical guide for biologists. Biol Rev 85:935-956

Navarro J, Oro D, Bertolero A, Genovart M, Delgado A, Forero MG (2010) Age and sexual

differences in the exploitation of two anthropogenic food resources for an

opportunistic seabird. Mar Biol 157:2453-2459

Naves LC, Cam E, Monnat JY (2007) Pair duration, breeding success and divorce in a long-

lived seabird: benefits of mate familiarity? Anim Behav 73:433-444

Neira FJ, Sporcic MI, Longmore AR (1999) Biology and fishery of pilchard, Sardinops sagax

(Clupeidae), within a large south-eastern Australian bay. Mar Freshwater Res 50:43-

Neira FJ, Lyle JM, Keane JP (2009) Shelf spawning habitat of Emmelichthys nitidus in south-

eastern Australia – Implications and suitability for egg-based biomass estimation.

Estuar Coast Shelf Sci 81:521-532

Nelson B (1978) The Sulidae: gannets and boobies. Oxford University Press, Oxford, UK

Newsome SD, Martinez del Rio C, Bearhop S, Phillips DL (2007) A niche for isotopic

ecology. Front Ecol Environ 5:429-436

205 Newsome SD, Tinker MT, Monson DH, Oftedal OT and others (2009) Using stable isotopes

to investigate individual diet specialization in California sea otters (Enhydra lutris

nereis). Ecology 90:961-974

Newsome SD, Tinker MT, Gill VA, Hoyt ZN, Doroff A, Nichol L, Bodkin JL (2015) The

interaction of intraspecific competition and habitat on individual diet specialization: a

near range-wide examination of sea otters. Oecologia 178:45-59

Norman F (1991) Distribution and abundance of seabirds off Phillip Island and within Port

Phillip Bay, Victoria, 1986-1988. Emu 91:377-394

Norman F, Menkhorst P (1995) Aspects of the breeding and feeding ecology of the

Australasian gannet Morus serrator in Port Phillip Bay, Victoria, 1988-92. Emu

95:23-40

Norman FI, Minton CDT, Bunce A, Govanstone AP (1998) Short communication: recent

changes in the status of Australasian gannets Morus serrator in Victoria. Emu 98:147-

150

Novak M, Tinker MT (2015) Timescales alter the inferred strength and temporal consistency

of intraspecific diet specialization. Oecologia 178:61-74

O'Sullivan D, Cullen J (1983) Food of the squid (Nototodarus gouldi) in Bass Strait. Mar

Freshwater Res 34:261-285

Orben RA, Paredes R, Roby DD, Irons DB, Shaffer SA (2015) Wintering North Pacific

black-legged kittiwakes balance spatial flexibility and consistency. Mov Ecol 3:36

Orians GH (1969) Age and hunting success in the brown pelican (Pelecanus occidentalis).

Anim Behav 17:316-319

Page RE, Waddington KD, Hunt GJ, Kim Fondrk M (1995) Genetic determinants of honey

bee foraging behaviour. Anim Behav 50:1617-1625

206 Paiva VH, Geraldes P, Ramirez I, Meirinho A, Garthe S, Ramos JA (2010) Foraging

plasticity in a pelagic seabird species along a marine productivity gradient. Mar Ecol

Prog Ser 398:259-274

Paiva VH, Geraldes P, Marques V, Rodríguez R, Garthe S, Ramos JA (2013) Effects of

environmental variability on different trophic levels of the North Atlantic food web.

Mar Ecol Prog Ser 477:15-28

Pankiw T, Tarpy DR, Page RE (2002) Genotype and rearing environment affect honeybee

perception and foraging behaviour. Anim Behav 64:663-672

Paradis E, Claude J, Strimmer K (2004) APE: analyses of phylogenetics and evolution in R

language. Bioinformatics 20:289-290

Parnell AC, Phillips DL, Bearhop S, Semmens BX and others (2013) Bayesian stable isotope

mixing models. Environmetrics 24:387-399

Parry G, Hobday D, Currie D, Officer R, Gason A (1995) The distribution, abundance and

diets of demersal fish in Port Phillip Bay. CSIRO Port Phillip Bay Environmental

Study, Technical Report 21:1-119

Patrick SC, Bearhop S, Gremillet D, Lescroel A and others (2014) Individual differences in

searching behaviour and spatial foraging consistency in a central place marine

predator. Oikos 123:33-40

Patrick SC, Weimerskirch H (2014) Consistency pays: sex differences and fitness

consequences of behavioural specialization in a wide-ranging seabird. Biol Lett

10:20140630

Patrick SC, Bearhop S, Bodey TW, Grecian WJ, Hamer KC, Lee J, Votier SC (2015)

Individual seabirds show consistent foraging strategies in response to predictable

fisheries discards. J Avian Biol 46:431-440

207 Patterson SA, Morris-Pocock JA, Friesen VL (2011) A multilocus phylogeny of the Sulidae

(Aves: Pelecaniformes). Mol Phylogenet Evol 58:181-191

Peat J, Goulson D (2005) Effects of experience and weather on foraging rate and pollen

versus nectar collection in the bumblebee, Bombus terrestris. Behav Ecol Sociobiol

58:152-156

Pecl GT, Hobday AJ, Frusher S, Sauer WHH, Bates AE (2014) Ocean warming hotspots

provide early warning laboratories for climate change impacts. Rev Fish Biol Fish

24:409-413

Pérez C, Granadeiro JP, Dias MP, Alonso H, Catry P (2014) When males are more inclined

to stay at home: insights into the partial migration of a pelagic seabird provided by

geolocators and isotopes. Behav Ecol 25:313-319

Péron C, Grémillet D (2013) Tracking through life stages: adult, immature and juvenile

autumn migration in a long-lived seabird. PLoS ONE 8:e72713

Perry G, Garland T (2002) Lizard home ranges revisited: effects of sex, body size, diet,

habitat, and phylogeny. Ecology 83:1870-1885

Pettex E, Lorentsen SH, Gremillet D, Gimenez O and others (2012) Multi-scale foraging

variability in Northern gannet (Morus bassanus) fuels potential foraging plasticity.

Mar Biol 159:2743-2756

Phillips DL, Koch PL (2002) Incorporating concentration dependence in stable isotope

mixing models. Oecologia 130:114-125

Phillips DL, Newsome SD, Gregg JW (2005a) Combining sources in stable isotope mixing

models: alternative methods. Oecologia 144:520-527

Phillips DL (2012) Converting isotope values to diet composition: the use of mixing models.

J Mammal 93:342-352

208 Phillips DL, Inger R, Bearhop S, Jackson AL and others (2014) Best practices for use of

stable isotope mixing models in food-web studies. Can J Zool 92:823-835

Phillips RA, Silk JRD, Croxall JP, Afanasyev V, Briggs DR (2004) Accuracy of geolocation

estimates for flying seabirds. Mar Ecol Prog Ser 266:265-272

Phillips RA, Silk JR, Croxall JP, Afanasyev V, Bennett VJ (2005b) Summer distribution and

migration of nonbreeding albatrosses: individual consistencies and implications for

conservation. Ecology 86:2386-2396

Phillips RA, Bearhop S, Mcgill RA, Dawson DA (2009) Stable isotopes reveal individual

variation in migration strategies and habitat preferences in a suite of seabirds during

the nonbreeding period. Oecologia 160:795-806

Phillips RA, Lewis S, González-Solís J, Daunt F (2017) Causes and consequences of

individual variability and specialization in foraging and migration strategies of

seabirds. Mar Ecol Prog Ser 578:117-150

Pinheiro J, Bates D, DebRoy S, Sarkar D, R Core Team (2014) nlme: Linear and Nonlinear

Mixed Effects Models. R package, version 3.1-131, https://CRAN.R-

project.org/package=nlme

Poloczanska ES, Brown CJ, Sydeman WJ, Kiessling W and others (2013) Global imprint of

climate change on marine life. Nat Clim Change 3:919

Porter JM, Sealy SG (1982) Dynamics of seabird multispecies feeding flocks: age-related

feeding behaviour. Behaviour 81:91-109

Post DM, Layman CA, Arrington DA, Takimoto G, Quattrochi J, Montaña CG (2007)

Getting to the fat of the matter: models, methods and assumptions for dealing with

lipids in stable isotope analyses. Oecologia 152:179-189

Price T (1987) Diet variation in a population of Darwin's finches. Ecology 68:1015-1028

209 Provencher JF, Elliott KH, Gaston AJ, Braune BM (2013) Networks of prey specialization in

an Arctic monomorphic seabird. J Avian Biol 44:551-560

Pyk T, Bunce A, Norman F (2007) The influence of age on reproductive success and diet in

Australasian gannets (Morus serrator) breeding at Pope's Eye, Port Phillip Bay,

Victoria. Aust J Zool 55:267-274

Pyk T, Weston M, Bunce A, Norman F (2013) Establishment and development of a seabird

colony: long-term trends in phenology, breeding success, recruitment, breeding

density and demography. J Ornithol 154:299-310

Pyke GH, Pulliam HR, Charnov EL (1977) Optimal Foraging: a selective review of theory

and tests. Q Rev Biol 52:137-154

Qasem L, Cardew A, Wilson A, Griffiths I and others (2012) Tri-axial dynamic acceleration

as a proxy for animal energy expenditure; should we be summing values or

calculating the vector? PLoS ONE 7:e31187

Quillfeldt P, McGill RA, Masello JF, Weiss F, Strange IJ, Brickle P, Furness RW (2008)

Stable isotope analysis reveals sexual and environmental variability and individual

consistency in foraging of thin-billed prions. Mar Ecol Prog Ser 373:137-148

Quillfeldt P, Ekschmitt K, Brickle P, McGill RAR, Wolters V, Dehnhard N, Masello JF

(2015) Variability of higher trophic level stable isotope data in space and time – a

case study in a marine ecosystem. Rapid Commun Mass Spectrom 29:667-674

R Core Team (2017) R: A language and environment for statistical computing. R Foundation

for Statistical Computing, Vienna, Austria. https://www.R-project.org

Radabaugh KR, Hollander DJ, Peebles EB (2013) Seasonal δ13C and δ15N isoscapes of fish

populations along a continental shelf trophic gradient. Cont Shelf Res 68:112-122

210 Ratcliffe N, Takahashi A, O’Sullivan C, Adlard S, Trathan PN, Harris MP, Wanless S (2013)

The roles of sex, mass and individual specialisation in partitioning foraging-depth

niches of a pursuit-diving predator. PLoS ONE 8:e79107

Rau GH, Sweeney RE, Kaplan IR (1982) Plankton 13C:12C ratio changes with latitude:

differences between northern and southern oceans. Deep-Sea Res A, Oceanogr Res

Pap 29:1035-1039

Raubenheimer D, Simpson SJ (1997) Integrative models of nutrient balancing: application to

insects and vertebrates. Nutr Res Rev 10:151-179

Reader SM (2015) Causes of individual differences in animal exploration and search. Top

Cogn Sci 7:451-468

Reid PJ, Shettleworth SJ (1992) Detection of cryptic prey: search image or search rate? J Exp

Psychol Anim Behav Process 18:273

Reid T, Hindell M, Eades D, Newman M (2002) Atlas of the Seabirds of Southeast Australia.

Birds Australia Monograph 4. Bird Australia, Melbourne, AU

Resano-Mayor J, Hernández-Matías A, Real J, Moleón M, Parés F, Inger R, Bearhop S

(2014) Multi-scale effects of nestling diet on breeding performance in a terrestrial top

predator inferred from stable isotope analysis. PLoS ONE 9:e95320

Richardson L, Mathews E, Heap A (2005) Geomorphology and sedimentology of the south

western planning area of Australia: review and synthesis of relevant literature in

support of Regional Marine Planning. Geoscience Australia, Record 2005/17. 124pp.

Ridgway KR (2007) Long-term trend and decadal variability of the southward penetration of

the East Australian Current. Geophys Res Lett 34:L13613

Robert D, Castonguay M, Fortier L (2008) Effects of intra- and inter-annual variability in

prey field on the feeding selectivity of larval Atlantic mackerel (Scomber scombrus). J

Plankton Res 30:673-688

211 Robertson A, Klumpp D (1983) Feeding habits of the southern Australian garfish

Hyporhamphus melanochir: a diurnal herbivore and nocturnal carnivore. Mar Ecol

Prog Ser 10:197-201

Robertson A, McDonald R, Delahay R, Kelly S, Bearhop S (2014) Individual foraging

specialisation in a social mammal: the European badger (Meles meles). Oecologia

176:409-421

Robertson HA, Baird K, Dowding JE, Elliott GP and others (2017) Conservation status of

New Zealand birds, 2016, Department of Conservation, Wellington

Ropert-Coudert Y, Grémillet D, Kato A, Ryan PG, Naito Y, Le Maho Y (2004) A fine-scale

time budget of Cape gannets provides insights into the foraging strategies of coastal

seabirds. Anim Behav 67:985-992

Ropert-Coudert Y, Daunt F, Kato A, Ryan PG and others (2009) Underwater wingbeats

extend depth and duration of plunge dives in northern gannets Morus bassanus. J

Avian Biol 40:380-387

Rosenblatt AE, Nifong JC, Heithaus MR, Mazzotti FJ and others (2015) Factors affecting

individual foraging specialization and temporal diet stability across the range of a

large “generalist” apex predator. Oecologia 178:5-16

Ruiz-Cooley RI, Markaida U, Gendron D, Aguíñiga S (2006) Stable isotopes in jumbo squid

(Dosidicus gigas) beaks to estimate its trophic position: comparison between stomach

contents and stable isotopes. J Mar Biol Assoc UK 86:437-445

Rumolo P, Basilone G, Fanelli E, Barra M and others (2017) Linking spatial distribution and

feeding behavior of Atlantic horse mackerel (Trachurus trachurus) in the Strait of

Sicily (Central Mediterranean Sea). J Sea Res 121:47-58

212 Sakamoto KQ, Sato K, Ishizuka M, Watanuki Y, Takahashi A, Daunt F, Wanless S (2009)

Can ethograms be automatically generated using body acceleration data from free-

ranging birds? PLoS ONE 4:e5379

Sakano H, Fujiwara E, Nohara S, Ueda H (2005) Estimation of nitrogen stable isotope

turnover rate of Oncorhynchus nerka. Environ Biol Fishes 72:13-18

Sampson J, Easton A, Singh M (2014) Port Phillip Bay. In: E. W (ed) Estuaries of Australia

in 2050 and beyond Estuaries of the World. Springer Science+Business Media

Dordrecht, p 49-68

Sánchez-Macouzet O, Rodríguez C, Drummond H (2014) Better stay together: pair bond

duration increases individual fitness independent of age-related variation. Proc R Soc

B-Biol Sci 281

Sandercock BK, Jaramillo A, Moore F (2002) Annual survival rates of wintering sparrows:

assessing demographic consequences of migration. Auk 119:149-165

Sandery PA, Kämpf J (2005) Winter-spring flushing of Bass Strait, south-eastern Australia: a

numerical modelling study. Estuar Coast Shelf Sci 63:23–31

Sandery PA, Kämpf J (2007) Transport timescales for identifying seasonal variation in Bass

Strait, south-eastern Australia. Estuar Coast Shelf Sci 74:684-696

Sargeant BL (2007) Individual foraging specialization: niche width versus niche overlap.

Oikos 116:1431-1437

Sawyer SL, Fogle SR (2013) Establishment of a new breeding colony of Australasian gannets

(Morus serrator) at Young Nick’s Head Peninsula. Notornis 60:180-182

Schahinger R (1987) Structure of coastal upwelling events observed off the south-east coast

of South Australia during February 1983 - April 1984. Mar Freshwater Res 38:439-

459

Schoener TW (1971) Theory of Feeding Strategies. Annu Rev Ecol Syst 2:369-404

213 Schreiber EA, Burger J (2001) Biology of marine birds. CRC Press, Boca Raton, US

Schuckard R, Melville D, Cook W, Machovsky-Capuska G (2012) Diet of the Australasian

gannet (Morus serrator) at Farewell Spit, New Zealand. Notornis 59:66-70

Schuett W, Dall SRX, Royle NJ (2011) Pairs of zebra finches with similar ‘personalities’

make better parents. Anim Behav 81:609-618

Scioscia G, Raya Rey A, Saenz Samaniego RA, Florentín O, Schiavini A (2014) Intra- and

interannual variation in the diet of the Magellanic penguin (Spheniscus magellanicus)

at Martillo Island, Beagle Channel. Polar Biol 37:1421-1433

Scott Forbes L, Jajam M, Kaiser GW (2000) Habitat constraints and spatial bias in seabird

colony distributions. Ecography 23:575-578

Sears J, Hatch SA, O’Brien DM (2009) Disentangling effects of growth and nutritional status

on seabird stable isotope ratios. Oecologia 159:41-48

Senior AM, Nakagawa S, Lihoreau M, Simpson SJ, Raubenheimer D (2015) An overlooked

consequence of dietary mixing: a varied diet reduces interindividual variance in

fitness. The American naturalist 186:649-659

Shaffer SA, Costa DP, Weimerskirch H (2003) Foraging effort in relation to the constraints

of reproduction in free-ranging albatrosses. Funct Ecol 17:66-74

Shealer DA (2001) Foraging behavior and food of seabirds. In: Schreiber E, Burger J (eds)

Biology of marine birds. CRC Press, p 137-177

Shoji A, Aris-Brosou S, Owen E, Bolton M and others (2016) Foraging flexibility and search

patterns are unlinked during breeding in a free-ranging seabird. Mar Biol 163:72

Sigman D, Casciotti K (2001) Nitrogen isotopes in the ocean. In: Steele JH, Turekian KK,

Thorpe SA (eds) Encyclopedia of ocean sciences. Academic, London, UK

214 Simpfendorfer CA, Goodreid AB, McAuley RB (2001) Size, sex and geographic variation in

the diet of the tiger shark, Galeocerdo cuvier, from Western Australian waters.

Environ Biol Fishes 61:37-46

Smith JL, C.P.H. Muler, and J.C. Ellis (2011) Seabirds as ecosystems engineers: nutrient

inputs and physical disturbance. In: Mulder CPH, Anderson, W.B., Towns, D.R. &

Bellingham, P.J (ed) Seabird Islands: Ecology, Invasion and Restoration, Vol Oxford

University Press, Oxford, UK, p 27-55

Snowberg Lisa K, Bolnick Daniel I (2008) Assortative mating by diet in a phenotypically

unimodal but ecologically variable population of stickleback. Am Nat 172:733-739

Soanes LM, Arnould JPY, Dodd SG, Sumner MD, Green JA (2013a) How many seabirds do

we need to track to define home-range area? J Appl Ecol 50:671-679

Soanes LM, Atkinson PW, Gauvain RD, Green JA (2013b) Individual consistency in the

foraging behaviour of Northern Gannets: implications for interactions with offshore

renewable energy developments. Mar Policy 38:507-514

Soanes LM, Arnould JPY, Dodd SG, Milligan G, Green JA (2014) Factors affecting the

foraging behaviour of the European shag: implications for seabird tracking studies.

Mar Biol 161:1335–1348

Spoon TR, Millam JR, Owings DH (2006) The importance of mate behavioural compatibility

in parenting and reproductive success by cockatiels, Nymphicus hollandicus. Anim

Behav 71:315-326

Srinivasan M, Dassis M, Benn E, Stockin KA, Martinez E, Machovsky-Capuska GE (2015)

Using non-systematic surveys to investigate effects of regional climate variability on

Australasian gannets in the Hauraki Gulf, New Zealand. J Sea Res 99:74-82

215 Staniland I, J. , Reid K, Boyd I, L. (2004) Comparing individual and spatial influences on

foraging behaviour in Antarctic fur seals Arctocephalus gazella. Mar Ecol Prog Ser

275:263-274

Stark KE (2008) Ecology of the Arrow Squid (Nototodarus gouldi) in southeastern Australian

Waters: a multi-scale investigation of spatial and temporal variability. PhD

dissertation, University of Tasmania

Stein PAS (1962) Winter sighting of gannets. Notornis 9:292-293

Stephens D, Krebs J (1986) Foraging theory. Princeton University Press, Princeton, US

Stephens DW (2008) Decision ecology: foraging and the ecology of animal decision making.

Cogn Affect Behav Neurosci 8:475-484

Stewart J, Hughes J, McAllister J, Lyle J, MacDonald C (2011) Australian salmon (Arripis

trutta): population structure, reproduction, diet and composition of commercial and

recreational catches, Cronulla Fisheries Research Centre of Excellence - Industry &

Investment NSW, Cronulla, AU

Stock BC, Semmens BX (2016a) Unifying error structures in commonly used biotracer

mixing models. Ecology 97:2562-2569

Stock BC, Semmens BX (2016b) MixSIAR GUI User Manual. R package, version 3.1.

https://github.com/brianstock/MixSIAR

Stock BC, Jackson AL, Ward EJ, Parnell AC, Phillips DL, Semmens BX (2018) Analyzing

mixing systems using a new generation of Bayesian tracer mixing models. PeerJ

6:e5096

Sugimoto T, Kimura S, Tadokoro K (2001) Impact of El Niño events and climate regime

shift on living resources in the western North Pacific. Prog Oceanogr 49:113-127

Sumner MD (2016) trip:Tools for the Analysis of Animal Track Data. R package, version

1.5.0. https://CRAN.R-project.org/package=trip

216 Svanbäck R, Bolnick DI (2007) Intraspecific competition drives increased resource use

diversity within a natural population. Proc R Soc B-Biol Sci 274:839-844

Sydeman WJ, Thompson SA, Kitaysky A (2012) Seabirds and climate change: roadmap for

the future. Mar Ecol Prog Ser 454:107-117

Symonds ME, Moussalli A (2011) A brief guide to model selection, multimodel inference

and model averaging in behavioural ecology using Akaike’s information criterion.

Behav Ecol Sociobiol 65:13-21

Tait A, Raubenheimer D, Stockin K, Merriman M, Machovsky-Capuska G (2014) Nutritional

geometry and macronutrient variation in the diets of gannets: the challenges in marine

field studies. Mar Biol:1-11

Taylor AR, Dann P, Arnould JPY (2013) Timing of breeding and diet of the Black-faced

Cormorant Phalacrocorax fuscescens. Mar Ornithol:23-27

Tieszen LL, Boutton TW, Tesdahl KG, Slade NA (1983) Fractionation and turnover of stable

carbon isotopes in animal tissues: implications for δ13C analysis of diet. Oecologia

57:32-37

Toscano BJ, Gownaris NJ, Heerhartz SM, Monaco CJ (2016) Personality, foraging behavior

and specialization: integrating behavioral and food web ecology at the individual

level. Oecologia 182:55-69

Tranquilla LAM, Montevecchi WA, Fifield DA, Hedd A, Gaston AJ, Robertson GJ, Phillips

RA (2014) Individual winter movement strategies in two species of murre (Uria spp.)

in the northwest Atlantic. PLoS ONE 9:13

Tremblay Y, Cherel Y (2003) Geographic variation in the foraging behaviour, diet and chick

growth of rockhopper penguins. Mar Ecol Prog Ser 251:279-297

217 Tremblay Y, Thiebault A, Mullers R, Pistorius P (2014) Bird-borne video-cameras show that

seabird movement patterns relate to previously unrevealed proximate environment,

not prey. PLoS ONE 9:e88424

Trueman CN, MacKenzie KM, St John Glew K (2017) Stable isotope-based location in a

shelf sea setting: accuracy and precision are comparable to light-based location

methods. Methods Ecol Evol 8:232-240

Vanderklift MA, Ponsard S (2003) Sources of variation in consumer-diet δ15N enrichment: a

meta-analysis. Oecologia 136:169-182

Votier SC, Bearhop S, Ratcliffe N, Furness RW (2004) Reproductive consequences for great

skuas specializing as seabird predators. Condor 106:275-287

Wakefield ED, Cleasby IR, Bearhop S, Bodey TW and others (2015) Long-term individual

foraging site fidelity - why some gannets don't change their spots. Ecology 96:3058-

3074

Walker SJ (1999) Coupled hydrodynamic and transport models of Port Phillip Bay, a semi-

enclosed bay in south-eastern Australia. Mar Freshwater Res 50:469-481

Ward EJ, Semmens BX, Schindler DE (2010) Includingsource uncertainty and prior

information in the analysis of stable isotope mixing models. Environ Sci Technol

44:4645-4650

Ward T, Goldsworthy S, Rogers P, Page B, McLeay L, Dimmlich W, AMM E (2008)

Ecological importance of small pelagic fishes in the Flinders Current System. Report

to Department of the Environment, Water, Heritage and the Arts. South Australian

Research and Development Institute (Aquatic Sciences), Adelaide. SARDI

Publication No. F2007/001194-1

Ward TM, McLeay LJ, Dimmlich WF, Rogers PJ and others (2006) Pelagic ecology of a

northern boundary current system: effects of upwelling on the production and

218 distribution of sardine (Sardinops sagax), anchovy (Engraulis australis) and southern

bluefin tuna (Thunnus maccoyii) in the Great Australian Bight. Fish Oceanogr 15:191-

207

Ward TM, Earl J, Ivey A (2015) Commonwealth Small Pelagic Fishery: Fishery Assessment

Report 2014. Report to the Australian Fisheries Management Authority. SARDI

Publication No. F2010/000270-6. SARDI Research Report Series No. 847. South

Australian Research and Development Institute (Aquatic Sciences). Adelaide, AU.

113pp

Warwick-Evans V, Atkinson PW, Gauvain RD, Robinson LA, Arnould JPY, Green JA

(2015) Time-in-area represents foraging activity in a wide-ranging pelagic forager.

Mar Ecol Prog Ser 527:233-246

Webb B, Grant C (1979) Age and Growth of Jack Mackerel (Trachurus declivis), from

South-eastern Australian Waters. Mar Freshwater Res 30:1-9

Weidel BC, Carpenter SR, Kitchell JF, Vander Zanden MJ (2011) Rates and components of

carbon turnover in fish muscle: insights from bioenergetics models and a whole-lake

13C addition. Can J Fish Aquat Sci 68:387-399

Weimerskirch H (1995) Regulation of foraging trips and incubation routine in male and

female wandering albatrosses. Oecologia 102:37-43

Weimerskirch H, Gault A, Cherel Y (2005a) Prey distribution and patchiness: factors in

foraging success and efficiency of wandering albatrosses. Ecology 86:2611-2622

Weimerskirch H, Le Corre M, Ropert-Coudert Y, Kato A, Marsac F (2005b) The three-

dimensional flight of red-footed boobies: adaptations to foraging in a tropical

environment? Proc R Soc Lond [Biol] 272:53-61

219 Weimerskirch H, Le Corre M, Ropert-Coudert Y, Kato A, Marsac F (2006) Sex-specific

foraging behaviour in a seabird with reversed sexual dimorphism: the red-footed

booby. Oecologia 146:681-691

Weimerskirch H (2007) Are seabirds foraging for unpredictable resources? Deep Sea Res

Part 2 Top Stud Oceanogr 54:211-223

Wells MR, Angel LP, Arnould JPY (2016) Habitat-specific foraging strategies in

Australasian gannets. Biol Open: bio.018085

Wingham E (1982) Breeding biology of the Australasian gannet Morus serrator (Gray) at

Motu Karamarama, Hauraki Gulf, New Zealand. II. Breeding success and chick

growth. Emu 84:211-224

Wingham E (1984) Breeding biology of the Australasian gannet Morus serrator (Gray) at

Motu Karamarama, Hauraki Gulf, New Zealand. I. The egg. Emu 84:129-136

Wolf M, van Doorn GS, Weissing FJ (2008) Evolutionary emergence of responsive and

unresponsive personalities. Proc Natl Acad Sci USA 105:15825-15830

Wolf M, Weissing FJ (2012) Animal personalities: consequences for ecology and evolution.

Trends Ecol Evol 27:452-461

Woo KJ, Elliott KH, Davidson M, Gaston AJ, Davoren GK (2008) Individual specialization

in diet by a generalist marine predator reflects specialization in foraging behaviour. J

Anim Ecol 77:1082-1091

Yamamoto T, Takahashi A, Oka N, Iida T, Katsumata N, Sato K, Trathan PN (2011)

Foraging areas of streaked shearwaters in relation to seasonal changes in the marine

environment of the Northwestern Pacific: inter-colony and sex-related differences.

Mar Ecol Prog Ser 424:191-204

Yamamoto T, Takahashi A, Sato K, Oka N, Yamamoto M, Trathan PN (2014) Individual

consistency in migratory behaviour of a pelagic seabird. Behaviour 151:683-701

220 Young JW, Jordan AR, Bobbi C, Johannes RE, Haskard K, Pullen G (1993) Seasonal and

interannual variability in krill (Nyctiphanes australis) stocks and their relationship to

the fishery for jack mackerel (Trachurus declivis) off eastern Tasmania, Australia.

Mar Biol 116:9-18

Yurkowski DJ, Ferguson S, Choy ES, Loseto LL and others (2016) Latitudinal variation in

ecological opportunity and intraspecific competition indicates differences in niche

variability and diet specialization of Arctic marine predators. Ecol Evol 6:1666-1678

Zaccarelli N, Bolnick DI, Mancinelli G (2013) RInSp: an R package for the analysis of

individual specialization in resource use. Methods Ecol Evol 4:1018-1023

Zanden MJV, Rasmussen JB (2001) Variation in δ15N and δ13C trophic fractionation:

Implications for aquatic food web studies. Limnol Oceanogr 46:2061-2066

Zavalaga CB, Halls J, Dell'Omo G (2010) Marine habitat use of Peruvian boobies: a

geographic and oceanographic comparison between inshore and offshore islands.

ICES J Mar Sci 67:940-951

Zuur AF, Ieno EN, Elphick CS (2010) A protocol for data exploration to avoid common

statistical problems. Methods Ecol Evol 1:3-14

221

222