Otolith growth chronologies: Investigating historic ... · Analysis of the growth chronologies held within the otoliths of fishes offers a means to achieve this aim. To date, most

Otolith growth chronologies:

Investigating historic impacts of climate change

on tropical fish

by

Joyce Jia Lin Ong

BSc (Hons) University of Melbourne

This thesis is presented for the degree of

Doctor of Philosophy

of

The University of Western Australia

Faculty of Science

School of Animal Biology

January 2017

iii

Declaration of authorship

I declare that this thesis is my own composition, all sources have been acknowledged

and my contribution is clearly identified in the thesis. This thesis has been substantially

completed during the course of enrolment at the University of Western Australia and

has not previously been accepted for a degree at any tertiary education institution. I

confirm that for any work in this thesis that has been co-published with other authors, I

have the permission of all co-authors to include this work in my thesis, and include a

signed declaration.

Joyce Jia Lin Ong

Date: 09 January 2017

v

Above: Tropical coral reef at the Rowley Shoals, Western Australia, in April 2013.

Below: Lutjanus bohar at Rowley Shoals, Western Australia, in April 2013.

Photo credits: Joyce Ong

vii

Abstract

Climate change driven by the burning of fossil fuels is having negative impacts on

marine ecosystems at global scales. An understanding of how marine fishes respond to

this process is critical in order to implement management strategies that will ensure the

future sustainability and resilience of fisheries on which billions of people depend on

for food. Because growth rates determine fisheries productivity and yield, identifying

the link between growth patterns and climate phenomena is a key element of this task.

Analysis of the growth chronologies held within the otoliths of fishes offers a means to

achieve this aim. To date, most chronologies have been developed for fishes from the

Northern Hemisphere and from temperate waters. However, fishes from the tropics

provide an important source of protein for almost one billion people, mostly in

developing countries. Moreover, many of the habitats in which these species live (e.g.

estuaries and coral reefs) are fragile and vulnerable to other anthropogenic impacts

including pollution and eutrophication. Hence, there is a pressing need to determine

how tropical fishes respond to climate change and growth chronologies hold promise in

this area. In this thesis, I focused on the task of developing and analysing growth

chronologies from the otoliths of tropical fishes in order to detect links between patterns

of growth and climate from local to large spatial scales.

In Chapter 2, I used dendrochronology techniques to develop a growth chronology for

both the adult and juvenile life history stages of the tropical snapper Lutjanus

argentimaculatus in Western Australia. This analysis compared the response to climate

change of two life history stages of a species with ontogenetic habitat shifts. I found a

strong correlation between the growth of adults and a measure of the El Niño-Southern

Oscillation (ENSO), whereas juvenile growth was correlated with rainfall. These

divergent patterns suggest that there is a within-region and within-species ontogenetic

component to climate change responses. In order to determine how the effects of

climate change might vary across a species’ range, in Chapter 3 I compared the drivers

of growth of the adults of two widely-separated populations of another common

snapper, Lutjanus bohar. Although large-scale climate indices were again found to be

important for both populations at similar latitudes, different aspects of this signal

influenced the growth of adults in their respective regions (north-west Australia and the

Great Barrier Reef). This implies that sampling must encompass a broad portion of a

species range if analyses using biochronologies are to identify key variables underlying

viii

responses to climate change. In Chapter 4, I then examined the influence of large-scale

climate phenomenon on taxa across both marine and terrestrial ecosystems. I did this by

combining and comparing growth chronologies developed from fishes, corals and trees

on the tropical coast of Western Australia. These analyses showed that the growth

patterns of these diverse groups responded in a synchronous manner to the ENSO

signal, implying that climate change alterations to ENSO will simultaneously affect

species in terrestrial and marine environments. Finally, I investigated the extent to

which growth patterns of both tropical and temperate fishes were driven by large-scale

climate signals, by comparing the otolith growth chronologies of six species that ranged

along 2300 km of the coastline of Western Australia. This analysis showed a

simultaneous and extensive effect of the Leeuwin Current and ENSO signal on the

growth of five of the six species. My study also identified a previously unsuspected

correlation of the growth of one species with the Pacific Decadal Oscillation and for the

remaining five species a relationship with the Leeuwin Current, which is strongly

associated with the western Pacific warm pool.

Combined, these analyses demonstrated that the ENSO signal had a far-reaching

influence on growth patterns of marine fishes across most of the coastline of Western

Australia, and also on benthic organisms and trees in this region. Thus, climate change

that alters the strength or frequency of El Niño and La Niña events, which are tightly

linked to the flow of the Leeuwin Current, is likely to have simultaneous and immediate

impacts on the growth, and potentially resilience of populations of fishes on this coast.

ix

Acknowledgements

To my supervisors, I have gotten this far because of your help and your foresight in

dreaming up this project in the first place. Jessica, thanks for your guidance, your

support and for your constant reassurances. In particular, thanks for honing my

presentation skills to the point that I can be confident and can even say that I enjoy

scientific communication. Adam, you are an incredible R genius and stats guru. Thanks

for teaching me (the hard way) how to make R do what I want, and for your continued

advice even after moving to the other side of the world. Pauline, thanks for being so

encouraging and for sharing your knowledge about trees and the climate. Mark, thanks

for the long hours spent sitting down with me and going through the manuscripts, for

teaching me to always define scales, and for all the grammar lessons (e.g. always write

in an active voice, never string too many adjectives consecutively, and for the relentless

corrections of ‘fish growth’ and ‘current strength’).

A group of fellow PhD mates have been my support throughout the 3.5 years. It was

heartening to know that we were on the same journey and the numerous coffee and

lunch breaks helped tremendously on this otherwise potentially lonely, journey. Thanks

Napo, Bev, Lu, Mike (in order of graduation), for the fun, the laughter and for sharing

my joy and frustrations. Thanks Napo for making me watch two movies back-to-back at

the cinema. Thanks Lu for showing me how easy and satisfying it is to make sushi and

for the accompaniment at the office on the weekends. Thanks Mike for letting me try to

beat you in badminton and for entertaining conversations. I am so grateful that I got to

travel through this journey with all of you. To the rest of the AIMS staff (especially

Mary, James, Becky, Michele, Olwyn, Louise and the other PhD students Pipi and

Milly) in the office, thank you for your friendships and for the wonderful morning tea

sessions and multiple cake celebrations, which kept my sugar levels up and ensured that

there would always be leftovers that I could prey on.

To my non blood-related family in Perth, I am incredibly blessed to have met you all.

All of you have supported me so warmly and lovingly, especially in the last few

months. Thank you Kel, Grace, Steve, Shuyan, Claude, Ruth, Jeremy, Rebecca, Ryan,

Bev, Raymond, Mel, Sulin, Chris, Kristi, Christie, Glenna and Euphina for your

fellowship, your advice, your prayers, your concern, your generosity, your help, your

love and your craziness. All my weekends have been bucket-loads of fun, food and

‘family’ bonding time. Special thanks to Bev for being an amazing friend and

x

housemate. It is only because of you that there has not been a dull moment in my life

over the past 3.5 years. Thanks for forcing me to be a neater and cleaner person than I

would ever choose to be, for the countless dinners you have cooked for me, and most

importantly, for deciding to follow your sharky dream so that I can pursue my fishy

dream.

My Tuesday nights were a joy with a bunch of friends from Alpha that I have known

since our days in Melbourne as young undergraduates. It is no coincidence that we

ended up in Perth and serving again in Alpha together. Thanks Hui, Nat, Fook Joo and

Alex for all your prayers and the wonderful fellowship moments. The miracles we have

seen in Alpha will never cease to amaze me.

To my close friends in Singapore, I thank you for the support and love that can be felt

across continents. Debbie and Hui hui, you two inspire me and I thank you both that in

the hard times, I could count on you two for prayers, godly advice and to lift my spirits.

To Mel, Han, Fiona, Sherry, Serene, Jemma, it has been an incredible 20+ years of

friendship and I look forward to the many more years of hilarious banter. To my ex-

colleagues from TMSI, thanks for being my close friends even though I left so long ago.

Thanks Michelle for being my ‘google’ and giving up your time to hear my

postgraduate rants. Iris, you always know how to make me laugh, thanks for the

constant supply of ‘lol’ moments. Jia Hui, thanks for the countless Margaret River

experiences on your annual trips to Perth.

I could not have done this without the support of my family in Singapore. Thank you

Mummy, Daddy and Jess for loving and supporting me in everything that I do. I have

missed you all a lot. Thank you for the financial support, timely encouragement, sound

advice and regular updates of Koby. Mummy and Daddy, thank you for the sacrifices

that you have made for me to be where I am today. I love you all so much and I am very

grateful and blessed to have you as my family.

Most importantly, I give thanks and praise to God, because He is the reason I have

successfully finished my PhD. God has been my source of comfort, encouragement, my

shelter and my inspiration. I have made it this far because I know that God is with me

and that ‘I can do all things through Christ who strengthens me’. Thank you God for the

relationship I have with you and for being my pillar of strength and my refuge.

xi

Publications from this thesis

Chapter 2:

Ong JJL, Rountrey AN, Meeuwig JJ, Newman SJ, Zinke J, Meekan MG (2015)

Contrasting environmental drivers of adult and juvenile growth in a marine fish:

implications for the effects of climate change. Scientific Reports 5:10859

Chapter 3:

Ong JJL, Rountrey AN, Marriott RJ, Newman SJ, Meeuwig JJ, Meekan MG (2016)

Cross-continent comparisons reveal differing environmental drivers of growth of the

coral reef fish, Lutjanus bohar. Coral Reefs, doi: 10.1007/s00338-016-1520-2.

Chapter 4:

Ong JJL, Rountrey AN, Zinke J, Meeuwig JJ, Grierson PF, O’Donnell AJ, Newman SJ,

Lough JM, Trougan M, Meekan MG (2016) Evidence for climate-driven synchrony of

marine and terrestrial ecosystems in northwest Australia. Global Change Biology 22:

2776-2786

Chapter 5:

Ong JJL, Rountrey AN, Coulson PG, Nguyen HM, Wakefield CB, Meeuwig JJ,

Newman SJ, Meekan MG (2016) A boundary current drives growth of marine fishes

across tropical and temperate latitudes. Nature Communications, in preparation.

xiii

Statement of candidate contributions

This thesis is presented as a series of four manuscripts in journal format, as well as the

general introduction and general discussion sections.

The samples were obtained from archived otolith collections. Samples from north-

western Australia were provided by Dr Stephen Newman from the Department of

Fisheries (Government of Western Australian). Samples from north-eastern Australia

were provided by Professor Colin Simpfendorfer from James Cook University.

All samples for Chapter 2 and 3 were imaged and processed by myself at the Centre for

Microscopy, Characterisation and Analyses, located in the University of Western

Australia. Environmental data were sourced from open-access repositories, with specific

details available in each chapter.

Existing data for Chapter 4 were obtained from Dr Janice Lough (Australian Institute of

Marine Science), Dr Alison O’Donnell (University of Western Australia) and Mélissa

Trougan (Natural Marine Park of Mayotte). Existing data for Chapter 5 were provided

by Dr Adam Rountrey (University of Michigan), Dr Peter Coulson (Murdoch

University), Dr Corey Wakefield (Department of Fisheries) and Hoang Minh Nguyen

(University of Texas).

The analyses described in all data chapters were carried out by myself and all chapters

were written by me with feedback from Professor Jessica Meeuwig (University of

Western Australia, all chapters), Dr Mark Meekan (Australian Institute of Marine

Science, all chapters), Dr Adam Rountrey (University of Michigan, Chapters 2–5), Dr

Stephen Newman (Department of Fisheries, Chapters 2–5), Professor Pauline Grierson

(University of Western Australia, Chapters 1, 4 and 6), Dr Jens Zinke (Freie Universität

Berlin, Chapters 2 and 4), Dr Alison O’Donnell (University of Western Australia,

Chapter 4), Dr Janice Lough (Australian Institute of Marine Science, Chapter 4), Dr

Ross Marriott (University of Western Australia, Chapter 3), Dr Peter Coulson (Murdoch

University, Chapter 5) and Dr Corey Wakefield (Department of Fisheries, Chapter 5).

xv

Co-author authorisation

By signing below, co-authors agree to the listed publication being included in the

candidate’s thesis and acknowledge that the candidate is the primary author, i.e.

contributed greater than 50% of the content and was primarily responsible for the

planning, execution and preparation of the work for publication.

Publication title: Contrasting environmental drivers of adult and juvenile growth in a

marine fish: implications for the effects of climate change

Co-authors:

Dr Adam Rountrey

(University of Michigan)

Professor Jessica Meeuwig

(University of Western Australia)

Dr Stephen Newman

(Department of Fisheries)

Dr Jens Zinke

(Freie Universität Berlin)

Dr Mark Meekan

(Australian Institute of Marine Science)

Publication title: Evidence for climate-driven synchrony of marine and terrestrial

ecosystems in northwest Australia

Co-authors:

Dr Adam Rountrey




xvi

Dr Stephen Newman


Dr Jens Zinke

(Freie Universität Berlin)

Dr Mark Meekan


Professor Pauline Grierson


Dr Alison O’Donnell


Dr Janice Lough


Mélissa Trougan

(Natural Marine Park of Mayotte)

Publication title: Cross-continent comparisons reveal differing environmental drivers of

growth of the coral reef fish, Lutjanus bohar

Co-authors:

Dr Adam Rountrey




Dr Stephen Newman


Dr Ross Marriott


Dr Mark Meekan


xvii

Publication title: A boundary current drives growth of marine fishes across tropical and

temperate latitudes

Co-authors:

Dr Adam Rountrey




Dr Stephen Newman


Dr Peter Coulson

Dr Corey Wakefield


Hoang Minh Nguyen

(University of Texas)

Dr Mark Meekan


xix

Table of Contents

Declaration of authorship ............................................................................................................ iii

Abstract ...................................................................................................................................... vii

Acknowledgements ..................................................................................................................... ix

Publications from this thesis ....................................................................................................... xi

Statement of candidate contributions ........................................................................................ xiii

Co-author authorisation ..............................................................................................................xv

Table of Contents ...................................................................................................................... xix

List of Figures ......................................................................................................................... xxiii

List of Tables ............................................................................................................................xxv

Chapter 1 General introduction ..............................................................................................1

1.1 Effects of climate change on marine ecosystems ..................................................2

1.2 Coral reef fish and climate change.........................................................................3

1.3 Importance of fish and fisheries .............................................................................4

1.4 Otoliths as recorders of growth ..............................................................................4

1.5 Emergence of biochronology techniques ...............................................................6

1.6 Tropical lutjanids as model species .......................................................................7

1.7 Aims .......................................................................................................................8

1.8 Thesis outline .........................................................................................................9

Chapter 2 Contrasting environmental drivers of adult and juvenile growth ........................13

2.1 Abstract ................................................................................................................13

2.2 Introduction ..........................................................................................................13

2.3 Methods ...............................................................................................................16

2.3.1 Study species .......................................................................................................16

2.3.2 Study site .............................................................................................................16

2.3.3 Sample collection .................................................................................................17

2.3.4 Image analysis......................................................................................................17

2.3.5 Otolith chronology development .........................................................................19

2.3.6 Correlations with environmental parameters .......................................................22

2.4 Results ..................................................................................................................22

2.4.1 Chronology development for adults and juveniles ..............................................22

2.4.2 Relationships between chronologies and environmental drivers .........................25

2.5 Discussion ............................................................................................................30

2.6 Acknowledgements ..............................................................................................33

xx

Chapter 3 Cross-continent comparisons reveal differing drivers of growth ........................ 35

3.1 Abstract ............................................................................................................... 35

3.2 Introduction ......................................................................................................... 35

3.3 Methods ............................................................................................................... 37

3.3.1 Study sites ........................................................................................................... 37

3.3.2 Study species ....................................................................................................... 38

3.3.3 Sampling methods ............................................................................................... 38

3.3.4 Microscopy methods ........................................................................................... 40

3.3.5 Growth increment data quality ............................................................................ 42

3.3.6 Environmental and size structure datasets ........................................................... 42

3.3.7 Statistical analyses ............................................................................................... 44

3.3.7.1 Regional differences in growth increments ................................................. 44

3.3.7.2 Regional drivers of growth increments ........................................................ 44

3.4 Results ................................................................................................................. 46

3.4.1 Growth chronology statistics ............................................................................... 46

3.4.2 Growth chronologies on NW and NE coasts ....................................................... 48

3.4.3 Factors influencing growth along NW coast ....................................................... 49

3.4.4 Factors influencing growth along NE coast ........................................................ 50

3.5 Discussion ........................................................................................................... 51

3.6 Acknowledgements ............................................................................................. 55

Chapter 4 Climate-driven synchrony of marine and terrestrial ecosystems ......................... 57

4.1 Abstract ............................................................................................................... 57

4.2 Introduction ......................................................................................................... 57

4.3 Methods ............................................................................................................... 60

4.3.1 Environmental drivers of marine and terrestrial regions ..................................... 60

4.3.2 Growth chronologies ........................................................................................... 62

4.3.2.1 Growth chronology of Lethrinus nebulosus ................................................ 63

4.3.2.2 Growth chronology of Lutjanus argentimaculatus ...................................... 64

4.3.2.3 Coral growth chronology ............................................................................. 65

4.3.2.4 Tree-ring chronology ................................................................................... 65

4.3.3 Climatic and environmental datasets ................................................................... 67

4.3.4 Data analyses ....................................................................................................... 67

4.4 Results ................................................................................................................. 68

4.4.1 Chronology statistics ........................................................................................... 68

4.4.2 Principal components analysis ............................................................................ 69

4.4.3 Relationships with ENSO .................................................................................... 71

4.4.4 Relationships with environmental variables ........................................................ 73

4.5 Discussion ........................................................................................................... 76

4.6 Acknowledgements ............................................................................................. 80

xxi

Chapter 5 A boundary current drives growth across tropical and temperate latitudes .........81

5.1 Abstract ................................................................................................................81

5.2 Introduction ..........................................................................................................81

5.3 Methods ...............................................................................................................83

5.3.1 Marine environment in Western Australia ...........................................................83

5.3.2 Otolith growth chronologies ................................................................................85

5.3.3 Environmental datasets ........................................................................................86

5.3.4 Statistical analyses ...............................................................................................87

5.4 Results ..................................................................................................................88

5.4.1 Chronology statistics............................................................................................88

5.4.2 Principal components analysis .............................................................................89

5.4.3 Relationships with environmental variables ........................................................90

5.5 Discussion ............................................................................................................95

5.6 Acknowledgements ..............................................................................................98

Chapter 6 General discussion ...............................................................................................99

6.1 Addressing the tropical knowledge gap .............................................................100

6.2 The varying effects of climate change ...............................................................101

6.2.1 Differing impacts at an intra-specific level ........................................................101

6.2.2 Synchronous climate impacts across ecosystems ..............................................102

6.2.3 Climate impacts on a continental scale ..............................................................103

6.3 Implications of varying climate impacts ............................................................104

6.4 Embracing new statistical methods ....................................................................105

6.5 Future directions ................................................................................................106

6.6 Conclusion .........................................................................................................108

References .................................................................................................................................109

xxiii

List of Figures

Figure 1.1 Schematic representation of the four data chapters and the relevant hypotheses

in those chapters...................................................................................................11

Figure 1.2 Schematic representation of the spatial scales of the thesis data chapters. ................11

Figure 2.1 Map of sampling sites of Lutjanus argentimaculatus in north-west Australia. .........19

Figure 2.2 Photomicrographs of the dorsal side of a L. argentimaculatus otolith section. .........21

Figure 2.3 Raw increment width time series, detrended and final chronologies used from

the otoliths of L. argentimaculatus. .....................................................................24

Figure 2.4 Pearson's correlation coefficients between L. argentimaculatus chronologies

and January to March values of significantly correlated environmental

variables. ..............................................................................................................27

Figure 2.5 Relationships between L. argentimaculatus chronologies and environmental

variables included in linear models......................................................................28

Figure 2.6 Map of spatial correlations of adult L. argentimaculatus chronology with

ocean heat content. ...............................................................................................29

Figure 2.7 Schematic diagram showing the environmental drivers that strongly influence

the juvenile and adult life history stages of L. argentimaculatus. .......................29

Figure 3.1 Sampling locations of Lutjanus bohar collected from tropical northern

Australia. ..............................................................................................................39

Figure 3.2 Photograph of the dorsal side of a L. bohar otolith section. ......................................41

Figure 3.3 Quality of L. bohar chronologies from tropical northern Australia. .........................47

Figure 3.4 Increment growth profiles of L. bohar from two regions in tropical northern

Australia. ..............................................................................................................48

Figure 3.5 Schematic diagram showing the environmental drivers that influence L. bohar

populations in northern Australia.........................................................................50

Figure 4.1 Sampling locations of growth chronologies for four taxa in north-west

Australia. ..............................................................................................................61

Figure 4.2 Growth increments of three diverse taxa from north-west Australia. ........................63

Figure 4.3 Raw and detrended increment width time-series from the otoliths of Lethrinus

nebulosus collected in north-west Australia. .......................................................66

Figure 4.4 Quality of chronology for coral Porites spp. collected in north-west Australia. .......69

Figure 4.5 Growth chronologies of four taxa with the respective leading principal

component (PC) scores. .......................................................................................71

Figure 4.6 Relationship between principal component (PC) scores and the Niño-4 index.........72

Figure 4.7 Spatial correlation maps between principal components (PC) scores and

environmental variables. ......................................................................................75

Figure 4.8 Schematic diagram showing the environmental drivers that influenced the

growth patterns of four taxa (two fishes, one coral and one tree) in north-

west Australia. .....................................................................................................76

Figure 5.1 General sampling locations of the six species of adult fishes collected along

the coast of Western Australia. ............................................................................84

Figure 5.2 Otolith growth chronologies of six marine fishes from Western Australia with

the respective leading principal component (PC) scores. ....................................89

xxiv

Figure 5.3 Significant correlations (p < 0.05) between the inverse of the first principal

component (PC1inv) and environmental variables averaged between January

and March. ........................................................................................................... 93


component (PC1inv) and environmental variables averaged quarterly

between April and December. ............................................................................. 94

Figure 5.5 Schematic diagram showing the influence of the Leeuwin Current (modulated

by ENSO) on the growth patterns of five species of fishes along the coast

of Western Australia. ........................................................................................... 95

Figure 6.1 Schematic diagram showing the multi-level effects of climate change found in

the four data chapters. ....................................................................................... 100

xxv

List of Tables

Table 2.1 Correlation coefficients and p-values of adult and juvenile Lutjanus

argentimaculatus with all 15 environmental variables. .......................................26

Table 3.1 Descriptive statistics for Lutjanus bohar collected from tropical northern

Australia. ..............................................................................................................40

Table 3.2 Explanatory variables used in the analyses of L. bohar from both the north-west

(NW) and north-east (NE) coast of Australia. .....................................................45

Table 3.3 Crossdating statistics for chronology development of L. bohar from tropical

northern Australia. ...............................................................................................46

Table 3.4 Models that involved the age-related growth trends of L. bohar from the two

regions. .................................................................................................................48

Table 3.5 Top five ranked models in two separate model selection processes for L. bohar

from tropical northern Australia. .........................................................................49

Table 3.6 Model parameters for the top-ranked model for L. bohar from north-west

Australia. ..............................................................................................................49

Table 3.7 Model parameters for the top-ranked model for L. bohar from north-east

Australia. ..............................................................................................................50

Table 4.1 Growth chronologies of fishes, coral and trees from north-west Australia. ...............62

Table 4.2 Loadings of the four taxa on the first and second principal component (PC)

scores. ..................................................................................................................70

Table 4.3 Pearson’s correlation matrix of environmental variables............................................70

Table 4.4 Models incorporated in the model selection process. .................................................74

Table 4.5 Selected first-ranked linear models. ............................................................................74

Table 5.1 Summary of the otolith growth chronologies of fishes from Western Australia. .......86

Table 5.2 Results of bootstrapped 𝒓 and expressed population signal (EPS) for the

detrended and standardized growth chronologies of six marine fishes from

Western Australia. ...............................................................................................88

Table 5.3 Loadings of the otolith biochronologies of six marine fishes on the first, second

and third principal component (PC) scores. .........................................................90

Table 5.4 Pearson’s correlation matrix of the four environmental variables (annual

means) over the years 1988 to 2003. ...................................................................91

Table 5.5 List of models in the model selection process. ...........................................................92

Table 5.6 First-ranked linear models that explain variation in the first three principal

component (PC) scores from the otolith growth chronologies of six marine

fishes in Western Australia. .................................................................................92

Table 5.7 Pearson’s correlation test results of the individual chronologies (entire

synchronous period) with Fremantle sea level. ....................................................92

1

Chapter 1 General introduction

This thesis describes the relationship between climate and growth of tropical fishes

along the coast of Australia over decadal time scales. Annual growth of fishes was

inferred from growth chronologies that were constructed from otoliths. These

biochronologies were used as proxies to allow the variation in somatic growth of fishes

to be compared to variability in climate and oceanography at annual scales.

Anthropogenic climate change is having a clear and growing impact on our world with

effects of this process expressed across ecosystems of all continents and oceans. The

concentration of atmospheric greenhouse gases (particularly carbon dioxide) are now

more than double the level compared to those at any point in the last 800,000 years and

the rate of climate warming since the 1950s, both in the atmosphere and oceans, is

unparalleled over recent centuries and even millennia (IPCC 2014). In addition, sea

levels and ocean acidity are rising, with a concurrent loss of glaciers, polar ice sheets

and snow cover (IPCC 2014). How these elements of climate change will impact the

biology and ecology of oceans is thus a critical issue. However, most studies of the

impacts of climate change have focussed on terrestrial environments (Walther et al.

2002; Parmesan and Yohe 2003; Rosenzweig et al. 2008). Changes to the ecology and

functioning of marine ecosystems are less well understood, largely due to the difficulty

of obtaining data from oceans and the lack of long-term studies (Hoegh-Guldberg and

Bruno 2010; Richardson et al. 2012).

In this introductory chapter, I provide background information on the importance of

understanding the effects of climate change on marine fishes in the tropics and show

how otoliths are a useful tool for this task, given their ability to record inter-annual

patterns of fish growth. I then describe a new technique to validate inter-annual growth

chronologies derived from otoliths and outline the rationale of the choice of study

species and study sites. This introduction thus forms the basis for the specific aims and

structure of the thesis.

Chapter 1: General introduction

2

1.1 Effects of climate change on marine ecosystems

The ocean covers 71% of our planet’s surface and plays a major role in regulating

Earth’s climate. The oceans have also been a major sink for rising atmospheric heat and

carbon dioxide in the atmosphere (Hoegh-Guldberg and Bruno 2010).The increased

ocean temperatures and acidification that have resulted from these anthropogenic

processes have major consequences for all biological components of marine

ecosystems, whether at levels of individuals, populations or communities (Roessig et al.

2004; Hoegh-Guldberg and Bruno 2010). Although studies in marine environments are

limited in number relative to research in terrestrial systems, they have documented

ocean warming (IPCC 2014), acidification (Orr et al. 2005; Doney et al. 2009),

increases in the amount of heat storage (Levitus et al. 2009) and changes in salinities

(Wong et al. 1999; Durack and Wijffels 2010). These changes can, in turn drive other

indirect impacts. For example, changes in the heat content of the oceans cause shifts in

temperature stratification that in turn affects primary productivity (Hoegh-Guldberg and

Bruno 2010). Increased heat content may also result in decreased oxygen concentrations

(Matear and Hirst 2003; Diaz and Rosenberg 2008; Keeling et al. 2010), alterations to

ocean circulation (Hayward 1997; Marshall et al. 2014) and drive sea level rise

(Vermeer and Rahmstorf 2009; Nicholls et al. 2011). Hence, there is a clear need to

understand how these direct and indirect effects of climate change will influence the

ecology of marine ecosystems.

Significant progress has been made over the past two decades towards understanding

and predicting the responses of fishes to climate change (Rijnsdorp et al. 2009; Pörtner

and Peck 2010; Cheung et al. 2012b). Because fish are ectotherms, much of this

assessment of the potential impacts of climate changes on fish has focused on

physiological changes at the individual level in relation to thermal and aerobic

tolerances (e.g. Pörtner and Peck 2010). Other studies have investigated population or

community level dynamics such as changes in timing of spawning (Sims et al. 2004),

recruitment (Asch 2015), food availability (Edwards and Richardson 2004) and

distribution shifts (Perry et al. 2005; Dulvy et al. 2008; Cheung et al. 2015). However,

these studies have been overwhelmingly conducted in the temperate regions of the

Northern Hemisphere, where major industrial and commercial fisheries are located

(Roessig et al. 2004; Pörtner and Peck 2010). Much less is understood of how

individuals and populations of fish respond to climate change in the tropics (Munday et

al. 2012).


3

1.2 Coral reef fish and climate change

Fishes in coral reef ecosystems, primarily in the tropics, are considered to be vulnerable

to climate change because of direct physiological responses (Munday et al. 2012). For

example, reduced aerobic capacity associated with warming temperatures has been

demonstrated for some coral reef fishes (Nilsson et al. 2009; Rummer et al. 2014),

leading to declines in somatic growth (Munday et al. 2008a) and reproductive

performance (Donelson et al. 2010). Metabolic performance has also been shown to be

reduced for fishes of both lower and higher trophic levels due to warming temperatures

(Johansen and Jones 2011; Johansen et al. 2015). Similarly, elevated levels of dissolved

carbon dioxide in the water column have been linked to reduced aerobic capacity in

coral reef fishes (Munday et al. 2009b). Increased levels of carbon dioxide have also

resulted in changes in olfactory-mediated behaviour of settlement-stage larvae (Dixson

et al. 2010) and auditory preferences of juvenile coral reef fish (Simpson et al. 2011).

Such changes in behaviour have in turn been linked to reduced recruitment and

increased mortality of these juveniles (Munday et al. 2010). Finally, warming

temperatures and ocean acidification can lead to the loss of structurally complex

habitats (coral reefs; Rogers et al. 2014), which can also negatively affect fish

populations (Pratchett et al. 2008).

Most coral reef fishes have complex life cycles and some species undergo ontogenetic

shifts in habitats, where larvae occur in the pelagic environment, juveniles settle in

nearshore nursery areas and adults tend to move to offshore reefs. Hence, the multiple

direct and indirect impacts of climate change are likely to vary over the different life

history stages and will have important consequences on population dynamics. The

majority of studies of climate change on coral reef fishes have tended to focus on

individual-level effects, particularly on the early life history stages such as larvae.

Several experiments have examined how elevated temperatures will affect the growth of

tropical fish larvae and newly settled recruits (McCormick and Molony 1995; Meekan

et al. 2003; Green and Fisher 2004; Sponaugle et al. 2006). However, less research has

focused on adult fishes (Munday et al. 2008b), with the exception of a study on the

adult damselfish Acanthochromis polyacanthus (Munday et al. 2008a). These studies

have revealed that temperature affects the growth of larvae at specific stages and the

short-term growth of adults on monthly time scales. However, it remains unknown how

subsequent stages of tropical fishes will respond to climate variations at longer time

scales (inter-annual and inter-decadal).


4

1.3 Importance of fish and fisheries

Recent estimates indicate that marine fisheries catches peaked at 130 million tonnes in

1996 and have since been declining (Pauly and Zeller 2016). This decline is of

considerable global concern given that an estimated 4.3 billion people depend on fish

for at least 15% of their protein intake (FAO 2014), and 10-12% of the world’s

population, predominantly in Asia, depends on fisheries and aquaculture as their source

of income (FAO 2014). Fish is also one of the most traded food commodities in the

world, with fisheries exports reported to be worth more than US $136 billion in 2013,

more than half of which occurs in developing economies (FAO 2014) that encompass

the majority of tropical fisheries (Andrew et al. 2007). Consequently, these fisheries

contribute significantly to foreign currency earnings, employment, food security and

nutrition in many developing nations. The latter two contributions are especially

important because subsistence fishing provides the primary source of animal protein for

developing countries in the tropics (Green et al. 2009) and the human population in this

region is expected to increase by about 2 billion by the year 2030 (McManus 1997;

Cohen 2006). Hence, sustainability of fisheries in the tropics is a critical issue for these

populations.

1.4 Otoliths as recorders of growth

Fish ear stones, or otoliths, are calcified paired structures found within the skull case of

fishes. Otoliths form growth increments at both daily and annual intervals, providing an

accurate estimate of age of a fish. In combination with body size, maturity and sex data,

otoliths have been used in fisheries management for over a century, providing insights

into the fundamental life history characteristics of fishes (Green et al. 2009). This

combination of age, length and sex data can be used to calculate sex-specific growth

rates of the population, estimate survival, mortality and age-specific recruitment

(Campana and Thorrold 2001; Fowler 2009). Growth, in particular, is important

because it affects the size-dependent survival of juveniles (Sogard 1997) and

reproductive success of adults (Donelson et al. 2008), which ultimately affects the

productivity of fisheries (Polunin 1996; Dutil and Brander 2003; Anderson et al. 2008).

Records of growth and elemental composition embedded within otoliths have also been

used to assess size-selective mortality for population dynamics (Sinclair et al. 2002),

determine length of pelagic larval duration (Pastén et al. 2003), identify species

(Campana 2004), distinguish separate populations (Secor et al. 2002) and to reconstruct


5

migration patterns in diadromous fishes (Tsukamoto and Arai 2001). However, the

application of otolith analyses to tropical species has lagged behind that of temperate

species for two main reasons: first, the misconception that reliable increments would not

be produced in the otoliths of tropical fish (Panella 1974); and second, the limited

research capabilities and hence research effort in many tropical (often developing)

countries (Fowler 2009). Indeed, while studies of the otoliths of temperate fishes began

in the late 19th century, application of this technique to tropical fishes largely

commenced around the mid-1990s (Fowler 2009). Otolith studies have revealed the

population dynamics and life histories of tropical fish, in particular, their longevity. For

example, many tropical fishes are known to have maximum ages beyond 30 years, as

demonstrated in a study on 10 species of acanthurid fishes (Choat and Axe 1996). Even

small Stegastes damselfishes are known to reach ages above 20 years (Meekan et al.

2001). The longevity of tropical fish has significant implications for the sustainability

and resilience of fisheries, because long-lived species are typically more vulnerable to

over-exploitation than fast-growing and shorter lived fishes (Musick 1999; Reynolds et

al. 2001).

Measurements of increment widths within otoliths have been used to infer individual

somatic growth rates for fishes on both a daily and annual basis, which is possible

because of the proportional relationship between the growth of the otolith and the

growth of the fish (e.g. Molony and Choat 1990; Rowell et al. 2008; Neuheimer et al.

2011; Stocks et al. 2011; Black et al. 2013). Examples of such studies have investigated

growth rates of juvenile fish in the southwest Pacific (Thresher et al. 2007) and how

temperature influences varying growth rates in multiple populations of temperate fish in

the Southern Hemisphere (Neuheimer et al. 2011). Similar approaches relating growth

increments to calendar dates and/or age have been successfully applied to other aquatic

taxa such as coral (De'ath et al. 2009) and mussels (Kendall et al. 2010). The biological

data (e.g. growth, environmental and physiological records) stored across time in fish,

corals and molluscs are known as ‘biochronologies’ (Morrongiello et al. 2012). These

potentially long-term records of several years to decades and occasionally centuries,

contribute to the understanding of how individuals and populations have responded

historically to the integration of multiple natural and anthropogenic factors that drive

the growth patterns that are recorded in these calcified structures.


6

1.5 Emergence of biochronology techniques

Over the last two decades, a new analysis of increment widths within otoliths of fish

(Black et al. 2005), mollusc shells (Marchitto et al. 2000) and coral cores (Hendy et al.

2003) has emerged as a way of assessing long-term changes and biological responses in

marine environments. This technique has applied the analytical methods largely

developed from dendrochronology (the study of tree rings) to biochronologies in marine

and freshwater organisms. Dendrochronology techniques to assess past environmental

conditions and to age trees were developed in the early 1900s and have mainly been

used to reconstruct records of historical climate change (e.g. in Australia see Cullen and

Grierson 2007; O'Donnell et al. 2015) or identify extreme events such as fires and

droughts from tree rings (Fritts and Swetnam 1989). Such studies are possible because

of the use of crossdating, a procedure that was originally applied to tree rings to assign

the correct calendar year by cross-matching growth patterns among multiple samples,

based on the assumption that the broader environment induces synchronous and time-

specific growth patterns in trees (Fritts 1971). Because of similarities in the annual

growth patterns of tree rings and otoliths, crossdating was successfully applied to the

otolith increments of the long-lived rockfish Sebastes diploproa, which enabled the

accurate dating of each annual increment for age validation purposes (Black et al.

2005). Since this innovative study, the crossdating technique has been used to clarify

relationships between climate and growth of fish from the Gulf of Mexico and the

Bering Sea in the Northern Hemisphere (Black et al. 2008; Matta et al. 2010; Black et

al. 2011).

Crossdating has also been successfully applied to the growth increments of mollusc

shells (Helama et al. 2006). Thus it is now possible to explore and compare the

responses of multiple taxa across marine (fish and mollusc) and terrestrial (tree)

ecosystems (Black 2009); the few studies that have attempted this to date have focussed

on the reconstruction of historical climate records (Black et al. 2009; Black et al. 2014).

Studies employing the use of crossdating on adult fish in the Southern Hemisphere only

began within the last four years and all have focused on temperate fishes (Gillanders et

al. 2012; Coulson et al. 2014; Rountrey et al. 2014; Nguyen et al. 2015). These studies

have revealed that water temperature is a strong driver of variations in the otolith

chronologies for most of these species. In parallel with the rise of the use of crossdating

techniques on temperate fishes in the Southern Hemisphere, biochronologies have also

been developed using non-detrended increment widths in mixed effects models


7

(Doubleday et al. 2015; Gillanders et al. 2015; see Morrongiello and Thresher 2015 for

detailed information on using mixed effects models). This method is useful because it

enables short-lived species to be analysed, which is one of the limitations of crossdating

(Weisberg et al. 2010; Morrongiello et al. 2011; Morrongiello and Thresher 2015).

1.6 Tropical lutjanids as model species

Lutjanid fishes, commonly known as snappers, are found in tropical environments

world-wide, typically in reef habitats (Carpenter and Niem 2001). Many species of

lutjanids such as Lutjanus johnii and Lutjanus argentimaculatus are highly valued for

commercial fisheries and mariculture in Southeast Asia (e.g. Kiso and Mahyam 2003;

Emata 2003 respectively), whereas Lutjanus fulviflamma is a species of major

commercial importance in southern Japan (Shimose and Tachihara 2005) and Tanzania

(Kamukuru et al. 2005). Similarly, Lutjanus guttatus (Amezcua et al. 2006) and

Lutjanus campechanus (Wilson and Nieland 2001) are targets of fisheries in Mexico

and several other lutjanids are valuable components of commercial and recreational

fisheries in Australia (Newman et al. 2000). Lutjanids are generally long-lived, with

maximum ages of around 30 years for many species, such as Lutjanus griseus (Fischer

et al. 2005), L. johnii (Cappo et al. 2013) and Lutjanus erythropterus (Newman et al.

2000), with two species, L. campechanus and Lutjanus bohar, having reported

maximum ages of around 55 years (Baker and Wilson 2001; Marriott and Mapstone

2006 respectively). These relatively long life spans mean that the otoliths of these

lutjanid fishes likely contain multi-decadal records of growth, making them good

candidates for biochronology studies.

The commercial importance of many lutjanid species, their long life spans and their

inclination to form aggregations (Hobson 1975; Newman and Williams 2001) have led

to concerns about the vulnerability of lutjanids to fisheries over-exploitation (e.g.

Kamukuru et al. 2005; Amezcua et al. 2006; Marriott et al. 2007; Heupel et al. 2010;

Newman et al. 2010). Hence, several studies have focused on an analysis of the

demographic characteristics of these species to help determine appropriate management

strategies (e.g. Wilson and Nieland 2001; Grandcourt et al. 2006; Heupel et al. 2010;

Yamada 2010). However, less research has focused on the effects of climate change on

the ecology and demographic characteristics of lutjanids. Commercial exploitation can

lead to reductions in demographic traits and geographic diversity of populations and

subsequently, increases in sensitivity of these populations to climate change (Brander


8

2007). To date, only one study has applied dendrochronology techniques to the otoliths

of lutjanids (L. campechanus and L. griseus) to examine their climate-growth

relationships in the Gulf of Mexico (Black et al. 2011). Similar studies are lacking for

other lutjanids and for tropical fish in general.

Many lutjanids tend to have complex life history cycles with ontogenetic changes in

habitat. Some examples include L. argentimaculatus (Russell et al. 2003;Elliott et al.

2007), L. bohar (Marriott et al. 2007), L. erythropterus and Lutjanus malabaricus

(Newman and Williams 1996). Ontogenetic migrations and the wide distributions of

tropical lutjanids imply that the effects of climate change may vary over different life

history stages and populations that inhabit different ocean basins.

1.7 Aims

The overall objective of this thesis is to describe climate-growth relationships for

tropical and commercially important fish in the Southern Hemisphere. Specifically, I

seek to:

1. Determine the usefulness of dendrochronology techniques for examining the

environmental drivers of the growth of the otoliths throughout the life history of

a tropical lutjanid, L. argentimaculatus;

2. Investigate intra-species variations in growth by comparing biochronologies for

two different populations of the tropical lutjanid, L. bohar;

3. Determine cross-taxa similarities in the growth responses of fishes, corals and

trees across marine and terrestrial ecosystems in relation to large-scale climate

influences; and

4. Compare the importance of climatic drivers on the growth chronologies of fishes

spanning both tropical and temperate marine environments.

My research focusses on the marine environment along the coastline of Western

Australia, which is an ideal system to explore the historic impacts of climate change on

tropical lutjanids. Sea surface temperatures along the coastline have been consistently

warming since 1950 at a rate faster than the global average (Pearce and Feng 2007).

Furthermore, surface temperatures along the tropical north-west coast are predicted to

increase by more than 2 °C by the year 2055 (Cheung et al. 2012a). Considerable

distribution shifts are also expected to occur in the eastern Indian Ocean, with an

increase in dominance of warmer-water species in temperate regions (Cheung et al.


9

2012a). The recruitment and settlement of marine organisms in Western Australia have

also been strongly linked to variations in El Niño-Southern Oscillation (ENSO) that

subsequently drive changes in current flows (Caputi et al. 1996; Feng et al. 2009).

Extreme changes in ENSO have also led to massive fish kills in the region, due to

abnormally high water temperatures, as documented in the summer of 2011 (Feng et al.

2013; Wernberg et al. 2013). Hence, climate changes in this region are likely to have

major influences on tropical lutjanids. Importantly, large-scale fisheries projects on

tropical lutjanids have been conducted in this region, resulting in a substantial archive

of lutjanid otoliths at the Western Australia Department of Fisheries (e.g. Newman and

Dunk 2003; Pember et al. 2005; Cappo et al. 2013). Long-term growth chronologies of

corals (Cooper et al. 2012) and trees (O'Donnell et al. 2015) have also been developed

for the north-west region of Australia, providing an opportunity to investigate

similarities in these responses across diverse taxa and ecosystems. To allow cross-

population comparisons, I also included the otoliths of L. bohar, which were collected

from north-east Queensland (Marriott and Mapstone 2006) at similar latitudes to those

collected in the north-west of Western Australia.

1.8 Thesis outline

In this thesis, I identify the links between growth patterns and climate by using

accurately dated and annually-resolved growth chronologies of fishes. In Chapter 2, I

describe and apply dendrochronology techniques to develop a high-resolution (annual)

growth chronology across two different life history stages (juvenile and adult) of a

tropical snapper L. argentimaculatus from north-west Australia. The usefulness of this

approach in obtaining precise and accurate growth data are assessed and the growth data

are then analysed with climate data to examine and compare how climate influences

juvenile and adult life history stages. I hypothesized that because the juveniles and

adults of L. argentimaculatus inhabit different environments (estuaries and mangroves

versus offshore reefs), the climatic drivers of growth would differ between these life

history stages (Figure 1.1).

In Chapters 3, 4 and 5, I apply the technique developed in Chapter 2 and further expand

this approach across populations, taxa and ecosystems at progressively larger spatial

scales (Figure 1.2). In Chapter 3, I compare the varying effects of climate change across

two separated populations of L. bohar on the east and west coasts of Australia. I tested

the hypothesis that the two populations will show differences in the key drivers of


10

growth because of the varying influences of climate and hence the oceanographic

environments to which they are exposed (Figure 1.1).

The synchrony of large-scale climate phenomenon across taxa from both the marine and

terrestrial ecosystems is examined in Chapter 4. Growth chronologies of two marine

fishes, a coral and a tree from north-west Australia are analysed to test the hypothesis

that the growth patterns of these diverse groups are similar because of the far-reaching

influences of climate change (Figure 1.1).

In Chapter 5, I explore the synchronous impact of climate change on fishes across the

entire Western Australia coastline, encompassing tropical, transitional and temperate

environments. The degree to which the effects of climate change acts simultaneously

over a broad spatial scale (3000 km of coastline) is tested using the growth chronologies

of six species of marine fishes across a range of depths, thermal limits and trophic levels

(Figure 1.1).

The main findings of my thesis are synthesized and discussed in Chapter 6. I consider

the suitability and benefits of dendrochronology techniques for the construction of

highly accurate and annually resolved growth chronologies from the otoliths of tropical

fishes, and how these can be applied to improve our understanding and thus

management of the impacts of climate changes on vulnerable populations of fishes.


11

Figure 1.1 Schematic representation of the four data chapters and the relevant hypotheses

in those chapters.

Figure 1.2 Schematic representation of the spatial scales of the thesis data chapters.

13

Chapter 2 Contrasting environmental

drivers of adult and juvenile

growth

2.1 Abstract

Many marine fishes have life history strategies that involve ontogenetic changes in the

use of coastal habitats. Such ontogenetic shifts may place these species at particular risk

from climate change, because the successive environments they inhabit can differ in the

type, frequency and severity of changes related to global warming. I used a

dendrochronology approach to examine the physical and biological drivers of growth of

adult and juvenile mangrove jack (Lutjanus argentimaculatus) from tropical north-

western Australia. Juveniles of this species inhabit estuarine environments and adults

reside on coastal reefs. The Niño-4 index, a measure of the status of the El Niño-

Southern Oscillation (ENSO) had the highest correlation with adult growth

chronologies, with La Niña years (characterised by warmer temperatures and lower

salinities) having positive impacts on growth. Atmospheric and oceanographic

phenomena operating at ocean-basin scales seem to be important correlates of the

processes driving growth in local coastal habitats. Conversely, terrestrial factors

influencing precipitation and river runoff were positively correlated with the growth of

juveniles in estuaries. These results show that the impacts of climate change on these

two life history stages are likely to be different, with implications for resilience and

management of populations.

2.2 Introduction

Many fishes of high commercial value have life history strategies in which successive

ontogenetic stages occupy different habitats. Changes in habitat can be relatively minor,

such as the transition from inshore to deeper offshore waters in Atlantic herring (Clupea

harengus; Whitehead 1984) and cod (Gadus morhua; Cohen et al. 1990), or more

extreme as observed in European seabass (Dicentrarchus labrax) and giant trevally

(Caranx ignobilis), where juveniles are found in estuarine nurseries, while adults are

found in fully marine habitats offshore (Elliott et al. 2007). Perhaps the most extreme

examples of habitat change between ontogenetic stages are shown by species that are

catadromous (e.g. barramundi, Lates calcarifer) and anadromous (e.g. chinook salmon,

Chapter 2: Contrasting environmental drivers of adult and juvenile growth

14

Oncorhynchus tshawytscha). In the case of the former, adults are found in brackish or

freshwater habitats while larvae and juveniles occupy marine habitats. In the latter,

larvae and juveniles are found in freshwater and the adults reside offshore (Elliott et al.

2007).

Species that undergo ontogenetic shifts of habitat are thought to do so to avoid predators

and/or to improve foraging, so that there is a selective balance between minimizing

mortality risk and maximising energy gains (Sutherland 1996). However, marine

habitats are changing due to anthropogenic processes such as global warming and ocean

acidification. The need to occupy a range of habitats during each life history stage could

make such species particularly vulnerable to increased climate variability, because

impacts could differ in type, frequency and severity in any of the successive habitats

they occupy. For example, sea-level rise and associated intrusions of salt water could

reduce the area of estuarine nursery habitats (Drinkwater et al. 2010) used by juveniles

of many species. Estuarine waters are also expected to undergo accelerated warming

because of their small size and shallow depths (Gillanders et al. 2011). Additionally,

changes in rainfall and terrestrial run-off patterns may alter freshwater inputs, changing

salinities and the stratification of the water column (Gillanders et al. 2011). These in

turn could lead to reduced water quality and hyper-saline conditions. In coastal waters,

sea surface temperatures are rising and patterns of current flow are changing. Although

warmer sea surface temperatures are likely to increase the growth rates of fishes living

at sub-optimal temperatures (Rountrey et al. 2014), for individuals and species living

closer to thermal optima, warmer waters might decrease growth rates (Neuheimer et al.

2011) and cause range shifts and contractions (Cheung et al. 2012a).

At present, what is known of the likely response to climate change by many marine

fishes tends to be based on observations of the adult life history stage (Munday et al.

2008b, but see Thresher et al. 2007 for an exception). Given that the consequences of

climate change may vary among habitats, obligate ontogenetic shifts in habitat may

make some species more vulnerable than others that are largely sedentary, complicating

predictions of species responses. Alternatively, migrant fish may be in better conditions

than sedentary fish, as found in a recent study on estuarine black bream, Acanthopagrus

butcheri, in South Australia (Gillanders et al. 2015). For this reason, studies are required

that assess the key physical and environmental drivers of growth at different life history

stages of marine fishes that undergo ontogenetic habitat shifts. This issue is of particular

concern in the tropics where fish assemblages in estuarine and shelf waters provide the


15

major source of protein for more than 600 million people in Asia (Stobutzki et al. 2006)

and more than 200 million people in Africa (Béné and Heck 2005), most of whom live

in developing countries. There are numerous examples of tropical species that spawn at

sea and enter estuaries as juveniles, including mangrove jack (Lutjanus

argentimaculatus), giant trevally (C. ignobilis), milkfish (Chanos chanos) and large-

scale mullet (Chelon macrolepis; Elliott et al. 2007). Such species are central to tropical

fisheries, largely because major population centres tend to be clustered on the coasts

that surround estuarine habitats.

Here, I examine the impacts of past environmental variation on different life history

stages of L. argentimaculatus as a means of assessing likely consequences of future

climate change. This species has a juvenile stage that occurs in mangroves, estuaries

and some freshwater habitats that are heavily influenced by ephemeral processes such as

river runoff and rainfall. In contrast, adults are found on tropical coastal and offshore

reefs, a more stable environment affected by oceanic factors. The physical and

biological processes within these different environments are expected to affect

individual fish via both physiological and ecological mechanisms that in turn, affect

their growth. Hence, I hypothesize that, in the warmer months of January to March (the

summer growing season), juvenile growth is likely to respond to factors that affect the

productivity of estuarine systems such as rainfall, whereas adult growth will respond to

factors that affect the productivity of coastal ecosystems such as sea surface temperature

(SST) and sea surface salinity (SSS). Both juvenile and adult growth are also likely to

be linked with large-scale climatic indices such as the Niño-4 index that affects SST

along the shelf of Western Australia (Feng et al. 2013; Zinke et al. 2014) and the Pacific

Decadal Oscillation (PDO), which is associated with precipitation (Mantua and Hare

2002).

I used a dendrochronology approach to generate records of growth for both the juvenile

and adult stages of L. argentimaculatus collected from north-west (NW) Australia.

These records were obtained from otoliths (fish ear stones), which contain annual

growth increments. Otolith growth and somatic growth are tightly linked, thus, time

series of annual growth can be assembled using otolith increment measurements

(Casselman 1990; Black et al. 2013). I generated a master growth chronology based on

records from a sample of individuals and tested for relationships with the physical

oceanographic and atmospheric variables outlined above. These analyses enabled the

identification of important drivers of growth for each life history stage. The results were


16

then evaluated with respect to the susceptibility of this species to predicted patterns of

climate change in both adult and juvenile habitats.

2.3 Methods

2.3.1 Study species

Lutjanus argentimaculatus is long-lived, with some individuals exceeding 50 years of

age (Pember et al. 2005). It is a prized target of commercial, artisanal and recreational

fisheries throughout its range, which includes much of the tropical Indian Ocean and

western Pacific. In Western Australia, estimated recreational catches of this species

were approximately 2.7 tonnes (Ryan et al. 2013) and commercial catches were

approximately 13 tonnes in 2011–2012 (Fletcher and Santoro 2012). Individuals exhibit

negative exponential growth rates early in ontogeny, with lengths of around 400 and

550 mm attained in 5 and 10 years respectively (Pember et al. 2005), and an observed

maximum length of around 800 mm from the sample collection. Juveniles have been

shown to inhabit estuaries and mangroves, before moving offshore at around 7 years old

(Pember et al. 2005).

2.3.2 Study site

The tropical NW coast of Australia is characterised by a large tidal range, a high

frequency of cyclones, and warm, oligotrophic surface waters. These low salinity waters

largely emanate from the Indonesian through-flow, which connects waters in the

western Pacific to the Indian Ocean (Meyers 1996). The NW coast consists of numerous

small barrier and fringing reefs in shallow water (Lough 1998). In this region, the

maximum rainfall occurs in summer, coinciding with the highest rate of tropical cyclone

occurrence. About 30 river basins drain into the NW coast, however, the majority of the

river flows enter the coastal waters north of ~18°S (Lough 1998). River flow in the

north-west has been reported by Lough (1998) to be more variable compared to river

flow on the east coast and across Australian rivers in general. Estuaries on the NW coast

are small and shallow compared to global standards (Gillanders et al. 2011), so are

likely to have variable environments heavily influenced by terrestrial inputs such as

river flows.


17

2.3.3 Sample collection

Archived otolith collections of L. argentimaculatus were obtained from the Department

of Fisheries (Government of Western Australia). These had been collected from the

Pilbara and Kimberley regions of NW Australia from commercial catches (fish trawls

and fish traps) and recreational line catches, supplemented with research sampling using

fish traps between 1996 and 2005 (Pember et al. 2005). Adult fish were caught from

offshore reefs across the continental shelf of NW Australia (Figure 2.1). The sagittal

otoliths of each fish were removed, cleaned and stored to dry in paper envelopes. One

sagittal otolith from each fish was embedded in epoxy resin and sectioned transversely

through the primordium in a direction perpendicular to the sulcus acusticus, using a low

speed saw with a diamond tipped blade (Buehler, United States of America, USA).

Sagittal otolith sections were cut into two or three 150-300 μm thick sections near the

core of the otolith to enhance interpretation. These sections were cleaned and mounted

on glass slides using DePeX mounting medium (Pember et al. 2005).

The formation of annual increments in the otoliths of L. argentimaculatus has been

validated using marginal increment analysis by a number of studies (Russell et al. 2003;

Pember et al. 2005). Each increment within the otolith consists of translucent and

adjacent opaque zones with the latter completed between October and November each

year (Pember et al. 2005). The spawning season takes place between September and

December each year (Pember et al. 2005), and growth is assumed to occur during the

warm summer months of January to March for both juveniles and adults. Only otoliths

from fish at least 23 years old (to ensure sufficient years for chronology development)

with clear increments (to increase accuracy) were selected for image analysis (n = 36),

with 3 fish discarded due to poor clarity. The selection of otoliths with clear increments

may lead to a possibility of sample bias, however the importance of being able to

correctly identify each increment and accurately assign calendar years was essential to

the study. Selected fish were collected over the years 1996–2005, with most collected in

2003 and 2004. The ages of fish ranged from 24–52 years with an average of 37 years.

The samples had a standard length range of 428–700 mm with an average of 558 mm.

2.3.4 Image analysis

Multiple sections of the same otolith were examined under a dissecting microscope to

assess the clarity of growth increments and the best section was used for increment

measurements. The region next to the sulcus on the dorsal side of each otolith was


18

imaged using an Olympus IX81 inverted microscope (Olympus Corporation, Tokyo,

Japan) with a ProScan II motorised stage (Prior Scientific, Japan). Multiple images were

captured in brightfield (transmitted light) using a Nikon Digital Sight DS-2Mv colour

camera (Nikon, Japan) with a 10x objective (Olympus UPlan 10x 0.3NA dry objective)

and stitched using NIS-Elements AR software (version 3.22 & 4.13, Nikon, Japan). This

region of the otolith was chosen because it was consistently the clearest region for

annual increments (Figure 2.2a). Increment widths were measured using a plugin

“IncMeas” (Rountrey 2009) written for the image analysis software, Image J (version

1.48, National Institutes of Health, USA). Three transects parallel to the direction of

growth were drawn on each image montage. The outer edge of the opaque zone of each

increment was marked along each transect (Figure 2.2b). Increment widths were

measured and recorded along with the inferred calendar year of formation, obtained by

working backwards from the year of capture and taking into consideration the

approximate timing of completion of each increment. This was part of the visual

crossdating process, which assumes that the environment induces synchronous, time-

specific growth patterns that can be matched among individuals. Beginning with the

clearest otolith sections in which all increments were easily identified, conspicuously

narrow or wide increments were noted (referred to as signature years that should be

synchronous among samples) and used to ensure the correct assignment of calendar

years to increments (Black et al. 2005).


19

Figure 2.1 Map of sampling sites of Lutjanus argentimaculatus in north-west Australia.

Fish were captured from various offshore sampling areas by research trapping, commercial

trapping or commercial fish trawls from 1996 to 2005. Points on the map show general

locations of capture.

2.3.5 Otolith chronology development

I used the program COFECHA, from the International Tree-Ring Data Bank Program

Library to assist with statistical crossdating (Black et al. 2005). Otolith increment series

from three transects per fish (obtained from Image J measurements) were loaded in

COFECHA and detrended to remove ontogenetic trends while retaining high frequency

variation. A spline rigidity parameter of 22 years was used, following the methods of

Black et al. (2005). Measurements were standardized by dividing by the spline fit, and

then the correlation between each standardized series and the average of all other

standardized series was calculated to allow assessment of series alignment. Samples

with low correlations were visually inspected for potential errors in increment boundary

placements (e.g. missed marking a year due to a faint opaque zone), and any obvious

errors were corrected.

For chronology development, I followed Black et al. (2013) in using a method similar to

regional curve standardization, which is used in tree-ring research to remove

ontogenetic trends while preserving low frequency variations in ring width. Increment


20

widths were aligned by fish age and the mean increment width at each age was

calculated. Each series was then divided by the mean-by-age series to obtain

standardized values. Multiple standardized series belonging to the same fish were

averaged before inclusion in the chronology.

Standardized measurements were classified as either “adult” or “juvenile” years and two

chronologies were created to allow the detection of different drivers of growth in the

two phases. The adult chronology was constructed with all 36 fish using increments

formed after age 7, based on an estimated age of 7 years for sexual maturity for more

than 50% of L. argentimaculatus caught in Western Australia (Pember et al. 2005).

Only years in the chronology with a sample depth of at least 20 fish were retained so

that variation would be more likely to reflect synchronous growth patterns among

individuals, rather than being strongly affected by variation in growth of only one or

two fish. Standardized series of increment width measurements of adults from the

selected years were averaged to create a single chronology representative of adult fish.

The juvenile chronology was constructed using data from age 1 to 9 since more than

90% of L. argentimaculatus caught in Western Australia are mature by 9 years of age

(Pember et al. 2005). Additionally, this produced more series overlap and a longer

chronology than if 7 years had been used. A minimum sample depth of 20 fish was not

possible for the juvenile dataset, hence the longest continuous sequence of years in

which sample depth exceeded 10 fish was selected, resulting in a sample size of 30 fish.

The quality of the chronology was assessed using the mean of pairwise series

correlations (�̅�), an estimate of fractional common variance, as well as expressed

population signal (EPS), a measure of how well the chronology represented the

theoretical population chronology (Wigley et al. 1984), using the package “dplR” (Bunn

2008) in R software (R Core Team 2015).


21

Figure 2.2 Photomicrographs of the dorsal side of a L. argentimaculatus otolith section.

(a) Dorsal section of the otolith was chosen for consistently clear annual increments; (b) close

up image of the same otolith with a transect line and increments labelled with corresponding

calendar years.


22

2.3.6 Correlations with environmental parameters

Pearson’s correlations and the linear models relating otolith chronologies to a number of

environmental factors were calculated using R software (R Core Team 2015).

Assumptions of normality, homoscedasticity and independence were verified for all

linear models and Durbin-Watson tests were used to check for any autocorrelation of

residuals. Five environmental factors were used in the analysis, with only January to

March values retained to coincide with the growing season for fish. Hence, a total of 15

variables were considered, three months (January, February and March) for each of the

five factors. These five environmental factors were large-scale climate indices such as

the Niño-4 index (Rayner et al. 2003), based on SST anomalies over the central Pacific

using HadISST1; Pacific Decadal Oscillation index (PDO; Kennedy et al. 2011), based

on SST anomalies over the North Pacific using HadSST3; as well as regional

environmental variables including SST (Rayner et al. 2003); sea surface salinity (SSS;

Ingleby and Huddleston 2007) and rainfall (Bureau of Meteorology, Australian

Government). Principal Components Analysis (PCA) was used to deal with collinearity

between the Niño-4 index, SSS and PDO over the period 1975–2003. Nine

environmental variables (January, February and March values for each of the three

collinear factors) were put into the PCA. Ocean heat content (incorporating temperature

and salinity from 0–750 m depth) from the Simple Ocean Data Assimilation (SODA)

reanalysis of ocean climate variability (Carton and Giese 2008) was used to examine the

interactions of temperature and salinity on the growth of adults. All data were obtained

from the Royal Netherlands Meteorological Institute (KNMI) Climate Explorer (Trouet

and Van Oldenborgh 2013), a web application for climate data (http://climexp.knmi.nl),

unless otherwise stated. Regional data were obtained from an area covering the latitudes

23°S to 14°S and longitudes 113°E to 127°E, which includes the Pilbara and Kimberley

regions of NW Australia (Figure 2.1).

2.4 Results

2.4.1 Chronology development for adults and juveniles

The chronology derived from the 36 adult fish covered the period from 1975 to 2003, a

29-year series (Figure 2.3). Measurements from at least 20 fish contributed to each year

value, with more than 28 fish contributing to the period between 1979–2002 and all fish

(n = 36) contributing to the years 1988–1995. Even though the common variance among

individuals was low (�̅� = 0.153 with s.e. = 0.016), the sample depth was sufficient to


23

show statistically significant, positive �̅� for all periods from 1975 to 2003, suggesting

that there were clear, synchronous growth patterns for these fish in that period. The

averaged expressed population signal (EPS) for the same period was 0.84, indicating

that the adult chronology well-represented the theoretical population chronology. In

most of our samples, the signature years of 1989–1990 had conspicuously wide

increments, while the years 1991–1992 had conspicuously narrow increments. As the

juvenile chronology only included increment width data from age one to nine, the series

were too short to obtain robust �̅� values, however a continuous 14-year series (1965–

1978; Figure 2.3) with a sample depth of at least 10 fish contributing to each year was

selected for analysis with environmental parameters.

24

Figure 2.3 Raw increment width time series, detrended and final chronologies used from the otoliths of L. argentimaculatus. (a) Raw increment widths of

adults and (b) juveniles; (c) detrended and final chronologies with standard error bars of adults and (d) juveniles.


25

2.4.2 Relationships between chronologies and environmental

drivers

Correlations were calculated between 15 variables (January, February and March values

for each of the five environmental factors) and both the adult and juvenile chronologies

across the relevant time periods. Following the approach of Black et al. (2011) we

reduced the level of significance from p < 0.05 to p < 0.03 to account for multiple

comparisons.

The growth chronology of adult L. argentimaculatus over the period 1975–2003 was

negatively correlated with the Niño-4 index, SSS and PDO (Figure 2.4a) and positively

correlated with rainfall in March (see Table 2.1 for results of all comparisons with

environmental variables). Because of collinearity among the Niño-4 index, SSS and

PDO, these variables were combined in a principal components analysis. The first

principal component (PC1) explained 60% of the total variation and the second (PC2)

18%. All variables had similar positive loadings on PC1 (range 0.310–0.357) and there

was a negative correlation between the adult chronology and PC1 (Figure 2.5a). A

linear regression model that related PC1 to the adult chronology was significant (p <

0.001) with an adjusted R2 = 0.424 (Figure 2.5b). As there was a significant correlation

between the adult chronology and rainfall in March (Table 2.1), this environmental

variable was fourth root transformed (due to the range of data) and included in the linear

model. However, the resulting model was not significant (p = 0.391). Strong positive

correlations between the adult chronology and ocean heat content were identified in the

map of spatial correlations, notably in the areas of the western Pacific warm pool, in

eastern Indonesian Seas and along the coast of NW Australia (Figure 2.6).

The juvenile chronology was significantly (p < 0.03) and negatively correlated with

PDO in February (Figure 2.4b, Figure 2.5c, Figure 2.5d) and there was a weaker

positive correlation with rainfall in the same month that was marginally non-significant

(p=0.037; Figure 2.4b, Figure 2.5e, Figure 2.5f). These two environmental variables

were not collinear (r = -0.096, p = 0.745), so both were included in a linear model with

the juvenile chronology. Although rainfall in February was marginally non-significant

in the correlation analyses (Figure 2.4b), the addition of this variable greatly improved

the strength of the significant (p = 0.010) linear model (adjusted R2 = 0.490 compared to

R2 = 0.318 without rainfall in February; Figure 2.5d). A schematic diagram shows the

results of the most highly correlated environmental variables with both life history

stages of L. argentimaculatus (Figure 2.7).


26

Table 2.1 Correlation coefficients and p-values of adult and juvenile Lutjanus

argentimaculatus with all 15 environmental variables. The adult chronology and relevant

environmental values were from 1975–2003 while the juvenile chronology and relevant

environmental values were from 1965–1978. Pearson’s correlation coefficients (R) were

calculated and a reduced level of significant (p < 0.03) was used to account for multiple

comparisons. Niño-4 = Niño-4 index, PDO = Pacific Decadal Oscillation, SSS = sea surface

salinity, SST = sea surface temperature.

Life stage Variable Month R p-values

Adult Niño-4 January -0.723 0.000009

Adult Niño-4 Febuary -0.716 0.00001

Adult Niño-4 March -0.659 0.0001

Adult SSS January -0.406 0.029

Adult SSS Febuary -0.389 0.037

Adult SSS March -0.465 0.011

Adult PDO January -0.364 0.052

Adult PDO Febuary -0.438 0.017

Adult PDO March -0.445 0.016

Adult Rainfall January 0.072 0.710

Adult Rainfall Febuary -0.106 0.585

Adult Rainfall March 0.435 0.018

Adult SST January 0.201 0.296

Adult SST Febuary 0.245 0.200

Adult SST March 0.114 0.557

Juvenile Niño-4 January 0.000 0.999

Juvenile Niño-4 Febuary 0.034 0.908

Juvenile Niño-4 March -0.026 0.930

Juvenile SSS January 0.072 0.807

Juvenile SSS Febuary 0.174 0.552

Juvenile SSS March 0.103 0.727

Juvenile PDO January -0.506 0.065

Juvenile PDO Febuary -0.609 0.021

Juvenile PDO March -0.530 0.051

Juvenile Rainfall January 0.135 0.645

Juvenile Rainfall Febuary 0.560 0.037

Juvenile Rainfall March -0.059 0.841

Juvenile SST January 0.425 0.129

Juvenile SST Febuary 0.457 0.101

Juvenile SST March 0.298 0.301


27

Figure 2.4 Pearson's correlation coefficients between L. argentimaculatus chronologies and

January to March values of significantly correlated environmental variables. (a) Adult

chronology from 1975–2003 with the Niño-4 index (Niño-4), sea surface salinity (SSS) and

Pacific Decadal Oscillation (PDO) and (b) juvenile chronology from 1965–1978 with PDO and

rainfall. Asterisks represent p-values of the Pearson’s correlation test, with p < 0.01 (***), p <

0.03 (**) and p < 0.04 (*)

28

Figure 2.5 Relationships between L. argentimaculatus chronologies and environmental variables included in linear models. (a) Adult chronology from

1975–20032 plotted with Principal Component 1 (PC1), which accounted for 60% of the variation for January to March values of the Niño-4 (Niño-4), sea

surface salinity (SSS) and Pacific Decadal Oscillation (PDO); (b) regression plot of adult chronology with PC1; (c) juvenile chronology from 1965–1978 with

February PDO index; (d) regression plot of juvenile chronology with February PDO index; (e) juvenile chronology from 1965–1978 with February rainfall (mm

month-1

); and (f) regression plot of juvenile chronology with fourth root transformed February rainfall values.


29

Figure 2.6 Map of spatial correlations of adult L. argentimaculatus chronology with ocean

heat content. January ocean heat content (0–750 m depth) was correlated with the adult

chronology from 1975–2003. Warmer colours indicate positive correlations, cooler colours

indicate negative correlations.

Figure 2.7 Schematic diagram showing the environmental drivers that strongly influence

the juvenile and adult life history stages of L. argentimaculatus.


30

2.5 Discussion

Our study is the first to develop crossdated growth chronologies from the otoliths of

both adult and juvenile stages of a tropical coastal fish. Although fractional common

variance (�̅�) was low for the adult chronology, we found significant environmental

drivers of otolith growth. The quality of the juvenile chronology could not be assessed

using �̅� values due to the relatively short length of each individual series (1–9 years),

but our results show that the growth patterns of both adult and juvenile (albeit with a

lesser amount of certainty) L. argentimaculatus show significant responses to

environmental variables.

The strongest correlation between the growth chronology of adults and environmental

signals involved a large-scale (1000s of km) SST variable (Niño-4 index). The Nino-4

index is calculated from the westernmost region between 5°N–5°S and 160°E–150°W

where the El Niño-Southern Oscillation (ENSO) variations lead to significant SST

changes. This implied that events occurring in the western Pacific affected adult growth,

presumably via the Indonesian through-flow, an oceanic connection between the

western Pacific and the Indian Ocean through current flow from the Indonesian Seas

(Meyers 1996). The strength of the Indonesian through-flow is greater during La Niña

conditions (in our datasets, notably during 1988–1989 and 1999–2000), when warmer

and lower salinity waters are transported to the NW coast of Australia. This is likely to

account for higher SST on the NW coast during La Niña years compared to El Niño

years (Meyers et al. 2007; Zinke et al. 2014). When La Niña events are exceptionally

strong, this may result in anomalously warm waters along the NW coast, as was the case

during the summer of 2010/2011 (Feng et al. 2013). For L. argentimaculatus, the

growth chronology suggested that conditions of warm and low salinity waters during

strong La Niña years were correlated with higher growth rates. Despite this pattern we

did not find any correlation of growth with SST at regional (100s of km) scales, a result

that contrasts with other studies that have recorded positive correlations of growth

chronologies with regional SST in temperate species such as flatfish (Limanda aspera;

Black et al. 2013) , red and grey snapper (L. campechanus and L. griseus; Black et al.

2011) luderick (Girella tricuspidata; Gillanders et al. 2012) and western blue groper

(Achoerodus gouldii; Rountrey et al. 2014). However, we did find strong correlations

with ocean heat content, which implies that a combination of both temperature and

salinity affects growth more strongly than either individual variable.


31

As fish are poikilotherms, it might be expected that they should grow faster when water

temperatures are warmer, assuming that they are not at thermal limits. Aside from

warmer temperatures during La Niña events, the decreased salinity may also influence

growth through osmoregulation or food conversion efficiency (Boeuf and Payan 2001).

Many studies have found higher growth rates of fish correlated with a lower standard

metabolic rate, which occurs at intermediate salinities below that of seawater (see

review by Boeuf and Payan 2001). Similarly, the lower salinity waters associated with

La Niña events might also lower metabolic costs and allow more energy to be allocated

to growth of adults in coastal waters. A study of Atlantic cod (Gadus morhua) found

higher growth rates at lower salinities, resulting from increased food conversion

efficiency (Lambert et al. 1994). This may be another possible mechanism influencing

growth of L. argentimaculatus in NW Australia, provided that the productivity during

La Niña events is also high (and thus the species is not food-limited). In any event, our

results suggest that the interaction of both temperature and salinity are likely to be

affecting adult growth.

The growth of juvenile L. argentimaculatus was influenced by environmental variables

that were linked both directly and indirectly (via the PDO) to rainfall patterns. Positive

correlations with rainfall were expected because lower salinities are likely to increase

growth rates (as noted above). This is consistent with the results of a study on the

estuarine black bream, A. butcheri, in South Australia (Doubleday et al. 2015). A

negative correlation with the PDO could reflect the influence of this variable on rainfall,

via SST anomalies in the Pacific that in turn generate precipitation and convection

heating anomalies (Meehl and Teng 2014). In eastern Australia, negative PDO values

generate cool and wet conditions (Power et al. 1999). However, we found that the PDO

and rainfall were not correlated in the NW region over the juvenile time period. It may

be possible that the PDO affects some other, unknown local factor that was not tested.

Irrespective of the link between growth and the PDO, the effects of rainfall on the

estuarine phase of the species are more obvious. Increases in rainfall lead to greater

river runoff and subsequent declines in salinity in estuaries, a process that will be

exacerbated by the small size and shallow depths of these habitats on the NW coast

(Gillanders et al. 2011). Previous studies on juvenile grey snapper (L. griseus) have

found that growth efficiencies decrease at higher salinities (Wuenschel et al. 2005) and

laboratory experiments have shown that they prefer lower salinities that minimize

osmoregulatory costs (Serrano et al. 2010). Juvenile turbot (Scophthalmus maximus)


32

have also shown better food conversion efficiencies at lower salinity levels (Gaumet et

al. 1995), so juveniles of L. argentimaculatus might respond in similar ways to changes

in salinity from increased rainfall. Additionally, there might be indirect effects of

salinity in estuarine habitats such as salinity-induced changes to prey, predator or

competitor abundances (Serrano et al. 2010). Changes in hydrological and temperature

regimes in estuarine environments are also likely to influence the inter-annual growth

rates of juveniles, which has previously been demonstrated for freshwater perch,

Macquaria ambigua, in south-east Australia (Morrongiello et al. 2011).

While mindful of the caveat that the adult and juvenile chronologies do not represent the

same time period, we show that the physical and biological phenomena that drive

growth patterns of L. argentimaculatus are likely to differ between juvenile and adult

life history stages. This result implies that the responses of the individuals to changing

climate conditions may vary with ontogeny. Rainfall in NW Australia has been

increasing from the 1970s to the present day and predicted increases in mean

precipitation during the Asian-Australian monsoon coupled with an intensification of

ENSO-induced rainfall variability (Christensen et al. 2013) might lead to better growth

rates for juveniles. However, the pattern of Indo-Pacific warming due to climate change

is also predicted to result in a greater frequency of El Niño-like conditions which might

reduce the strength of the Indonesian through-flow (Meyers 1996), and could negatively

affect the growth of adult L. argentimaculatus. These contrasting effects complicate

predictions of the response of this species to climate change in the future. As the

patterns of ontogenetic movement of L. argentimaculatus are common to many species,

our study suggests that it may be very difficult to generalise on the likely outcomes of

climate change for a large suite of fishes in tropical coastal environments.

Because of the ubiquity of ontogenetic changes in habitat by fishes in nearly all marine

environments, our results have implications for many species. The degree of complexity

of the effect of climate change on a species may be dependent on the scale, degree and

timing of change in habitats between life history stages, and the impacts could be

reduced for those species that show only limited patterns of movement. However, even

those species that do not change habitat are likely to change diet and trophic role with

age. Although the biochronology approach addresses only population growth anomalies

inferred from otoliths, these studies are now required to determine if trophic changes

add a similar layer of complexity to the prediction of the effects of climate change as

those of ontogenetic movement. In addition, the climate-growth relationships obtained


33

from chronology-based studies will be a critical component of ecosystem-based fishery

management.

2.6 Acknowledgements

I acknowledge the facilities, and the scientific and technical assistance of the Australian

Microscopy & Microanalysis Research Facility at the Centre for Microscopy,

Characterisation & Analysis, the University of Western Australia, a facility funded by

the University, State and Commonwealth Governments. In particular, I am grateful to

Professor Paul Rigby and Miss Alysia Buckley for the generous help and advice they

have provided for the microscopy work. This work was funded by the Australian

National Network in Marine Science, the Australian Institute of Marine Science and the

Center for Marine Futures at the University of Western Australia. This work was made

possible with a scholarship from the Australian Postgraduate Awards. I would also like

to thank Steve Newman for providing the samples and Jens Zinke for his advice.

35

Chapter 3 Cross-continent comparisons

reveal differing drivers of growth

3.1 Abstract

Biochronologies provide important insights into the growth responses of fishes to past

variability in physical and biological environments and in so doing, allow modelling of

likely responses to climate change in the future. I examined spatial variability in the key

drivers of growth patterns of a widespread, tropical snapper, Lutjanus bohar at similar

tropical latitudes on the north-western and north-eastern coasts of the continent of

Australia. For this study I developed biochronologies from otoliths that provided

proxies of somatic growth and these were analysed using mixed effects models to

examine the historical drivers of growth. The analyses demonstrated that growth

patterns of fish were driven by different climatic and biological factors in each region,

including Pacific and Indian ocean climate indices, regional sea level and the size

structure of the fish assemblage. The results showed that the local oceanographic and

biological context of reef systems strongly influenced the growth of L. bohar and that a

single age-related growth trend cannot be assumed for separate populations of this

species that are likely to experience different environmental conditions. Generalized

predictions about the growth response of fishes to climate change will thus require

adequate characterisation of the spatial variability in growth patterns likely to be found

throughout the range of species that have cosmopolitan distributions.

3.2 Introduction

Growth is one of the key parameters determining the productivity and yield of fish

populations. For this reason, an understanding of the factors driving growth patterns is

central to management strategies that seek to mitigate future impacts and improve the

resilience of fisheries to climate change (Wilson et al. 2010; Morrongiello et al. 2012;

Rountrey et al. 2014). As biochronologies of increment widths within otoliths can serve

as proxies of somatic growth rates of fishes (Rowell et al. 2008; Neuheimer et al. 2011;

Black et al. 2013), this allows chronologies to be compared to time series of climate and

other environmental variables in order to identify the key drivers of variation in growth

in the past (e.g. Black et al. 2011; Black et al. 2013) and to predict growth rates and

demographic traits under different scenarios of climate change (Morrongiello et al.

Chapter 3: Cross-continent comparisons reveal differing drivers of growth

36

2012; Rountrey et al. 2014). However, the variables underlying growth patterns are

likely to differ across the distribution of a species. This is particularly the case for

widespread, cosmopolitan fishes that occur throughout a range of depths and

environments subject to variable oceanographic regimes and climate patterns.

At present, there have been few attempts to examine variability in the drivers of growth

of marine fishes across large (1000s km) spatial scales and the likely impacts of this

variability on responses to climate change. Some exceptions are studies on temperate

fishes from the south and south-eastern regions of Australia, which showed that

different populations of fishes had varying growth patterns that were linked to

environmental variables such as temperature (Neuheimer et al. 2011; Doubleday et al.

2015; Morrongiello and Thresher 2015). Growth responses to climate change is

particularly critical in the tropics, because many species inhabit areas where

temperatures are already near the upper physiological limits for growth, making fishes

very sensitive to changes in environmental regimes (Nilsson et al. 2009; Munday et al.

2012). Coastal fishes are also the major source of protein for many millions of people

living in the developing countries that border tropical seas and oceans (e.g. Stobutzki et

al. 2006) and it is thus essential that we have the ability to model and predict the likely

impacts of climate change on exploited stocks.

Here, I examine spatial variability in the drivers of growth of a tropical lutjanid,

Lutjanus bohar. This snapper is an important target for artisanal, small-scale and

commercial fisheries across coral reef ecosystems in the Indo-Pacific (Allen 1985;

Wright et al. 1986). It has a long life span (up to 55 years; Marriott and Mapstone

2006), making it an ideal candidate for the development of otolith biochronologies. I

use this approach to compare growth patterns of the species in waters of the tropical

north-western and north-eastern coasts of the Australian continent, regions that lie on

the margins of the Indian and Pacific ocean basins. Although both regions have marine

environments dominated by southward-flowing, warm-water boundary currents, the

current on the eastern coast (East Australian Current) has a stronger flow than the

current on the western coast (Leeuwin Current) and is most evident during summer

(Ridgway and Godfrey 1997), whereas the Leeuwin Current is strongest in winter (Feng

et al. 2003). In the west, the Leeuwin Current is modulated by the El Niño-Southern

Oscillation (ENSO), strengthening during La Niña and weakening during El Niño

phases (Caputi et al. 1996). In the east, there is a complex relationship between the East

Australian Current and ENSO, so that during El Niño summers, flow strengthens due to


37

enhanced oceanic inflows in the south, whereas to the north, changes in atmospheric

conditions (higher sea level pressure and suppressed surface wind anomalies) lead to

weaker summer monsoons, reduced cyclone activity, rainfall and more radiation

(Redondo-Rodriguez et al. 2012).

These contrasting links of ENSO to the flow of boundary current systems suggests that

populations of L. bohar from the tropical north-west (NW) and north-east (NE) coasts

of Australia (at equivalent latitudes) may show differences in the key drivers of growth

patterns and thus in responses to climate change. I examine this hypothesis by

comparing and contrasting the drivers of biochronologies of growth extracted from the

otoliths of this species in each of these regions. Possible drivers of growth include a

number of local environmental variables as well as some large-scale climate indices

relevant to the region that subsequently influences the local environmental variables

(detailed in Section 3.3.6). I also explored the influence of commercial fisheries (using

an index of size structure of the fish assemblage, detailed in Section 3.3.6) on the

growth of these species at both localities.

3.3 Methods

3.3.1 Study sites

The arid coast of NW Australia provides little terrigenous input to adjacent reef systems

and major coral reefs in the region are isolated atolls on the shelf edge or small barrier

and fringing reefs in shallow water (Lough 1998). In contrast, along the NE Australian

coast, the Great Barrier Reef forms an almost continuous network of barrier reefs across

the shelf. These range from environments that experience terrestrial inputs from

freshwater run-off close to the coast, to shelf-edge locations that are bathed by oceanic

waters (Lough 1998). Fisheries for medium sized (30–89 cm) reef-associated fishes

such as L. bohar have historically been smaller off the NW coast (mean annual catch of

1,550 tonnes; range of 790 to 2240 tonnes from 1970–1990) than the NE coast (mean

annual catch of 3,640 tonnes; range of 2220 to 4240 tonnes over the same time period).

Furthermore, based on the extent of available habitat (shelf and inshore), reconstructed

catches of medium-sized reef fish (Pauly and Zeller 2016) off the NE coast have been

consistently 2 to 3 times higher than catches on the NW coast.


38

3.3.2 Study species

Lutjanus bohar is a valuable component of fisheries in areas where the perceived risk of

ciguateria is low (e.g. Seychelles, Marriott and Mapstone 2006). It is recognised as a

top-order predator in most coral reef environments (Kulbicki et al. 2005; Stevenson et

al. 2007). Individuals have high longevity (maximum of 55 years), slow growth and

attain relatively large sizes (maximum fork length 78 cm in Australia; S Newman pers.

comm.). Females may mature at an older age (9 years) than males (2 years), although no

differences have been found in growth trajectories of males and females sampled from

the NE coast of Australia (Marriott et al. 2007). It has also been reported that this

species moves to deeper offshore areas as they age (Wright et al. 1986; Marriott et al.

2007).

3.3.3 Sampling methods

For the NW sites, otoliths of L. bohar were obtained from fish collected by the

Department of Fisheries (Western Australia) between 1995 and 2009 from deep-water

surveys of the outer reef slopes of the Rowley Shoals and Scott Reef (both atolls rising

from about 200 m depths; Figure 3.1) using fish traps and lines. All samples were

caught in depths ranging from 80 to 180 m, where the well-mixed surface layer extends

to around 100 m depth in winter (Wolanski and Deleersnijder 1998). The sagittal

otoliths of each fish were removed, cleaned and stored to dry in paper envelopes.

For the NE sites, samples of L. bohar were collected from the northern (Lizard Island)

and central (Townsville reefs) sections of the Great Barrier Reef (Figure 3.1) along the

NE coast of Australia between October 1995 and November 2001 (Marriott and

Mapstone 2006). Fish were collected by line fishing from 5–30 m depths, as the depths

of the inshore reefs in this region are typically less than 50m (Choukroun et al. 2010)

and are relatively well-mixed (Condie and Dunn 2006). The sagittal otoliths were

removed, cleaned and stored dry.

Sectioning methods to prepare thin sections (between 0.15–0.19 mm) from otoliths

followed standard procedures (Ferreira and Russ 1994; Marriott and Mapstone 2006;

Wakefield et al. 2015), in which one sagittal otolith was embedded in epoxy resin and

sectioned transversely through the primordium with a Buehler Isomet low speed saw.

Sections were mounted on glass slides with a cover slip using either Crystal Bond 509

adhesive (samples from NE Australia) or casting resin (samples from NW Australia).


39

To obtain a long growth chronology, I selected the otoliths of older individuals (over 15

years old) from the larger collection for increment analysis. This selection included 55

fish from both Scott Reefs and Rowley Shoals on the NW coast and 39 fish from both

Lizard Island and Townsville reefs on the NE coast (see Table 3.1 for descriptive

statistics). This may have led to biases in the growth chronology because the earlier

years of records were selectively composed of fish that had survived to an old age, and

such fish may not provide an accurate representation of the average growth pattern of

the population. However, these older fish were collected over more than a decade

(1995–2009; Table 3.1), and in order to deal with any potential bias, I included an

‘individual’ random effect in the mixed effects model to account for differences in mean

growth rates among individuals.

Figure 3.1 Sampling locations of Lutjanus bohar collected from tropical northern

Australia. Fish were collected between the years 1995–2009. Boxes indicate the environmental

grids that were used to obtain the regional environmental variables. Triangles and circles

identify centroids of sampling sites: closed triangle = Scott Reef; open triangle = Rowley

Shoals; closed circle = Lizard Island cluster; open circle = centroid of Townsville reef cluster.


40

Table 3.1 Descriptive statistics for Lutjanus bohar collected from tropical northern

Australia. NW= north-west, NE = north-east, FL = fork length at capture, Min = minimum,

Max = maximum, n = sample size. FL data from NE has two samples with missing values.

Region NW NE

Site Rowley Shoals Scott Reef Lizard Island Townsville

Fish (n) 24 31 16 23

Female fish (n) 13 17 4 10

Male fish (n) 11 14 3 2

Unknown sexes (n) 0 0 9 11

Mean FL(cm) 64.6 64.8 59.5 59.1

Min FL (cm) 50.1 49.4 46.8 48.8

Max FL (cm) 74.5 78.1 71.8 67.5

Mean age 31.0 32.6 27.4 33.7

Min age 16 19 15 20

Max age 47 41 46 55

First year collected 1997 1997 1995 1995

Last year collected 2008 2009 2000 2001

3.3.4 Microscopy methods

Imaging and increment marking methods followed the procedure described in Section

2.3.4. The region next to the sulcus on the dorsal side of each otolith section was

imaged using the same microscope, camera and software as described earlier. This

region was chosen because it was consistently the clearest region for discrimination of

annual increments (Figure 3.2a). Increment widths were measured using the same

technique and software as described earlier, with three transects drawn and the outer

edge of the opaque zone for each increment marked (Figure 3.2b). The same visual

crossdating process was used, as described in Section 2.3.4.


41

Figure 3.2 Photograph of the dorsal side of a L. bohar otolith section. (a) The dorsal section

had consistently clear annual increments and (b) a close up image of the same otolith section

with transect line and increments marked with the corresponding calendar years.


42

3.3.5 Growth increment data quality

Statistical crossdating was undertaken separately for fish from the NW and the NE of

Australia, using the same methods in the program COFECHA, as detailed in Section

2.3.5. Potential errors in increment boundary placements were visually inspected before

any corrections were made. After the statistical crossdating process, multiple transects

belonging to the same fish were averaged. For all fish, measurements from the first

seven years of growth, in which increment widths change rapidly, were excluded. This

age cut-off was also consistent with observations from a previous study showing that

most females were sexually mature at around 9 years of age, whereas the males were

sexually mature at around 2 years of age (Marriott et al. 2007). Hence, only adult data

(i.e. increments widths from 8 years of age onwards) were used for the biochronology.

The quality of the chronology was assessed using series inter-correlation (overall

average of correlations between each series and the average of all other series) and

mean sensitivity (measure of year-to-year variability among pairs of successive

increments with higher values indicating stronger variability) in COFECHA (Grissino-

Mayer 2001). In addition, I assessed the mean of pairwise series correlations (�̅�; Wigley

et al. 1984) and expressed population signal (EPS; Wigley et al. 1984) using the same

software as described in Section 2.3.5. The EPS is a function of �̅� and sample depth, and

a minimum value of 0.85 is generally accepted in dendrochronology as an indication of

the reliability of the chronology (Wigley et al. 1984).

I also calculated the bootstrap 95% confidence intervals for �̅� for the chronology of fish

from the NW coast (1963–2007) and the NE coast (1962–1999) following the methods

of Rountrey et al. (2014). Raw data were detrended using a spline with a rigidity of 0.67

of the series length (Cook et al. 1990) and the bootstrap confidence intervals were

calculated using 15-year intervals with a 14-year overlap. Periods in which the 95%

confidence intervals for �̅� did not include zero were considered to contain a synchronous

signal (Rountrey et al. 2014). For the following statistical analyses, only years from

periods shown to have a synchronous signal with a common overlap were used for fish

from both NW and NE Australia. This overlap corresponded to the years 1968 to 1999,

a 32-year period.

3.3.6 Environmental and size structure datasets

Annual mean values of the climatic indices Multivariate ENSO Index (Wolter and

Timlin 1993), Pacific Decadal Oscillation (PDO; constructed from sea surface


43

temperature anomalies over the North Pacific; Mantua and Hare 2002) and Dipole

Mode Index (DMI; difference between sea surface temperature anomalies in the western

and eastern equatorial Indian Ocean; Saji et al. 1999) were obtained from the Royal

Netherlands Meteorological Institute (KNMI) Climate Explorer (Trouet and Van

Oldenborgh 2013), a web application for accessing and analysing climate data

(http://climexp.knmi.nl). The PDO and DMI were included because of their known

influence over the circulation patterns of the tropical Pacific and Indian Ocean,

respectively (Saji et al. 1999; Newman et al. 2016a). I also obtained regional (NW or

NE coast) environmental data for sea surface temperature (SST; based on HadISST

data; Rayner et al. 2003) and precipitation (PPT; CRU TS3.23; Jones and Harris 2008)

from KNMI Climate Explorer. The grid cell values for SST and PPT were averaged for

a grid box (Figure 3.1) covering the NW coast (12°S–20°S, 116°E–124°E) and the NE

coast (12.5°S–20.5°S, 143°E–151°E). For both SST and PPT, the summer (January–

March) and winter (July–September) mean values for each year were used. Seasonal

minimum (January–February) and maximum (May–June) sea level values from

Fremantle (a well-known proxy for the Leeuwin Current strength; Feng et al. 2003),

were obtained from the Bureau of Meteorology (Australian Government) website,

http://www.bom.gov.au. I also included Darwin sea level (January–March) because sea

level during the austral summer in the Gulf of Carpentaria is known to be a contributor

to seasonality in strength of the Leeuwin Current (Ridgway and Godfrey 2015).On the

NE coast of Australia, I used the seasonal minimum (August–September) and maximum

(February–March) values of Townsville sea level as a proxy for the strength of the

current flow in the Queensland region, based on the expectation that sea levels along the

east coast will be correlated with the strength of the adjacent boundary current and

because earlier work has shown that sea level at Fort Denison is a reliable proxy of the

strength of the East Australian Current along the New South Wales coast (Holbrook et

al. 2011). Seasonal minimum and maximum values for sea level were used to indicate

the relative strengths of the average current in each year and also because growth

increment data were measured at annual scales. Sea level estimates were recorded by

tide gauges and represented heights in metres above tide gauge zero. Two indicators of

the size structure of the fish assemblage, mean maximum length of catch and region-

based marine trophic index (illustrates changes in biodiversity of large fishes over time

while accounting for the expansion and contraction of fishing fleets in each region;

Kleisner et al. 2014) were obtained from the Sea Around Us project,

http://www.seaaroundus.org/. In the absence of available studies that might more


44

accurately depict the size structure of the fish assemblages, I used reconstructed data to

examine the potential changes to the size structure of the fish assemblages in those

regions. Prior to statistical analyses, all environmental variables were normalized (mean

= 0, variance = 1) over the common period of overlap (1968–1999) in the

biochronology.

3.3.7 Statistical analyses

3.3.7.1 Regional differences in growth increments

To determine if the growth increments of L. bohar differed between the NW and NE

coast, the natural log of raw increment widths were analysed using a generalized

additive mixed model (GAMM) in the R package ‘gamm4’ (Wood and Scheipl 2015).

Three models were considered in the model selection approach, a model with a single

age smooth term and no representation of region (NW or NE), a model with a single age

smooth term and region as a fixed effect, and a model with different age smooth terms

for each region. All models included individual fish as a random effect (intercept only)

to account for individual differences and a nested random effect, year nested in region

(intercept only), to account for inter-annual environmental variability. Second-order

Akaike information criterion (AICC) was used to rank the models. Simple linear

dependence was assumed for the variables other than age because the use of smooth

terms for these variables did not improve the models.

3.3.7.2 Regional drivers of growth increments

The data for NW and NE were analysed separately in order to describe the factors

influencing the growth increments of L. bohar in each region. The respective

explanatory variables used in GAMMs are shown in Table 3.2. Model selection using

the function ‘dredge’ in package ‘MuMIn’ (Barton 2015) was applied to determine

which variables were important in explaining variations in growth. These GAMMs

included a smooth term for age (k = 6), with fish (individual) as a random effect

(intercept only). In the model selection process, I limited the number of variables to a

maximum of five (to minimize errors in parameter estimation), used AICC to rank the

models, fixed inclusion of the age smooth term so it would always be present and

ensured that collinear variables (r > |0.40| and p < 0.05) would not appear in the same

model. Models were assessed using differences in AICC (ΔAICC) and model

probabilities (Akaike weights). Models in the model selection process were fitted using

maximum likelihood estimation, whereas the top-ranked models were refitted using


45

restricted maximum likelihood estimation for unbiased parameter estimates (Zuur et al.

2009). Model validation was included for all top-ranked models identified in the model

selection process to ensure that assumptions of homogeneity, normality and temporal

autocorrelation were not violated. All statistical analyses were completed in R version

3.1.3 (R Core Team 2015).

Table 3.2 Explanatory variables used in the analyses of L. bohar from both the north-west

(NW) and north-east (NE) coast of Australia. ROW = Rowley Shoals, SCO = Scott Reef, LIZ

= Lizard Island, TVS = Townsville reefs, ENSO = El Niño-Southern Oscillation, SST = sea

surface temperature, PPT = precipitation. Mean and standard deviation (SD) values were

included for the environmental factors, climate indices and size structure indicators (for the

general fish assemblage). These values refer to the annual means (calculated from monthly

means) over the years 1968–1999. Boreal summer (January to March) averages and winter (July

to September) averages were used for the temperature and precipitation dataset. A full dataset

for sex was only available for samples from the NW coast.

Category Variable Mean SD

Intrinsic factors Age (years) 32.24 (NW)

31.15 (NE)

6.82

10.0

Sex NA NA

Geographical

factors

NW region: ROW & SCO (sites)

NE region: LIZ & TVS (sites)

NA

NA

NA

NA

Climate indices Multivariate ENSO Index (MEI) 0.208 (annual) 0.845

Pacific Decadal Oscillation (PDO) 0.219 (annual) 0.838

Dipole Mode Index (DMI) -0.01 (annual) 0.236

Environmental

factors

Summer SST (°C)

Winter SST (°C)

29.25 (NW)

28.71 (NE)

26.01 (NW)

24.60 (NE)

0.399

0.363

0.433

0.420

Summer PPT (mm per month)

Winter PPT (mm per month)

145.9 (NW)

247.6 (NE)

3.77 (NW)

22.4 (NE)

60.64

79.62

4.98

10.23

Seasonal maximum sea level (m)

Seasonal minimum sea level (m)

4.220 (Darwin)

0.851 (Fremantle)

2.027 (Townsville)

0.666 (Fremantle)

1.836 (Townsville)

0.070

0.065

0.049

0.074

0.038

Size structure

indicators

Mean maximum length of catch (mm) 61.58 (NW)

66.54 (NE)

11.04

8.70

Region-based marine trophic index 3.07 (NW)

3.03 (NE)

0.145

0.156


46

3.4 Results

3.4.1 Growth chronology statistics

Crossdating statistics for the full time series (1958 to 2007 for the NW coast and 1952

to 2000 for the NE coast) showed that both datasets had clear, synchronous patterns of

growth during the respective time periods (Table 3.3). For fish from the NW coast, the

results of the bootstrap 95% confidence interval for �̅� (Figure 3.3) showed that the

period with �̅� > 0 and expressed population signal, EPS > 0.5 spanned the years 1967 to

2007 with a minimum sample depth of seven fish. For fish from the NE coast, this

period spanned the years 1967 to 1999 with a minimum sample depth of nine fish

(Figure 3.3). Conservatively, I used a minimum sample depth of nine fish for both

regions, resulting in the common overlap period of 1968 to 1999. The fractional

common variance (�̅�) was moderate for both populations and the EPS was well

approximated (Table 3.3), indicating that the growth of all fish within each region

responded in a similar manner to environmental signals.

Table 3.3 Crossdating statistics for chronology development of L. bohar from tropical

northern Australia. Average �̅� and expressed population signal (EPS) were based on a running

window length of 15 years with a 14 year overlap for the years 1968–1999. NW = north-west,

NE = north-east.

Region NW NE

No. of series 114 98

Years 1958–2007 1952–2000

Series inter-correlation 0.328 0.315

Average mean sensitivity 0.158 0.165

Average �̅� (1968–1999) 0.234 0.235

Average EPS (1968–1999) 0.909 0.880


47

Figure 3.3 Quality of L. bohar chronologies from tropical northern Australia. Bootstrap

95% confidence interval for �̅� and expressed population signal (EPS) from (a) north-west and

(b) north-east Australia. Data were calculated using 15-year intervals such that the x-axis

consists of the midpoint of each 15-year period.


48

3.4.2 Growth chronologies on NW and NE coasts

The model with separate age smooth terms for the two regions (NW and NE) had the

smallest AICC value (Table 3.4). Hence, the fish from the two regions exhibited age-

related growth trends that were different (Figure 3.4), and this difference was better

represented as a difference in the age smooths than as an intercept term (i.e. a fixed

offset in increment width throughout life was not supported).

Table 3.4 Models that involved the age-related growth trends of L. bohar from the two

regions. Fish were from the north-west and north-east regions of Australia. A model selection

approach was used to determine the smooth terms that best represented the otolith growth data

from the two regions. All models were run with the same response variable (natural log of the

otolith growth increments) and the same intercept-only random effects: individual fish (to

account for individual variability) and year nested in region (to account for environmental

variability). DF = degrees of freedom, AICC = second-order Akaike information criterion.

Model Fixed effects DF AICC

1 Single age smooth term 7 -1183

2 Single age smooth term and Region 8 -1189

3 Separate age smooth terms (by Region) 9 -1234

Figure 3.4 Increment growth profiles of L. bohar from two regions in tropical northern

Australia. Each point represents the average increment width (µm) at each age for all fish from

the north-west (NW) and north-east (NE) coast of Australia. Error bars indicate the standard

error of means.


49

3.4.3 Factors influencing growth along NW coast

The top-ranked model for growth of fish from the NW coast (Akaike weight = 0.42, 867

models evaluated) was four times more likely than the second-ranked model (Table

3.5). In this model, positive winter precipitation (PPT) anomalies were associated with

wider growth increments, whereas positive anomalies in Darwin sea level (maximum in

summer), Pacific Decadal Oscillation (PDO) index and mean maximum length of

catches (ML) were associated with narrower growth increments of L. bohar from the

NW coast (Table 3.6; Figure 3.5).

Table 3.5 Top five ranked models in two separate model selection processes for L. bohar

from tropical northern Australia. Fish were from the north-west (NW) and north-east (NE)

regions of Australia, with data from the years 1968–1999. Models were ranked using the

second-order Akaike information criterion (AICC). Age was represented as a smooth term in the

models. DF = degrees of freedom, Ak. wt. = Akaike model weights, GI = growth increments,

SSTW = winter sea surface temperature, SSTS = summer sea surface temperature, PPTW = winter

precipitation, PPTS = summer precipitation, SLDAR = maximum Darwin sea level, SLFRE =

maximum Fremantle sea level, SLTVS = minimum Townsville sea level, DMI = Dipole Mode

Index, PDO = Pacific Decadal Oscillation, ML = mean maximum length of catches.

Region Rank Model Equation DF ΔAICC Ak. wt.

NW 1 GI ~ Age + PPTW + SLDAR + PDO + ML 9 0.00 0.42

NW 2 GI ~ Age + PPTW + SLDAR + PDO + Sex 9 2.88 0.10

NW 3 GI ~ Age + PPTW + SLDAR + PDO 8 3.77 0.06

NW 4 GI ~ Age + PPTW + SLDAR + PDO + SSTW 9 3.96 0.06

NW 5 GI ~ Age + PPTW + SLDAR + PDO + SLFRE 9 4.19 0.05

NE 1 GI ~ Age + SLTVS + DMI + ML + Site 9 0.00 0.43

NE 2 GI ~ Age + SLTVS + DMI + ML + SSTS 9 3.28 0.08

NE 3 GI ~ Age + SLTVS + SSTS + ML + Site 9 4.60 0.04

NE 4 GI ~ Age + SLTVS + DMI + ML 8 4.79 0.04

NE 5 GI ~ Age + SLTVS + DMI + PPTS + Site 9 4.91 0.04

Table 3.6 Model parameters for the top-ranked model for L. bohar from north-west

Australia. Data was from the years 1968–1999. SE = standard error, PPT = precipitation, PDO

= Pacific Decadal Oscillation, ML = mean maximum length of catches. Sea level refers to mean

Darwin sea level over the months of January to March (seasonal maximum).

Coefficients Estimate SE t-value p-value

Intercept 3.99 0.020 204.4 <0.01

Sea level -0.03 0.006 -5.27 <0.01

PDO -0.03 0.007 -4.20 <0.01

Winter PPT 0.03 0.007 3.82 <0.01

ML -0.02 0.007 -2.41 0.02


50

Figure 3.5 Schematic diagram showing the environmental drivers that influence L. bohar

populations in northern Australia. Environmental drivers are shown in decreasing order of

importance, based on absolute t-values.

3.4.4 Factors influencing growth along NE coast

The top-ranked model of growth of fish from the NE coast (Akaike weight = 0.43, 424

models evaluated) was five times more likely than the second-ranked model (Table 3.5).

The top-ranked model showed that positive anomalies in Townsville sea level

(minimum in winter), mean maximum length of catches (ML) and Dipole Mode Index

(DMI; a measure of the Indian Ocean Dipole) were associated with wider growth

increments in otoliths of L. bohar from the NE coast (Table 3.7; Figure 3.5).

Furthermore, fish from Townsville reefs had narrower growth increments compared to

fish from the Lizard Island reefs (Table 3.7).

Table 3.7 Model parameters for the top-ranked model for L. bohar from north-east

Australia. Data was from the years 1968–1999. SE = standard error, SL = sea level, DMI =

dipole mode index, ML = mean maximum length of catches. Sea level refers to mean

Townsville sea level over the months August to September (seasonal minimum).

Coefficients Estimate SE t-value p-value

Intercept 3.76 0.029 131.4 <0.01

ML 0.02 0.006 3.42 <0.01

Sea level 0.02 0.006 3.35 <0.01

Annual DMI 0.01 0.005 2.74 <0.01

Site (Townsville) -0.10 0.037 -2.66 <0.01


51

3.5 Discussion

Populations of L. bohar on the NW and NE coasts of Australia experienced different

environmental conditions that influenced the widths of otolith increment (a proxy of

somatic growth rates) throughout the life span of the fish we sampled. The results

showed that extrinsic factors including climate, regional-scale environmental variables,

size structure of fish assemblages and location were important influences on growth

trends.

Different climate indices influenced the growth of NW and NE populations of L. bohar.

The Pacific Decadal Oscillation (PDO), an index of sea surface temperatures (SST) in

the northern Pacific, was an important factor for fish from the NW coast, whereas the

Dipole Mode Index (DMI), largely driven by SST in the Indian Ocean, was important

for fish from the NE coast. This result seems counter-intuitive given that the NW

population is on the margin of the Indian Ocean while the NE population is on the

margin of the Pacific Ocean. However, the PDO signal that oscillates over decadal time

scales (low frequencies) has also been regarded as a reddened response to ENSO where

the ENSO signal re-emerges in the subsequent year (Newman et al. 2003; Shakun and

Shaman 2009; Newman et al. 2016a). Hence, a recent review concluded that the PDO

represents a combination of processes that originate from the tropics (Newman et al.

2016a). ENSO events in the tropical Pacific are acknowledged to have strong influences

on oceanographic conditions (Feng et al. 2009) and the recruitment of marine fauna

(Caputi et al. 1996) along the coast of Western Australia. As L. bohar is a long-lived

species, it is possible that low frequency anomalies (on decadal time scales) in the

tropical Pacific, via the Indonesian through-flow (Meyers 1996), could influence the

growth of the NW population (Chen and Wallace 2015). Alternatively, the growth of

the NW population is responding to the PDO because it is a delayed ENSO signal,

meaning that the ENSO signal only appears in the growth of L. bohar a year later.

Most studies of DMI relate to its influence on rainfall in eastern Australia (e.g. Cai et al.

2009; Risbey et al. 2009), although in the results precipitation was not a factor

identified by the model selection process. Consequently, the relationship of growth of L.

bohar with the DMI might be due to other factors that are linked to SST in the Indian

Ocean, such as the strength of the Pacific trade winds (Alory et al. 2007) or thermocline

variability in the western Pacific (Wijffels and Meyers 2004). The former may generate

wind-driven counter currents that could increase productivity within the otherwise

oligotrophic East Australian Current on the NE coast, similar to the situation on the NW


52

coast (Hanson et al. 2005). On inter-annual timescales, the Indonesian region transmits

thermocline variability from the western Pacific into the Indian Ocean (Wijffels and

Meyers 2004). Hence, correlations of the NE population to the DMI might also be

related to changes in the thermocline depth of the western Pacific, which could

influence vertical mixing, productivity and ultimately, growth.

Sea level (SL) influenced the growth of L. bohar in both regions. However, the timing

of correlations differed (summer for NW and winter for NE) as did responses, with

increases in SL associated with a decrease in growth of the NW population, whereas

growth of the NE population increased. Along the NW coast of Australia, it has been

found that increases in SL lead to an increased flow of warmer and lower productivity

waters (Pearce et al. 2003). Given that SST was not significantly correlated with

growth, it may be that changes in productivity caused by variability in SL have a

stronger influence on the growth of L. bohar on the NW coast than water temperature.

For example, other studies to the south of the NW study site have shown that an

increase in flow of the dominant current (as indicated by the increase in SL) in summer

suppressed a counter current (the Ningaloo Current) and hence local upwelling events

(Hanson et al. 2005). Thus lowered productivity may have reduced the growth of L.

bohar along the NW coast because of flow-on effects up the food web. On the other

side of the continent, L. bohar reside in the vicinity of the Coral Sea, with a complex

bathymetry consisting of numerous reefs, islands, channels and ridges (Choukroun et al.

2010; Wolanski et al. 2013). Oceanic circulation in the Coral Sea is strongly influenced

by wind-driven currents, notably the strong south-easterly winds that dominate during

the Austral winter. The results suggest that the strong wind-driven currents in winter (as

shown by increased SL) were associated with larger growth increments of L. bohar off

the NE coast. Since SST again did not seem to be an important factor, changes in

productivity might play a more important role for the association between winter

currents and growth of the NE population. For example, stronger boundary currents

associated with a shallower coastal thermocline are thought to lead to enhanced primary

production and increases in abundance anchovy (Engraulis spp.) off the coasts of both

California and Peru (Chavez et al. 2003). A similar mechanism might explain the

positive association between current strength and the growth of L. bohar along the NE

coast.

Responses to mean maximum length of catches (ML) also differed between the NW and

NE populations of L. bohar. Years with lower ML (i.e. fish assemblages had smaller


53

sizes), seemed to be associated with an increase in growth increments for L. bohar in

the NW coast (although ML was not as important as the other factors), whereas the

opposite response occurred on the NE coast. One possible explanation for these

differences could be the varying levels of fishing pressure in the two populations.

Overall, fishing mortality has historically been greater off the NE than the NW coasts

(Pauly and Zeller 2016). Fishing mortality can be size-selective, causing changes in the

demographic structure (growth rates and recruitment) of affected populations (Law

1991; Conover and Munch 2002). Another well-documented effect of fishing is the

removal of top-order predators (e.g. DeMartini et al. 2008), which may influence

condition (and growth rates) of L. bohar in those fish communities (Walsh et al. 2012).

However, the exact mechanisms of how the growth of L. bohar populations are affected

by the changes in size structure of the fish assemblages are unknown because of the

complex food webs and indirect demographic effects (e.g. Ruttenberg et al. 2011).

Studies of the relative rates of fishing intensity and diet of this species might help to

explain these findings.

The geographical factor site was important for fish from the NE coast but not for fish

from the NW coast. This suggests that environmental conditions influencing L. bohar

on the NW coast were similar and were acting at a spatial scale that included both sites

on the NW coast. However, site (Townsville/Lizard Island) had an effect on the otolith

growth of fish from NE coast, with L. bohar from Townsville reefs having generally

smaller increments than those from Lizard Island. This possibly reflects the differing

cross-shelf positions of these reefs and the oceanographic environments that surround

them. For example, water temperatures were consistently warmer at Lizard Island than

on Townsville reefs and the former reefs are also closer to the coast, thus will be more

strongly influenced by terrestrial inputs (Marriott 2005). A previous analysis of this

dataset found that average lengths of the most frequently sampled age class (4+ years)

of fish from the Townsville reef cluster were significantly higher than those from the

Lizard Island cluster, although any evidence for a significant difference in the fit of the

von Bertanlanffy growth model to length-at-age data from the different sites was

equivocal (Marriott 2005). Distance across the shelf is also known to be an important

driver of growth in reef fishes (Gust et al. 2002; Kingsford and Hughes 2005; Williams

et al. 2007). In addition, the complex bathymetry along the NE coast (mentioned above)

is likely to result in varying environmental conditions throughout the region, hence


54

contributing to the result of between-site differences in growth patterns of the NE

population.

Both populations (NW and NE) of L. bohar responded to climate signals in their

respective regions. This suggests that the minimum sample depth of nine fish

contributing to each year value of otolith increment width and the sample sizes attained

in this study were sufficient for the analyses of growth patterns. Furthermore, the

sensitivity of this species to climatic influences reveals the suitability of L. bohar as a

potential indicator of climate change (notably in terms of productivity changes and low

frequency anomalies in SST). These conclusions must be viewed in light of the caveat

that fish were sampled from different depths in the NW (80–180 m) and NE (5–30 m)

regions. This was an unavoidable artefact of the bathymetry in these areas, with

offshore reefs sampled along the NW coast, whereas the reefs along the NE coast were

part of an extensive network of the Great Barrier Reef that lies on the shallow coastal

shelf. However, such differences in habitat structure were unlikely to have a major

effect on growth patterns, since the study species is a wide-ranging top-order predator in

coral reef systems (Kulbicki et al. 2005; Stevenson et al. 2007). Tagging work at the

Scott Reefs and Rowley Shoals (Meekan et al. unpublished data) shows that L. bohar

displays limited site-attachment and moves over areas encompassing several kilometres

and depth ranges, typically occurring in depths of 10–70 m (Sommer et al. 1996).

Because of the mobility of this species, depth is unlikely to strongly confound the

results since, unlike very territorial species and herbivorous fishes where depth may

strongly influence productivity of algal food sources, growth patterns are more likely to

be integrated across depth ranges. This is particularly the case in the well-mixed reef

waters that surround coral reefs on both the NW and NE coasts.

In summary, I have shown that the drivers of growth patterns differed across the range

of a tropical predatory fish, thus a single age-related growth trend should not be

assumed for separate populations of this species. These findings also suggest that the

responses of species to environmental factors at both large and local spatial scales are

complex and likely to be strongly influenced by the local oceanographic and biological

context of reef systems. Generalizations about species responses to climate change will

require adequate characterisation of this variability, especially across different ocean

basins.


55


I am grateful to Professor Paul Rigby and Miss Alysia Buckley for the generous help

and advice they have provided for the microscopy work. This work was funded by the

Australian National Network in Marine Science, the Australian Institute of Marine

Science and the Center for Marine Futures at the University of Western Australia. This

work was made possible with a scholarship from the Australian Postgraduate Awards. I

would like to thank Steve Newman and Ross Marriott for their assistance in providing

the samples and for their helpful comments about the methods and results.

57

Chapter 4 Climate-driven synchrony of

marine and terrestrial ecosystems

4.1 Abstract

The effects of climate change are difficult to predict for many marine species because

little is known of their response to climate variations in the past. However, long-term

chronologies of growth, a variable that integrates multiple physical and biological

factors, are now available for several marine taxa. These allow us to search for climate-

driven synchrony in growth across multiple taxa and ecosystems, identifying the key

processes driving biological responses at very large spatial scales. I hypothesized that in

north-west (NW) Australia, a region that is predicted to be strongly influenced by

climate change, the El Niño-Southern Oscillation (ENSO) phenomenon would be an

important factor influencing the growth patterns of organisms in both marine and

terrestrial environments. To test this idea, I analysed existing growth chronologies of

the marine fish Lutjanus argentimaculatus, the coral Porites spp. and the tree Callitris

columellaris and developed a new chronology for another marine fish, Lethrinus

nebulosus. Principal components analysis and linear model selection showed evidence

of ENSO-driven synchrony in growth among all four taxa at inter-annual time scales,

the first such result for the Southern Hemisphere. Rainfall, sea surface temperatures and

sea surface salinities, which are linked to the ENSO system, influenced the annual

growth of fishes, trees and corals. All four taxa had negative relationships with the

Niño-4 index (a measure of ENSO status), with positive growth patterns occurring

during strong La Niña years. This finding implies that future changes in the strength and

frequency of ENSO events are likely to have major consequences for both marine and

terrestrial taxa. Strong similarities in the growth patterns of fish and trees offer the

possibility of using tree-ring chronologies, which span longer time periods than those of

fish, to aid understanding of both historical and future responses of fish populations to

climate variation.

4.2 Introduction

Research efforts focused on the effects of climate change on organisms in both

terrestrial and marine ecosystems (Rosenzweig et al. 2008; Hoegh-Guldberg and Bruno

2010) have mostly examined single species or groups of species in common

Chapter 4: Climate-driven synchrony of marine and terrestrial ecosystems

58

environments. Although it is recognised that terrestrial and marine ecosystems are

intimately linked (e.g. Dai and Wigley 2000), the isolated nature of many studies means

that the effects of a climate phenomenon across different ecosystems have not been

fully explored. Our understanding of these connections has been further hampered by a

lack of long-term (decades to centuries) records of the responses of marine taxa to

climate variations (Rosenzweig et al. 2008; Richardson et al. 2012). Chronologies of

growth are now being developed for an expanding suite of marine organisms including

corals, molluscs and fishes, all of which have annual cycles of growth within their hard

parts (see review by Morrongiello et al. 2012). These chronologies provide powerful

insights into the effects of climate change, since growth is a variable that integrates the

effects of multiple physical and biological factors (Morrongiello et al. 2012) and these

taxa are relatively long-lived (typically many decades).

Initial attempts to compare growth of taxa across ecosystems have shown evidence for

links between oceanic/atmospheric variation and growth, with some studies revealing

climate-driven synchrony in growth across multiple taxa. For example, the growth of

freshwater fish and trees were correlated in the United States because of similar

responses of these taxa to rainfall and river discharge (Guyette and Rabeni 1995).

Synchronous growth patterns of trees, marine fish and bivalves in the north-east Pacific

have been linked to ENSO through the influence this phenomenon has on sea surface

temperatures (SST), land temperatures and precipitation (Black et al. 2009). An

understanding of the factors and mechanisms that drive such linkages provides us with

an improved capacity to hind- and forecast the effects of climate change on the growth

of aquatic organisms.

Additionally, growth chronologies derived from taxa that are sensitive to climate

variations can be utilised to reconstruct past patterns of climate. In Australia, long-term

(multi-decadal) growth records from trees and corals have been used to extend records

of rainfall (Cullen and Grierson 2009; Lough 2011; O'Donnell et al. 2015) and SST

(Hendy et al. 2002; Zinke et al. 2014; Zinke et al. 2015) to times prior to instrumental

records. Where connections between oceanic and atmospheric processes lead to

synchronous growth responses among marine and terrestrial taxa, multi-proxy

reconstructions of broad-scale climate phenomena can be developed. For example, tree

and coral growth increments and ice core stratigraphy spanning the Pacific basin have

been found to be synchronously responsive to the influence of the ENSO phenomenon

on regional temperatures and precipitation. These chronologies were subsequently used


59

to develop a robust, multi-proxy reconstruction of ENSO variability over the last ~450

years (Braganza et al. 2009). Such reconstructions have greatly extended instrument

records and furthered our knowledge of the amplitude and frequency of variation in

climate through time.

Linked biological responses of taxa across terrestrial and marine ecosystems could also

enable the use of terrestrial chronologies (which are generally available over longer time

scales than marine records) as proxies for estimating the likely responses of marine taxa

to climate change. For example, synchrony in the growth of trees, marine fish and the

breeding success of seabirds has been linked to the influence of sea level pressure on

upwelling and precipitation in the north-east Pacific (Black et al. 2014). This strong

connection between oceanic and atmospheric processes has enabled the use of growth

chronologies from trees to develop a robust ~600-year reconstruction of upwelling

intensity (California Current Winter Index) along the California coast (Black et al.

2014). Similarly, other coastal ecosystems with strong links to atmospheric processes

that influence trees may benefit from this method of hindcasting historic ecosystem

states beyond available instrumental records.

Here, I present the first regional comparison of the climatic drivers of the growth of

fishes, corals and trees from the Southern Hemisphere. I focus on the marine and

terrestrial environments of north-west (NW) Australia. Western Australia (WA) has

been identified as a potential ‘hotspot’ of climate change (Pearce and Feng 2007), where

water temperatures along the NW coast are predicted to increase by more than 2°C by

the year 2055 (Cheung et al. 2012a). In this region, large–scale drivers (i.e., over

hundreds to thousands of kilometres) such as the ENSO interact with regional Indian

Ocean processes to influence the marine environment on the NW coast (Marshall et al.

2015; Zinke et al. 2015). The combination of these interactions can result in phenomena

such as the ‘Ningaloo Niño’, an anomalous warming of surface waters that has caused

widespread fish kills and coral bleaching (Feng et al. 2013).

Long-term growth chronologies have already been developed from trees (O'Donnell et

al. 2015), corals (Cooper et al. 2012) and fish (Chapter 2) in this region, providing an

opportunity to investigate linked biological responses to climate patterns across taxa and

ecosystems. These earlier studies have revealed that growth of trees in NW Australia

show a strong positive response to rainfall because water is a limiting resource

(O'Donnell et al. 2015). Similarly, coral growth is influenced by regional changes in

SST that affect calcification rates (Cooper et al. 2012), while adult fish respond to


60

changes in SST and sea surface salinity (SSS) because of changes in metabolic rates,

osmoregulation or food conversion efficiencies (Chapter 2). Given that ENSO drives

regional environmental and climate variables such as SST, SSS and rainfall in

Australia’s NW region, I hypothesized that the growth of these taxa will exhibit similar

patterns. Additionally, I identify the key environmental variables driving patterns in

growth among taxa.

4.3 Methods

4.3.1 Environmental drivers of marine and terrestrial regions

The NW coast of Australia includes two major marine bioregions (as defined by

Fletcher and Santoro 2014): the North Coast, which includes coastal areas of the Pilbara

and Kimberley regions, and the more southerly Gascoyne Coast from Exmouth Gulf to

Shark Bay (Figure 4.1). The warm, low salinity waters off the North Coast are of

Pacific origin, entering the region via the Indonesian through-flow and interacting with

waters of the Indian Ocean (Meyers 1996). The North Coast bioregion is entirely

tropical while the Gascoyne Coast bioregion is subtropical and is a transition zone

between the tropics to the north and the temperate zone to the south (Fletcher and

Santoro 2014). The marine environment off the Gascoyne Coast is influenced by the

Leeuwin Current, a pole-ward flowing, eastern boundary current (Cresswell and

Golding 1980; Feng et al. 2009) that transports warm tropical waters southwards along

the coast of WA (Fletcher and Santoro 2014) and is strongly influenced by ENSO on

inter-annual time scales (Feng et al. 2009). In the tropical marine waters of the North

Coast, SST in summer averages 28.8°C with a maximum of ca. 30°C while average SST

in winter drops to a monthly minimum of ca. 24°C (1970–2010 seasonal averages;

Rayner et al. 2003). In this region, the intra-annual variability of SSS is low, with an

average of 34.8 PSU (practical salinity units) and a range of ~0.3 PSU (1970–2010

seasonal average; Good et al. 2013). In the Gascoyne Coast region, average SST in

summer is slightly lower than the North Coast (25.2°C; range of ~1.1°C) while SSS is

slightly higher with an average of 35.4 PSU and range of 0.3 PSU. Both the North Coast

and the Gascoyne Coast are seasonally influenced by summer tropical cyclones

(Fletcher and Santoro 2014) and the Kimberley region of the North Coast, in particular,

is affected by river outflows from summer rainfall (Lough 1998).

In the semi-arid and arid terrestrial environments of NW Australia, biological processes

are principally driven by rainfall (Cullen et al. 2008). This is shown by strong


61

correlations of the growth of Callitris columellaris trees with rainfall and humidity

(Cullen and Grierson 2007; Cullen et al. 2008; O'Donnell et al. 2015). In NW Australia,

rainfall is extremely variable both within and among years. Most rain falls during the

summer months (average of 102 mm per month from January to March over the years

1970–2010; Jones and Harris 2008) and is associated with tropical cyclones or rain-

bearing low pressure systems (Gentilli 1971). In contrast, the austral winter to spring

months of June to November average only 12 mm per month (data from 1970–2010;

Jones and Harris 2008).

Figure 4.1 Sampling locations of growth chronologies for four taxa in north-west

Australia. Chronologies were for the period from 1984 to 2003. LA = Lutjanus

argentimaculatus, LN = Lethrinus nebulosus, all corals were Porites spp. and trees were

Callitris columellaris. L. argentimaculatus locations are approximate sampling areas within the

boxes. Terrestrial regions follow the Interim Biogeographic Regionalization for Australia

(IBRA) version 7, modified from the Department of Environment (Australian Government).


62

4.3.2 Growth chronologies

Growth chronologies from Lutjanus argentimaculatus, Porites spp. and C. columellaris

were obtained from earlier studies as mentioned above (Table 4.1; Figure 4.2). These

were supplemented with a new growth chronology developed from otoliths of another

tropical fish, the spangled emperor (Lethrinus nebulosus). For all species, we only used

data for the years 1984 to 2003, which were common to chronologies from all taxa. The

quality of the chronologies was assessed using the mean of pairwise series correlations

(�̅�) and expressed population signal (EPS) using the same software as described in

Section 2.3.5.

Table 4.1 Growth chronologies of fishes, coral and trees from north-west Australia. LA =

Lutjanus argentimaculatus, LN = Lethrinus nebulosus, all corals were Porites spp. and trees

were Callitris columellaris.

Taxa Type of data Length of

chronology

Location

(sample size)

Source

Fish (LA) Annual growth increments

from otoliths

1975–2003 Kimberley (15)

Pilbara (15)

Gascoyne (6)

Chapter 2

Fish (LN) Annual growth increments

from otoliths

1984–2003 Gascoyne (23) This study

Coral Annual calcification rate

from coral cores

1900–2010 Clerke Reef (5)

Imperieuse Reef (4)

Bundegi (4)

Tantabiddi (7)

Coral Bay (4)

Cooper et al.

2012

Tree Ring-width chronology 1802–2012 Hamersley Range,

inland Pilbara (27)

O'Donnell et al.

2015


63

Figure 4.2 Growth increments of three diverse taxa from north-west Australia. Sections

were obtained from the otoliths of a fish, tree core and coral core.

4.3.2.1 Growth chronology of Lethrinus nebulosus

Archived collections of the otoliths of L. nebulosus were obtained from the Department

of Fisheries (Government of Western Australia). These otoliths came from fish

collected in the Gascoyne Coast region of WA (Figure 4.1) from 2006–2010 (Marriott

et al. 2010). The sagittal otoliths of each fish were cleaned and one otolith was

embedded in epoxy resin. Two to three thin transverse sections were made near the

primordium in a direction perpendicular to the sulcus acusticus with a low speed saw

containing a diamond-wafering blade, following the methods of Marriott et al. (2010).

The sections were then washed by agitating in 2% hydrochloric acid for up to 10

seconds (to remove calcium build-up), followed by rinsing in water. Dry sections were

then mounted on microscope slides using casting resin.


64

The otoliths from 23 fish aged 24–32 years old with sufficiently clear increments were

used for image analysis. The region next to the sulcus acusticus on the dorsal side of

each otolith was imaged using an Aperio Scanscope Digital Slide Scanner (Leica

Biosystems, Nussloch, Germany) with a motorized stage system. Images were captured

using transmitted light with a 20x objective. Increment widths were measured on the

otolith images using a plugin (“IncMeas”; Rountrey 2009) written for ImageJ, an open

source image processing program (version 1.48, National Institutes of Health,

Maryland, USA). Two to three transects parallel to the growth axis were drawn, and the

outer edge of the opaque zones were marked (along the transects) from the edge of the

otolith to the core. The calendar years were also recorded for each marked increment by

working backwards from the date of capture and taking into consideration the timing of

completion of the opaque zone (austral summer; Marriott et al. 2010), as part of the

visual crossdating process. Crossdating assumes that the environment induces

synchronous, time-specific growth patterns that can be matched among individuals

(Fritts 1971; Gillanders et al. 2012). Averages of increment widths from the multiple

transects per sample were calculated and used if the inter-transect correlations were

greater than 0.9. Statistical crossdating was used to check the correct assignments of

calendar years to increments (Black et al. 2005) and any errors were visually inspected

before measurements were changed.

To produce the overall chronology, the increment widths were aligned by fish age and

the mean increment width at each age was calculated, following the methods of Black et

al. (2013). Each series was then divided by the mean-by-age series to obtain

standardized series that removed ontogenetic trends, and the standardized series were

averaged by calendar year to create a single overall chronology (see Figure 4.3 for raw,

detrended and averaged series). Only years with a sample depth of more than eight fish

(1984–2003) were used for analysis. EPS and �̅� were calculated using only one time

series for each individual fish for the period from 1984–2003.

4.3.2.2 Growth chronology of Lutjanus argentimaculatus

I used existing detrended (ontogenetic trends removed by dividing the raw series with

the mean-by-age series) growth increment series for 36 L. argentimaculatus that were

collected between 1996 and 2005 at various sites along the NW coast (Figure 4.1). The

detrended increment series from the 36 fish were averaged to obtain a single growth

chronology. The chronology consisted of increment data from 1975 to 2003 with a

sample depth of at least 20 fish contributing to each year value (Chapter 2).


65

4.3.2.3 Coral growth chronology

The coral chronology was a record of annual calcification (calculated as the product of

linear extension and skeletal density; Lough and Cooper 2011) from 24 cores of Porites

spp. (Cooper et al. 2012) collected between October 2008 and September 2010 from

five reefs (Table 4.1) along the NW coast (Figure 4.1). Available data spanned the

period 1900 to 2010. To obtain a standardized growth index, the annual calcification

rates were normalized by first subtracting the mean for the period 1961–1990 and

subsequently dividing by the standard deviation of this period. Normalized calcification

rates were calculated for each of the 24 coral cores from all five reefs. The 24 time

series were averaged to obtain a single coral chronology for the NW coast.

4.3.2.4 Tree-ring chronology

I used a ring-width chronology developed from 27 C. columellaris trees (O'Donnell et

al. 2015) from the Hamersley Ranges of the inland Pilbara region (Figure 4.1). The

chronology had been detrended using the signal-free method (Melvin and Briffa 2008)

to improve the retention of medium frequency (representing time-scales of decades to a

century) variance, reduce trend distortion at the ends of the chronology and remove age-

related trends (O'Donnell et al. 2015). The ring-width chronology covered the period

1802–2012 and was constructed using 41 series from the 27 trees.


66

Figure 4.3 Raw and detrended increment width time-series from the otoliths of Lethrinus

nebulosus collected in north-west Australia. (a) Raw increment width time-series from all

transects of the 23 fish used and (b) detrended increment width time-series and the final

chronology (in black line) with the associated standard error of mean.


67

4.3.3 Climatic and environmental datasets

The results from Chapter 2 shows that ENSO (represented by the Niño-4 index) and

SSS are important drivers of the growth of L. argentimaculatus, while coral growth has

been correlated with decadal trends in SST (Cooper et al. 2012). The growth of C.

columellaris trees in the Pilbara mainly responds to rainfall in the austral summer from

December to May (Cullen et al. 2008; O'Donnell et al. 2015). I compared growth

patterns to the Niño-4 index (based on SST in the Western Pacific between 5°N–5°S

and 160°E–150°W; Rayner et al. 2003), SST (HadISST; Rayner et al. 2003), SSS

(Good et al. 2013) and rainfall (Jones and Harris 2008). All environmental data were

obtained from the Royal Netherlands Meteorological Institute (KNMI) Climate

Explorer (Trouet and Van Oldenborgh 2013), a web application for climate data

(http://climexp.knmi.nl). The SST, SSS and rainfall values were averaged for a grid box

covering the NW coast from the Kimberley south to Shark Bay (14°S–28°S, 110°E–

127°E). For each environmental variable, January to March averages were used because

the growing season for fishes, corals and trees in NW Australia usually occurs in the

austral summer (Chapter 2; Lough and Barnes 2000; O'Donnell et al. 2015

respectively). In addition to the January to March averages for each regional

environmental variable from 1984–2003, I also used the previous year’s values (i.e.

1983–2002) for SST, SSS and rainfall from the same grid, and for the Niño-4 index to

allow for possible lagged responses. Austral winter (June to August) SST values were

used in the higher resolution spatial correlation maps detailed below.

4.3.4 Data analyses

All four chronologies were standardized (μ = 0, σ2 = 1) and analysed using principal

components analysis (PCA). The scores for the principal components that accounted for

the majority of the variance (PC1 and PC2) were tested for significant correlations

(using Pearson’s correlation) with current and lagged Niño-4 index. The principal

component scores were subsequently included as response variables in linear regression

models to assess the importance (based on information-theoretic methods) of current

and previous year’s SST, SSS and rainfall as drivers of growth. The rainfall values were

square root transformed (due to the large range of values from 30–200 mm per month)

before insertion into the linear models used in the model selection process, to satisfy the

assumptions of homogeneity for linear models. Collinearity between all six

environmental variables (|r| > 0.5, p < 0.01) was evaluated. The R package ‘MuMIn’

(Barton 2015) was used for model selection using the second-order Akaike information


68

criterion (AICC) based on Kullback-Leibler information loss and accounting for small

sample sizes (Burnham and Anderson 2004). Differences in AICC values (ΔAICC) were

used to assess the different models. Adjusted R2 values, F-statistic, t-statistic and p-

values were reported. Model validation was carried out to ensure that the models

conformed to the assumptions of linear models and tested for autocorrelation. All

statistical analyses were completed in R version 3.1.3 (R Core Team 2015). After the

model selection process, spatial correlation maps of the significant regional variables

were made in the web application KNMI Climate Explorer to show the relationships at

a higher spatial resolution.

4.4 Results

4.4.1 Chronology statistics

The growth chronology of L. nebulosus included the years from 1984 to 2009 (Figure

4.3). Measurements from more than eight fish contributed to each yearly value, with 22

out of the 23 fish contributing to the period between 1988 and 2003. Although the

fractional common variance (�̅� = 0.14) and EPS value (0.78) were low relative to tree-

ring data, indicating that variability among individuals was high, the mean chronology

from 1984 to 2003 did relate to environmental variables as evidenced by significant

correlations with January to March SST around the northern Gascoyne Coast (21°S–

23°S, 112°E–115°E; r = 0.60, p = 0.005) and marginally significant correlations with

average rainfall from January to March over the entire NW area (r = 0.44, p = 0.05).

The published chronology of L. argentimaculatus from 1975 to 2003 had �̅� = 0.153 and

EPS = 0.84 for the entire period (Chapter 2). Bootstrapped �̅� and EPS values (Rountrey

et al. 2014) were calculated for the 24 coral cores of Porites spp. for all possible 15-year

intervals from 1950–2003 (Figure 4.4) and showed that there was weak but significant

synchronicity among corals from the year 1980 onwards (�̅� ~ 0.05, EPS ~ 0.6). The

published ring-width chronology of C. columellaris trees had a running �̅� (greater than

0.4) and EPS (greater than 0.85) for 51-year intervals with 25 year overlaps (O'Donnell

et al. 2015).


69

Figure 4.4 Quality of chronology for coral Porites spp. collected in north-west Australia.

Bootstrapped �̅� (with 95% confidence intervals) and expressed population signal (EPS) for the

24 coral cores collected over the years 2008–2010.

4.4.2 Principal components analysis

The standardized growth chronologies of all four taxa (Figure 4.5) from 1984 to 2003

were analysed using a PCA. The first principal component (PC1) accounted for 41% of

the variance and PC2 accounted for 33%. The third and fourth principal components

each accounted for less than 15% of the variance and were not included in any further

analyses. Three of the taxa (fishes and trees) had similar negative loadings on PC1

(Table 4.2), indicating the similarities in growth patterns of these three taxa (Figure

4.5a). The coral series had the strongest loading on PC2, followed by L.

argentimaculatus (Table 4.2), with Figure 4.5b showing the strong synchrony between

the coral series and PC2. Inverse values of both PC1 (PC1inv) and PC2 (PC2inv) were

used in further analyses because the strongest loadings were negative (Table 4.2).


70

Table 4.2 Loadings of the four taxa on the first and second principal component (PC)

scores. Chronologies of the four taxa (two fishes, one coral and one tree) included the years

1984 to 2003.

Taxa Loading on PC1 Loading on PC2

Fish (Lethrinus nebulosus) -0.60 +0.29

Tree (Callitris columellaris) -0.58 +0.31

Fish (Lutjanus argentimaculatus) -0.54 -0.47

Coral (Porites spp.) -0.13 -0.78

Table 4.3 Pearson’s correlation matrix of environmental variables. The six variables are

from a grid covering the north-west Australian coast (14°S–28°S, 110°E–127°E). Each variable

consists of January to March averages from the years 1984–2003. SST = sea surface

temperature, SSS = sea surface salinity, lag = data with a one year lag, * represents p-values <

0.05, ** represents p-values < 0.01.

Environmental variables SST lag SSS SSS lag Rainfall Rainfall lag

SST 0.4 <0.1 <0.1 0.2 0.4

SST lag 1 -0.6** <0.1 0.5* 0.2

SSS -0.6** 1 0.5* -0.4 -0.4*

SSS lag <0.1 0.5* 1 -0.2 -0.4

Rainfall 0.5* -0.4 -0.2 1 <0.1

Rainfall lag 0.2 -0.4* -0.4 <0.1 1


71

Figure 4.5 Growth chronologies of four taxa with the respective leading principal

component (PC) scores. Chronologies of four taxa were detrended and standardized (μ = 0, σ2

= 1). (a) PC1inv with L. argentimaculatus, L. nebulosus and C. columellaris chronology and (b)

PC2inv with L. argentimaculatus and Porites spp. chronology. The inverse PC scores were used

because the stronger loading taxa were negatively loaded on both PC1 and PC2.

4.4.3 Relationships with ENSO

PC1inv was negatively correlated with the Niño-4 index (average January to March

values) with no lag (r = -0.65, p = 0.002; Figure 4.6a, b). PC2inv was negatively

correlated with the Niño-4 index (average January to March values) in the previous

years (r = -0.52, p = 0.02; Figure 4.6c, d).

72

Figure 4.6 Relationship between principal component (PC) scores and the Niño-4 index. PC scores were constructed from the growth chronologies of four

taxa (two fishes, one coral and one tree) in north-west Australia and the Niño-4 index was calculated from the average of January to March values. (a) PC1inv and

the Niño-4 index over the same years; (b) regression plot of PC1inv and Niño-4 index; (c) PC2inv and the lagged Niño-4 index (average January to March values

from the previous year) and (d) regression plot of PC2inv with lagged Niño-4 index. The inverse PC scores were used because the stronger loading taxa were

negatively loaded on both PC1 and PC2.


73

4.4.4 Relationships with environmental variables

Because of some collinearity among the six environmental variables (|r| > 0.5; Table

4.3) and the low number of observations (n = 20), models using a maximum of two non-

collinear variables were constructed. These 17 models (Table 4.4) were evaluated in the

model selection process for PC1inv and PC2inv separately. The model selection process

involving PC1inv and the 17 possible combinations of environmental variables found

that the first-ranked model (i.e. lowest AICC; Table 4.4) was one that related PC1inv with

rainfall and SST from the current year (Table 4.5). This first-ranked model was

considered to be substantially better than the second model (ΔAICC = 8.7; Table 4.4).

The linear model relating PC1inv with rainfall and SST from the current year explained

70% of the variation in PC1inv (Table 4.5), which largely reflected the growth of fishes

and C. columellaris trees. In this linear model, both variables were highly significant (p

< 0.01) with rainfall having a positive t-value of 4.92 and SST a positive t-value of 3.72.

Spatial correlation maps (using higher resolution environmental variables) show the

positive relationship between PC1inv and these two significant variables, rainfall and

SST from the current year (Figure 4.7).

The second model selection process involving PC2inv and the 17 possible combinations

of environmental variables identified a first-ranked model (Table 4.4) that related PC2inv

with SSS and rainfall from the current year (Table 4.5). This first-ranked model was not

considered to be significantly better than the second model that only included SSS

(ΔAICC = 1.5; Table 4.4), hence I chose initially to keep both variables. The linear

model relating PC2inv with SSS and rainfall from the current year explained 44% of the

variation in PC2inv (Table 4.5), however, SSS was the only significant variable (t = -

4.07, p = 0.0008). PC2inv (mainly reflecting variation in growth of corals) had a negative

relationship with SSS (Figure 4.7c). A spatial correlation map for PC2inv and SST from

June to August of the previous year (using higher resolution environmental variables)

also showed a strong positive relationship between PC2inv and offshore waters along the

NW coast, in addition to the waters around the Indonesian region (Figure 4.7d). Local

environmental variables and the links to ENSO, which influenced all four taxa, are

shown in Figure 4.8.


74

Table 4.4 Models incorporated in the model selection process. Each model had a maximum

of two non-collinear explanatory variables. Explanatory variables consist of January to March

averages from the years 1984 to 2003. Response variables are the inverse of the first two

principal components (PC) scores from the growth chronologies of four taxa (two fishes, one

coral and one tree) in north-west Australia. ΔAICC = difference in second order Akaike

information criterion, SST = sea surface temperature, SSS = sea surface salinity, lag = data with

a one year lag. Model 1 shows the first ranked model for PC1inv and model 2 is the first ranked

model for PC2inv.

Model Explanatory variables ΔAICC for PC1inv ΔAICC for PC2inv

1 SST + Rainfall 0.00 13.40

2 Rainfall + SSS 11.89 0.00

3 SST + SSS 17.67 3.80

4 Rainfall + SSS lag 11.47 12.30

5 SST lag + SSS lag 12.99 9.10

6 SST lag + Rainfall lag 13.00 9.60

7 SST + SSS lag 17.67 12.50

8 SST + Rainfall lag 17.13 13.50

9 SSS lag + Rainfall lag 24.97 10.00

10 Rainfall + Rainfall lag 11.82 13.50

11 SSS 21.81 1.50

12 Rainfall 8.75 10.48

13 SST 14.56 10.42

14 SSS lag 23.22 9.52

15 Rainfall lag 23.00 10.48

16 SST lag 9.84 7.13

17 Intercept only 20.43 7.86

Table 4.5 Selected first-ranked linear models. Response variables are the inverse of the first

two principal components (PC) scores from the growth chronologies of four taxa (two fishes,

one coral and one tree) in north-west Australia. Environmental variables are January to March

averages and chronologies are from the years 1984 to 2003. SST = sea surface temperature, SSS

= sea surface salinity.

Model equation Adjusted R2 F-statistic Model p-value

PC1inv ~ Rainfall + SST 0.70 23.3 0.00001

PC2inv ~ SSS + Rainfall 0.44 8.4 0.003


75

Figure 4.7 Spatial correlation maps between principal components (PC) scores and

environmental variables. Significant correlations (p < 0.05) shown with inverse PC scores

constructed from the growth chronologies of four taxa (two fishes, one coral and one tree). (a)

PC1inv and rainfall (mm per month); (b) PC1inv and sea surface temperature (°C); (c) PC2inv and

sea surface salinity (PSU) and (d) PC2inv and sea surface temperature (°C) from June to August

in the previous year. All data were from the years 1984–2003 and environmental variables from

(a) to (c) were over the January to March period of the same year.


76

Figure 4.8 Schematic diagram showing the environmental drivers that influenced the

growth patterns of four taxa (two fishes, one coral and one tree) in north-west Australia.

4.5 Discussion

This study revealed that the growth patterns of taxa from both marine and terrestrial

ecosystems in NW Australia were coupled to large-scale, oceanographic and

atmospheric processes. Growth of the study species (two fishes, one coral and one tree)

had significant inverse relationships with the ENSO phenomenon (as measured by the

Niño-4 index) over two decades, so that when the index was positive (where sustained,

strongly positive values indicate an El Niño phase), growth slowed, whereas at times

when the index was negative (where sustained, strongly negative values indicate a La

Niña phase), growth rates increased.

These strong relationships between ENSO and growth responses of all taxa can be

explained by the influence this phenomenon has on the temperature and salinity of

coastal waters and on rainfall patterns in the water-limited terrestrial ecosystems of the

NW region. During the La Niña phase of ENSO there is greater transport of warmer and

less saline waters from the western Pacific towards the coast of NW Australia via the

Indonesian through-flow (Meyers et al. 2007; Zinke et al. 2014). The stronger

Indonesian through-flow subsequently drives a stronger Leeuwin Current that increases

the transport of warmer and less saline waters along the coast of WA. Warmer waters

have been shown to positively influence growth of corals and fish on the WA coast

(Cooper et al. 2012; Rountrey et al. 2014 respectively), while lower salinities may


77

increase fish growth through various metabolic pathways that result in reduced

metabolic costs (see review by Boeuf and Payan 2001) or by increasing food conversion

efficiency (Lambert et al. 1994). Furthermore, Hanson et al. (2005) found much higher

rates of primary productivity along the coastal Gascoyne region in austral summer,

when there were strong correlations between the growth of all taxa and ENSO. The

Leeuwin Current is weakest during austral summer, when southerly winds that favour

coastal upwelling prevail and generate a system of inshore counter-currents that flow

toward the Equator (the Ningaloo Current and Capes Current; Hanson et al. 2005).

These localized upwelling events enhance primary production in otherwise oligotrophic

waters and might play an important role in the increased growth of the study organisms

that was observed during the austral summer.

The La Niña phase of ENSO is also typically associated with higher rainfall over inland

NW Australia. La Niña tends to strengthen the Australian monsoon by influencing

SSTs, low-level winds, vertical motion and convection north of Australia (Wang et al.

2003). This enhanced monsoon causes higher rainfall over NW (and much of northern

and eastern) Australia, which in turn stimulates tree growth in NW Australia (Cullen et

al. 2008; O'Donnell et al. 2015). The ENSO phenomenon also influences NW

Australian rainfall through its effect on the activity of tropical cyclones off the NW

coast of Australia (Denniston et al. 2015), where tropical cyclone activity is enhanced in

La Niña and suppressed in El Niño conditions (Liu and Chan 2012). Tropical cyclones

(and other closed low pressure systems) cause intense rain events over inland NW

Australia and contribute to more than half of the region’s annual rainfall (Lavender and

Abbs 2013). Along the North Coast where there is higher rainfall, it is possible for river

outflows to directly link terrestrial and marine systems, however, the tree and some fish

data were mostly collected around the Gascoyne region, an area subject to very sporadic

patterns of rainfall and river outflow (Lough 1998). Hence, it is more likely that the La

Niña phase of ENSO positively influences the growth of both fishes and trees in NW

Australia, due to its indirect links with climatic conditions likely to favour growth (i.e.,

warmer, less saline sea water in the eastern Indian Ocean and greater rainfall over NW

Australia).

The correlations between ENSO and growth patterns of the study species occurred

despite the fact that the fractional common variance of the growth chronology of L.

nebulosus was relatively low compared to trees and some fishes (e.g. Cullen and

Grierson 2007; Gillanders et al. 2012). Such low fractional common variances appear to


78

be a feature of fishes sampled from the coast of WA (Rountrey et al. 2014; Nguyen et

al. 2015), but it is important to note that all WA fishes for which growth chronologies

have been constructed have displayed significant correlations with regional

environmental factors such as SST.

The strong correlations found between PC2inv (largely reflecting coral growth) and SSS

were unexpected, given the small range of changes in salinity that occur in the NW

region and the results of an earlier study that suggested that decadal growth rates of

corals were most strongly correlated with SST (Cooper et al. 2012). However, there was

strong collinearity between SSS and lagged SST at higher spatial resolution scales,

implying that the latter (or perhaps some other unmeasured factor) may be causing the

apparent correlation between SSS and coral growth. The results also showed that SST

from June to August in the previous year had strong positive correlations with PC2inv.

The lag in this relationship may reflect the fact that coral calcification values were based

on a year defined by annual density minima, which were presumed to occur in the

austral winter months of June to August (Cantin and Lough 2014). Hence, a year in the

coral chronology was based on calcification rates from August of the previous calendar

year to August of the current calendar year. Alternatively, changes in salinity, in

particular anomalous lows, were responsible for around 30% of the unusual

enhancement of the Leeuwin Current transport during the marine heatwave event in the

austral summer of 2010/2011 (Feng et al. 2015). This observation suggests that salinity

may have a more general influence on the growth rates of marine taxa in the NW

region.

The importance of ENSO along the coastline of WA is well recognised. In this region,

the inter-annual variability of this phenomenon has been linked to the survival of

various life history stages of marine taxa, with La Niña years (stronger Leeuwin

Current) showing a greater transport of nutrients into the euphotic zone (Thompson et

al. 2011) that accounts for greater phytoplankton biomass (Koslow et al. 2008) and

increased fisheries recruitment (Caputi 2008). These findings show the influence of

ENSO on the growth rates of adult fish and corals, increasing our knowledge of the far-

reaching impacts of ENSO on a range of life history stages of marine taxa and across

different trophic levels. In addition to strong correlations between growth of all taxa and

the current year’s ENSO, I also found significant, albeit slightly weaker, correlations

between growth and the ENSO signal in the previous year. This suggests that the


79

influence of the ENSO system on growth may carry over between years, or it may have

shown up because the ENSO signal is autocorrelated.

Overall, the strong negative relationship between the growth responses of all four taxa

with ENSO has important implications for the future. Predicted increases in rainfall

(Christensen et al. 2013) and SST (Cheung et al. 2012a) for NW Australia suggest that

growth rates of the study taxa will continue to increase in WA until thermal limits are

reached. However, the strong La Niña conditions (with peak SST reaching 5°C above

average) over the summer of 2010/2011 led to fish kills and widespread coral bleaching

(Feng et al. 2013), suggesting that the thermal limits of fishes and corals are relatively

close to present day conditions on the NW coast. Extreme La Niña events typically

follow strong El Niño conditions and both are predicted to occur more frequently in the

future (Cai et al. 2014; Cai et al. 2015), which may create greater year-to-year variation

in the productivity and yield of fisheries and the likelihood of bleaching in coral

communities along the NW coast. The magnitude of SST changes in the future (along

with the frequencies of El Niño and La Niña events) is likely to have major

consequences on both marine and terrestrial taxa and will need to be carefully

monitored.

The similarities in the growth patterns of the fish and tree species used in this study

suggest that it may be possible to use tree-ring chronologies to hindcast/reconstruct the

growth responses of fish where archives of otoliths do not exist. In many coastal

locations worldwide, tree-ring chronologies now extend centuries into the past, while

the most comprehensive otolith archives are generally the product of fisheries

management studies with a relatively recent history (less than 60 years in most cases).

This study shows that where strong links between the growth of fishes and trees can be

established, chronologies of tree growth may provide a proxy to understand the

response of fish populations to climate change, both in the past and the future.

In summary, I have provided the first empirical evidence for climate-driven synchrony

between marine and terrestrial ecosystems in the Southern Hemisphere at annual time-

scales. These links occur through the influence of ENSO events on regional

environmental variables that affect the annual growth of fishes, corals and trees

throughout the region. Although the available data did not include an overlap of all

taxonomic groups across the entire region, this is a common limitation of any program

that seeks to access legacy datasets where researchers have no control over the intensity

and location of sampling in the past. The large historical archives of fish otoliths


80

(Campana and Thorrold 2001), coral (Tierney et al. 2015) and tree-ring (St. George

2014) records held by institutions and organizations worldwide offer a major

opportunity to expand the scale and resolution of this approach. This will improve both

our understanding of the effect of climate fluctuations on ecosystems in the past and the

likely impact of climate change on both marine and terrestrial ecosystems in the future.


I acknowledge the facilities and the scientific and technical assistance of the Australian

Microscopy & Microanalysis Research Facility at the Centre for Microscopy,

Characterisation & Analysis, the University of Western Australia, a facility funded by

the University, State and Commonwealth Governments. In particular, I am grateful to

Professor Paul Rigby and Miss Alysia Buckley for the generous help and advice they

have provided for the microscopy work. This work was funded by the Australian

National Network in Marine Science, the Australian Institute of Marine Science and the

Center for Marine Futures at the University of Western Australia. This work was made

possible with a scholarship from the Australian Postgraduate Awards. I would also like

to thank the many collaborators who provided data and advice (Alison O’Donnell,

Janice Lough, Jens Zinke, Steve Newman and Mélissa Trougan).

81

Chapter 5 A boundary current drives

growth across tropical and

temperate latitudes

5.1 Abstract

Predictions of the effects of climate change on the productivity and dynamics of marine

fishes are hampered by a lack of historical data on growth patterns. I use otolith

biochronologies to show that the strength of a regional boundary current, largely

modulated by the El Niño-Southern Oscillation, accounted for almost half of the

variability in decadal growth patterns of six species of tropical and temperate marine

fishes across 23° of latitude and 3000 km of continental shelf in Western Australia.

Stronger flow during La Niña years drove increased growth of five of the six species,

whereas weaker flow during El Niño years reduced growth. This work is the first to link

the growth patterns of adult fishes at large spatial scales across multiple climate zones,

habitat types and depth ranges, and has important implications for the potential scale,

timing and outcomes of climate change impacts along continental shelves.

5.2 Introduction

Changes to the world’s climate driven by the release of carbon dioxide into the

atmosphere due to human activities are having serious consequences for marine

ecosystems and these effects are predicted to increase in coming decades (Lehodey et al.

2006; Munday et al. 2008b; Hoegh-Guldberg and Bruno 2010; Cheung et al. 2012b).

For marine fishes, research on climate change has largely focused on the implications of

a warming ocean for thermal limits and distributions of species (Harley et al. 2006;

Pörtner and Peck 2010). Many of these studies have identified “hotspots” of change in

marine environments where pronounced temperature anomalies are now occurring and

these serve as a focal point for the prediction of impacts likely to occur in other regions

(e.g. Cheung et al. 2012b; Poloczanska et al. 2013). However, alterations to the

distributional limits of fishes are only part of the potential effects of climate change; this

process is also likely to modify current flows, patterns of productivity and population

dynamics of species in coastal and oceanic ecosystems (Bakun et al. 2015; García-

Reyes et al. 2015; Fordham et al. 2013 respectively). Because warming affects both

regional (10-1000s km) and large scale (1000s km) climate phenomena, changes could

Chapter 5: A boundary current drives growth across tropical and temperate latitudes

82

occur simultaneously and dramatically at multiple spatial scales, rather than just in

hotspots or progressively at the fringes of species distributions. Evidence for this

possibility is shown by the links between the ecology and growth patterns of many

species within regional environments to climatic features such as the El Niño-Southern

Oscillation (ENSO) events and the flows of regional boundary currents (Chavez et al.

2003; Black et al. 2008; Helser et al. 2012; Black et al. 2014).

The construction and analysis of biochronologies provides a powerful technique to

assess the extent to which growth patterns of marine fishes are linked at large spatial

scales and to determine the likely impact of global warming on the drivers of these

patterns (see review by Morrongiello et al. 2012). Growth integrates the effects of both

physical and biological processes that can vary with climate change, allowing insight

into their effects on key parameters (such as size-at-age, etc.) that determine the

productivity and yields of species targeted by fisheries. Recently, this approach has

been used to show how increasing sea surface temperature (SST) can affect the growth

of temperate marine fishes (Gillanders et al. 2012; Coulson et al. 2014; Rountrey et al.

2014) and how the drivers of growth patterns can even be linked across divergent taxa

(trees, corals, fish and bivalves) in the same environment (Black 2009; Chapter 4).

Here, I use a biochronology approach to identify key climate drivers of growth of

marine fishes from a variety of habitats and trophic guilds across 23° of latitude

encompassing tropical, transitional, and warm and cool temperate zones. I then examine

the vulnerability to and likely scale of impact of the predicted effects of global warming

on the growth patterns of these species. The marine environments of the continental

shelf of Australia provide an ideal model to examine broad scale impacts of climate

change on the growth of fishes. Off the coast of Western Australia (WA), the shelf is

dominated by the Leeuwin Current, a poleward-flowing current that suppresses

productivity and creates an oligotrophic marine environment (Feng et al. 2009). This

contrasts with the highly productive and equatorward eastern boundary currents present

in other parts of the world (Canary, California, Humbolt and Benguela currents), all of

which contribute significantly to global catches of marine fishes (Fréon et al. 2009). The

Leeuwin Current extends more than 2000 km transporting warm, nutrient-poor water

southward along the continental shelf from the tropics to cool temperate coasts. It is

ecologically important for many species in the region (e.g. Caputi et al. 1996; Caputi

2008) and extends the distribution of some tropical taxa southward (Maxwell and


83

Cresswell 1981). As is the case for other boundary currents, the strength of the Leeuwin

Current is strongly linked to the ENSO phenomenon (Feng et al. 2009).

Earlier studies of otolith chronologies of fishes along the coast of WA have shown that

growth of some shallow-water, temperate species is driven by SST (e.g. Coulson et al.

2014; Rountrey et al. 2014). In contrast, the growth of tropical fishes is correlated with

ENSO variability (as seen in Chapter 2 and Chapter 4), whereas the growth of a deep-

water fish off the coast of the south-west is correlated with the strength of the Leeuwin

Current (Nguyen et al. 2015). Given that these observed drivers are inter-related, the

growth patterns of fishes might be similar across relatively large spatial scales, from

tropical to temperate zones off WA. Thus, the effects of climate change could manifest

simultaneously and synoptically across environments where fishes share connected

drivers of growth. I predict that strength of the Leeuwin Current, which is strongly

associated to ENSO fluctuations and constrains regional environmental variables such

as temperature, will be an important driver of the growth of adult fishes throughout the

coastal marine waters of WA. I also examine the implications of this link for growth

and productivity of fish populations under scenarios of future climate change.

5.3 Methods

5.3.1 Marine environment in Western Australia

The coastline of WA encompasses four major climate regions: tropical (North Coast),

tropical-transitional (Gascoyne Coast), warm temperate (West Coast) and cool

temperate (South Coast) coasts (Fletcher and Santoro 2015; Figure 5.1). The North

Coast includes the coastal areas of the Kimberley and Pilbara regions and is entirely

tropical. The warm, low salinity waters in this region originate in the Pacific and enter

via the Indonesian through-flow (Meyers 1996), continuing southwards via the seasonal

Holloway Current, which is strongest in autumn (D’Adamo et al. 2009; see Figure 5.1).

The Gascoyne Coast, which extends from Exmouth Gulf to Shark Bay and includes the

Ningaloo Reef, is a subtropical transition zone. The Leeuwin Current becomes evident

along the Gascoyne Coast and flows southward along the narrow continental shelf

(Cresswell and Golding 1980; Feng et al. 2009; see Figure 5.1). It is strongest during

winter and weakest during summer when the prevailing southerly winds generate

inshore counter-currents such as the Ningaloo Current, which is associated with

localised upwelling (Hanson et al. 2005). The temperate West Coast extends between

Kalbarri and Augusta and is heavily influenced by the warm, oligotrophic waters of the


84

Leeuwin Current. Another inshore counter-current, the Capes Current, occurs

predominantly in summer and contributes to localised upwelling (Hanson et al. 2005).

The cool temperate South Coast extends from Augusta in the south-west to the border

with South Australia in the east. The marine environment off the South Coast is

influenced by the Southern Ocean and, despite being temperate, is low in nutrients and

warmer than expected due to the influence of the Leeuwin Current (Fletcher and

Santoro 2015).

Figure 5.1 General sampling locations of the six species of adult fishes collected along the

coast of Western Australia. LA (Δ) = Lutjanus argentimaculatus, LB () = Lutjanus bohar,

LN () = Lethrinus nebulosus, AG () = Achoerodus gouldii, PO () = Polyprion oxygeneios,

SA () = Scorpis aequipinnis. The marine regions and the major current flows along the

coastline are also indicated.


85

5.3.2 Otolith growth chronologies

The otoliths of fishes grow in proportion to body size and thus provide a proxy for

measures of somatic growth (Rowell et al. 2008; Neuheimer et al. 2011; Stocks et al.

2011; Black et al. 2013). Existing increment data from the otoliths of five coastal fishes

(Lutjanus argentimaculatus, Lutjanus bohar, Lethrinus nebulosus, Achoerodus gouldii

and Polyprion oxygeneios) in WA were obtained from earlier studies and a new

chronology for Scorpis aequipinnis was constructed (Table 5.1). These six species,

which all have relatively long lifespans, were sampled in different climatic zones

(Figure 5.1) and depth ranges (Table 5.1). Sampling and otolith sectioning methods for

S. aequipinnis have been described by Coulson et al. (2012). Only those sectioned

otoliths of S. aequipinnis that had clearly defined opaque zones (which form annually;

Coulson et al. 2012) were considered for inclusion in the analyses. These 28 individuals

ranged in age from 37–69 years (median = 45.5), including year-classes between 1940

and 1973. Following the methods of Rountrey et al. (2014) and using a plugin

(“IncMeas”; Rountrey 2009) written for Image J (National Institutes of Health, USA),

polyline transects were drawn on the digital otolith image, dorsal to the sulcus acusticus

and parallel to the direction of growth. The earliest two to six increments were excluded

because these increments typically had diffuse boundaries that prevented accurate

measurements. Visual crossdating was complicated by the very small variations in

increment width that were extremely difficult to distinguish by eye. Consequently,

COFECHA software (Holmes 1983) was employed for statistical crossdating, using the

dates of capture to anchor each increment time series, following the methods of Black et

al. (2005). Statistical crossdating was used to check the correct assignments of calendar

years to increments (Black et al. 2016) and any errors were inspected visually before

measurements were changed.

Raw otolith increment widths for all individuals from each species were obtained and

detrended using the method of Black et al. (2013). The quality of the detrended

increment series (for fish of the same species) was then assessed with the mean of pair-

wise series correlations (�̅�) and the expressed population signal (EPS) using the same

software as described in Section 2.3.5. To ensure that the chronologies contained

synchronous growth signals, bootstrapped 95% confidence intervals for �̅� (with 15-year

intervals) were estimated as described in Section 3.3.5. These were calculated using a

modified version of the package “dplR” (Bunn 2008) in R software (R Core Team

2015). Only periods with �̅� > 0 and EPS > 0.5 were used to construct the growth


86

chronology. If the detrended increment series for any species did not contain periods

with �̅� > 0 and EPS > 0.5, the detrending methods used in the original publications were

adopted. The growth series for L. argentimaculatus, L. bohar, L. nebulosus and S.

aequipinnis were detrended following Black et al. (2013), while A. gouldii was

detrended using the double-detrending technique (Rountrey et al. 2014) and P.

oxygeneois was detrended using a nine year spline (Nguyen et al. 2015). The double-

detrending and spline methods may have been more successful in emphasizing signal

and removing noise for A. gouldii and P. oxygeneois, respectively, due to higher levels

of individual variation in year-adjusted mean increment width and/or more medium

frequency variance caused by local factors rather than regional drivers. However

explicit testing of detrending methods was beyond the scope of this study. The final

growth chronologies for each species were constructed from the average of all

detrended increment series for all individuals within each species. The time span of

overlap of chronologies for all six species covered 16 years, from 1988 to 2003.

Table 5.1 Summary of the otolith growth chronologies of fishes from Western Australia.

LA = Lutjanus argentimaculatus, LB = Lutjanus bohar, LN = Lethrinus nebulosus, AG =

Achoerodus gouldii, PO = Polyprion oxygeneios, SA = Scorpis aequipinnis, NC = North Coast,

GC = Gascoyne Coast, WC = West Coast, SC = South Coast, ENSO = El Niño-Southern

Oscillation, SSS = sea surface salinity, PDO = Pacific Decadal Oscillation, SL = sea level, SST

= sea surface temperature, n = sample size. Lag indicates that the growth chronologies were

significantly correlated with environmental variables in the previous year. Drivers of growth are

those identified in the publications listed.

Taxa Length of

chronology

Location (n) Capture

depth (m)

Publication source Drivers of growth

LA 1975–2004 NC, GC (36) 50–120 Chapter 2 ENSO, SSS, PDO

LB 1962–2007 NC (55) 80–180 Chapter 3 Rainfall, SL, PDO

LN 1980–2009 GC (23) 5–80 Chapter 4 SST, Rainfall

SA 1952–2009 SC (28) 5–50 -- --

AG 1952–2003 SC (56) 5–100 Rountrey et al. 2014 SST

PO 1979–2009 WC, SC (44) 200–450 Nguyen et al. 2015 SL (lag), SST (lag)

5.3.3 Environmental datasets

Annual mean values of the sea level at Fremantle Port (FSL), a well-known and well-

justified proxy for the strength of the Leeuwin Current (Feng et al. 2003) were obtained

from the Bureau of Meteorology (Australian Government) website,

http://www.bom.gov.au. Annual means of three climate indices; the multivariate ENSO

index (Wolter and Timlin 1993), Pacific Decadal Oscillation (PDO; based on SST

anomalies over the North Pacific; Mantua and Hare 2002) and the Dipole Mode Index


87

(DMI; difference in SST anomalies between the western and eastern equatorial Indian

Ocean; Saji et al. 1999) were obtained from the Royal Netherlands Meteorological

Institute, KNMI Climate Explorer (Trouet and Van Oldenborgh 2013), a web

application for climate data (http://climexp.knmi.nl). These climate indices were chosen

because the results of Chapter 2, Chapter 3 and Chapter 4 found links between ENSO

and PDO indices and otolith growth in WA. In addition, the mode of the Indian Ocean

dipole (described by the DMI) affects the amount of austral winter and spring rainfall

(Ashok et al. 2003; Cai et al. 2009). Quarterly spatial correlation maps were constructed

to show the relationship between growth of adult fish in WA and ocean heat content

from 0–750 m depth, obtained from the Simple Ocean Data Assimilation reanalysis of

ocean climate variability (Carton and Giese 2008). This variable was chosen because

SST (Rountrey et al. 2014; Figure 4.7) and ocean heat content (Figure 2.6) were

previously identified as influencing fish growth in WA, and ocean heat is also a

predictor of the strength of the Leeuwin Current (Hendon and Wang 2010). Quarterly

spatial correlation maps were constructed to illustrate the relationship between growth

and ocean mean temperatures from 0–700 m depths, obtained from National Oceanic

and Atmospheric Administration World Ocean Atlas (Locarnini et al. 2013). All spatial

correlation maps were constructed in KNMI Climate Explorer.

5.3.4 Statistical analyses

All chronologies were standardized (mean = 0, variance = 1) and analysed using

principal components analysis (PCA). The scores from the principal components that

accounted for the majority of the variance were included as response variables in linear

regression models to assess the influence (based on information-theoretic methods) of

FSL and the climate indices including the multivariate ENSO index, PDO and DMI.

Collinearity between the environmental variables (|r| > 0.5, p < 0.01) was evaluated to

ensure that collinear variables were not included in the same model. The R package

“MuMIn” (Barton 2015) was used for model selection using the second-order Akaike

information criterion (AICC) based on Kullback-Leibler information loss and

accounting for small sample sizes (Burnham and Anderson 2004). Differences in AICC

values (ΔAICC), model probabilities (Akaike weights) and relative variable importance

(using the sum of Akaike weights over all models) were used to assess the models.

Adjusted R2 values, F-statistics and p-values were reported for the top-ranked models.

All top-ranked models identified in the model selection process were validated to ensure

that assumptions of homogeneity, normality and temporal autocorrelation were not


88

violated. Linear dependence was assumed for the environmental variables because the

use of smoothers for the variables did not improve the fit of the models. All statistical

analyses were completed in R software (R Core Team 2015). After the model selection

process, spatial correlation maps of the first principal component with quarterly ocean

heat content and ocean mean temperatures were constructed to illustrate the spatial

relationships. Pearson’s correlation tests were carried out on the synchronous periods

for each species chronology with the environmental variable selected by the top-ranked

model involving the first principal component, to ensure accuracy of the model results.

5.4 Results

5.4.1 Chronology statistics

The bootstrapped 95% confidence intervals for �̅� and the expressed population signal

(EPS) for the detrended data of all six species showed the periods in which there was a

synchronous growth signal among fish for each species (where �̅� > 0 and EPS > 0.5;

Table 5.2). The period of overlap for all species was 1988–2003, a 16-year period and

the standardized growth chronologies over these years are shown in Figure 5.2.

Table 5.2 Results of bootstrapped �̅� and expressed population signal (EPS) for the

detrended and standardized growth chronologies of six marine fishes from Western

Australia. The period with synchronous growth signals (synchronous period) was determined

using the conditions �̅� > 0 and EPS > 0.5. The values of �̅�, 95% confidence interval (CI) for �̅�

and EPS represent the average values for the chronologies during this time (16 years).

Species Synchronous

period

Average

�̅�

95% CI for

�̅�

EPS

Lutjanus argentimaculatus 1975–2004 0.153 0.122–0.182 0.838

Lutjanus bohar 1970–2007 0.090 0.043–0.137 0.768

Lethrinus nebulosus 1981–2006 0.112 0.034–0.184 0.690

Scorpis aequipinnis 1985–2008 0.104 0.074–0.134 0.744

Achoerodus gouldii 1969–2003 0.021 0.007–0.035 0.540

Polyprion oxygeneios 1988–2003 0.030 0.002–0.057 0.553


89

Figure 5.2 Otolith growth chronologies of six marine fishes from Western Australia with

the respective leading principal component (PC) scores. Chronologies were detrended and

standardized (mean=0, variance =1). (a) Lutjanus argentimaculatus, Polyprion oxygeneios and

Achoerodus gouldii chronologies with PC1inv, (b) Scorpis aequipinnis and Lutjanus bohar

chronologies with PC2inv and (c) Lethrinus nebulosus chronology with PC3. The inverse of PC1

and PC2 were used because the respective taxa were negatively loaded on both principal

components.

5.4.2 Principal components analysis

The first three principal component (PC) scores accounted for 82% of the total variance

in the data set. Of these, the first accounted for 41% of the variance, the second for 24%

and the third, 17%. The remaining principal components were not included in further

analyses. Growth of three of the species (L. argentimaculatus, A. gouldii and P.

oxygeneios) had negative loadings on PC1 (Table 5.3) and displayed similarities in

temporal patterns of growth (Figure 5.2a). Chronologies of two of the remaining species

(S. aequipinnis and L. nebulosus) also had negative loadings on PC1, whereas L. bohar


90

had a positive loading (Table 5.3). Chronologies for L. bohar and S. aequipinnis had

negative loadings on PC2 (Table 5.3) and displayed similar temporal patterns in growth

(Figure 5.2b). Lethrinus nebulosus had the strongest positive loading on PC3 (Table 5.3,

Figure 5.2c). Inverse values of PC1 (PC1inv) and PC2 (PC2inv) scores were used in

further analyses.

Table 5.3 Loadings of the otolith biochronologies of six marine fishes on the first, second

and third principal component (PC) scores. Fishes were sampled from coastal Western

Australia and the otolith growth chronologies were from the years 1988 to 2003.

Species PC1 loading PC2 loading PC3 loading

Lutjanus argentimaculatus -0.55 -0.09 -0.38

Lutjanus bohar +0.17 -0.66 -0.32

Lethrinus nebulosus -0.37 -0.15 +0.74

Scorpis aequipinnis -0.32 -0.60 -0.08

Achoerodus gouldii -0.49 -0.01 +0.17

Polyprion oxygeneios -0.44 +0.43 -0.41

5.4.3 Relationships with environmental variables

Due to collinearity among the four environmental variables (|r| > 0.5; Table 5.4) and the

low number of observations (n = 16), only one variable was used in the construction of

each model. The resulting five models (one for each of the four environmental variables

and an intercept-only model) were evaluated separately in the model selection process

for the response variables PC1inv, PC2inv and PC3 (Table 5.5). The model selection

process involving PC1inv found that the first-ranked model (i.e. lowest AICC) was one

that related PC1inv with Fremantle sea level (FSL) and was approximately six times

more likely than the second-ranked model (Table 5.5). The first-ranked model showed a

positive relationship between PC1inv and FSL, and it explained 48% of the variance in

PC1inv (Table 5.6).

The influence of the Leeuwin Current along the coastline of WA, and the relationship

between the Leeuwin Current and the western Pacific warm pool were evident in the

spatial correlation maps of PC1inv with ocean heat content from January to March

(Figure 5.3a; see Figure 5.4a–c for spatial correlation maps of the other months). A

strong link between the coastal waters off WA and western Pacific waters via the

Indonesian through-flow was also evident (Figure 5.3a). The spatial correlation map of

PC1inv and ocean temperatures from January to March (Figure 5.3b; see Figure 5.4d–f

for spatial correlation maps of the other months) showed that higher temperatures were

correlated with wider otolith increments for five of the six study species. This result was


91

consistent with the results for ocean heat content, which also showed positive

correlations with growth when higher amounts of heat were stored in these waters. The

influence of the Leeuwin Current was further supported by the moderate to high

correlation coefficients found in four of the six study species (Table 5.7). In particular,

even though S. aequipinnis did not have high loadings on PC1 (Table 5.3), the dataset

including the whole synchronous period had strong correlations with FSL (proxy for the

strength of the Leeuwin Current) as shown in Table 5.7. The influence of the Leeuwin

Current, itself modulated by ENSO, on five of the six study species (as indicated by the

PC1 loadings) are shown in Figure 5.5.

The first-ranked model for PC2inv identified a possible relationship between growth and

the Pacific Decadal Oscillation (PDO, Table 5.5). It was only approximately twice as

likely as the second-ranked, intercept-only model (Table 5.5) and was not considered to

be a substantial improvement as the ΔAICC was less than 2 (Burnham and Anderson

2004). The relative variable importance of the PDO was 0.56, which was much greater

than other variables (0.06–0.07). The linear model showed that the negative relationship

between PC2inv and the PDO was marginally significant (Table 5.6) and explained 20%

of the variation in PC2inv, which largely reflected the growth chronologies of L. bohar

and S. aequipinnis (Table 5.3).

The first-ranked model for PC3 included the Dipole Mode Index (DMI) and was

approximately six times more likely than the second-ranked model (Table 5.5). The

linear model showed a positive relationship between PC3 and the DMI, which explained

30% of the variation in PC3 (Table 5.6) and mainly reflected the growth of L. nebulosus

(Table 5.3).

Table 5.4 Pearson’s correlation matrix of the four environmental variables (annual

means) over the years 1988 to 2003. FSL = Fremantle sea level, MEI = Multivariate El Niño-

Southern Oscillation Index, PDO = Pacific Decadal Oscillation, DMI = Dipole Mode Index.

Variable MEI PDO DMI

FSL -0.86 -0.60 -0.21

MEI – +0.69 +0.29

PDO +0.69 – -0.05

DMI +0.29 -0.05 –


92

Table 5.5 List of models in the model selection process. Explanatory variables were annual

means over the years 1988–2003. Response variables were the first three principal component

(PC) scores from six marine fishes in Western Australia. ENSO = El Niño-Southern Oscillation,

ΔAICC = difference in second-order Akaike information criterion.

Model Explanatory variable ΔAICC

(PC1inv)

ΔAICC

(PC2inv)

ΔAICC

(PC3)

1 Fremantle Sea level 0.00 4.13 6.04

2 Pacific Decadal Oscillation 6.11 0.00 6.42

3 Dipole Mode Index 11.53 4.16 0.00

4 Multivariate ENSO Index 3.64 4.41 6.71

5 Intercept only 8.47 1.65 3.64

Table 5.6 First-ranked linear models that explain variation in the first three principal

component (PC) scores from the otolith growth chronologies of six marine fishes in

Western Australia. Environmental variables are annual averages and standardized growth

chronologies were from the years 1988 to 2003. FSL = Fremantle sea level, PDO = Pacific

Decadal Oscillation, DMI = Dipole Mode Index. The inverse of PC1 and PC2 were used

because the respective taxa were negatively loaded on both PC scores.

Model equation Adjusted R2 F-statistic Model p-value

PC1 inv ~ FSL 0.48 14.81 0.002

PC2inv ~ PDO 0.20 4.81 0.05

PC3 ~ DMI 0.30 7.30 0.02

Table 5.7 Pearson’s correlation test results of the individual chronologies (entire

synchronous period) with Fremantle sea level.

Species Correlation coefficient p-value

Lutjanus argentimaculatus 0.457 0.011

Lutjanus bohar -0.038 0.823

Lethrinus nebulosus 0.211 0.301

Scorpis aequipinnis 0.556 0.005

Achoerodus gouldii 0.489 0.003

Polyprion oxygeneios 0.419 0.084


93


component (PC1inv) and environmental variables averaged between January and March.

PC1inv was constructed from the standardized growth chronologies of six fishes from Western

Australia and inverse values were plotted as five of the six species were negatively loaded on

PC1. (a) PC1inv and ocean heat content (0–750 m depth) and (b) PC1inv and ocean mean

temperature (0–700 m depth). Warmer colours indicate positive correlations, cooler colours

indicate negative correlations. All data were from the years 1988 to 2003 and spatial correlation

maps were obtained and modified from KNMI Climate Explorer.

94

Figure 5.4 Significant correlations (p < 0.05) between the inverse of the first principal component (PC1inv) and environmental variables averaged

quarterly between April and December. PC1inv was constructed from the standardized chronologies of six fishes from Western Australia. Maps show

correlations between PC1inv and ocean heat content (0–750 m depth) from (a) April to June, (b) July to September, (c) October to December. (d) to (f) show

correlations between PC1inv and ocean temperature (0–700 m depth) from the same months as (a) to (c). Warmer colours indicate positive correlations, cooler

colours indicate negative correlations. All data were from the years 1988 to 2003 and spatial correlation maps were obtained and modified from KNMI Climate

Explorer.


95

Figure 5.5 Schematic diagram showing the influence of the Leeuwin Current (modulated

by ENSO) on the growth patterns of five species of fishes along the coast of Western

Australia.

5.5 Discussion

The strength of the Leeuwin Current, linked to inter-annual variations of ENSO (Feng

et al. 2009), was an influential driver of the growth of five of six species of adult fishes

across more than 3000 km of coastline in the eastern Indian Ocean. These fishes were

collected from a shelf that encompassed tropical, subtropical, warm temperate and cool

temperate environments over ~23° of latitude and from both shallow coastal waters and

deeper continental slopes. Fremantle sea level (a proxy for strength of the Leeuwin

Current) and composite growth of five of the six study species were positively

correlated, indicating that in years of high flow (La Niña phases of ENSO), growth rates

of adult fishes tended to increase, whereas in years of low flow (El Niño phases of

ENSO), growth declined. Overall, current strength accounted for almost half of the

variability in the composite growth chronologies that represented five of the six species.


96

This work is the first to provide evidence of a link between the growth patterns of adult

fishes, flow of a boundary current and ENSO at spatial scales that include multiple

climate zones, habitat types and depth ranges. Results from the earlier chapters (Chapter

2 and Chapter 4) have found strong relationships between growth and ENSO indices for

single species in north-western Australia. However, the results of this chapter show that

these links can occur for multiple species along both tropical and temperate coasts and

across the shelf in waters influenced by the Leeuwin Current, an area encompassing a

third of the entire coastline of the Australian continent.

There are several possible mechanisms that could account for the positive relationship

between growth (as measured by the PC1inv index) and the strength of the Leeuwin

Current. Firstly, an increase in current flow results in a greater input of warmer and

lower salinity waters from the western Pacific to the WA coast (Meyers et al. 2007). As

fish are poikilotherms, these warmer waters may increase metabolic rates and aid

growth (e.g. Rountrey et al. 2014), provided individuals do not exceed thermal limits

and sufficient food is available to support energy demands. Lower salinities may also be

beneficial to fish growth because of reduced metabolic costs (Boeuf and Payan 2001).

Secondly, stronger flows of the Leeuwin Current may create intense mesoscale eddies

that draw nutrient-rich water from deeper waters into the photic zone, increasing

primary production (Koslow et al. 2008) that then flows up the food chain to fishes at

higher trophic levels (Nguyen et al. 2015). Finally, a stronger Leeuwin Current is also

associated with higher rainfall over north-western Australia due to the influences of La

Niña over SST, winds and convection north of the continent (Wang et al. 2003). The

input of terrestrial nutrient via runoff at these times might be an important factor for the

growth of fish that live in or near large embayments or estuaries such as juvenile L.

argentimaculatus or L. nebulosus, as previously demonstrated in Chapter 2 and Chapter

4 respectively.

The Pacific Decadal Oscillation (PDO) was identified as a driver of adult growth of L.

bohar, S. aequipinnis and, to a lesser extent, P. oxygeneios. Although it might seem

incongruous that the physical oceanography of the northern Pacific could influence

growth of fishes in the eastern Indian Ocean, this teleconnection is not without

precedent. The PDO describes a combination of oceanographic processes that span both

the tropics and the extra-tropics and is strongly linked to the oceanography of the

tropical western Pacific (Chen and Wallace 2015; Newman et al. 2016a). It is also

regarded as a reddened response to ENSO where the ENSO signal re-emerges in the


97

subsequent year (Newman et al. 2003; Shakun and Shaman 2009). Hence, correlations

with the PDO might have occurred via two mechanisms. Firstly, growth variations in

the three species could be linked to the lagged ENSO signal. Alternatively, the growth

patterns were responding to low frequency (ie. decadal) variations of SST anomalies in

the tropical Pacific (Chen and Wallace 2015). In support of the second explanation,

multi-decadal variations of the PDO in the tropical Pacific have been found to influence

low frequency variability of the Fremantle sea level, (a proxy for the strength of the

Leeuwin Current; Feng et al. 2003), a process that is recorded in the growth patterns of

corals along the WA coast (Zinke et al. 2014).

The growth of adult L. nebulosus was strongly influenced by the Dipole Mode Index

(DMI), an indicator of the strength of the Indian Ocean Dipole. This likely reflects the

correlation between the DMI and rainfall in north-western Australia (Risbey et al. 2009)

as the majority of individuals of this species were caught in Shark Bay, a large, shallow

marine embayment that is partially hypersaline (Fletcher and Santoro 2015). In such an

environment, changes in salinity due to rainfall are highly likely to have major impacts

on the growth of marine taxa (e.g. Boeuf and Payan 2001) and this result is consistent

with earlier work on this species, which also tied growth patterns to salinities (Chapter

4).

The importance of ENSO and the Leeuwin Current for the settlement and recruitment of

marine invertebrates and fishes along parts of the coastline of WA is well-recognised

(Caputi et al. 1996). The strength of the current also influences the presence and

persistence of corals and their symbionts at high latitudes (Collins et al. 1993;

Silverstein et al. 2011) and the abundance of very large planktivores, such as whale

sharks (Sleeman et al. 2010). My study shows that the growth of adult fishes is also

driven by these large-scale atmospheric and oceanographic phenomena, irrespective of

the thermal range of a species (tropical or temperate), the depth of habitat or trophic role

it occupies. This result has important implications for the potential scale and ubiquity of

climate change impacts along the WA coast. Climate forecasts suggest that El Niño and

La Niña events are likely to become more frequent and intense (Cai et al. 2014; Cai et

al. 2015). Strong El Niño events slow the flow of the Leeuwin Current (Feng et al.

2009) and my study shows that such conditions may lead to concurrent declines in the

growth of many fishes along the entire coastline of WA. This has the potential to alter

key life history traits (e.g. size at age; Rountrey et al. 2014) with subsequent impacts on

population demography. Although moderate La Niña conditions are likely to positively


98

affect growth rates, as demonstrated by my study, extreme events can produce

anomalous water temperatures that may surpass the thermal limits of fishes. For

example, the marine heat wave that occurred in the summer of 2011 on the coastline of

WA was driven by an unusually strong Leeuwin Current, itself a result of a strong La

Niña event (Feng et al. 2013) and caused major fish kills along the coast (Wernberg et

al. 2013), reductions in abundances of commercially important invertebrates such as the

abalone Haliotis roei and scallops Amusium balloti (Caputi et al. 2016) and coral

bleaching (Depczynski et al. 2013). Thus, extreme phases of both El Niño and La Niña

are likely to negatively impact fish populations in a synoptic manner along the entire

coastline of WA, irrespective of distribution ranges, thermal limits, depth ranges and

trophic levels.

These recent marine heat waves highlight the way in which the ENSO signal can

introduce extreme variability in physical environments inhabited by marine fishes.

Ultimately, these events may complicate the ability of fishes to adapt to climate change

due to rapid and frequent shifts in physical conditions beyond the upper physiological

limits for growth. Under such conditions management strategies may need to account

for the likely lowered resilience of exploited populations.


This work was made possible with a scholarship from the Australian Postgraduate

Awards. I would also like to thank the many collaborators who provided data and

advice (Peter Coulson, Thomas Nguyen, Corey Wakefield, Stephen Newman).

99

Chapter 6 General discussion

The aim of my thesis was to determine how climate change influences the growth of

tropical fishes using biochronologies derived from otoliths and validated using

crossdating techniques. Tropical fishes represent a key gap in climate change research,

where the focus has typically been on commercially important species that support

temperate fisheries in the Northern Hemisphere. Tropical otoliths have also been under-

utilized due to the misconception that otoliths in tropical fish did not form reliable

annual bands. The use of validated otolith biochronologies to determine climate-growth

relationships is a relatively new field and, within the Southern Hemisphere, has only

been applied to temperate fishes. In this thesis, I widened the application of the

approach by showing its usefulness for growth analysis of two different life history

stages and populations of tropical lutjanids in the Southern Hemisphere. I then extended

this approach, along with different analytical methods, to other fishes, marine benthic

invertebrates (corals) and terrestrial species (trees) at increasing spatial scales in order to

assess the influence of climate change on multiple taxa across Australia.

I found that biochronologies of fish with ontogenetic shifts in habitat (Chapter 2) and

those of adult populations of the same species residing in separate ocean basins

(Chapter 3) differed in their response to key climate variables. However, for species

across the marine and terrestrial ecosystems within the same geographic region (Chapter

4) and for five species of adult fishes along the coastline of Western Australia in the

eastern Indian Ocean (Chapter 5), there was evidence of climate-driven synchrony in

biochronologies of growth. This evidence for multi-level effects of climate change

across life history stages, populations, multiple taxa and ecosystems (Figure 6.1) were

extracted through the use of biochronologies that were validated through crossdating,

resulting in precisely dated, high-resolution data sets. The exploration of these climate-

growth relationships across a wide range of taxa enables more accurate predictions of

the impacts of climate changes on economically important and vulnerable tropical

species and help in the development of management strategies to improve resilience and

maintain sustainability of these populations in the face of our rapidly changing climate.

Chapter 6: General discussion

100

Figure 6.1 Schematic diagram showing the multi-level effects of climate change found in

the four data chapters. ENSO = El Niño-Southern Oscillation. The different colours (green

and red) for climatic and oceanographic factors in Chapter 3 represent different sets of factors.

6.1 Addressing the tropical knowledge gap

My thesis sets out to address knowledge gaps on the likely effects of climate change on

tropical fishes, and shows that validated otolith biochronologies are capable of

capturing such links. The forecast for coral reef ecosystems in the tropics is not an

optimistic one, with studies predicting that about 35% of global coral reefs will be in

critical decline and close to extinction within the next two to three decades (Carpenter et

al. 2008; Wilkinson 2008; Hoegh-Guldberg 2011). Such declines are of concern for

human populations living along tropical coastlines, because of their reliance on fish

extracted from these reef systems as a major source of animal protein (Pauly et al.

2002). Compounding this problem is the fact that the majority of these populations are

in developing countries where human population growth is doubling at a faster rate than

the global mean and more than 20% live in poverty (Montgomery 2011). Thus, for the

hundreds of millions of people dependent on coral reef ecosystems for food and

livelihoods (Hoegh-Guldberg 2011), knowledge of the impacts of climate change and

other anthropogenic disturbances on tropical reef-associated fish is crucial for

management strategies that seek to ensure conservation and sustainability.


101

6.2 The varying effects of climate change

The results of my thesis show that the effects of climate change differ at an intra-

specific level (Figure 6.1) both with respect to different life history stages and among

populations within the geographical range of a species. However, when assessing cross-

taxa climate signals in adults, there were similarities in the growth responses of multiple

species to the same climate signals, at both ecosystem and continental scales (Figure

6.1). These similarities likely reflect the strong atmospheric and oceanographic links

between the marine and terrestrial ecosystems, and the influence of the major boundary

current along the continental shelf in the study regions.

6.2.1 Differing impacts at an intra-specific level

Most marine fishes in the tropics have a complex life history with ontogenetic changes

in habitat (Thresher 1984; Leis and McCormick 2002). Such ontogenetic variation adds

to the difficulty of predicting the effects of climate change on a single species. My

thesis demonstrated that there were contrasting drivers of growth in the juvenile and

adult stages of a tropical lutjanid (Chapter 2) that displays an ontogenetic shift in habitat

between juvenile and adult life history stages. This result indicates that observations or

predictions about the effects of climate change that apply to one ontogenetic stage

cannot be assumed to apply across the entire life history of a species and contributes to

the existing body of research on life-history specific responses to climate change. For

instance, an increasing number of studies consider the effects of climate change on

larval biology and performance (see Munday et al. 2009a for review). These studies

have shown that warming temperatures and ocean acidification affect growth and

behaviour of fish larvae (e.g. McCormick and Molony 1995; Green and Fisher 2004;

Dixson et al. 2010). Adult-focused studies have demonstrated that elevated

temperatures and acidification reduces aerobic performance of some coral reef fishes

(Munday et al. 2008a; Munday et al. 2009b). However, fewer studies have attempted to

simultaneously explore the effects of climate change on multiple life history stages of a

tropical fish. One exception is the study by Munday et al. (2008a), which examined the

effects of elevated temperatures on the growth of both the juvenile and adult stages of

Acanthochromis polyacanthus. The contrast between the drivers of growth of juveniles

and adults described in Chapter 2 confirms the need to investigate climate-driven

responses across all life history stages of tropical fishes (Harley et al. 2006; Rijnsdorp et

al. 2009). When clear, multi-level climate implications can be elucidated across the

major life history stages of a model species, these predictions can potentially be applied


102

to other species of fish that display similar patterns of larval dispersal and use of

juvenile and adult habitats.

In addition to the different environmental drivers across ontogenetic stages of a tropical

fish, I show in Chapter 3 that there are differences in drivers of growth across the adults

of two populations in different ocean basins. This result implies that there is a need for

caution when making generalizations of species responses across geographically

separated populations. In the last decade, numerous studies have made predictions

regarding the changes in the distributions of fishes due to climate change (e.g. Perry et

al. 2005; Nye et al. 2009; Cheung et al. 2012a; Poloczanska et al. 2013; Brown et al.

2016). However, these predictive studies can involve observations collected from

particular localities that may not be representative of the entire species range

(Poloczanska et al. 2013; Brown et al. 2016). Moreover, research focusing on inter-

oceanic comparisons of tropical fish has tended to examine demographic characteristics

of different populations across a species range, but with contrasting results (Trip et al.

2008; Wakefield et al. 2015). For example, growth rates and adult sizes of the

surgeonfish Ctenochaetus striatus did not differ between populations in the Indian and

Pacific ocean (Trip et al. 2008) whereas growth, longevity and mortality of the eightbar

grouper Hyporthodus octofasciatus varied between the two ocean basins (Wakefield et

al. 2015). Clearly, there is a need for studies that include observations across ocean

basins, in order to improve predictions of the responses of tropical fishes to climate

change.

6.2.2 Synchronous climate impacts across ecosystems

Studies of the impacts of the El Niño-Southern Oscillation (ENSO) phenomenon across

ecosystems are rare (but see Black 2009; Braganza et al. 2009) despite ENSO being

recognised as the most influential climate event on the planet, with the inter-annual

fluctuations expressed across marine and terrestrial ecosystems world-wide (McPhaden

et al. 2006). The two studies by Black (2009) and Braganza et al. (2009) demonstrated

synchronous responses to the ENSO signal across diverse taxa (fish, trees, bivalves) in

the Pacific Ocean. Even though the impacts of ENSO on physical oceanography and the

atmosphere have been recorded world-wide, such studies have not been replicated in

regions beyond the Pacific Ocean. This is particularly important for tropical regions at

the margins between the Pacific and the Indian oceans, since these encompass areas

with the highest diversity of fishes (Roberts et al. 2002), yet are the most vulnerable and


103

under-studied areas in terms of the likely impacts of climate change (Munday et al.

2008b; Rosenzweig et al. 2008; Richardson et al. 2012).

Chapter 4 described synchronous patterns in growth across diverse taxa from both

marine and terrestrial ecosystems in the eastern Indian Ocean, demonstrating the far-

reaching and consistent influence of the ENSO signal and the strong links between the

eastern Indian and Pacific oceans. This finding reinforces the potential significance of

strong ENSO events in the tropical Indo-Pacific and highlights the need for expansion

of this work across the region. ENSO events are forecast to increase in strength and

frequency with climate change (Cai et al. 2014; Cai et al. 2015) and these major pulse

events are likely to drive extreme oceanographic conditions, with significant impacts

across multiple taxa. One recent example is the global coral bleaching event caused by

strong El Niño conditions (Eakin et al. 2016). The extreme temperature fluctuations that

resulted in the bleaching phenomenon would also have impacted other marine

organisms, even though the consequences of this event are still unknown. To date, most

climate change research in tropical ecosystems has focused on the effects of gradual

increases in land and sea temperatures and/or ocean acidification (Munday et al. 2008b;

Rummer and Munday 2016). However, in addition to gradual processes of climate

change, the consequences of strong ENSO events need to be addressed because, as my

thesis shows, the phenomenon has a pervasive influence on demographic traits of

diverse taxa spread across multiple ecosystems.

6.2.3 Climate impacts on a continental scale

Most studies that have examined demographic or life history characteristics of tropical

fishes have shown differences across latitudes (typically, because of changes in water

temperature), whether at regional spatial scales within, for instance, the Great Barrier

Reef in eastern Australia (Newman and Williams 1996; Gust et al. 2002; Williams et al.

2003), within ocean basins (Robertson et al. 2005; Williams et al. 2012), or across

ocean basins (Trip et al. 2008). The majority of these studies focus on single species and

have reported that regional (10s to 100s of km) differences in temperature, habitat and

other factors influence abundance and demographic characteristics (such as longevity,

size and growth rates) of different populations. However, the results of Chapter 5 show

that changes in the inter-annual strength of the ENSO signal, which drives the flow of

the major boundary current in Western Australia, simultaneously affected adult

populations of fishes over 23° of latitude across thermal limits, depths, demography and

trophic levels. The extensive influence of the Leeuwin Current on fishes occupying such


104

a wide range of habitats and niches has implications for other marine systems. Climate-

driven synchrony in growth across diverse taxa has already been recorded in the region

influenced by the California Current, another strong boundary current, enabling

construction of a multi-proxy, 576-year history of the winter upwelling index (Black et

al. 2014). This result strengthens the possibility that populations of fishes (and other

marine taxa) that reside in habitats strongly influenced by other boundary currents such

as the Benguela, Humboldt, Agulhas and Kuroshio currents may also show synchronous

responses to climate signals that affect key oceanographic drivers. Given that fisheries

within these current systems are major contributors to the global food production (e.g.

Fréon et al. 2009; García-Reyes et al. 2015), this has important implications for food

security world-wide.

6.3 Implications of varying climate impacts

The multi-faceted impacts of climate change on different life history stages and

populations across the range of a species indicate that a nuanced understanding of

impacts is required. Studies should include various life history stages and for species

that have ranges extending across different ocean basins or seas, region-specific studies

need to be established. Once baseline data for a species have been collected (Chapter 2

and Chapter 3), extension of the implications can be made to other similar taxa within

the same ecosystem if strong atmospheric and oceanographic links exist under the major

influence of a climate phenomenon such as the ENSO (Chapter 4 and Chapter 5).

Alternatively, it may be possible to use existing datasets to infer responses of multiple

taxa along the same continental shelf where a strong boundary current is present. Such

inferences will be valuable for data-deficient and/or threatened species, and for under-

studied species in the tropical regions with limited resources and research funding. One

such region of particular interest is the tropical Indo-Pacific where the dominant current,

the Indonesian through-flow, is strongly linked to the ENSO system (Meyers 1996;

Wijffels and Meyers 2004). Annual fluctuations of ENSO, which affect the strength of

the Indonesian through-flow, might create similar climate-driven responses of multiple

taxa. Demonstration of synchrony across diverse taxa will also allow for the prediction

of how large-scale climate changes such as ENSO could affect the many species that are

currently data-deficient. Synchronous effects of climate change will be of great value in

addressing the knowledge gaps in these vulnerable coral reef ecosystems.


105

6.4 Embracing new statistical methods

A number of emerging analytical techniques have the potential to enhance the use of

otolith biochronologies in climate change studies. The majority of the studies on

validated otolith biochronologies use a dimensionless growth index after detrending for

age-related effects and average across individuals to construct an overall biochronology.

However, this approach reduces sample sizes and does not take individual variation into

account. Mixed effects models such as generalized additive mixed effects models can

analyse multiple intrinsic (e.g. age) and extrinsic (e.g. climate) drivers of growth

simultaneously and account for temporal and spatial autocorrelation (e.g. Weisberg et

al. 2010; Rountrey et al. 2014; Morrongiello and Thresher 2015). In particular, the

paper by Morrongiello and Thresher (2015) is very useful because it provides a detailed

appraisal of the use of mixed effects models in the analysis of otolith biochronologies.

The paper shows researchers how to develop model frameworks and is able to develop

nested spatially and temporally explicit models that allow for quantitative comparisons

of within- and among-individual growth rates. Another promising option is Wavelet

analysis, which is a time series analysis that overcomes the problem of non-stationarity

in time series data (Cazelles et al. 2008). Wavelet analysis can track changes in

frequency over time, a shortcoming of standard Fourier analysis (Cazelles et al. 2008).

Wavelet analysis has been used to detect the synergistic effects of fishing and climate

on abundance of the European hake Merluccius merluccius in the Mediterranean Sea

(Hidalgo et al. 2011), and might prove useful for studies using validated otolith

biochronologies to investigate climate-growth relationships.

Missing values are a common problem when dealing with long time-series of data. The

use of dynamic factor analysis, a dimension reduction technique similar to principal

component analysis, offers one means of addressing this issue (Zuur et al. 2003). The

main advantage of this technique is that it can handle missing values, hence

multivariate, historical time-series data can be analysed without the need to truncate

data (as is required in principal components analysis). The dynamic factor analysis

method was used successfully to estimate common trends in the time series of squid

Loligo forbesi abundance in Scottish (UK) waters and to relate these trends to climate

variables (Zuur and Pierce 2004). Likewise, the potential of this method to tease out

common trends from otolith biochronologies should be explored.


106

6.5 Future directions

Building on the results of this thesis, knowledge of how climate change affects life

history stages of multiple populations of ‘model’ tropical species will prove to be

critical to reduce uncertainty in the forecasting of climate impacts (Munday et al.

2009a). This is achievable through collaborative work and the synthesis of datasets that

examine the different life history stages of the same species. Large historical archives of

fish otoliths (Campana and Thorrold 2001) are also available worldwide and offer a

resource to identify climate-driven synchronies across multiple populations and species

and this may be particularly applicable in ecosystems that are driven by major current

flows. Demonstrably, fish records can also be supplemented with relevant coral, tree or

mollusc records to build up a comprehensive, multi-species analysis across ecosystems,

a need previously highlighted by Doney et al. (2012). The existence of such

representative cross-taxa synchronies allow for the extension of predicted responses to

other, data-limited species, an important generalisation given the high diversity of

tropical fishes, coupled with the lack of time, funding and data for research. That my

study species show spatial variability in responses across populations suggests a strong

need to also extend analyses regionally. This is a clear gap given that a review of the

number of publications on coral reef fishes from 1999–2009 showed that the

overwhelming majority of studies were completed in Australia and fewer in the

developing countries that contain coral reefs (Table 1, Montgomery 2011). This is

despite the fact that the ‘Coral Triangle’ of Southeast Asia is the centre of biodiversity,

with the highest species richness for all tropical reef taxa (Roberts et al. 2002).

Relatively little research has been done in this area because of the lack of well-equipped

and well-funded research facilities (Montgomery 2011). Thus, the development of

techniques used in this study holds promise for extending the analyses to these regions.

Clearly, this knowledge gap also has to be addressed given the importance of tropical

fisheries for the livelihoods of more than 20 million people who live in Southeast Asia

(Cesar et al. 1997; Sadovy 2005).

There is also a need to extend the approaches developed in this thesis to deep-water

species that can be affected by climate change (Nguyen et al. 2015) and are increasingly

important to food security as shallow-water populations are depleted. In particular,

tropical deep-water fishes include three groups that are highly valued by fisheries;

lutjanids (snappers), epinephelids (groupers) and lethrinids (emperors; Newman et al.

2016b). These taxa occur at depths of between 100–500 m and have high commercial,


107

subsistence and artisanal values throughout the Indo-Pacific region (Dalzell et al. 1996;

Wakefield et al. 2015; Newman et al. 2016b). Vulnerability of these deep-water fishes is

expected to be high because of their longevity, slow growth and low natural mortality

rates (Koslow et al. 2000; Newman and Dunk 2003; Fry et al. 2006; Andrews et al.

2012; Wakefield et al. 2015). However, the limited biological data that exists for these

species (Wakefield et al. 2015; Newman et al. 2016b) coupled with evidence of

depletion of some populations (Koslow et al. 2000; Perez et al. 2009) indicates that

there is an urgent need for biological data on fisheries and climate-induced impacts in

order to implement sustainable management strategies (Morato et al. 2006; Williams et

al. 2013; Newman et al. 2016b). The biochronology approach used in this thesis has the

potential to reveal their response to climate-driven variability in addition to fundamental

demographic information for these long-lived species such as age validation (Black et

al. 2005).

Another critical research direction is to develop a better understanding of the

relationship between growth, as manifest in the otolith biochronologies, and fisheries

productivity. First, the relationship between records of growth in otoliths and somatic

growth as individual fish age needs elucidation. Although otolith growth is used as a

proxy for somatic growth in this study and elsewhere (e.g. Rowell et al. 2008;

Neuheimer et al. 2011; Stocks et al. 2011; Black et al. 2013), it is important to note that

for sexually mature fishes, energy is allocated to reproduction rather than to increasing

size (Calow 1979; Ware 1982; McBride et al. 2015). Hence, changes in otolith growth

of older fishes might translate to changes in body condition or weight rather than

changes to somatic growth. The body condition of sexually mature adults is important

because life history stages are inter-dependent, with female condition affecting egg

quality (Gagliano and McCormick 2007) and size (Donelson et al. 2008), which can

translate to larval quality (McCormick 2003). As larval quality subsequently influences

larval survival (Gagliano et al. 2007) and recruitment (Bergenius et al. 2002; Venturelli

et al. 2009), this will influence juvenile survival (Vigliola and Meekan 2009) and

ultimately, adult abundance.

Second, it will be important to determine how climate-driven effects on growth impact

fisheries productivity. It can be difficult to disentangle both types of impacts (Nye et al.

2009), and it may thus be valuable to consider the combined effects of climate and

fishing on ecosystems rather than responses to these factors individually (Doney et al.

2012). For instance, a recent study by Quetglas et al. (2013) found a synchronous


108

response in population fluctuation of six exploited species in the Mediterranean brought

about by the interaction between fishing impact and climate variability. There is also

growing evidence that over-exploited species vary more strongly with climate than

unexploited species because fisheries-impacted recruitment is more variable (Hsieh et

al. 2006; Anderson et al. 2008). Indeed, this suggests that exploitation can reduce the

resilience and increase the sensitivity of populations to climate change (Brander 2007;

Botsford et al. 2011). Conversely, climate variations can reduce the productivity of

fisheries because of negative influences on growth, survival or reproductive output of

these populations (Brander 2007). Hence, these synergistic effects play an important

role in structuring populations, communities and ecosystems.

6.6 Conclusion

This thesis has demonstrated that the construction and analysis of biochronologies of

growth from the otoliths of fishes offer a means to examine and identify the key climate

drivers of growth patterns both at inter-annual and decadal time scales, and at spatial

scales from individuals reefs to regions encompassing both tropical and temperate

environments (~ 23° of latitude) on continental and ocean-basin scales. Such

applications of the methods and analyses using crossdated otolith biochronologies are

relatively new in tropical fisheries and the scope for expansion is exciting. Data-poor

regions of tropical fisheries, predominantly along the coastlines of developing countries

where inhabitants are heavily reliant on such fisheries for food and livelihoods,

represent a major challenge for sustainable management, particularly given issues such

as rapid climate change, over-exploitation and the growth of human populations in the

tropics. Given the value and vulnerability of tropical fisheries, it is imperative that we

further our understanding of the impacts of climate change in order to conserve and

sustain such ecosystems.

109

References

Allen GR (1985) FAO species catalogue. Vol. 6. Snappers of the world. An annotated and

illustrated catalogue of lutjanid species known to date. FAO fisheries synopsis 125, 208.

Alory G, Wijffels S, Meyers G (2007) Observed temperature trends in the Indian Ocean over

1960–1999 and associated mechanisms. Geophysical Research Letters 34, L02606, doi:

10.1029/2006GL028044.

Amezcua F, Soto-Avila C, Green-Ruiz Y (2006) Age, growth, and mortality of the spotted rose

snapper Lutjanus guttatus from the southeastern Gulf of California. Fisheries Research

77, 293-300.

Anderson CNK, Hsieh C-h, Sandin SA et al. (2008) Why fishing magnifies fluctuations in fish

abundance. Nature 452, 835-839.

Andrew NL, Béné C, Hall SJ, Allison EH, Heck S, Ratner BD (2007) Diagnosis and

management of small-scale fisheries in developing countries. Fish and Fisheries 8, 227-

240.

Andrews AH, DeMartini EE, Brodziak J, Nichols RS, Humphreys RL (2012) A long-lived life

history for a tropical, deepwater snapper (Pristipomoides filamentosus): bomb

radiocarbon and lead–radium dating as extensions of daily increment analyses in

otoliths. Canadian Journal of Fisheries and Aquatic Sciences 69, 1850-1869.

Asch RG (2015) Climate change and decadal shifts in the phenology of larval fishes in the

California Current ecosystem. Proceedings of the National Academy of Sciences 112,

E4065-E4074.

Ashok K, Guan Z, Yamagata T (2003) Influence of the Indian Ocean Dipole on the Australian

winter rainfall. Geophysical Research Letters 30, L1821.

Baker SJM, Wilson CA (2001) Use of bomb radiocarbon to validate otolith section ages of red

snapper Lutjanus campechanus from the northern Gulf of Mexico. Limnology and

Oceanography 46, 1819-1824.

Bakun A, Black BA, Bograd SJ, García-Reyes M, Miller AJ, Rykaczewski RR, Sydeman WJ

(2015) Anticipated Effects of Climate Change on Coastal Upwelling Ecosystems.

Current Climate Change Reports 1, 85-93.

Barton K (2015) Multi-model inference. R package version 1.13.4. Available at:

http://CRAN.R-project.org/package=MuMIn

Béné C, Heck S (2005) Fish and food security in Africa. pp 1-13, Cairo, Egypt, WorldFish

Center Quaterly.

Bergenius MA, Meekan MG, Robertson RD, McCormick MI (2002) Larval growth predicts the

recruitment success of a coral reef fish. Oecologia 131, 521-525.

Black BA, Boehlert GW, Yoklavich MM (2005) Using tree-ring crossdating techniques to

validate annual growth increments in long-lived fishes. Canadian Journal of Fisheries

and Aquatic Sciences 62, 2277-2284.

Black BA, Boehlert GW, Yoklavich MM (2008) Establishing climate-growth relationships for

yelloweye rockfish (Sebastes ruberrimus) in the northeast Pacific using a

dendrochronological approach. Fisheries Oceanography 17, 368-379.

Black BA, Copenheaver CA, Frank DC, Stuckey MJ, Kormanyos RE (2009) Multi-proxy

reconstructions of northeastern Pacific sea surface temperature data from trees and

Pacific geoduck. Palaeogeography, Palaeoclimatology, Palaeoecology 278, 40-47.

Black BA (2009) Climate-driven synchrony across tree, bivalve, and rockfish growth-increment

chronologies of the northeast Pacific. Marine Ecology Progress Series 378, 37-46.

Black BA, Allman RJ, Schroeder ID, Schirripa MJ (2011) Multidecadal otolith growth histories

for red and gray snapper (Lutjanus spp.) in the northern Gulf of Mexico, USA.

Fisheries Oceanography 20, 347-356.

Black BA, Matta ME, Helser TE, Wilderbuer TK (2013) Otolith biochronologies as

multidecadal indicators of body size anomalies in yellowfin sole (Limanda aspera).

Fisheries Oceanography 22, 523-532.

Black BA, Sydeman WJ, Frank DC et al. (2014) Six centuries of variability and extremes in a

coupled marine-terrestrial ecosystem. Science 345, 1498-1502.

References

110

Black BA, Griffin D, van der Sleen P et al. (2016) The value of crossdating to retain high-

frequency variability, climate signals, and extreme events in environmental proxies.

Global Change Biology 22, 2582-2595.

Boeuf G, Payan P (2001) How should salinity influence fish growth? Comparative Biochemistry

and Physiology Part C 130, 411-423.

Botsford LW, Holland MD, Samhouri JF, White JW, Hastings A (2011) Importance of age

structure in models of the response of upper trophic levels to fishing and climate

change. ICES Journal of Marine Science 68, 1270-1283.

Braganza K, Gergis JL, Power SB, Risbey JS, Fowler AM (2009) A multiproxy index of the El

Niño–Southern Oscillation, A.D. 1525–1982. Journal of Geophysical Research:

Atmospheres 114, D05106.

Brander KM (2007) Global fish production and climate change. Proceedings of the National

Academy of Sciences 104, 19709-19714.

Brown CJ, O'Connor MI, Poloczanska ES et al. (2016) Ecological and methodological drivers

of species’ distribution and phenology responses to climate change. Global Change

Biology 22, 1548-1560.

Bunn AG (2008) A dendrochronology program library in R (dplR). Dendrochronologia 26,

115-124.

Burnham KP, Anderson DR (2004) Multimodel inference: understanding AIC and BIC in

model selection. Sociological Methods & Research 33, 261-304.

Cai W, Cowan T, Sullivan A (2009) Recent unprecedented skewness towards positive Indian

Ocean Dipole occurrences and its impact on Australian rainfall. Geophysical Research

Letters 36, L11705, doi: 10.1029/2009GL037604.

Cai W, Borlace S, Lengaigne M et al. (2014) Increasing frequency of extreme El Niño events

due to greenhouse warming. Nature Climate Change 4, 111-116.

Cai W, Wang G, Santoso A et al. (2015) Increasing frequency of extreme La Niña events under

greenhouse warming. Nature Climate Change 5, 132-137.

Calow P (1979) The cost of reproduction - a physiological approach. Biological Reviews 54, 23-

40.

Campana SE, Thorrold SR (2001) Otoliths, increments, and elements: keys to a comprehensive

understanding of fish populations? Canadian Journal of Fisheries and Aquatic Sciences

58, 30-38.

Campana SE (2004) Photographic Atlas of Fish Otoliths of the Northwest Atlantic Ocean

Ottawa, Canada, NRC Research Press.

Cantin NE, Lough JM (2014) Surviving coral bleaching events: Porites growth anomalies on

the Great Barrier Reef. PLoS ONE 9, e88720.

Cappo M, Marriott RJ, Newman SJ (2013) James's rule and causes and consequences of a

latitudinal cline in the demography of John's Snapper (Lutjanus johnii) in coastal waters

of Australia. Fishery Bulletin 111, 309-324.

Caputi N, Fletcher WJ, Pearce A, Chubb CF (1996) Effect of the Leeuwin Current on the

recruitment of fish and invertebrates along the Western Australian coast. Marine and

Freshwater Research 47, 147-155.

Caputi N (2008) Impact of the Leeuwin Current on the spatial distribution of the puerulus

settlement of the western rock lobster (Panulirus cygnus) and implications for the

fishery of Western Australia. Fisheries Oceanography 17, 147-152.

Caputi N, Kangas M, Denham A, Feng M, Pearce A, Hetzel Y, Chandrapavan A (2016)

Management adaptation of invertebrate fisheries to an extreme marine heat wave event

at a global warming hot spot. Ecology and Evolution 6, 3583-3593.

Carpenter KE, Niem VH (2001) FAO species identification guide for fishery purposes. In: The

living marine resources of the Western Central Pacific. Bony fishes part 3 (Menidae to

Pomacentridae). FAO Library.

Carpenter KE, Abrar M, Aeby G et al. (2008) One-Third of Reef-Building Corals Face Elevated

Extinction Risk from Climate Change and Local Impacts. Science 321, 560-563.

Carton JA, Giese BS (2008) A reanalysis of ocean climate using Simple Ocean Data

Assimilation (SODA). Monthly Weather Review 136, 2999-3017.

References

111

Casselman JM (1990) Growth and relative size of calcified structures of fish. Transactions of

the American Fisheries Society 119, 673-388.

Cazelles B, Chavez M, Berteaux D, Ménard F, Vik JO, Jenouvrier S, Stenseth NC (2008)

Wavelet analysis of ecological time series. Oecologia 156, 287-304.

Cesar H, Lundin CG, Bettencourt S, Dixon J (1997) Indonesian coral reefs - an economic

analysis of a precious but threatened resource. Ambio 26, 345-350.

Chavez FP, Ryan J, Lluch-Cota SE, Ñiquen C. M (2003) From Anchovies to Sardines and

Back: Multidecadal Change in the Pacific Ocean. Science 299, 217-221.

Chen X, Wallace JM (2015) ENSO-Like Variability: 1900–2013. Journal of Climate 28, 9623-

9641.

Cheung WWL, Meeuwig JJ, Feng M et al. (2012a) Climate-change induced tropicalisation of

marine communities in Western Australia. Marine and Freshwater Research 63, 415.

Cheung WWL, Sarmiento JL, Dunne J et al. (2012b) Shrinking of fishes exacerbates impacts of

global ocean changes on marine ecosystems. Nature Climate Change 3, 254-258.

Cheung WWL, Brodeur RD, Okey TA, Pauly D (2015) Projecting future changes in

distributions of pelagic fish species of Northeast Pacific shelf seas. Progress in

Oceanography 130, 19-31.

Choat JH, Axe LM (1996) Growth and longevity in acanthurid fishes; an analysis of otolith

increments. Marine Ecology Progress Series 134, 15-26.

Choukroun S, Ridd PV, Brinkman R, McKinna LIW (2010) On the surface circulation in the

western Coral Sea and residence times in the Great Barrier Reef. Journal of

Geophysical Research: Oceans 115, C06013, doi: 10.1029/2009JC005761.

Christensen JH, Krishna Kumar K, Aldrian E et al. (2013) Climate phenomena and their

relevance for future regional climate change. In: Climate Change 2013: The physical

science basis. (eds Stocker TF, Qin D, Plattner G-K, Tignor M, Allen SK, Boschung J,

Nauels A, Xia Y, Bex V, Midgley PM) pp 1217-1308, Cambridge, UK and New York,

USA, Cambridge University Press.

Cohen B (2006) Urbanization in developing countries: Current trends, future projections, and

key challenges for sustainability. Technology in Society 28, 63-80.

Cohen DM, Inada T, Iwamoto T, Scialabba N (1990) Gadiform fishes of the world (Order

Gadiformes). In: FAO Fisheries Synopsis No. 125. pp 442, Rome.

Collins LB, Zhu ZR, Wyrwoll KH et al. (1993) Late Quaternary evolution of coral reefs on a

cool-water carbonate margin: the Abrolhos Carbonate Platforms, southwest Australia.

Marine Geology 110, 203-212.

Condie SA, Dunn JR (2006) Seasonal characteristics of the surface mixed layer in the

Australasian region: implications for primary production regimes and biogeography.

Marine and Freshwater Research 57, 569-590.

Conover DO, Munch SB (2002) Sustaining Fisheries Yields Over Evolutionary Time Scales.

Science 297, 94-96.

Cook ER, Briffa KR, Shiyatov S, Mazepa A, Jones PD (1990) Data analysis. In: Methods of

Dendrochronology: Applications in the Environmental Sciences. (eds Cook ER,

Kairiukstis LA) pp 97-162. Dordrecht, Netherlands, Kluwer Academic Publishers.

Cooper TF, O'Leary RA, Lough JM (2012) Growth of Western Australian corals in the

Anthropocene. Science 335, 593-596.

Coulson PG, Potter IC, Hall NG (2012) The biological characteristics of Scorpis aequipinnis

(Kyphosidae), including relevant comparisons with those of other species and

particularly of a heavily exploited congener. Fisheries Research 125–126, 272-282.

Coulson PG, Black BA, Potter IC, Hall NG (2014) Sclerochronological studies reveal that

patterns of otolith growth of adults of two co-occurring species of Platycephalidae are

synchronised by water temperature variations. Marine Biology 161, 383-393.

Cresswell GR, Golding TJ (1980) Observations of a southward flowing current in the south-

eastern Indian Ocean. Deep-Sea Research 27A, 449-466.

Cullen LE, Grierson PF (2007) A stable oxygen, but not carbon, isotope chronology of Callitris

columellaris reflects recent climate change in north-western Australia. Climatic Change

85, 213-229.

References

112

Cullen LE, Adams MA, Anderson MJ, Grierson PF (2008) Analyses of δ13C and δ18O in tree

rings of Callitris columellaris provide evidence of a change in stomatal control of

photosynthesis in response to regional changes in climate. Tree Physiology 28, 1525-

1533.

Cullen LE, Grierson PF (2009) Multi-decadal scale variability in autumn-winter rainfall in

south-western Australia since 1655 AD as reconstructed from tree rings of Callitris

columellaris. Climate Dynamics 33, 433-444.

D’Adamo N, Fandry C, Buchan S, Domingues C (2009) Northern sources of the Leeuwin

current and the “Holloway Current” on the North West Shelf. Journal of the Royal

Society of Western Australia 92, 53-66.

Dai A, Wigley TML (2000) Global patterns of ENSO-induced precipitation. Geophysical

Research Letters 27, 1283-1286.

Dalzell P, Adams T, Polunin N (1996) Coastal fisheries in the Pacific Islands. Oceanography

and marine biology: an annual review 34, 395-537.

De'ath G, Lough JM, Fabricius KE (2009) Declining coral calcification on the Great Barrier

Reef. Science 323, 116-119.

DeMartini EE, Friedlander AM, Sandin SA, Sala E (2008) Differences in fish-assemblage

structure between fished and unfished atolls in the northern Line Islands, central Pacific.

Marine Ecology Progress Series 365, 199-215.

Denniston RF, Villarini G, Gonzales AN et al. (2015) Extreme rainfall activity in the Australian

tropics reflects changes in the El Niño/Southern Oscillation over the last two millennia.

Proceedings of the National Academy of Sciences 112, 4576-4581.

Depczynski M, Gilmour JP, Ridgway T et al. (2013) Bleaching, coral mortality and subsequent

survivorship on a West Australian fringing reef. Coral Reefs 32, 233-238.

Diaz RJ, Rosenberg R (2008) Spreading Dead Zones and Consequences for Marine Ecosystems.

Science 321, 926-929.

Dixson DL, Munday PL, Jones GP (2010) Ocean acidification disrupts the innate ability of fish

to detect predator olfactory cues. Ecology Letters 13, 68-75.

Donelson JM, McCormick MI, Munday PL (2008) Parental condition affects early life-history

of a coral reef fish. Journal of Experimental Marine Biology and Ecology 360, 109-116.

Donelson JM, Munday PL, McCormick MI, Pankhurst NW, Pankhurst PM (2010) Effects of

elevated water temperature and food availability on the reproductive performance of a

coral reef fish. Marine Ecology Progress Series 401, 233-243.

Doney SC, Fabry VJ, Feely RA, Kleypas JA (2009) Ocean Acidification: The Other CO2

Problem. Annual Review of Marine Science 1, 169-192.

Doney SC, Ruckelshaus M, Duffy JE et al. (2012) Climate change impacts on marine

ecosystems. Annual Review of Marine Science 4, 11-37.

Doubleday ZA, Izzo C, Haddy JA, Lyle JM, Ye Q, Gillanders BM (2015) Long-term patterns in

estuarine fish growth across two climatically divergent regions. Oecologia 179, 1079-

1090.

Drinkwater KF, Beaugrand G, Kaeriyama M et al. (2010) On the processes linking climate to

ecosystem changes. Journal of Marine Systems 79, 374-388.

Dulvy NK, Rogers SI, Jennings S, Stelzenmüller V, Dye SR, Skjoldal HR (2008) Climate

change and deepening of the North Sea fish assemblage: a biotic indicator of warming

seas. Journal of Applied Ecology 45, 1029-1039.

Durack PJ, Wijffels SE (2010) Fifty-Year Trends in Global Ocean Salinities and Their

Relationship to Broad-Scale Warming. Journal of Climate 23, 4342-4362.

Dutil J-D, Brander K (2003) Comparing productivity of North Atlantic cod (Gadus morhua)

stocks and limits to growth production. Fisheries Oceanography 12, 502-512.

Eakin CM, Liu G, Gomez AM et al. (2016) Global Coral Bleaching 2014-2017: Status and an

Appeal for Observations. In: Reef Encounter. pp 20-26.

Edwards M, Richardson AJ (2004) Impact of climate change on marine pelagic phenology and

trophic mismatch. Nature 430, 881-884.

Elliott M, Whitfield AK, Potter IC, Blaber SJM, Cyrus DP, Nordlie FG, Harrison TD (2007)

The guild approach to categorizing estuarine fish assemblages: a global review. Fish

and Fisheries 8, 241-268.

References

113

Emata AC (2003) Reproductive performance in induced and spontaneous spawning of the

mangrove red snapper, Lutjanus argentimaculatus: a potential candidate species for

sustainable aquaculture. Aquaculture Research 34, 849-857.

FAO (2014) The State of World Fisheries and Aquaculture - 2014 (SOFIA). pp 223, Rome,

Food and Agriculture Organization of the United Nations.

Feng M, Meyers G, Pearce A, Wijffels S (2003) Annual and interannual variations of the

Leeuwin Current at 32°S. Journal of Geophysical Research: Oceans 108, 3355.

Feng M, Waite AM, Thompson PA (2009) Climate variability and ocean production in the

Leeuwin Current system off the west coast of Western Australia. Journal of the Royal

Society of Western Australia 92, 67-81.

Feng M, McPhaden MJ, Xie SP, Hafner J (2013) La Niña forces unprecedented Leeuwin

Current warming in 2011. Scientific Reports 3, 1277.

Feng M, Benthuysen J, Zhang N, Slawinski D (2015) Freshening anomalies in the Indonesian

throughflow and impacts on the Leeuwin Current during 2010–2011. Geophysical

Research Letters 42, 8555-8562.

Ferreira BP, Russ GR (1994) Age validation and estimation of growth rate of the coral trout,

Plectropomus leopardus, (Lavepede 1802) from Lizard Island, Northern Great Barrier

Reef. Fishery Bulletin 92, 42-57.

Fischer AJ, Baker Jr MS, Wilson CA, Nieland DL (2005) Age, growth, mortality, and

radiometric age validation of gray snapper (Lutjanus griseus) from Louisiana. Fishery

Bulletin 103, 307-319.

Fletcher WJ, Santoro K (2012) Status Reports of the Fisheries and Aquatic Resources of

Western Australia 2011/2012: The state of the fisheries. pp 371, Western Australia,

Department of Fisheries.

Fletcher WJ, Santoro K (2014) Status reports of the Fisheries and Aquatic resources of Western

Australia 2013/14: The state of the Fisheries. pp 360, Western Australia, Department of

Fisheries.

Fletcher WJ, Santoro K (2015) Status Reports of the Fisheries and Aquatic Resources of

Western Australia 2014/15: The state of the Fisheries. pp 354, Western Australia,

Department of Fisheries.

Fordham DA, Mellin C, Russell BD et al. (2013) Population dynamics can be more important

than physiological limits for determining range shifts under climate change. Global

Change Biology 19, 3224-3237.

Fowler AJ (2009) Age in Years from Otoliths of Adult Tropical Fish. In: Tropical Fish

Otoliths: Information for Assessment, Management and Ecology. (eds Green BS,

Mapstone BD, Carlos G, Begg GA) pp 55-92. Dordrecht, Springer Netherland.

Fréon P, Barange M, Aristegui J (2009) Eastern boundary upwelling ecosystems: integrative

and comparative approaches. Progress in Oceanography 83, 1-14.

Fritts HC (1971) Dendroclimatology and dendroecology. Quaternary Research 1, 419-449.

Fritts HC, Swetnam TW (1989) Dendroecology: A Tool for Evaluating Variations in Past and

Present Forest Environments. In: Advances in Ecological Research. (eds Begon M,

Fitter AH, Ford ED, Macfadyen A) pp 111-188. Academic Press.

Fry GC, Brewer DT, Venables WN (2006) Vulnerability of deepwater demersal fishes to

commercial fishing: Evidence from a study around a tropical volcanic seamount in

Papua New Guinea. Fisheries Research 81, 126-141.

Gagliano M, McCormick MI (2007) Maternal condition influences phenotypic selection on

offspring. Journal of Animal Ecology 76, 174-182.

Gagliano M, McCormick MI, Meekan MG (2007) Survival against the odds: ontogenetic

changes in selective pressure mediate growth-mortality trade-offs in a marine fish.

Proceedings of the Royal Society of London B: Biological Sciences 274, 1575-1582.

García-Reyes M, Sydeman WJ, Schoeman DS, Rykaczewski RR, Black BA, Smit AJ, Bograd

SJ (2015) Under Pressure: Climate Change, Upwelling, and Eastern Boundary

Upwelling Ecosystems. Frontiers in Marine Science 2, Article 109.

Gaumet F, Boeuf G, Severe A, Le Roux A, Mayer-Gostan N (1995) Effects of salinity on the

ionic balance and growth of juvenile turbot. Journal of Fish Biology 47, 865-876.

Gentilli J (1971) Climate of Australia and New Zealand, Amsterdam, Elsevier.

References

114

Gillanders BM, Elsdon TS, Halliday IA, Jenkins GP, Robins JB, Valesini FJ (2011) Potential

effects of climate change on Australian estuaries and fish utilising estuaries: a review.


Gillanders BM, Black BA, Meekan MG, Morrison MA (2012) Climatic effects on the growth of

a temperate reef fish from the Southern Hemisphere: a biochronological approach.

Marine Biology 159, 1327-1333.

Gillanders BM, Izzo C, Doubleday ZA, Ye Q (2015) Partial migration: growth varies between

resident and migratory fish. Biology Letters 11, 20140850.

Good SA, Martin MJ, Rayner NA (2013) EN4: Quality controlled ocean temperature and

salinity profiles and monthly objective analyses with uncertainty estimates. Journal of

Geophysical Research 118, 6704-6716.

Grandcourt EM, Abdessalaam TZA, Francis F (2006) Age, growth, mortality and reproduction

of the blackspot snapper, Lutjanus fulviflamma (Forsskål, 1775), in the southern

Arabian Gulf. Fisheries Research 78, 203-210.

Green BS, Fisher R (2004) Temperature influences swimming speed, growth and larval duration

in coral reef fish larvae. Journal of Experimental Marine Biology and Ecology 299,

115-132.

Green BS, Mapstone BD, Carlos G, Begg GA (2009) Introduction to Otoliths and Fisheries in

the Tropics. In: Tropical Fish Otoliths: Informaiton for Assessment, Management and

Ecology. (eds Green BS, Mapstone BD, Carlos G, Begg GA) pp 1-22. Dordrecht,

Springer Netherland.

Grissino-Mayer HD (2001) Evaluating crossdating accuracy: a manual and tutorial for the

computer program COFECHA. Tree-ring Research 57, 205-221.

Gust N, Choat J, Ackerman J (2002) Demographic plasticity in tropical reef fishes. Marine

Biology 140, 1039-1051.

Guyette RP, Rabeni CF (1995) Climate response among growth increments of fish and trees.

Oecologia 104, 272-279.

Hanson CE, Pattiaratchi CB, Waite AM (2005) Sporadic upwelling on a downwelling coast:

Phytoplankton responses to spatially variable nutrient dynamics off the Gascoyne

region of Western Australia. Continental Shelf Research 25, 1561-1582.

Harley CDG, Randall Hughes A, Hultgren KM et al. (2006) The impacts of climate change in

coastal marine systems. Ecology Letters 9, 228-241.

Hayward TL (1997) Pacific Ocean climate change: atmospheric forcing, ocean circulation and

ecosystem response. Trends in Ecology & Evolution 12, 150-154.

Helama S, Schöne BR, Black BA, Dunca E (2006) Constructing long-term proxy series for

aquatic environments with absolute dating control using a sclerochronological

approach: introduction and advanced applications. Marine and Freshwater Research

57, 591-599.

Helser TE, Lai H-l, Black BA (2012) Bayesian hierarchical modeling of Pacific geoduck growth

increment data and climate indices. Ecological Modelling 247, 210-220.

Hendon HH, Wang G (2010) Seasonal prediction of the Leeuwin Current using the POAMA

dynamical seasonal forecast model. Climate Dynamics 34, 1129-1137.

Hendy EJ, Gagan MK, Alibert CA, McCulloch MT, Lough JM, Isdale PJ (2002) Abrupt

decrease in tropical pacific sea surface salinity at end of little ice age. Science 295,

1511-1514.

Hendy EJ, Gagan MK, Lough JM (2003) Chronological control of coral records using

luminescent lines and evidence for non-stationary ENSO teleconnections in northeast

Australia. The Holocene 13, 187-199.

Heupel MR, Williams AJ, Welch DJ et al. (2010) Demographic characteristics of exploited

tropical lutjanids: a comparative analysis. Fishery Bulletin 108, 420-432.

Hidalgo M, Rouyer T, Molinero JC, Massutí E, Moranta J, Guijarro B, Stenseth NC (2011)

Synergistic effects of fishing-induced demographic changes and climate variation on

fish population dynamics. Marine Ecology Progress Series 426, 1-12.

Hobson ES (1975) Feeding patterns among tropical reef fishes. American Scientist 63, 382-392.

Hoegh-Guldberg O, Bruno JF (2010) The impact of climate change on the world's marine

ecosystems. Science 328, 1523-1528.

References

115

Hoegh-Guldberg O (2011) The Impact of Climate Change on Coral Reef Ecosystems. In: Coral

Reefs: An Ecosystem in Transition. (eds Dubinsky Z, Stambler N) pp 391-403.

Netherlands, Springer Netherlands.

Holbrook NJ, Goodwin ID, McGregor S, Molina E, Power SB (2011) ENSO to multi-decadal

time scale changes in East Australian Current transports and Fort Denison sea level:

Oceanic Rossby waves as the connecting mechanism. Deep Sea Research Part II:

Topical Studies in Oceanography 58, 547-558.

Holmes RL (1983) Computer-assisted quality control in tree-ring dating and measurement.

Tree-ring Bulletin 43, 69-78.

Hsieh C-h, Reiss CS, Hunter JR, Beddington JR, May RM, Sugihara G (2006) Fishing elevates

variability in the abundance of exploited species. Nature 443, 859-862.

Ingleby B, Huddleston M (2007) Quality control of ocean temperature and salinity profiles -

historical and real-time data. Journal of Marine Systems 65, 158-175.

IPCC (2014) Climate Change 2014: Synthesis Report. In: Contribution of Working Groups I, II

and III to the Fifth Assessment Report of the Intergovernmental Panel on Climate

Change. (eds Core Writing Team, Pachauri RJ, Meyer L) pp 151, Geneva, Switzerland.

Johansen JL, Jones GP (2011) Increasing ocean temperature reduces the metabolic performance

and swimming ability of coral reef damselfishes. Global Change Biology 17, 2971-

2979.

Johansen JL, Pratchett MS, Messmer V, Coker DJ, Tobin AJ, Hoey AS (2015) Large predatory

coral trout species unlikely to meet increasing energetic demands in a warming ocean.

Scientific Reports 5, 13830.

Jones PD, Harris I (2008) Climatic Research Unit (CRU) time-series datasets of variations in

climate with variations in other phenomena. University of East Anglia Climatic

Research Unit, Harwell Oxford, UK, NCAS British Atmospheric Data Centre.

Kamukuru AT, Hecht T, Mgaya YD (2005) Effects of exploitation on age, growth and mortality

of the blackspot snapper, Lutjanus fulviflamma, at Mafia Island, Tanzania. Fisheries

Management and Ecology 12, 45-55.

Keeling RF, Körtzinger A, Gruber N (2010) Ocean Deoxygenation in a Warming World.

Annual Review of Marine Science 2, 199-229.

Kendall NW, Rich HB, Jensen LR, Quinn TP (2010) Climate effects on inter-annual variation in

growth of the freshwater mussel (Anodonta beringiana) in an Alaskan lake. Freshwater

Biology 55, 2339-2346.

Kennedy JJ, Rayner NA, Smith RO, Parker DE, Saunby M (2011) Reassessing biases and other

uncertainties in sea-surface temperature observations measured in situ since 1850, part

1: measurement and sampling uncertainties. Journal of Geophysical Research 116,

D14103.

Kingsford MJ, Hughes JM (2005) Patterns of growth, mortality, and size of the tropical

damselfish Acanthochromis polyacanthus across the continental shelf of the Great

Barrier Reef. Fishery Bulletin 103, 561-573.

Kiso K, Mahyam M-I (2003) Distribution and feeding habits of juvenile and young John's

snapper Lutjanus johnii in the Matang mangrove estuary, west coast of Peninsular

Malaysia. Fisheries Science 69, 563-568.

Kleisner K, Mansour H, Pauly D (2014) Region-based MTI: resolving geographic expansion in

the Marine Trophic Index. Marine Ecology Progress Series 512, 185-199.

Koslow JA, Boehlert GW, Gordon JDM, Haedrich RL, Lorance P, Parin N (2000) Continental

slope and deep-sea fisheries: implications for a fragile ecosystem. ICES Journal of

Marine Science 57, 548-557.

Koslow JA, Pesant S, Feng M et al. (2008) The effect of the Leeuwin Current on phytoplankton

biomass and production off Southwestern Australia. Journal of Geophysical Research:

Oceans 113, C07050.

Kulbicki M, Bozec Y-M, Labrosse P, Letourneur Y, Mou-Tham G, Wantiez L (2005) Diet

composition of carnivorous fishes from coral reef lagoons of New Caledonia. Aquatic

Living Resources 18, 231-250.

References

116

Lambert Y, Dutil J-D, Munro J (1994) Effects of intermediate and low salinity conditions on

growth rate and food conversion of Atlantic cod (Gadus morhua). Canadian Journal of

Fisheries and Aquatic Sciences 51, 1569-1576.

Lavender SL, Abbs DJ (2013) Trends in Australian rainfall: contribution of tropical cyclones

and closed lows. Climate Dynamics 40, 317-326.

Law R (1991) On the Quantitative Genetics of Correlated Characters under Directional

Selection in Age-Structured Populations. Philosophical Transactions of the Royal

Society of London, Series B: Biological Sciences 331, 213-223.

Lehodey P, Alheit J, Barange M et al. (2006) Climate Variability, Fish, and Fisheries. Journal

of Climate 19, 5009-5030.

Leis JM, McCormick MI (2002) The Biology, Behavior, and Ecology of the Pelagic, Larval

stage of Coral Reef Fishes. In: Coral Reef Fishes: Dynamics and Diversity in a

Complex Ecosystem. (ed Sale PF) pp 171-199. USA, Elsevier Science.

Levitus S, Antonov JI, Boyer TP, Locarnini RA, Garcia HE, Mishonov AV (2009) Global ocean

heat content 1955–2008 in light of recently revealed instrumentation problems.

Geophysical Research Letters 36, L07608.

Liu KS, Chan JCL (2012) Interannual variation of Southern Hemisphere tropical cyclone

activity and seasonal forecast of tropical cyclone number in the Australian region.

International Journal of Climatology 32, 190-202.

Locarnini RA, Mishonov AV, Antonov JI et al. (2013) World Ocean Atlas 2013, Volume 1:

Temperature. In: NOAA Atlas NESDIS 73. (eds Levitus S, Mishonov A) pp 40, NOAA.

Lough JM (1998) Coastal climate of northwest Australia and comparisons with the Great

Barrier Reef: 1960 to 1992. Coral Reefs 17, 351-367.

Lough JM, Barnes DJ (2000) Environmental controls on growth of the massive coral Porites.

Journal of Experimental Marine Biology and Ecology 245, 225-243.

Lough JM, Cooper TF (2011) New insights from coral growth band studies in an era of rapid

environmental change. Earth-Science Reviews 108, 170-184.

Lough JM (2011) Great Barrier Reef coral luminescence reveals rainfall variability over

northeastern Australia since the 17th century. Paleoceanography 26, PA2201.

Mantua NJ, Hare SR (2002) The Pacific Decadal Oscillation. Journal of Oceanography 58, 35-

44.

Marchitto TM, Jones GA, Goodfriend GA, Weidman CR (2000) Precise Temporal Correlation

of Holocene Mollusk Shells Using Sclerochronology. Quaternary Research 53, 236-

246.

Marriott RJ (2005) Population biology, dynamics and their implications for management of red

bass: a large, long-lived reef fish. James Cook University, PhD thesis.

Marriott RJ, Mapstone BD (2006) Geographic influences on and the accuracy and precision of

age estimates for the red bass, Lutjanus bohar (Forsskal 1775): A large tropical reef

fish. Fisheries Research 80, 322-328.

Marriott RJ, Mapstone BD, Begg GA (2007) Age-specific demographic parameters, and their

implications for management of the red bass, Lutjanus bohar (Forsskal 1775): A large,

long-lived reef fish. Fisheries Research 83, 204-215.

Marriott RJ, Adams DJ, Jarvis NDC, Moran MJ, Newman SJ, Craine M (2010) Age-based

demographic assessment of fished stocks of Lethrinus nebulosus in the Gascoyne

Bioregion of Western Australia. Fisheries Management and Ecology 18, 89-103.

Marshall A, Hendon H, Feng M, Schiller A (2015) Initiation and amplification of the Ningaloo

Niño. Climate Dynamics 45, 2367-2385.

Marshall J, Armour KC, Scott JR et al. (2014) The ocean's role in polar climate change:

asymmetric Arctic and Antarctic responses to greenhouse gas and ozone forcing.

Philosophical Transactions of the Royal Society of London A: Mathematical, Physical

and Engineering Sciences 372, 20130040.

Matear RJ, Hirst AC (2003) Long-term changes in dissolved oxygen concentrations in the ocean

caused by protracted global warming. Global Biogeochemical Cycles 17, 1125.

Matta EM, Black BA, Wilderbuer TK (2010) Climate-driven synchrony in otolith growth-

increment chronologies for three Bering Sea flatfish species. Marine Ecology Progress

Series 413, 137-145.

References

117

Maxwell J, Cresswell G (1981) Dispersal of tropical marine fauna to the Great Australian Bight

by the Leeuwin Current. Marine and Freshwater Research 32, 493-500.

McBride RS, Somarakis S, Fitzhugh GR et al. (2015) Energy acquisition and allocation to egg

production in relation to fish reproductive strategies. Fish and Fisheries 16, 23-57.

McCormick MI, Molony BW (1995) Influence of water temperature during the larval stage on

size, age and body condition of a tropical reef fish at settlement. Marine Ecology

Progress Series 118, 59-68.

McCormick MI (2003) Consumption of coral propagules after mass spawning enhances larval

quality of damselfish through maternal effects. Oecologia 136, 37-45.

McManus WJ (1997) Tropical marine fisheries and the future of coral reefs: a brief review with

emphasis on Southeast Asia. Coral Reefs 16, S121-S127.

McPhaden MJ, Zebiak SE, Glantz MH (2006) ENSO as an Integrating Concept in Earth

Science. Science 314, 1740-1745.

Meehl GA, Teng H (2014) Regional precipitation simulations for the mid-1970s shift and early-

2000s hiatus. Geophysical Research Letters 41, 7658-7665.

Meekan MG, Ackerman JL, Wellington GM (2001) Demography and age structures of coral

reef damselfishes in the tropical eastern Pacific Ocean. Marine Ecology Progress Series

212, 223-232.

Meekan MG, Carleton JH, McKinnon AD, Flynn K, Furnas M (2003) What determines the

growth of tropical reef fish larvae in the plankton: food or temperature? Marine Ecology


Melvin TM, Briffa KR (2008) A "signal-free" approach to dendroclimatic standardisation.

Dendrochronologia 26, 71-86.

Meyers G (1996) Variation of Indonesian throughflow and the El Niño-Southern Oscillation.

Journal of Geophysical Research 101, 12255-12263.

Meyers G, McIntosh P, Pigot L, Pook M (2007) The Years of El Niño, La Niña, and

Interactions with the Tropical Indian Ocean. Journal of Climate 20, 2872-2880.

Molony BW, Choat JH (1990) Otolith increment widths and somatic growth rate: the presence

of a time-lag. Journal of Fish Biology 37, 541-551.

Montgomery WL (2011) Coral Reef Fishes: Opportunities, Challenges and Concerns. In: Coral

Reefs: An Ecosystem in Transition. (eds Dubinsky Z, Stambler N) pp 327-346.

Netherlands, Springer Netherlands.

Morato T, Watson R, Pitcher TJ, Pauly D (2006) Fishing down the deep. Fish and Fisheries 7,

24-34.

Morrongiello JR, Crook DA, King AJ, Ramsey DSL, Brown P (2011) Impacts of drought and

predicted effects of climate change on fish growth in temperate Australian lakes. Global


Morrongiello JR, Thresher RE, Smith DC (2012) Aquatic biochronologies and climate change.

Nature Climate Change 2, 849-857.

Morrongiello JR, Thresher RE (2015) A statistical framework to explore ontogenetic growth

variation among individuals and populations: a marine fish example. Ecological

Monographs 85, 93-115.

Munday PL, Kingsford MJ, O’Callaghan M, Donelson JM (2008a) Elevated temperature

restricts growth potential of the coral reef fish Acanthochromis polyacanthus. Coral

Reefs 27, 927-931.

Munday PL, Jones GP, Pratchett MS, Williams AJ (2008b) Climate change and the future for

coral reef fishes. Fish and Fisheries 9, 261-285.

Munday PL, Leis JM, Lough JM, Paris CB, Kingsford MJ, Berumen ML, Lambrechts J (2009a)

Climate change and coral reef connectivity. Coral Reefs 28, 379-395.

Munday PL, Crawley NE, Nilsson GE (2009b) Interacting effects of elevated temperature and

ocean acidification on the aerobic performance of coral reef fishes. Marine Ecology


Munday PL, Dixson DL, McCormick MI, Meekan M, Ferrari MCO, Chivers DP (2010)

Replenishment of fish populations is threatened by ocean acidification. Proceedings of

the National Academy of Sciences 107, 12930-12934.

References

118

Munday PL, McCormick MI, Nilsson GE (2012) Impact of global warming and rising CO2

levels on coral reef fishes: what hope for the future? Journal of Experimental Biology

215, 3865-3873.

Musick JA (1999) Ecology and conservation of long-lived marine animals. In: Life in the slow

lane: Ecology and conservation of long-lived marine animals. (ed Musick JA) pp 7.

Maryland, USA, American Fisheries Society.

Neuheimer AB, Thresher RE, Lyle JM, Semmens JM (2011) Tolerance limit for fish growth

exceeded by warming waters. Nature Climate Change 1, 110-113.

Newman M, Compo GP, Alexander MA (2003) ENSO-Forced Variability of the Pacific

Decadal Oscillation. Journal of Climate 16, 3853-3857.

Newman M, Alexander MA, Ault TR et al. (2016a) The Pacific Decadal Oscillation, Revisited.

Journal of Climate 29, 4399-4427.

Newman SJ, Williams DM (1996) Variation in reef associated assemblages of the Lutjanidae

and Lethrinidae at different distances offshore in the central Great Barrier Reef.

Environmental Biology of Fishes 46, 123-138.

Newman SJ, Cappo M, Williams DM (2000) Age, growth, mortality rates and corresponding

yield estimates using otoliths of the tropical red snappers, Lutjanus erythropterus, L.

malabaricus and L. sebae, from the central Great Barrier Reef. Fisheries Research 48,

1-14.

Newman SJ, Williams DM (2001) Spatial and temporal variation in assemblages of Lutjanidae,

Lethrinidae and associated fish species among mid-continental shelf reefs in the central

Great Barrier Reef. Marine and Freshwater Research 52, 843-851.

Newman SJ, Dunk IJ (2003) Age validation, growth, mortality, and additional population

parameters of the goldband snapper (Pristipomoides multidens) off the Kimberley coast

of northwestern Australia. Fishery Bulletin 101, 116-128.

Newman SJ, Skepper CL, Wakefield CB (2010) Age estimation and otolith characteristics of an

unusually old, red emperor snapper (Lutjanus sebae) captured off the Kimberley coast

of north-western Australia. Journal of Applied Ichthyology 26, 120-122.

Newman SJ, Williams AJ, Wakefield CB, Nicol SJ, Taylor BM, O’Malley JM (2016b) Review

of the life history characteristics, ecology and fisheries for deep-water tropical demersal

fish in the Indo-Pacific region. Reviews in Fish Biology and Fisheries 26, 537-562.

Nguyen HM, Rountrey AN, Meeuwig JJ et al. (2015) Growth of a deep-water, predatory fish is

influenced by the productivity of a boundary current system. Scientific Reports 5, 9044.

Nicholls RJ, Marinova N, Lowe JA et al. (2011) Sea-level rise and its possible impacts given a

‘beyond 4°C world’ in the twenty-first century. Philosophical Transactions of the Royal

Society of London A: Mathematical, Physical and Engineering Sciences 369, 161-181.

Nilsson GE, Crawley N, Lunde IG, Munday PL (2009) Elevated temperature reduces the

respiratory scope of coral reef fishes. Global Change Biology 15, 1405-1412.

Nye JA, Link JS, Hare JA, Overholtz WJ (2009) Changing spatial distribution of fish stocks in

relation to climate and population size on the Northeast United States continental shelf.

Marine Ecology Progress Series 393, 111-129.

O'Donnell AJ, Cook ER, Palmer JG, Turney CSM, Page GFM, Grierson PF (2015) Tree rings

show recent high summer-autumn precipitation in Northwest Australia is unprecedented

within the last two centuries. PLoS ONE 10, e0128533.

Orr JC, Fabry VJ, Aumont O et al. (2005) Anthropogenic ocean acidification over the twenty-

first century and its impact on calcifying organisms. Nature 437, 681-686.

Panella G (1974) Otolith growth patterns: an aid in age determination in temperate and tropical

fishes. In: The ageing of fish. (ed Bagenal TB) pp 28-39. London, UK, Unwin Brothers

Ltd.

Parmesan C, Yohe G (2003) A globally coherent fingerprint of climate change impacts across

natural systems. Nature 421, 37-42.

Pastén GP, Katayama S, Omori M (2003) Timing of Parturition, Planktonic Duration, and

Settlement Patterns of the Black Rockfish, Sebastes inermis. Environmental Biology of

Fishes 68, 229-239.

Pauly D, Christensen V, Guenette S et al. (2002) Towards sustainability in world fisheries.

Nature 418, 689-695.

References

119

Pauly D, Zeller D (2016) Catch reconstructions reveal that global marine fisheries catches are

higher than reported and declining. Nature Communications 7: 10244.

Pearce A, Buchan S, Chiffings T et al. (2003) A review of the oceanography of the Dampier

Archipelago, Western Australia. In: The marine flora and fauna of Dampier, Western

Australia. (eds Wells FE, Walker DI, Jones DS) pp 13-50. Perth, Western Australia,

Western Australian Museum.

Pearce A, Feng M (2007) Observations of warming on the Western Australian continental shelf.


Pember MB, Newman SJ, Hesp SA, Young GC, Skepper CL, Hall NG, Potter IC (2005)

Biological parameters for managing the fisheries for Blue and King Threadfin Salmons,

Estuary Rockcod, Malabar Grouper and Mangrove jack in north-western Australia.

Western Australia, Centre for Fish and Fisheries Research, Murdoch University.

Perez JAA, Pezzuto PR, Wahrlich R, de Souza Soares AL (2009) Deep-water fisheries in

Brazil: history, status and perspectives. Latin American Journal of Aquatic Research

37, 513-541.

Perry AL, Low PJ, Ellis JR, Reynolds JD (2005) Climate change and distribution shifts in

marine fishes. Science 308, 1912-1915.

Poloczanska ES, Brown CJ, Sydeman WJ et al. (2013) Global imprint of climate change on

marine life. Nature Clim. Change 3, 919-925.

Polunin NVC (1996) Trophodynamics of reef fisheries productivity. In: Reef Fisheries. (eds

Polunin NVC, Roberts CM) pp 113-135. Dordrecht, Springer Netherlands.

Pörtner HO, Peck MA (2010) Climate change effects on fishes and fisheries: towards a cause-

and-effect understanding. Journal of Fish Biology 77, 1745-1779.

Power S, Tseitkin F, Mehta W, Lavery B, Torok S, Holbrook N (1999) Decadal climate

variability in Australia during the twentieth century. International Journal of

Climatology 19, 169-184.

Pratchett MS, Munday P, Wilson SK et al. (2008) Effects of climate-induced coral bleaching on

coral-reef fishes. Oceanography and marine biology: an annual review 46, 251-296.

Quetglas A, Ordines F, Hidalgo M et al. (2013) Synchronous combined effects of fishing and

climate within a demersal community. ICES Journal of Marine Science: Journal du

Conseil 70, 319-328.

R Core Team (2015) R: A language and environment for statistical computing. 3.1.3. Available

at: http://www.R-project.org/

Rayner NA, Parker DE, Horton EB et al. (2003) Global analyses of sea surface temperature, sea

ice, and night marine air temperature since the late nineteenth century. Journal of

Geophysical Research: Atmospheres 108, 4407.

Redondo-Rodriguez A, Weeks SJ, Berkelmans R, Hoegh-Guldberg O, Lough JM (2012)

Climate variability of the Great Barrier Reef in relation to the tropical Pacific and El

Niño-Southern Oscillation. Marine and Freshwater Research 63, 34-47.

Reynolds JD, Jenning S, Dulvy N (2001) Life histories of fishes and population responses to

exploitation. In: Conservation of exploited species. (eds Reynolds JD, Mace GM,

Redford KH, Robinson JG) pp 147-168. Cambridge, UK, Cambridge University Press.

Richardson AJ, Brown CJ, Brander K et al. (2012) Climate change and marine life. Biology

Letters 8, 907-909.

Ridgway KR, Godfrey JS (1997) Seasonal cycle of the East Australian Current. Journal of

Geophysical Research: Oceans 102, 22921-22936.

Ridgway KR, Godfrey JS (2015) The source of the Leeuwin Current seasonality. Journal of

Geophysical Research: Oceans 120, 6843-6864.

Rijnsdorp AD, Peck MA, Engelhard GH, Möllmann C, Pinnegar JK (2009) Resolving the effect

of climate change on fish populations. ICES Journal of Marine Science 66, 1570-1583.

Risbey JS, Pook MJ, McIntosh PC, Wheeler MC, Hendon HH (2009) On the Remote Drivers of

Rainfall Variability in Australia. Monthly Weather Review 137, 3233-3253.

Roberts CM, McClean CJ, Veron JEN et al. (2002) Marine Biodiversity Hotspots and

Conservation Priorities for Tropical Reefs. Science 295, 1280-1284.

References

120

Robertson DR, Ackerman JL, Choat JH, Posada JM, Pitt J (2005) Ocean surgeonfish

Acanthurus bahianus. I. The geography of demography. Marine Ecology Progress

Series 295, 229-244.

Roessig JM, Woodley CM, Cech JJ, Hansen LJ (2004) Effects of global climate change on

marine and estuarine fishes and fisheries. Reviews in Fish Biology and Fisheries 14,

251-275.

Rogers A, Blanchard Julia L, Mumby Peter J (2014) Vulnerability of Coral Reef Fisheries to a

Loss of Structural Complexity. Current Biology 24, 1000-1005.

Rosenzweig C, Karoly D, Vicarelli M et al. (2008) Attributing physical and biological impacts

to anthropogenic climate change. Nature 453, 353-357.

Rountrey AN (2009) Life histories of juvenile woolly mammoths from Siberia: stable isotope

and elemental analyses of tooth dentin. PhD Thesis. The University of Michigan.

Rountrey AN, Coulson PG, Meeuwig JJ, Meekan MG (2014) Water temperature and fish

growth: otoliths predict growth patterns of a marine fish in a changing climate. Global


Rowell K, Flessa KW, Dettman DL, Román MJ, Gerber LR, Findley LT (2008) Diverting the

Colorado River leads to a dramatic life history shift in an endangered marine fish.

Biological Conservation 141, 1138-1148.

Rummer JL, Couturier CS, Stecyk JAW, Gardiner NM, Kinch JP, Nilsson GE, Munday PL

(2014) Life on the edge: thermal optima for aerobic scope of equatorial reef fishes are

close to current day temperatures. Global Change Biology 20, 1055-1066.

Rummer JL, Munday PL (2016) Climate change and the evolution of reef fishes: past and

future. Fish and Fisheries doi: 10.1111/faf.12164.

Russell DJ, McDougall AJ, Fletcher AS, Ovenden JR, Street R (2003) Biology, management

and genetic stock structure of mangrove jack (Lutjanus argentimaculatus) in Australia.

pp 189, Queensland Department of Primary Industries.

Ruttenberg BI, Hamilton SL, Walsh SM et al. (2011) Predator-Induced Demographic Shifts in

Coral Reef Fish Assemblages. PLoS ONE 6, e21062.

Ryan KL, Wise BS, Hall NG, Pollock KH, Sulin EH, Gaughan DJ (2013) An integrated system

to survey boat-based recreational fishing in Western Australia 2011/12. In: Fisheries

Research Report No. 249. pp 168, Western Australia, Department of Fisheries.

Sadovy Y (2005) Trouble on the reef: the imperative for managing vulnerable and valuable

fisheries. Fish and Fisheries 6, 167-185.

Saji NH, Goswami BN, Vinayachandran PN, Yamagata T (1999) A dipole mode in the tropical

Indian Ocean. Nature 401, 360-363.

Secor DH, Campana SE, Zdanowicz VS, Lam JWH, Yang L, Rooker JR (2002) Inter-laboratory

comparison of Atlantic and Mediterranean bluefin tuna otolith microconstituents. ICES

Journal of Marine Science 59, 1294-1304.

Serrano X, Grosell M, Serafy JE (2010) Salinity selection and preference of the grey snapper

Lutjanus griseus: field and laboratory observations. Journal of Fish Biology 76, 1592-

1608.

Shakun JD, Shaman JCL (2009) Tropical origins of North and South Pacific decadal variability.

Geophysical Research Letters 36, L19711.

Shimose T, Tachihara K (2005) Age, growth and maturation of the blackspot snapper Lutjanus

fulviflammus around Okinawa Island, Japan. Fisheries Science 71, 48-55.

Silverstein RN, Correa AMS, LaJeunesse TC, Baker AC (2011) Novel algal symbiont

(Symbiodinium spp.) diversity in reef corals of Western Australia. Marine Ecology


Simpson SD, Munday PL, Wittenrich ML, Manassa R, Dixson DL, Gagliano M, Yan HY

(2011) Ocean acidification erodes crucial auditory behaviour in a marine fish. Biology

Letters 7, 917-920.

Sims DW, Wearmouth VJ, Genner MJ, Southward AJ, Hawkins SJ (2004) Low-temperature-

driven early spawning migration of a temperate marine fish. Journal of Animal Ecology

73, 333-341.

References

121

Sinclair AF, Swain DP, Hanson JM (2002) Measuring changes in the direction and magnitude

of size-selective mortality in a commercial fish population. Canadian Journal of

Fisheries and Aquatic Sciences 59, 361-371.

Sleeman JC, Meekan MG, Fitzpatrick BJ, Steinberg CR, Ancel R, Bradshaw CJA (2010)

Oceanographic and atmospheric phenomena influence the abundance of whale sharks at

Ningaloo Reef, Western Australia. Journal of Experimental Marine Biology and

Ecology 382, 77-81.

Sogard SM (1997) Size-Selective Mortality in the Juvenile Stage of Teleost Fishes: A Review.

Bulletin of Marine Science 60, 1129-1157.

Sommer C, Schneider W, Poutiers JM (1996) The living marine resources of Somalia. In: FAO

species identification guide for fishery purposes. pp 376, Rome, Food and Agriculture

Organization of the United Nations.

Sponaugle S, Grorud-Colvert K, Pinkard D (2006) Temperature-mediated variation in early life

history traits and recruitment success of the coral reef fish Thalassoma bifasciatum in

the Florida Keys. Marine Ecology Progress Series 308, 1-15.

St. George SR (2014) The global network of tree-ring widths and its applications to

paleoclimatology. PAGES Magazine 22, 16-17.

Stevenson C, Katz LS, Micheli F et al. (2007) High apex predator biomass on remote Pacific

islands. Coral Reefs 26, 47-51.

Stobutzki IC, Silvestre GT, Garces LR (2006) Key issues in coastal fisheries in South and

Southeast Asia, outcomes of a regional initiative. Fisheries Research 78, 109-118.

Stocks J, Stewart J, Gray CA, West RJ (2011) Using otolith increment widths to infer spatial,

temporal and gender variation in the growth of sand whiting Sillago ciliata. Fisheries

Management and Ecology 18, 121-131.

Sutherland WJ (1996) From individual behaviour to population ecology, Oxford, UK, Oxford

University Press.

Thompson PA, Wild-Allen K, Lourey M, Rousseaux C, Waite AM, Feng M, Beckley LE

(2011) Nutrients in an oligotrophic boundary current: Evidence of a new role for the

Leeuwin Current. Progress in Oceanography 91, 345-359.

Thresher RE (1984) Reproduction in reef fishes, Neptune City, T.F.H. Publications Inc Ltd.

Thresher RE, Koslow JA, Morison AK, Smith DC (2007) Depth-mediated reversal of the

effects of climate change on long-term growth rates of exploited marine fish.

Proceedings of the National Academy of Sciences 104, 7461-7465.

Tierney JE, Abram NJ, Anchukaitis KJ et al. (2015) Tropical sea surface temperatures for the

past four centuries reconstructed from coral archives. Paleoceanography 30, 226-252.

Trip EL, Choat JH, Wilson DT, Robertson DR (2008) Inter-oceanic analysis of demographic

variation in a widely distributed Indo-Pacific coral reef fish. Marine Ecology Progress

Series 373, 97-109.

Trouet V, Van Oldenborgh GJ (2013) KNMI Climate Explorer: a web-based research tool for

high-resolution paleoclimatology. Tree-ring Research 69, 3-13.

Tsukamoto K, Arai T (2001) Facultative catadromy of the eel Anguilla japonica between

freshwater and seawater habitats. Marine Ecology Progress Series 220, 265-276.

Venturelli PA, Shuter BJ, Murphy CA (2009) Evidence for harvest-induced maternal influences

on the reproductive rates of fish populations. Proceedings of the Royal Society B-

Biological Sciences 276, 919-924.

Vermeer M, Rahmstorf S (2009) Global sea level linked to global temperature. Proceedings of

the National Academy of Sciences 106, 21527-21532.

Vigliola L, Meekan MG (2009) The Back-Calculation of Fish Growth From Otoliths. In:

Tropical Fish Otoliths: Information for Assessment, Management and Ecology. (eds

Green BS, Mapstone BD, Carlos G, Begg GA) pp 174-211. Dordrecht, Springer

Netherland.

Wakefield CB, Williams AJ, Newman SJ, Bunel M, Boddington DK, Vourey E, Fairclough DV

(2015) Variations in growth, longevity and natural mortality for the protogynous

hermaphroditic eightbar grouper Hyporthodus octofasciatus between the Indian and

Pacific Oceans. Fisheries Research 172, 26-33.

References

122

Walsh SM, Hamilton SL, Ruttenberg BI, Donovan MK, Sandin SA (2012) Fishing top predators

indirectly affects condition and reproduction in a reef-fish community. Journal of Fish

Biology 80, 519-537.

Walther G-R, Post E, Convey P et al. (2002) Ecological responses to recent climate change.

Nature 416, 389-395.

Wang B, Wu R, Li T (2003) Atmosphere–Warm Ocean Interaction and Its Impacts on Asian–

Australian Monsoon Variation. Journal of Climate 16, 1195-1211.

Ware DM (1982) Power and Evolutionary Fitness of Teleosts. Canadian Journal of Fisheries

and Aquatic Sciences 39, 3-13.

Weisberg S, Spangler G, Richmond LS (2010) Mixed effects models for fish growth. Canadian

Journal of Fisheries and Aquatic Sciences 67, 269-277.

Wernberg T, Smale DA, Tuya F et al. (2013) An extreme climatic event alters marine

ecosystem structure in a global biodiversity hotspot. Nature Climate Change 3, 78-82.

Whitehead PJP (1984) Clupeidae. In: Fishes of the north-eastern Atlantic and the

Mediterranean. (eds Whitehead PJP, Bauchot ML, Hureau J-C, Nielsen J, Tortonese E)

pp 28-281. Paris, France, UNESCO.

Wigley TML, Briffa KR, Jones PD (1984) On the average value of correlated time series, with

applications in dendroclimatology and hydrometeorology. Journal of Climate and

Applied Meteorology 23, 201-213.

Wijffels S, Meyers G (2004) An Intersection of Oceanic Waveguides: Variability in the

Indonesian Throughflow Region. Journal of Physical Oceanography 34, 1232-1253.

Wilkinson CR (2008) Status of coral reefs of the world. pp 298, Townsville, Australia, Global

Coral Reef Monitoring Network and Reef and Rainforest Research Centre.

Williams AJ, Davies CR, Mapstone BD, Russ GR (2003) Scales of spatial variation in

demography of a large coral-reef fish: an exception to the typical model? Fishery

Bulletin 101, 673-683.

Williams AJ, Mapstone BD, Davies CR (2007) Spatial and interannual patterns in growth of an

exploited coral-reef fish. Journal of Fish Biology 71, 970-992.

Williams AJ, Farley JH, Hoyle SD, Davies CR, Nicol SJ (2012) Spatial and Sex-Specific

Variation in Growth of Albacore Tuna (Thunnus alalunga) across the South Pacific

Ocean. PLoS ONE 7, e39318.

Williams AJ, Loeun K, Nicol SJ et al. (2013) Population biology and vulnerability to fishing of

deep-water Eteline snappers. Journal of Applied Ichthyology 29, 395-403.

Wilson CA, Nieland DL (2001) Age and growth of red snapper, Lutjanus campechanus, from

the northern Gulf of Mexico off Louisiana. Fishery Bulletin 99, 653-665.

Wilson SK, Adjeroud M, Bellwood DR et al. (2010) Crucial knowledge gaps in current

understanding of climate change impacts on coral reef fishes. The Journal of

Experimental Biology 213, 894-900.

Wolanski E, Deleersnijder E (1998) Island-generated internal waves at Scott Reef, Western

Australia. Continental Shelf Research 18, 1649-1666.

Wolanski E, Lambrechts J, Thomas C, Deleersnijder E (2013) The net water circulation through

Torres strait. Continental Shelf Research 64, 66-74.

Wolter K, Timlin MS (1993) Monitoring ENSO in COADS with a seasonally adjusted principal

component index. In: Proceedings of the 17th Climate Diagnostics Workshop. pp 52-

57.

Wong APS, Bindoff NL, Church JA (1999) Large-scale freshening of intermediate waters in the

Pacific and Indian oceans. Nature 400, 440-443.

Wood S, Scheipl F (2015) Generalized additive mixed models using mgcv and lme4. R package

version 0.2-3. Available at: http://CRAN.R-project.org/package=gamm4

Wright A, Dalzell PJ, Richards AH (1986) Some aspects of the biology of the red bass, Lutjanus

bohar (Forsskal), from the Tigak Islands, Papua New Guinea. Journal of Fish Biology

28, 533-544.

Wuenschel MJ, Jugovich AR, Hare JA (2005) Metabolic response of juvenile gray snapper

(Lutjanus griseus) to temperature and salinity: physiological cost of different

environments. Journal of Experimental Marine Biology and Ecology 321, 145-154.

References

123

Yamada H (2010) Age and growth during immature stages of the mangrove red snapper

Lutjanus argentimaculatus in waters around Ishigaki Island, southern Japan. Fisheries

Science 76, 445-450.

Zinke J, Rountrey AN, Feng M et al. (2014) Corals record long-term Leeuwin current

variability including Ningaloo Niño/Niña since 1975. Nature Communications 5, 3607.

Zinke J, Hoell A, Lough JM et al. (2015) Coral record of southeast Indian Ocean marine

heatwaves with intensified Western Pacific temperature gradient. Nature

Communications 6, 8562.

Zuur A, Pierce G (2004) Common trends in northeast Atlantic squid time series. Journal of Sea

Research 52, 57-72.

Zuur AF, Fryer R, Jolliffe I, Dekker R, Beukema J (2003) Estimating common trends in

multivariate time series using dynamic factor analysis. Environmetrics 14, 665-685.

Zuur AF, Ieno EN, Walker NJ, Saveliev AA, Smith GM (2009) Mixed effects models and

extensions in ecology with R, New York, USA, Springer.

Documents

Otolith growth chronologies: Investigating historic ... · Analysis of the growth chronologies held within the otoliths of fishes offers a means to achieve this aim. To date, most